linux/mm/memcontrol.c

7198 lines
186 KiB
C
Raw Normal View History

treewide: Replace GPLv2 boilerplate/reference with SPDX - rule 157 Based on 3 normalized pattern(s): this program is free software you can redistribute it and or modify it under the terms of the gnu general public license as published by the free software foundation either version 2 of the license or at your option any later version this program is distributed in the hope that it will be useful but without any warranty without even the implied warranty of merchantability or fitness for a particular purpose see the gnu general public license for more details this program is free software you can redistribute it and or modify it under the terms of the gnu general public license as published by the free software foundation either version 2 of the license or at your option any later version [author] [kishon] [vijay] [abraham] [i] [kishon]@[ti] [com] this program is distributed in the hope that it will be useful but without any warranty without even the implied warranty of merchantability or fitness for a particular purpose see the gnu general public license for more details this program is free software you can redistribute it and or modify it under the terms of the gnu general public license as published by the free software foundation either version 2 of the license or at your option any later version [author] [graeme] [gregory] [gg]@[slimlogic] [co] [uk] [author] [kishon] [vijay] [abraham] [i] [kishon]@[ti] [com] [based] [on] [twl6030]_[usb] [c] [author] [hema] [hk] [hemahk]@[ti] [com] this program is distributed in the hope that it will be useful but without any warranty without even the implied warranty of merchantability or fitness for a particular purpose see the gnu general public license for more details extracted by the scancode license scanner the SPDX license identifier GPL-2.0-or-later has been chosen to replace the boilerplate/reference in 1105 file(s). Signed-off-by: Thomas Gleixner <tglx@linutronix.de> Reviewed-by: Allison Randal <allison@lohutok.net> Reviewed-by: Richard Fontana <rfontana@redhat.com> Reviewed-by: Kate Stewart <kstewart@linuxfoundation.org> Cc: linux-spdx@vger.kernel.org Link: https://lkml.kernel.org/r/20190527070033.202006027@linutronix.de Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
2019-05-27 06:55:06 +00:00
// SPDX-License-Identifier: GPL-2.0-or-later
/* memcontrol.c - Memory Controller
*
* Copyright IBM Corporation, 2007
* Author Balbir Singh <balbir@linux.vnet.ibm.com>
*
* Copyright 2007 OpenVZ SWsoft Inc
* Author: Pavel Emelianov <xemul@openvz.org>
*
* Memory thresholds
* Copyright (C) 2009 Nokia Corporation
* Author: Kirill A. Shutemov
*
memcg: kmem controller infrastructure Introduce infrastructure for tracking kernel memory pages to a given memcg. This will happen whenever the caller includes the flag __GFP_KMEMCG flag, and the task belong to a memcg other than the root. In memcontrol.h those functions are wrapped in inline acessors. The idea is to later on, patch those with static branches, so we don't incur any overhead when no mem cgroups with limited kmem are being used. Users of this functionality shall interact with the memcg core code through the following functions: memcg_kmem_newpage_charge: will return true if the group can handle the allocation. At this point, struct page is not yet allocated. memcg_kmem_commit_charge: will either revert the charge, if struct page allocation failed, or embed memcg information into page_cgroup. memcg_kmem_uncharge_page: called at free time, will revert the charge. Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:56 +00:00
* Kernel Memory Controller
* Copyright (C) 2012 Parallels Inc. and Google Inc.
* Authors: Glauber Costa and Suleiman Souhlal
*
* Native page reclaim
* Charge lifetime sanitation
* Lockless page tracking & accounting
* Unified hierarchy configuration model
* Copyright (C) 2015 Red Hat, Inc., Johannes Weiner
*/
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
#include <linux/page_counter.h>
#include <linux/memcontrol.h>
#include <linux/cgroup.h>
#include <linux/pagewalk.h>
#include <linux/sched/mm.h>
#include <linux/shmem_fs.h>
#include <linux/hugetlb.h>
memcg: handle swap caches SwapCache support for memory resource controller (memcg) Before mem+swap controller, memcg itself should handle SwapCache in proper way. This is cut-out from it. In current memcg, SwapCache is just leaked and the user can create tons of SwapCache. This is a leak of account and should be handled. SwapCache accounting is done as following. charge (anon) - charged when it's mapped. (because of readahead, charge at add_to_swap_cache() is not sane) uncharge (anon) - uncharged when it's dropped from swapcache and fully unmapped. means it's not uncharged at unmap. Note: delete from swap cache at swap-in is done after rmap information is established. charge (shmem) - charged at swap-in. this prevents charge at add_to_page_cache(). uncharge (shmem) - uncharged when it's dropped from swapcache and not on shmem's radix-tree. at migration, check against 'old page' is modified to handle shmem. Comparing to the old version discussed (and caused troubles), we have advantages of - PCG_USED bit. - simple migrating handling. So, situation is much easier than several months ago, maybe. [hugh@veritas.com: memcg: handle swap caches build fix] Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Tested-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Hugh Dickins <hugh@veritas.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:07:56 +00:00
#include <linux/pagemap.h>
#include <linux/vm_event_item.h>
memory cgroup enhancements: add status accounting function for memory cgroup Add statistics account infrastructure for memory controller. All account information is stored per-cpu and caller will not have to take lock or use atomic ops. This will be used by memory.stat file later. CACHE includes swapcache now. I'd like to divide it to PAGECACHE and SWAPCACHE later. This patch adds 3 functions for accounting. * __mem_cgroup_stat_add() ... for usual routine. * __mem_cgroup_stat_add_safe ... for calling under irq_disabled section. * mem_cgroup_read_stat() ... for reading stat value. * renamed PAGECACHE to CACHE (because it may include swapcache *now*) [akpm@linux-foundation.org: coding-style fixes] [akpm@linux-foundation.org: fix smp_processor_id-in-preemptible] [akpm@linux-foundation.org: uninline things] [akpm@linux-foundation.org: remove dead code] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Signed-off-by: YAMAMOTO Takashi <yamamoto@valinux.co.jp> Cc: Balbir Singh <balbir@linux.vnet.ibm.com> Cc: Pavel Emelianov <xemul@openvz.org> Cc: Paul Menage <menage@google.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Cc: Kirill Korotaev <dev@sw.ru> Cc: Herbert Poetzl <herbert@13thfloor.at> Cc: David Rientjes <rientjes@google.com> Cc: Vaidyanathan Srinivasan <svaidy@linux.vnet.ibm.com> Cc: Kirill Korotaev <dev@sw.ru> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Cc: Paul Menage <menage@google.com> Cc: Pavel Emelianov <xemul@openvz.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Vaidyanathan Srinivasan <svaidy@linux.vnet.ibm.com> Cc: YAMAMOTO Takashi <yamamoto@valinux.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-07 08:14:24 +00:00
#include <linux/smp.h>
Memory controller: memory accounting Add the accounting hooks. The accounting is carried out for RSS and Page Cache (unmapped) pages. There is now a common limit and accounting for both. The RSS accounting is accounted at page_add_*_rmap() and page_remove_rmap() time. Page cache is accounted at add_to_page_cache(), __delete_from_page_cache(). Swap cache is also accounted for. Each page's page_cgroup is protected with the last bit of the page_cgroup pointer, this makes handling of race conditions involving simultaneous mappings of a page easier. A reference count is kept in the page_cgroup to deal with cases where a page might be unmapped from the RSS of all tasks, but still lives in the page cache. Credits go to Vaidyanathan Srinivasan for helping with reference counting work of the page cgroup. Almost all of the page cache accounting code has help from Vaidyanathan Srinivasan. [hugh@veritas.com: fix swapoff breakage] [akpm@linux-foundation.org: fix locking] Signed-off-by: Vaidyanathan Srinivasan <svaidy@linux.vnet.ibm.com> Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com> Cc: Pavel Emelianov <xemul@openvz.org> Cc: Paul Menage <menage@google.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Cc: Kirill Korotaev <dev@sw.ru> Cc: Herbert Poetzl <herbert@13thfloor.at> Cc: David Rientjes <rientjes@google.com> Cc: <Valdis.Kletnieks@vt.edu> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-07 08:13:53 +00:00
#include <linux/page-flags.h>
#include <linux/backing-dev.h>
Memory controller: memory accounting Add the accounting hooks. The accounting is carried out for RSS and Page Cache (unmapped) pages. There is now a common limit and accounting for both. The RSS accounting is accounted at page_add_*_rmap() and page_remove_rmap() time. Page cache is accounted at add_to_page_cache(), __delete_from_page_cache(). Swap cache is also accounted for. Each page's page_cgroup is protected with the last bit of the page_cgroup pointer, this makes handling of race conditions involving simultaneous mappings of a page easier. A reference count is kept in the page_cgroup to deal with cases where a page might be unmapped from the RSS of all tasks, but still lives in the page cache. Credits go to Vaidyanathan Srinivasan for helping with reference counting work of the page cgroup. Almost all of the page cache accounting code has help from Vaidyanathan Srinivasan. [hugh@veritas.com: fix swapoff breakage] [akpm@linux-foundation.org: fix locking] Signed-off-by: Vaidyanathan Srinivasan <svaidy@linux.vnet.ibm.com> Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com> Cc: Pavel Emelianov <xemul@openvz.org> Cc: Paul Menage <menage@google.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Cc: Kirill Korotaev <dev@sw.ru> Cc: Herbert Poetzl <herbert@13thfloor.at> Cc: David Rientjes <rientjes@google.com> Cc: <Valdis.Kletnieks@vt.edu> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-07 08:13:53 +00:00
#include <linux/bit_spinlock.h>
#include <linux/rcupdate.h>
#include <linux/limits.h>
#include <linux/export.h>
memcg: mem+swap controller core This patch implements per cgroup limit for usage of memory+swap. However there are SwapCache, double counting of swap-cache and swap-entry is avoided. Mem+Swap controller works as following. - memory usage is limited by memory.limit_in_bytes. - memory + swap usage is limited by memory.memsw_limit_in_bytes. This has following benefits. - A user can limit total resource usage of mem+swap. Without this, because memory resource controller doesn't take care of usage of swap, a process can exhaust all the swap (by memory leak.) We can avoid this case. And Swap is shared resource but it cannot be reclaimed (goes back to memory) until it's used. This characteristic can be trouble when the memory is divided into some parts by cpuset or memcg. Assume group A and group B. After some application executes, the system can be.. Group A -- very large free memory space but occupy 99% of swap. Group B -- under memory shortage but cannot use swap...it's nearly full. Ability to set appropriate swap limit for each group is required. Maybe someone wonder "why not swap but mem+swap ?" - The global LRU(kswapd) can swap out arbitrary pages. Swap-out means to move account from memory to swap...there is no change in usage of mem+swap. In other words, when we want to limit the usage of swap without affecting global LRU, mem+swap limit is better than just limiting swap. Accounting target information is stored in swap_cgroup which is per swap entry record. Charge is done as following. map - charge page and memsw. unmap - uncharge page/memsw if not SwapCache. swap-out (__delete_from_swap_cache) - uncharge page - record mem_cgroup information to swap_cgroup. swap-in (do_swap_page) - charged as page and memsw. record in swap_cgroup is cleared. memsw accounting is decremented. swap-free (swap_free()) - if swap entry is freed, memsw is uncharged by PAGE_SIZE. There are people work under never-swap environments and consider swap as something bad. For such people, this mem+swap controller extension is just an overhead. This overhead is avoided by config or boot option. (see Kconfig. detail is not in this patch.) TODO: - maybe more optimization can be don in swap-in path. (but not very safe.) But we just do simple accounting at this stage. [nishimura@mxp.nes.nec.co.jp: make resize limit hold mutex] [hugh@veritas.com: memswap controller core swapcache fixes] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:08:00 +00:00
#include <linux/mutex.h>
#include <linux/rbtree.h>
#include <linux/slab.h>
#include <linux/swap.h>
#include <linux/swapops.h>
#include <linux/spinlock.h>
#include <linux/eventfd.h>
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
#include <linux/poll.h>
#include <linux/sort.h>
#include <linux/fs.h>
#include <linux/seq_file.h>
memcg: add memory.pressure_level events With this patch userland applications that want to maintain the interactivity/memory allocation cost can use the pressure level notifications. The levels are defined like this: The "low" level means that the system is reclaiming memory for new allocations. Monitoring this reclaiming activity might be useful for maintaining cache level. Upon notification, the program (typically "Activity Manager") might analyze vmstat and act in advance (i.e. prematurely shutdown unimportant services). The "medium" level means that the system is experiencing medium memory pressure, the system might be making swap, paging out active file caches, etc. Upon this event applications may decide to further analyze vmstat/zoneinfo/memcg or internal memory usage statistics and free any resources that can be easily reconstructed or re-read from a disk. The "critical" level means that the system is actively thrashing, it is about to out of memory (OOM) or even the in-kernel OOM killer is on its way to trigger. Applications should do whatever they can to help the system. It might be too late to consult with vmstat or any other statistics, so it's advisable to take an immediate action. The events are propagated upward until the event is handled, i.e. the events are not pass-through. Here is what this means: for example you have three cgroups: A->B->C. Now you set up an event listener on cgroups A, B and C, and suppose group C experiences some pressure. In this situation, only group C will receive the notification, i.e. groups A and B will not receive it. This is done to avoid excessive "broadcasting" of messages, which disturbs the system and which is especially bad if we are low on memory or thrashing. So, organize the cgroups wisely, or propagate the events manually (or, ask us to implement the pass-through events, explaining why would you need them.) Performance wise, the memory pressure notifications feature itself is lightweight and does not require much of bookkeeping, in contrast to the rest of memcg features. Unfortunately, as of current memcg implementation, pages accounting is an inseparable part and cannot be turned off. The good news is that there are some efforts[1] to improve the situation; plus, implementing the same, fully API-compatible[2] interface for CONFIG_MEMCG=n case (e.g. embedded) is also a viable option, so it will not require any changes on the userland side. [1] http://permalink.gmane.org/gmane.linux.kernel.cgroups/6291 [2] http://lkml.org/lkml/2013/2/21/454 [akpm@linux-foundation.org: coding-style fixes] [akpm@linux-foundation.org: fix CONFIG_CGROPUPS=n warnings] Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Acked-by: Kirill A. Shutemov <kirill@shutemov.name> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: Mel Gorman <mgorman@suse.de> Cc: Glauber Costa <glommer@parallels.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Luiz Capitulino <lcapitulino@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Leonid Moiseichuk <leonid.moiseichuk@nokia.com> Cc: KOSAKI Motohiro <kosaki.motohiro@gmail.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> Cc: John Stultz <john.stultz@linaro.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-04-29 22:08:31 +00:00
#include <linux/vmpressure.h>
#include <linux/mm_inline.h>
#include <linux/swap_cgroup.h>
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
#include <linux/cpu.h>
#include <linux/oom.h>
#include <linux/lockdep.h>
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
#include <linux/file.h>
memcg: punt high overage reclaim to return-to-userland path Currently, try_charge() tries to reclaim memory synchronously when the high limit is breached; however, if the allocation doesn't have __GFP_WAIT, synchronous reclaim is skipped. If a process performs only speculative allocations, it can blow way past the high limit. This is actually easily reproducible by simply doing "find /". slab/slub allocator tries speculative allocations first, so as long as there's memory which can be consumed without blocking, it can keep allocating memory regardless of the high limit. This patch makes try_charge() always punt the over-high reclaim to the return-to-userland path. If try_charge() detects that high limit is breached, it adds the overage to current->memcg_nr_pages_over_high and schedules execution of mem_cgroup_handle_over_high() which performs synchronous reclaim from the return-to-userland path. As long as kernel doesn't have a run-away allocation spree, this should provide enough protection while making kmemcg behave more consistently. It also has the following benefits. - All over-high reclaims can use GFP_KERNEL regardless of the specific gfp mask in use, e.g. GFP_NOFS, when the limit was breached. - It copes with prio inversion. Previously, a low-prio task with small memory.high might perform over-high reclaim with a bunch of locks held. If a higher prio task needed any of these locks, it would have to wait until the low prio task finished reclaim and released the locks. By handing over-high reclaim to the task exit path this issue can be avoided. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Michal Hocko <mhocko@kernel.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:46:11 +00:00
#include <linux/tracehook.h>
mm, memcg: throttle allocators when failing reclaim over memory.high We're trying to use memory.high to limit workloads, but have found that containment can frequently fail completely and cause OOM situations outside of the cgroup. This happens especially with swap space -- either when none is configured, or swap is full. These failures often also don't have enough warning to allow one to react, whether for a human or for a daemon monitoring PSI. Here is output from a simple program showing how long it takes in usec (column 2) to allocate a megabyte of anonymous memory (column 1) when a cgroup is already beyond its memory high setting, and no swap is available: [root@ktst ~]# systemd-run -p MemoryHigh=100M -p MemorySwapMax=1 \ > --wait -t timeout 300 /root/mdf [...] 95 1035 96 1038 97 1000 98 1036 99 1048 100 1590 101 1968 102 1776 103 1863 104 1757 105 1921 106 1893 107 1760 108 1748 109 1843 110 1716 111 1924 112 1776 113 1831 114 1766 115 1836 116 1588 117 1912 118 1802 119 1857 120 1731 [...] [System OOM in 2-3 seconds] The delay does go up extremely marginally past the 100MB memory.high threshold, as now we spend time scanning before returning to usermode, but it's nowhere near enough to contain growth. It also doesn't get worse the more pages you have, since it only considers nr_pages. The current situation goes against both the expectations of users of memory.high, and our intentions as cgroup v2 developers. In cgroup-v2.txt, we claim that we will throttle and only under "extreme conditions" will memory.high protection be breached. Likewise, cgroup v2 users generally also expect that memory.high should throttle workloads as they exceed their high threshold. However, as seen above, this isn't always how it works in practice -- even on banal setups like those with no swap, or where swap has become exhausted, we can end up with memory.high being breached and us having no weapons left in our arsenal to combat runaway growth with, since reclaim is futile. It's also hard for system monitoring software or users to tell how bad the situation is, as "high" events for the memcg may in some cases be benign, and in others be catastrophic. The current status quo is that we fail containment in a way that doesn't provide any advance warning that things are about to go horribly wrong (for example, we are about to invoke the kernel OOM killer). This patch introduces explicit throttling when reclaim is failing to keep memcg size contained at the memory.high setting. It does so by applying an exponential delay curve derived from the memcg's overage compared to memory.high. In the normal case where the memcg is either below or only marginally over its memory.high setting, no throttling will be performed. This composes well with system health monitoring and remediation, as these allocator delays are factored into PSI's memory pressure calculations. This both creates a mechanism system administrators or applications consuming the PSI interface to trivially see that the memcg in question is struggling and use that to make more reasonable decisions, and permits them enough time to act. Either of these can act with significantly more nuance than that we can provide using the system OOM killer. This is a similar idea to memory.oom_control in cgroup v1 which would put the cgroup to sleep if the threshold was violated, but it's also significantly improved as it results in visible memory pressure, and also doesn't schedule indefinitely, which previously made tracing and other introspection difficult (ie. it's clamped at 2*HZ per allocation through MEMCG_MAX_HIGH_DELAY_JIFFIES). Contrast the previous results with a kernel with this patch: [root@ktst ~]# systemd-run -p MemoryHigh=100M -p MemorySwapMax=1 \ > --wait -t timeout 300 /root/mdf [...] 95 1002 96 1000 97 1002 98 1003 99 1000 100 1043 101 84724 102 330628 103 610511 104 1016265 105 1503969 106 2391692 107 2872061 108 3248003 109 4791904 110 5759832 111 6912509 112 8127818 113 9472203 114 12287622 115 12480079 116 14144008 117 15808029 118 16384500 119 16383242 120 16384979 [...] As you can see, in the normal case, memory allocation takes around 1000 usec. However, as we exceed our memory.high, things start to increase exponentially, but fairly leniently at first. Our first megabyte over memory.high takes us 0.16 seconds, then the next is 0.46 seconds, then the next is almost an entire second. This gets worse until we reach our eventual 2*HZ clamp per batch, resulting in 16 seconds per megabyte. However, this is still making forward progress, so permits tracing or further analysis with programs like GDB. We use an exponential curve for our delay penalty for a few reasons: 1. We run mem_cgroup_handle_over_high to potentially do reclaim after we've already performed allocations, which means that temporarily going over memory.high by a small amount may be perfectly legitimate, even for compliant workloads. We don't want to unduly penalise such cases. 2. An exponential curve (as opposed to a static or linear delay) allows ramping up memory pressure stats more gradually, which can be useful to work out that you have set memory.high too low, without destroying application performance entirely. This patch expands on earlier work by Johannes Weiner. Thanks! [akpm@linux-foundation.org: fix max() warning] [akpm@linux-foundation.org: fix __udivdi3 ref on 32-bit] [akpm@linux-foundation.org: fix it even more] [chris@chrisdown.name: fix 64-bit divide even more] Link: http://lkml.kernel.org/r/20190723180700.GA29459@chrisdown.name Signed-off-by: Chris Down <chris@chrisdown.name> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Nathan Chancellor <natechancellor@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-09-23 22:34:55 +00:00
#include <linux/psi.h>
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
#include <linux/seq_buf.h>
memcg: synchronized LRU A big patch for changing memcg's LRU semantics. Now, - page_cgroup is linked to mem_cgroup's its own LRU (per zone). - LRU of page_cgroup is not synchronous with global LRU. - page and page_cgroup is one-to-one and statically allocated. - To find page_cgroup is on what LRU, you have to check pc->mem_cgroup as - lru = page_cgroup_zoneinfo(pc, nid_of_pc, zid_of_pc); - SwapCache is handled. And, when we handle LRU list of page_cgroup, we do following. pc = lookup_page_cgroup(page); lock_page_cgroup(pc); .....................(1) mz = page_cgroup_zoneinfo(pc); spin_lock(&mz->lru_lock); .....add to LRU spin_unlock(&mz->lru_lock); unlock_page_cgroup(pc); But (1) is spin_lock and we have to be afraid of dead-lock with zone->lru_lock. So, trylock() is used at (1), now. Without (1), we can't trust "mz" is correct. This is a trial to remove this dirty nesting of locks. This patch changes mz->lru_lock to be zone->lru_lock. Then, above sequence will be written as spin_lock(&zone->lru_lock); # in vmscan.c or swap.c via global LRU mem_cgroup_add/remove/etc_lru() { pc = lookup_page_cgroup(page); mz = page_cgroup_zoneinfo(pc); if (PageCgroupUsed(pc)) { ....add to LRU } spin_lock(&zone->lru_lock); # in vmscan.c or swap.c via global LRU This is much simpler. (*) We're safe even if we don't take lock_page_cgroup(pc). Because.. 1. When pc->mem_cgroup can be modified. - at charge. - at account_move(). 2. at charge the PCG_USED bit is not set before pc->mem_cgroup is fixed. 3. at account_move() the page is isolated and not on LRU. Pros. - easy for maintenance. - memcg can make use of laziness of pagevec. - we don't have to duplicated LRU/Active/Unevictable bit in page_cgroup. - LRU status of memcg will be synchronized with global LRU's one. - # of locks are reduced. - account_move() is simplified very much. Cons. - may increase cost of LRU rotation. (no impact if memcg is not configured.) Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:08:01 +00:00
#include "internal.h"
#include <net/sock.h>
#include <net/ip.h>
#include "slab.h"
#include <linux/uaccess.h>
#include <trace/events/vmscan.h>
cgroup: clean up cgroup_subsys names and initialization cgroup_subsys is a bit messier than it needs to be. * The name of a subsys can be different from its internal identifier defined in cgroup_subsys.h. Most subsystems use the matching name but three - cpu, memory and perf_event - use different ones. * cgroup_subsys_id enums are postfixed with _subsys_id and each cgroup_subsys is postfixed with _subsys. cgroup.h is widely included throughout various subsystems, it doesn't and shouldn't have claim on such generic names which don't have any qualifier indicating that they belong to cgroup. * cgroup_subsys->subsys_id should always equal the matching cgroup_subsys_id enum; however, we require each controller to initialize it and then BUG if they don't match, which is a bit silly. This patch cleans up cgroup_subsys names and initialization by doing the followings. * cgroup_subsys_id enums are now postfixed with _cgrp_id, and each cgroup_subsys with _cgrp_subsys. * With the above, renaming subsys identifiers to match the userland visible names doesn't cause any naming conflicts. All non-matching identifiers are renamed to match the official names. cpu_cgroup -> cpu mem_cgroup -> memory perf -> perf_event * controllers no longer need to initialize ->subsys_id and ->name. They're generated in cgroup core and set automatically during boot. * Redundant cgroup_subsys declarations removed. * While updating BUG_ON()s in cgroup_init_early(), convert them to WARN()s. BUGging that early during boot is stupid - the kernel can't print anything, even through serial console and the trap handler doesn't even link stack frame properly for back-tracing. This patch doesn't introduce any behavior changes. v2: Rebased on top of fe1217c4f3f7 ("net: net_cls: move cgroupfs classid handling into core"). Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Neil Horman <nhorman@tuxdriver.com> Acked-by: "David S. Miller" <davem@davemloft.net> Acked-by: "Rafael J. Wysocki" <rjw@rjwysocki.net> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Peter Zijlstra <peterz@infradead.org> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Ingo Molnar <mingo@redhat.com> Acked-by: Li Zefan <lizefan@huawei.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Serge E. Hallyn <serue@us.ibm.com> Cc: Vivek Goyal <vgoyal@redhat.com> Cc: Thomas Graf <tgraf@suug.ch>
2014-02-08 15:36:58 +00:00
struct cgroup_subsys memory_cgrp_subsys __read_mostly;
EXPORT_SYMBOL(memory_cgrp_subsys);
struct mem_cgroup *root_mem_cgroup __read_mostly;
#define MEM_CGROUP_RECLAIM_RETRIES 5
/* Socket memory accounting disabled? */
static bool cgroup_memory_nosocket;
/* Kernel memory accounting disabled? */
static bool cgroup_memory_nokmem;
/* Whether the swap controller is active */
#ifdef CONFIG_MEMCG_SWAP
bool cgroup_memory_noswap __read_mostly;
#else
#define cgroup_memory_noswap 1
mm: memcontrol: make swap tracking an integral part of memory control Without swap page tracking, users that are otherwise memory controlled can easily escape their containment and allocate significant amounts of memory that they're not being charged for. That's because swap does readahead, but without the cgroup records of who owned the page at swapout, readahead pages don't get charged until somebody actually faults them into their page table and we can identify an owner task. This can be maliciously exploited with MADV_WILLNEED, which triggers arbitrary readahead allocations without charging the pages. Make swap swap page tracking an integral part of memcg and remove the Kconfig options. In the first place, it was only made configurable to allow users to save some memory. But the overhead of tracking cgroup ownership per swap page is minimal - 2 byte per page, or 512k per 1G of swap, or 0.04%. Saving that at the expense of broken containment semantics is not something we should present as a coequal option. The swapaccount=0 boot option will continue to exist, and it will eliminate the page_counter overhead and hide the swap control files, but it won't disable swap slot ownership tracking. This patch makes sure we always have the cgroup records at swapin time; the next patch will fix the actual bug by charging readahead swap pages at swapin time rather than at fault time. v2: fix double swap charge bug in cgroup1/cgroup2 code gating [hannes@cmpxchg.org: fix crash with cgroup_disable=memory] Link: http://lkml.kernel.org/r/20200521215855.GB815153@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Alex Shi <alex.shi@linux.alibaba.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Roman Gushchin <guro@fb.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Naresh Kamboju <naresh.kamboju@linaro.org> Link: http://lkml.kernel.org/r/20200508183105.225460-16-hannes@cmpxchg.org Debugged-by: Hugh Dickins <hughd@google.com> Debugged-by: Michal Hocko <mhocko@kernel.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:02:14 +00:00
#endif
writeback, memcg: Implement foreign dirty flushing There's an inherent mismatch between memcg and writeback. The former trackes ownership per-page while the latter per-inode. This was a deliberate design decision because honoring per-page ownership in the writeback path is complicated, may lead to higher CPU and IO overheads and deemed unnecessary given that write-sharing an inode across different cgroups isn't a common use-case. Combined with inode majority-writer ownership switching, this works well enough in most cases but there are some pathological cases. For example, let's say there are two cgroups A and B which keep writing to different but confined parts of the same inode. B owns the inode and A's memory is limited far below B's. A's dirty ratio can rise enough to trigger balance_dirty_pages() sleeps but B's can be low enough to avoid triggering background writeback. A will be slowed down without a way to make writeback of the dirty pages happen. This patch implements foreign dirty recording and foreign mechanism so that when a memcg encounters a condition as above it can trigger flushes on bdi_writebacks which can clean its pages. Please see the comment on top of mem_cgroup_track_foreign_dirty_slowpath() for details. A reproducer follows. write-range.c:: #include <stdio.h> #include <stdlib.h> #include <unistd.h> #include <fcntl.h> #include <sys/types.h> static const char *usage = "write-range FILE START SIZE\n"; int main(int argc, char **argv) { int fd; unsigned long start, size, end, pos; char *endp; char buf[4096]; if (argc < 4) { fprintf(stderr, usage); return 1; } fd = open(argv[1], O_WRONLY); if (fd < 0) { perror("open"); return 1; } start = strtoul(argv[2], &endp, 0); if (*endp != '\0') { fprintf(stderr, usage); return 1; } size = strtoul(argv[3], &endp, 0); if (*endp != '\0') { fprintf(stderr, usage); return 1; } end = start + size; while (1) { for (pos = start; pos < end; ) { long bread, bwritten = 0; if (lseek(fd, pos, SEEK_SET) < 0) { perror("lseek"); return 1; } bread = read(0, buf, sizeof(buf) < end - pos ? sizeof(buf) : end - pos); if (bread < 0) { perror("read"); return 1; } if (bread == 0) return 0; while (bwritten < bread) { long this; this = write(fd, buf + bwritten, bread - bwritten); if (this < 0) { perror("write"); return 1; } bwritten += this; pos += bwritten; } } } } repro.sh:: #!/bin/bash set -e set -x sysctl -w vm.dirty_expire_centisecs=300000 sysctl -w vm.dirty_writeback_centisecs=300000 sysctl -w vm.dirtytime_expire_seconds=300000 echo 3 > /proc/sys/vm/drop_caches TEST=/sys/fs/cgroup/test A=$TEST/A B=$TEST/B mkdir -p $A $B echo "+memory +io" > $TEST/cgroup.subtree_control echo $((1<<30)) > $A/memory.high echo $((32<<30)) > $B/memory.high rm -f testfile touch testfile fallocate -l 4G testfile echo "Starting B" (echo $BASHPID > $B/cgroup.procs pv -q --rate-limit 70M < /dev/urandom | ./write-range testfile $((2<<30)) $((2<<30))) & echo "Waiting 10s to ensure B claims the testfile inode" sleep 5 sync sleep 5 sync echo "Starting A" (echo $BASHPID > $A/cgroup.procs pv < /dev/urandom | ./write-range testfile 0 $((2<<30))) v2: Added comments explaining why the specific intervals are being used. v3: Use 0 @nr when calling cgroup_writeback_by_id() to use best-effort flushing while avoding possible livelocks. v4: Use get_jiffies_64() and time_before/after64() instead of raw jiffies_64 and arthimetic comparisons as suggested by Jan. Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-08-26 16:06:56 +00:00
#ifdef CONFIG_CGROUP_WRITEBACK
static DECLARE_WAIT_QUEUE_HEAD(memcg_cgwb_frn_waitq);
#endif
/* Whether legacy memory+swap accounting is active */
static bool do_memsw_account(void)
{
return !cgroup_subsys_on_dfl(memory_cgrp_subsys) && !cgroup_memory_noswap;
}
#define THRESHOLDS_EVENTS_TARGET 128
#define SOFTLIMIT_EVENTS_TARGET 1024
/*
* Cgroups above their limits are maintained in a RB-Tree, independent of
* their hierarchy representation
*/
struct mem_cgroup_tree_per_node {
struct rb_root rb_root;
struct rb_node *rb_rightmost;
spinlock_t lock;
};
struct mem_cgroup_tree {
struct mem_cgroup_tree_per_node *rb_tree_per_node[MAX_NUMNODES];
};
static struct mem_cgroup_tree soft_limit_tree __read_mostly;
/* for OOM */
struct mem_cgroup_eventfd_list {
struct list_head list;
struct eventfd_ctx *eventfd;
};
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
/*
* cgroup_event represents events which userspace want to receive.
*/
struct mem_cgroup_event {
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
/*
* memcg which the event belongs to.
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
*/
struct mem_cgroup *memcg;
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
/*
* eventfd to signal userspace about the event.
*/
struct eventfd_ctx *eventfd;
/*
* Each of these stored in a list by the cgroup.
*/
struct list_head list;
/*
* register_event() callback will be used to add new userspace
* waiter for changes related to this event. Use eventfd_signal()
* on eventfd to send notification to userspace.
*/
int (*register_event)(struct mem_cgroup *memcg,
struct eventfd_ctx *eventfd, const char *args);
/*
* unregister_event() callback will be called when userspace closes
* the eventfd or on cgroup removing. This callback must be set,
* if you want provide notification functionality.
*/
void (*unregister_event)(struct mem_cgroup *memcg,
struct eventfd_ctx *eventfd);
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
/*
* All fields below needed to unregister event when
* userspace closes eventfd.
*/
poll_table pt;
wait_queue_head_t *wqh;
wait_queue_entry_t wait;
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
struct work_struct remove;
};
static void mem_cgroup_threshold(struct mem_cgroup *memcg);
static void mem_cgroup_oom_notify(struct mem_cgroup *memcg);
/* Stuffs for move charges at task migration. */
/*
* Types of charges to be moved.
*/
#define MOVE_ANON 0x1U
#define MOVE_FILE 0x2U
#define MOVE_MASK (MOVE_ANON | MOVE_FILE)
/* "mc" and its members are protected by cgroup_mutex */
static struct move_charge_struct {
memcg: avoid deadlock between move charge and try_charge() __mem_cgroup_try_charge() can be called under down_write(&mmap_sem)(e.g. mlock does it). This means it can cause deadlock if it races with move charge: Ex.1) move charge | try charge --------------------------------------+------------------------------ mem_cgroup_can_attach() | down_write(&mmap_sem) mc.moving_task = current | .. mem_cgroup_precharge_mc() | __mem_cgroup_try_charge() mem_cgroup_count_precharge() | prepare_to_wait() down_read(&mmap_sem) | if (mc.moving_task) -> cannot aquire the lock | -> true | schedule() Ex.2) move charge | try charge --------------------------------------+------------------------------ mem_cgroup_can_attach() | mc.moving_task = current | mem_cgroup_precharge_mc() | mem_cgroup_count_precharge() | down_read(&mmap_sem) | .. | up_read(&mmap_sem) | | down_write(&mmap_sem) mem_cgroup_move_task() | .. mem_cgroup_move_charge() | __mem_cgroup_try_charge() down_read(&mmap_sem) | prepare_to_wait() -> cannot aquire the lock | if (mc.moving_task) | -> true | schedule() To avoid this deadlock, we do all the move charge works (both can_attach() and attach()) under one mmap_sem section. And after this patch, we set/clear mc.moving_task outside mc.lock, because we use the lock only to check mc.from/to. Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <balbir@linux.vnet.ibm.com> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: <stable@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-11-24 20:57:06 +00:00
spinlock_t lock; /* for from, to */
memcg: relocate charge moving from ->attach to ->post_attach Hello, So, this ended up a lot simpler than I originally expected. I tested it lightly and it seems to work fine. Petr, can you please test these two patches w/o the lru drain drop patch and see whether the problem is gone? Thanks. ------ 8< ------ If charge moving is used, memcg performs relabeling of the affected pages from its ->attach callback which is called under both cgroup_threadgroup_rwsem and thus can't create new kthreads. This is fragile as various operations may depend on workqueues making forward progress which relies on the ability to create new kthreads. There's no reason to perform charge moving from ->attach which is deep in the task migration path. Move it to ->post_attach which is called after the actual migration is finished and cgroup_threadgroup_rwsem is dropped. * move_charge_struct->mm is added and ->can_attach is now responsible for pinning and recording the target mm. mem_cgroup_clear_mc() is updated accordingly. This also simplifies mem_cgroup_move_task(). * mem_cgroup_move_task() is now called from ->post_attach instead of ->attach. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Debugged-and-tested-by: Petr Mladek <pmladek@suse.com> Reported-by: Cyril Hrubis <chrubis@suse.cz> Reported-by: Johannes Weiner <hannes@cmpxchg.org> Fixes: 1ed1328792ff ("sched, cgroup: replace signal_struct->group_rwsem with a global percpu_rwsem") Cc: <stable@vger.kernel.org> # 4.4+
2016-04-21 23:09:02 +00:00
struct mm_struct *mm;
struct mem_cgroup *from;
struct mem_cgroup *to;
unsigned long flags;
unsigned long precharge;
unsigned long moved_charge;
unsigned long moved_swap;
struct task_struct *moving_task; /* a task moving charges */
wait_queue_head_t waitq; /* a waitq for other context */
} mc = {
.lock = __SPIN_LOCK_UNLOCKED(mc.lock),
.waitq = __WAIT_QUEUE_HEAD_INITIALIZER(mc.waitq),
};
2009-09-23 22:56:39 +00:00
/*
* Maximum loops in mem_cgroup_hierarchical_reclaim(), used for soft
* limit reclaim to prevent infinite loops, if they ever occur.
*/
#define MEM_CGROUP_MAX_RECLAIM_LOOPS 100
#define MEM_CGROUP_MAX_SOFT_LIMIT_RECLAIM_LOOPS 2
2009-09-23 22:56:39 +00:00
enum charge_type {
MEM_CGROUP_CHARGE_TYPE_CACHE = 0,
MEM_CGROUP_CHARGE_TYPE_ANON,
memcg: handle swap caches SwapCache support for memory resource controller (memcg) Before mem+swap controller, memcg itself should handle SwapCache in proper way. This is cut-out from it. In current memcg, SwapCache is just leaked and the user can create tons of SwapCache. This is a leak of account and should be handled. SwapCache accounting is done as following. charge (anon) - charged when it's mapped. (because of readahead, charge at add_to_swap_cache() is not sane) uncharge (anon) - uncharged when it's dropped from swapcache and fully unmapped. means it's not uncharged at unmap. Note: delete from swap cache at swap-in is done after rmap information is established. charge (shmem) - charged at swap-in. this prevents charge at add_to_page_cache(). uncharge (shmem) - uncharged when it's dropped from swapcache and not on shmem's radix-tree. at migration, check against 'old page' is modified to handle shmem. Comparing to the old version discussed (and caused troubles), we have advantages of - PCG_USED bit. - simple migrating handling. So, situation is much easier than several months ago, maybe. [hugh@veritas.com: memcg: handle swap caches build fix] Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Tested-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Hugh Dickins <hugh@veritas.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:07:56 +00:00
MEM_CGROUP_CHARGE_TYPE_SWAPOUT, /* for accounting swapcache */
MEM_CGROUP_CHARGE_TYPE_DROP, /* a page was unused swap cache */
NR_CHARGE_TYPE,
};
memcg: mem+swap controller core This patch implements per cgroup limit for usage of memory+swap. However there are SwapCache, double counting of swap-cache and swap-entry is avoided. Mem+Swap controller works as following. - memory usage is limited by memory.limit_in_bytes. - memory + swap usage is limited by memory.memsw_limit_in_bytes. This has following benefits. - A user can limit total resource usage of mem+swap. Without this, because memory resource controller doesn't take care of usage of swap, a process can exhaust all the swap (by memory leak.) We can avoid this case. And Swap is shared resource but it cannot be reclaimed (goes back to memory) until it's used. This characteristic can be trouble when the memory is divided into some parts by cpuset or memcg. Assume group A and group B. After some application executes, the system can be.. Group A -- very large free memory space but occupy 99% of swap. Group B -- under memory shortage but cannot use swap...it's nearly full. Ability to set appropriate swap limit for each group is required. Maybe someone wonder "why not swap but mem+swap ?" - The global LRU(kswapd) can swap out arbitrary pages. Swap-out means to move account from memory to swap...there is no change in usage of mem+swap. In other words, when we want to limit the usage of swap without affecting global LRU, mem+swap limit is better than just limiting swap. Accounting target information is stored in swap_cgroup which is per swap entry record. Charge is done as following. map - charge page and memsw. unmap - uncharge page/memsw if not SwapCache. swap-out (__delete_from_swap_cache) - uncharge page - record mem_cgroup information to swap_cgroup. swap-in (do_swap_page) - charged as page and memsw. record in swap_cgroup is cleared. memsw accounting is decremented. swap-free (swap_free()) - if swap entry is freed, memsw is uncharged by PAGE_SIZE. There are people work under never-swap environments and consider swap as something bad. For such people, this mem+swap controller extension is just an overhead. This overhead is avoided by config or boot option. (see Kconfig. detail is not in this patch.) TODO: - maybe more optimization can be don in swap-in path. (but not very safe.) But we just do simple accounting at this stage. [nishimura@mxp.nes.nec.co.jp: make resize limit hold mutex] [hugh@veritas.com: memswap controller core swapcache fixes] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:08:00 +00:00
/* for encoding cft->private value on file */
enum res_type {
_MEM,
_MEMSWAP,
_OOM_TYPE,
memcg: kmem accounting basic infrastructure Add the basic infrastructure for the accounting of kernel memory. To control that, the following files are created: * memory.kmem.usage_in_bytes * memory.kmem.limit_in_bytes * memory.kmem.failcnt * memory.kmem.max_usage_in_bytes They have the same meaning of their user memory counterparts. They reflect the state of the "kmem" res_counter. Per cgroup kmem memory accounting is not enabled until a limit is set for the group. Once the limit is set the accounting cannot be disabled for that group. This means that after the patch is applied, no behavioral changes exists for whoever is still using memcg to control their memory usage, until memory.kmem.limit_in_bytes is set for the first time. We always account to both user and kernel resource_counters. This effectively means that an independent kernel limit is in place when the limit is set to a lower value than the user memory. A equal or higher value means that the user limit will always hit first, meaning that kmem is effectively unlimited. People who want to track kernel memory but not limit it, can set this limit to a very high number (like RESOURCE_MAX - 1page - that no one will ever hit, or equal to the user memory) [akpm@linux-foundation.org: MEMCG_MMEM only works with slab and slub] Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:47 +00:00
_KMEM,
_TCP,
};
#define MEMFILE_PRIVATE(x, val) ((x) << 16 | (val))
#define MEMFILE_TYPE(val) ((val) >> 16 & 0xffff)
memcg: mem+swap controller core This patch implements per cgroup limit for usage of memory+swap. However there are SwapCache, double counting of swap-cache and swap-entry is avoided. Mem+Swap controller works as following. - memory usage is limited by memory.limit_in_bytes. - memory + swap usage is limited by memory.memsw_limit_in_bytes. This has following benefits. - A user can limit total resource usage of mem+swap. Without this, because memory resource controller doesn't take care of usage of swap, a process can exhaust all the swap (by memory leak.) We can avoid this case. And Swap is shared resource but it cannot be reclaimed (goes back to memory) until it's used. This characteristic can be trouble when the memory is divided into some parts by cpuset or memcg. Assume group A and group B. After some application executes, the system can be.. Group A -- very large free memory space but occupy 99% of swap. Group B -- under memory shortage but cannot use swap...it's nearly full. Ability to set appropriate swap limit for each group is required. Maybe someone wonder "why not swap but mem+swap ?" - The global LRU(kswapd) can swap out arbitrary pages. Swap-out means to move account from memory to swap...there is no change in usage of mem+swap. In other words, when we want to limit the usage of swap without affecting global LRU, mem+swap limit is better than just limiting swap. Accounting target information is stored in swap_cgroup which is per swap entry record. Charge is done as following. map - charge page and memsw. unmap - uncharge page/memsw if not SwapCache. swap-out (__delete_from_swap_cache) - uncharge page - record mem_cgroup information to swap_cgroup. swap-in (do_swap_page) - charged as page and memsw. record in swap_cgroup is cleared. memsw accounting is decremented. swap-free (swap_free()) - if swap entry is freed, memsw is uncharged by PAGE_SIZE. There are people work under never-swap environments and consider swap as something bad. For such people, this mem+swap controller extension is just an overhead. This overhead is avoided by config or boot option. (see Kconfig. detail is not in this patch.) TODO: - maybe more optimization can be don in swap-in path. (but not very safe.) But we just do simple accounting at this stage. [nishimura@mxp.nes.nec.co.jp: make resize limit hold mutex] [hugh@veritas.com: memswap controller core swapcache fixes] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:08:00 +00:00
#define MEMFILE_ATTR(val) ((val) & 0xffff)
/* Used for OOM nofiier */
#define OOM_CONTROL (0)
memcg: mem+swap controller core This patch implements per cgroup limit for usage of memory+swap. However there are SwapCache, double counting of swap-cache and swap-entry is avoided. Mem+Swap controller works as following. - memory usage is limited by memory.limit_in_bytes. - memory + swap usage is limited by memory.memsw_limit_in_bytes. This has following benefits. - A user can limit total resource usage of mem+swap. Without this, because memory resource controller doesn't take care of usage of swap, a process can exhaust all the swap (by memory leak.) We can avoid this case. And Swap is shared resource but it cannot be reclaimed (goes back to memory) until it's used. This characteristic can be trouble when the memory is divided into some parts by cpuset or memcg. Assume group A and group B. After some application executes, the system can be.. Group A -- very large free memory space but occupy 99% of swap. Group B -- under memory shortage but cannot use swap...it's nearly full. Ability to set appropriate swap limit for each group is required. Maybe someone wonder "why not swap but mem+swap ?" - The global LRU(kswapd) can swap out arbitrary pages. Swap-out means to move account from memory to swap...there is no change in usage of mem+swap. In other words, when we want to limit the usage of swap without affecting global LRU, mem+swap limit is better than just limiting swap. Accounting target information is stored in swap_cgroup which is per swap entry record. Charge is done as following. map - charge page and memsw. unmap - uncharge page/memsw if not SwapCache. swap-out (__delete_from_swap_cache) - uncharge page - record mem_cgroup information to swap_cgroup. swap-in (do_swap_page) - charged as page and memsw. record in swap_cgroup is cleared. memsw accounting is decremented. swap-free (swap_free()) - if swap entry is freed, memsw is uncharged by PAGE_SIZE. There are people work under never-swap environments and consider swap as something bad. For such people, this mem+swap controller extension is just an overhead. This overhead is avoided by config or boot option. (see Kconfig. detail is not in this patch.) TODO: - maybe more optimization can be don in swap-in path. (but not very safe.) But we just do simple accounting at this stage. [nishimura@mxp.nes.nec.co.jp: make resize limit hold mutex] [hugh@veritas.com: memswap controller core swapcache fixes] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:08:00 +00:00
/*
* Iteration constructs for visiting all cgroups (under a tree). If
* loops are exited prematurely (break), mem_cgroup_iter_break() must
* be used for reference counting.
*/
#define for_each_mem_cgroup_tree(iter, root) \
for (iter = mem_cgroup_iter(root, NULL, NULL); \
iter != NULL; \
iter = mem_cgroup_iter(root, iter, NULL))
#define for_each_mem_cgroup(iter) \
for (iter = mem_cgroup_iter(NULL, NULL, NULL); \
iter != NULL; \
iter = mem_cgroup_iter(NULL, iter, NULL))
memcg: killed threads should not invoke memcg OOM killer If a memory cgroup contains a single process with many threads (including different process group sharing the mm) then it is possible to trigger a race when the oom killer complains that there are no oom elible tasks and complain into the log which is both annoying and confusing because there is no actual problem. The race looks as follows: P1 oom_reaper P2 try_charge try_charge mem_cgroup_out_of_memory mutex_lock(oom_lock) out_of_memory oom_kill_process(P1,P2) wake_oom_reaper mutex_unlock(oom_lock) oom_reap_task mutex_lock(oom_lock) select_bad_process # no victim The problem is more visible with many threads. Fix this by checking for fatal_signal_pending from mem_cgroup_out_of_memory when the oom_lock is already held. The oom bypass is safe because we do the same early in the try_charge path already. The situation migh have changed in the mean time. It should be safe to check for fatal_signal_pending and tsk_is_oom_victim but for a better code readability abstract the current charge bypass condition into should_force_charge and reuse it from that path. " Link: http://lkml.kernel.org/r/01370f70-e1f6-ebe4-b95e-0df21a0bc15e@i-love.sakura.ne.jp Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: David Rientjes <rientjes@google.com> Cc: Kirill Tkhai <ktkhai@virtuozzo.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-03-05 23:46:47 +00:00
static inline bool should_force_charge(void)
{
return tsk_is_oom_victim(current) || fatal_signal_pending(current) ||
(current->flags & PF_EXITING);
}
memcg: add memory.pressure_level events With this patch userland applications that want to maintain the interactivity/memory allocation cost can use the pressure level notifications. The levels are defined like this: The "low" level means that the system is reclaiming memory for new allocations. Monitoring this reclaiming activity might be useful for maintaining cache level. Upon notification, the program (typically "Activity Manager") might analyze vmstat and act in advance (i.e. prematurely shutdown unimportant services). The "medium" level means that the system is experiencing medium memory pressure, the system might be making swap, paging out active file caches, etc. Upon this event applications may decide to further analyze vmstat/zoneinfo/memcg or internal memory usage statistics and free any resources that can be easily reconstructed or re-read from a disk. The "critical" level means that the system is actively thrashing, it is about to out of memory (OOM) or even the in-kernel OOM killer is on its way to trigger. Applications should do whatever they can to help the system. It might be too late to consult with vmstat or any other statistics, so it's advisable to take an immediate action. The events are propagated upward until the event is handled, i.e. the events are not pass-through. Here is what this means: for example you have three cgroups: A->B->C. Now you set up an event listener on cgroups A, B and C, and suppose group C experiences some pressure. In this situation, only group C will receive the notification, i.e. groups A and B will not receive it. This is done to avoid excessive "broadcasting" of messages, which disturbs the system and which is especially bad if we are low on memory or thrashing. So, organize the cgroups wisely, or propagate the events manually (or, ask us to implement the pass-through events, explaining why would you need them.) Performance wise, the memory pressure notifications feature itself is lightweight and does not require much of bookkeeping, in contrast to the rest of memcg features. Unfortunately, as of current memcg implementation, pages accounting is an inseparable part and cannot be turned off. The good news is that there are some efforts[1] to improve the situation; plus, implementing the same, fully API-compatible[2] interface for CONFIG_MEMCG=n case (e.g. embedded) is also a viable option, so it will not require any changes on the userland side. [1] http://permalink.gmane.org/gmane.linux.kernel.cgroups/6291 [2] http://lkml.org/lkml/2013/2/21/454 [akpm@linux-foundation.org: coding-style fixes] [akpm@linux-foundation.org: fix CONFIG_CGROPUPS=n warnings] Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Acked-by: Kirill A. Shutemov <kirill@shutemov.name> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: Mel Gorman <mgorman@suse.de> Cc: Glauber Costa <glommer@parallels.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Luiz Capitulino <lcapitulino@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Leonid Moiseichuk <leonid.moiseichuk@nokia.com> Cc: KOSAKI Motohiro <kosaki.motohiro@gmail.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> Cc: John Stultz <john.stultz@linaro.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-04-29 22:08:31 +00:00
/* Some nice accessors for the vmpressure. */
struct vmpressure *memcg_to_vmpressure(struct mem_cgroup *memcg)
{
if (!memcg)
memcg = root_mem_cgroup;
return &memcg->vmpressure;
}
struct cgroup_subsys_state *vmpressure_to_css(struct vmpressure *vmpr)
{
return &container_of(vmpr, struct mem_cgroup, vmpressure)->css;
}
mm: introduce CONFIG_MEMCG_KMEM as combination of CONFIG_MEMCG && !CONFIG_SLOB Introduce new config option, which is used to replace repeating CONFIG_MEMCG && !CONFIG_SLOB pattern. Next patches add a little more memcg+kmem related code, so let's keep the defines more clearly. Link: http://lkml.kernel.org/r/153063053670.1818.15013136946600481138.stgit@localhost.localdomain Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Tested-by: Shakeel Butt <shakeelb@google.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Guenter Roeck <linux@roeck-us.net> Cc: "Huang, Ying" <ying.huang@intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Josef Bacik <jbacik@fb.com> Cc: Li RongQing <lirongqing@baidu.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Matthias Kaehlcke <mka@chromium.org> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Michal Hocko <mhocko@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Philippe Ombredanne <pombredanne@nexb.com> Cc: Roman Gushchin <guro@fb.com> Cc: Sahitya Tummala <stummala@codeaurora.org> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Waiman Long <longman@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:25 +00:00
#ifdef CONFIG_MEMCG_KMEM
memcg: allocate memory for memcg caches whenever a new memcg appears Every cache that is considered a root cache (basically the "original" caches, tied to the root memcg/no-memcg) will have an array that should be large enough to store a cache pointer per each memcg in the system. Theoreticaly, this is as high as 1 << sizeof(css_id), which is currently in the 64k pointers range. Most of the time, we won't be using that much. What goes in this patch, is a simple scheme to dynamically allocate such an array, in order to minimize memory usage for memcg caches. Because we would also like to avoid allocations all the time, at least for now, the array will only grow. It will tend to be big enough to hold the maximum number of kmem-limited memcgs ever achieved. We'll allocate it to be a minimum of 64 kmem-limited memcgs. When we have more than that, we'll start doubling the size of this array every time the limit is reached. Because we are only considering kmem limited memcgs, a natural point for this to happen is when we write to the limit. At that point, we already have set_limit_mutex held, so that will become our natural synchronization mechanism. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:38 +00:00
/*
slab: embed memcg_cache_params to kmem_cache Currently, kmem_cache stores a pointer to struct memcg_cache_params instead of embedding it. The rationale is to save memory when kmem accounting is disabled. However, the memcg_cache_params has shrivelled drastically since it was first introduced: * Initially: struct memcg_cache_params { bool is_root_cache; union { struct kmem_cache *memcg_caches[0]; struct { struct mem_cgroup *memcg; struct list_head list; struct kmem_cache *root_cache; bool dead; atomic_t nr_pages; struct work_struct destroy; }; }; }; * Now: struct memcg_cache_params { bool is_root_cache; union { struct { struct rcu_head rcu_head; struct kmem_cache *memcg_caches[0]; }; struct { struct mem_cgroup *memcg; struct kmem_cache *root_cache; }; }; }; So the memory saving does not seem to be a clear win anymore. OTOH, keeping a pointer to memcg_cache_params struct instead of embedding it results in touching one more cache line on kmem alloc/free hot paths. Besides, it makes linking kmem caches in a list chained by a field of struct memcg_cache_params really painful due to a level of indirection, while I want to make them linked in the following patch. That said, let us embed it. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Dave Chinner <david@fromorbit.com> Cc: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-12 22:59:20 +00:00
* This will be the memcg's index in each cache's ->memcg_params.memcg_caches.
* The main reason for not using cgroup id for this:
* this works better in sparse environments, where we have a lot of memcgs,
* but only a few kmem-limited. Or also, if we have, for instance, 200
* memcgs, and none but the 200th is kmem-limited, we'd have to have a
* 200 entry array for that.
memcg: allocate memory for memcg caches whenever a new memcg appears Every cache that is considered a root cache (basically the "original" caches, tied to the root memcg/no-memcg) will have an array that should be large enough to store a cache pointer per each memcg in the system. Theoreticaly, this is as high as 1 << sizeof(css_id), which is currently in the 64k pointers range. Most of the time, we won't be using that much. What goes in this patch, is a simple scheme to dynamically allocate such an array, in order to minimize memory usage for memcg caches. Because we would also like to avoid allocations all the time, at least for now, the array will only grow. It will tend to be big enough to hold the maximum number of kmem-limited memcgs ever achieved. We'll allocate it to be a minimum of 64 kmem-limited memcgs. When we have more than that, we'll start doubling the size of this array every time the limit is reached. Because we are only considering kmem limited memcgs, a natural point for this to happen is when we write to the limit. At that point, we already have set_limit_mutex held, so that will become our natural synchronization mechanism. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:38 +00:00
*
* The current size of the caches array is stored in memcg_nr_cache_ids. It
* will double each time we have to increase it.
memcg: allocate memory for memcg caches whenever a new memcg appears Every cache that is considered a root cache (basically the "original" caches, tied to the root memcg/no-memcg) will have an array that should be large enough to store a cache pointer per each memcg in the system. Theoreticaly, this is as high as 1 << sizeof(css_id), which is currently in the 64k pointers range. Most of the time, we won't be using that much. What goes in this patch, is a simple scheme to dynamically allocate such an array, in order to minimize memory usage for memcg caches. Because we would also like to avoid allocations all the time, at least for now, the array will only grow. It will tend to be big enough to hold the maximum number of kmem-limited memcgs ever achieved. We'll allocate it to be a minimum of 64 kmem-limited memcgs. When we have more than that, we'll start doubling the size of this array every time the limit is reached. Because we are only considering kmem limited memcgs, a natural point for this to happen is when we write to the limit. At that point, we already have set_limit_mutex held, so that will become our natural synchronization mechanism. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:38 +00:00
*/
static DEFINE_IDA(memcg_cache_ida);
int memcg_nr_cache_ids;
memcg: add rwsem to synchronize against memcg_caches arrays relocation We need a stable value of memcg_nr_cache_ids in kmem_cache_create() (memcg_alloc_cache_params() wants it for root caches), where we only hold the slab_mutex and no memcg-related locks. As a result, we have to update memcg_nr_cache_ids under the slab_mutex, which we can only take on the slab's side (see memcg_update_array_size). This looks awkward and will become even worse when per-memcg list_lru is introduced, which also wants stable access to memcg_nr_cache_ids. To get rid of this dependency between the memcg_nr_cache_ids and the slab_mutex, this patch introduces a special rwsem. The rwsem is held for writing during memcg_caches arrays relocation and memcg_nr_cache_ids updates. Therefore one can take it for reading to get a stable access to memcg_caches arrays and/or memcg_nr_cache_ids. Currently the semaphore is taken for reading only from kmem_cache_create, right before taking the slab_mutex, so right now there's no much point in using rwsem instead of mutex. However, once list_lru is made per-memcg it will allow list_lru initializations to proceed concurrently. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Dave Chinner <david@fromorbit.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Greg Thelen <gthelen@google.com> Cc: Glauber Costa <glommer@gmail.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-12 22:59:01 +00:00
/* Protects memcg_nr_cache_ids */
static DECLARE_RWSEM(memcg_cache_ids_sem);
void memcg_get_cache_ids(void)
{
down_read(&memcg_cache_ids_sem);
}
void memcg_put_cache_ids(void)
{
up_read(&memcg_cache_ids_sem);
}
memcg: allocate memory for memcg caches whenever a new memcg appears Every cache that is considered a root cache (basically the "original" caches, tied to the root memcg/no-memcg) will have an array that should be large enough to store a cache pointer per each memcg in the system. Theoreticaly, this is as high as 1 << sizeof(css_id), which is currently in the 64k pointers range. Most of the time, we won't be using that much. What goes in this patch, is a simple scheme to dynamically allocate such an array, in order to minimize memory usage for memcg caches. Because we would also like to avoid allocations all the time, at least for now, the array will only grow. It will tend to be big enough to hold the maximum number of kmem-limited memcgs ever achieved. We'll allocate it to be a minimum of 64 kmem-limited memcgs. When we have more than that, we'll start doubling the size of this array every time the limit is reached. Because we are only considering kmem limited memcgs, a natural point for this to happen is when we write to the limit. At that point, we already have set_limit_mutex held, so that will become our natural synchronization mechanism. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:38 +00:00
/*
* MIN_SIZE is different than 1, because we would like to avoid going through
* the alloc/free process all the time. In a small machine, 4 kmem-limited
* cgroups is a reasonable guess. In the future, it could be a parameter or
* tunable, but that is strictly not necessary.
*
* MAX_SIZE should be as large as the number of cgrp_ids. Ideally, we could get
memcg: allocate memory for memcg caches whenever a new memcg appears Every cache that is considered a root cache (basically the "original" caches, tied to the root memcg/no-memcg) will have an array that should be large enough to store a cache pointer per each memcg in the system. Theoreticaly, this is as high as 1 << sizeof(css_id), which is currently in the 64k pointers range. Most of the time, we won't be using that much. What goes in this patch, is a simple scheme to dynamically allocate such an array, in order to minimize memory usage for memcg caches. Because we would also like to avoid allocations all the time, at least for now, the array will only grow. It will tend to be big enough to hold the maximum number of kmem-limited memcgs ever achieved. We'll allocate it to be a minimum of 64 kmem-limited memcgs. When we have more than that, we'll start doubling the size of this array every time the limit is reached. Because we are only considering kmem limited memcgs, a natural point for this to happen is when we write to the limit. At that point, we already have set_limit_mutex held, so that will become our natural synchronization mechanism. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:38 +00:00
* this constant directly from cgroup, but it is understandable that this is
* better kept as an internal representation in cgroup.c. In any case, the
* cgrp_id space is not getting any smaller, and we don't have to necessarily
memcg: allocate memory for memcg caches whenever a new memcg appears Every cache that is considered a root cache (basically the "original" caches, tied to the root memcg/no-memcg) will have an array that should be large enough to store a cache pointer per each memcg in the system. Theoreticaly, this is as high as 1 << sizeof(css_id), which is currently in the 64k pointers range. Most of the time, we won't be using that much. What goes in this patch, is a simple scheme to dynamically allocate such an array, in order to minimize memory usage for memcg caches. Because we would also like to avoid allocations all the time, at least for now, the array will only grow. It will tend to be big enough to hold the maximum number of kmem-limited memcgs ever achieved. We'll allocate it to be a minimum of 64 kmem-limited memcgs. When we have more than that, we'll start doubling the size of this array every time the limit is reached. Because we are only considering kmem limited memcgs, a natural point for this to happen is when we write to the limit. At that point, we already have set_limit_mutex held, so that will become our natural synchronization mechanism. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:38 +00:00
* increase ours as well if it increases.
*/
#define MEMCG_CACHES_MIN_SIZE 4
#define MEMCG_CACHES_MAX_SIZE MEM_CGROUP_ID_MAX
memcg: allocate memory for memcg caches whenever a new memcg appears Every cache that is considered a root cache (basically the "original" caches, tied to the root memcg/no-memcg) will have an array that should be large enough to store a cache pointer per each memcg in the system. Theoreticaly, this is as high as 1 << sizeof(css_id), which is currently in the 64k pointers range. Most of the time, we won't be using that much. What goes in this patch, is a simple scheme to dynamically allocate such an array, in order to minimize memory usage for memcg caches. Because we would also like to avoid allocations all the time, at least for now, the array will only grow. It will tend to be big enough to hold the maximum number of kmem-limited memcgs ever achieved. We'll allocate it to be a minimum of 64 kmem-limited memcgs. When we have more than that, we'll start doubling the size of this array every time the limit is reached. Because we are only considering kmem limited memcgs, a natural point for this to happen is when we write to the limit. At that point, we already have set_limit_mutex held, so that will become our natural synchronization mechanism. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:38 +00:00
memcg: infrastructure to match an allocation to the right cache The page allocator is able to bind a page to a memcg when it is allocated. But for the caches, we'd like to have as many objects as possible in a page belonging to the same cache. This is done in this patch by calling memcg_kmem_get_cache in the beginning of every allocation function. This function is patched out by static branches when kernel memory controller is not being used. It assumes that the task allocating, which determines the memcg in the page allocator, belongs to the same cgroup throughout the whole process. Misaccounting can happen if the task calls memcg_kmem_get_cache() while belonging to a cgroup, and later on changes. This is considered acceptable, and should only happen upon task migration. Before the cache is created by the memcg core, there is also a possible imbalance: the task belongs to a memcg, but the cache being allocated from is the global cache, since the child cache is not yet guaranteed to be ready. This case is also fine, since in this case the GFP_KMEMCG will not be passed and the page allocator will not attempt any cgroup accounting. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:40 +00:00
/*
* A lot of the calls to the cache allocation functions are expected to be
* inlined by the compiler. Since the calls to memcg_kmem_get_cache are
* conditional to this static branch, we'll have to allow modules that does
* kmem_cache_alloc and the such to see this symbol as well
*/
DEFINE_STATIC_KEY_FALSE(memcg_kmem_enabled_key);
memcg: infrastructure to match an allocation to the right cache The page allocator is able to bind a page to a memcg when it is allocated. But for the caches, we'd like to have as many objects as possible in a page belonging to the same cache. This is done in this patch by calling memcg_kmem_get_cache in the beginning of every allocation function. This function is patched out by static branches when kernel memory controller is not being used. It assumes that the task allocating, which determines the memcg in the page allocator, belongs to the same cgroup throughout the whole process. Misaccounting can happen if the task calls memcg_kmem_get_cache() while belonging to a cgroup, and later on changes. This is considered acceptable, and should only happen upon task migration. Before the cache is created by the memcg core, there is also a possible imbalance: the task belongs to a memcg, but the cache being allocated from is the global cache, since the child cache is not yet guaranteed to be ready. This case is also fine, since in this case the GFP_KMEMCG will not be passed and the page allocator will not attempt any cgroup accounting. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:40 +00:00
EXPORT_SYMBOL(memcg_kmem_enabled_key);
memcg: use static branches when code not in use We can use static branches to patch the code in or out when not used. Because the _ACTIVE bit on kmem_accounted is only set after the increment is done, we guarantee that the root memcg will always be selected for kmem charges until all call sites are patched (see memcg_kmem_enabled). This guarantees that no mischarges are applied. Static branch decrement happens when the last reference count from the kmem accounting in memcg dies. This will only happen when the charges drop down to 0. When that happens, we need to disable the static branch only on those memcgs that enabled it. To achieve this, we would be forced to complicate the code by keeping track of which memcgs were the ones that actually enabled limits, and which ones got it from its parents. It is a lot simpler just to do static_key_slow_inc() on every child that is accounted. Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:09 +00:00
struct workqueue_struct *memcg_kmem_cache_wq;
#endif
mm, memcg: assign memcg-aware shrinkers bitmap to memcg Imagine a big node with many cpus, memory cgroups and containers. Let we have 200 containers, every container has 10 mounts, and 10 cgroups. All container tasks don't touch foreign containers mounts. If there is intensive pages write, and global reclaim happens, a writing task has to iterate over all memcgs to shrink slab, before it's able to go to shrink_page_list(). Iteration over all the memcg slabs is very expensive: the task has to visit 200 * 10 = 2000 shrinkers for every memcg, and since there are 2000 memcgs, the total calls are 2000 * 2000 = 4000000. So, the shrinker makes 4 million do_shrink_slab() calls just to try to isolate SWAP_CLUSTER_MAX pages in one of the actively writing memcg via shrink_page_list(). I've observed a node spending almost 100% in kernel, making useless iteration over already shrinked slab. This patch adds bitmap of memcg-aware shrinkers to memcg. The size of the bitmap depends on bitmap_nr_ids, and during memcg life it's maintained to be enough to fit bitmap_nr_ids shrinkers. Every bit in the map is related to corresponding shrinker id. Next patches will maintain set bit only for really charged memcg. This will allow shrink_slab() to increase its performance in significant way. See the last patch for the numbers. [ktkhai@virtuozzo.com: v9] Link: http://lkml.kernel.org/r/153112549031.4097.3576147070498769979.stgit@localhost.localdomain [ktkhai@virtuozzo.com: add comment to mem_cgroup_css_online()] Link: http://lkml.kernel.org/r/521f9e5f-c436-b388-fe83-4dc870bfb489@virtuozzo.com Link: http://lkml.kernel.org/r/153063056619.1818.12550500883688681076.stgit@localhost.localdomain Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Tested-by: Shakeel Butt <shakeelb@google.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Guenter Roeck <linux@roeck-us.net> Cc: "Huang, Ying" <ying.huang@intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Josef Bacik <jbacik@fb.com> Cc: Li RongQing <lirongqing@baidu.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Matthias Kaehlcke <mka@chromium.org> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Michal Hocko <mhocko@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Philippe Ombredanne <pombredanne@nexb.com> Cc: Roman Gushchin <guro@fb.com> Cc: Sahitya Tummala <stummala@codeaurora.org> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Waiman Long <longman@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:37 +00:00
static int memcg_shrinker_map_size;
static DEFINE_MUTEX(memcg_shrinker_map_mutex);
static void memcg_free_shrinker_map_rcu(struct rcu_head *head)
{
kvfree(container_of(head, struct memcg_shrinker_map, rcu));
}
static int memcg_expand_one_shrinker_map(struct mem_cgroup *memcg,
int size, int old_size)
{
struct memcg_shrinker_map *new, *old;
int nid;
lockdep_assert_held(&memcg_shrinker_map_mutex);
for_each_node(nid) {
old = rcu_dereference_protected(
mem_cgroup_nodeinfo(memcg, nid)->shrinker_map, true);
/* Not yet online memcg */
if (!old)
return 0;
new = kvmalloc_node(sizeof(*new) + size, GFP_KERNEL, nid);
mm, memcg: assign memcg-aware shrinkers bitmap to memcg Imagine a big node with many cpus, memory cgroups and containers. Let we have 200 containers, every container has 10 mounts, and 10 cgroups. All container tasks don't touch foreign containers mounts. If there is intensive pages write, and global reclaim happens, a writing task has to iterate over all memcgs to shrink slab, before it's able to go to shrink_page_list(). Iteration over all the memcg slabs is very expensive: the task has to visit 200 * 10 = 2000 shrinkers for every memcg, and since there are 2000 memcgs, the total calls are 2000 * 2000 = 4000000. So, the shrinker makes 4 million do_shrink_slab() calls just to try to isolate SWAP_CLUSTER_MAX pages in one of the actively writing memcg via shrink_page_list(). I've observed a node spending almost 100% in kernel, making useless iteration over already shrinked slab. This patch adds bitmap of memcg-aware shrinkers to memcg. The size of the bitmap depends on bitmap_nr_ids, and during memcg life it's maintained to be enough to fit bitmap_nr_ids shrinkers. Every bit in the map is related to corresponding shrinker id. Next patches will maintain set bit only for really charged memcg. This will allow shrink_slab() to increase its performance in significant way. See the last patch for the numbers. [ktkhai@virtuozzo.com: v9] Link: http://lkml.kernel.org/r/153112549031.4097.3576147070498769979.stgit@localhost.localdomain [ktkhai@virtuozzo.com: add comment to mem_cgroup_css_online()] Link: http://lkml.kernel.org/r/521f9e5f-c436-b388-fe83-4dc870bfb489@virtuozzo.com Link: http://lkml.kernel.org/r/153063056619.1818.12550500883688681076.stgit@localhost.localdomain Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Tested-by: Shakeel Butt <shakeelb@google.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Guenter Roeck <linux@roeck-us.net> Cc: "Huang, Ying" <ying.huang@intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Josef Bacik <jbacik@fb.com> Cc: Li RongQing <lirongqing@baidu.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Matthias Kaehlcke <mka@chromium.org> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Michal Hocko <mhocko@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Philippe Ombredanne <pombredanne@nexb.com> Cc: Roman Gushchin <guro@fb.com> Cc: Sahitya Tummala <stummala@codeaurora.org> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Waiman Long <longman@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:37 +00:00
if (!new)
return -ENOMEM;
/* Set all old bits, clear all new bits */
memset(new->map, (int)0xff, old_size);
memset((void *)new->map + old_size, 0, size - old_size);
rcu_assign_pointer(memcg->nodeinfo[nid]->shrinker_map, new);
call_rcu(&old->rcu, memcg_free_shrinker_map_rcu);
}
return 0;
}
static void memcg_free_shrinker_maps(struct mem_cgroup *memcg)
{
struct mem_cgroup_per_node *pn;
struct memcg_shrinker_map *map;
int nid;
if (mem_cgroup_is_root(memcg))
return;
for_each_node(nid) {
pn = mem_cgroup_nodeinfo(memcg, nid);
map = rcu_dereference_protected(pn->shrinker_map, true);
if (map)
kvfree(map);
rcu_assign_pointer(pn->shrinker_map, NULL);
}
}
static int memcg_alloc_shrinker_maps(struct mem_cgroup *memcg)
{
struct memcg_shrinker_map *map;
int nid, size, ret = 0;
if (mem_cgroup_is_root(memcg))
return 0;
mutex_lock(&memcg_shrinker_map_mutex);
size = memcg_shrinker_map_size;
for_each_node(nid) {
map = kvzalloc_node(sizeof(*map) + size, GFP_KERNEL, nid);
mm, memcg: assign memcg-aware shrinkers bitmap to memcg Imagine a big node with many cpus, memory cgroups and containers. Let we have 200 containers, every container has 10 mounts, and 10 cgroups. All container tasks don't touch foreign containers mounts. If there is intensive pages write, and global reclaim happens, a writing task has to iterate over all memcgs to shrink slab, before it's able to go to shrink_page_list(). Iteration over all the memcg slabs is very expensive: the task has to visit 200 * 10 = 2000 shrinkers for every memcg, and since there are 2000 memcgs, the total calls are 2000 * 2000 = 4000000. So, the shrinker makes 4 million do_shrink_slab() calls just to try to isolate SWAP_CLUSTER_MAX pages in one of the actively writing memcg via shrink_page_list(). I've observed a node spending almost 100% in kernel, making useless iteration over already shrinked slab. This patch adds bitmap of memcg-aware shrinkers to memcg. The size of the bitmap depends on bitmap_nr_ids, and during memcg life it's maintained to be enough to fit bitmap_nr_ids shrinkers. Every bit in the map is related to corresponding shrinker id. Next patches will maintain set bit only for really charged memcg. This will allow shrink_slab() to increase its performance in significant way. See the last patch for the numbers. [ktkhai@virtuozzo.com: v9] Link: http://lkml.kernel.org/r/153112549031.4097.3576147070498769979.stgit@localhost.localdomain [ktkhai@virtuozzo.com: add comment to mem_cgroup_css_online()] Link: http://lkml.kernel.org/r/521f9e5f-c436-b388-fe83-4dc870bfb489@virtuozzo.com Link: http://lkml.kernel.org/r/153063056619.1818.12550500883688681076.stgit@localhost.localdomain Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Tested-by: Shakeel Butt <shakeelb@google.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Guenter Roeck <linux@roeck-us.net> Cc: "Huang, Ying" <ying.huang@intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Josef Bacik <jbacik@fb.com> Cc: Li RongQing <lirongqing@baidu.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Matthias Kaehlcke <mka@chromium.org> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Michal Hocko <mhocko@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Philippe Ombredanne <pombredanne@nexb.com> Cc: Roman Gushchin <guro@fb.com> Cc: Sahitya Tummala <stummala@codeaurora.org> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Waiman Long <longman@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:37 +00:00
if (!map) {
memcg_free_shrinker_maps(memcg);
ret = -ENOMEM;
break;
}
rcu_assign_pointer(memcg->nodeinfo[nid]->shrinker_map, map);
}
mutex_unlock(&memcg_shrinker_map_mutex);
return ret;
}
int memcg_expand_shrinker_maps(int new_id)
{
int size, old_size, ret = 0;
struct mem_cgroup *memcg;
size = DIV_ROUND_UP(new_id + 1, BITS_PER_LONG) * sizeof(unsigned long);
old_size = memcg_shrinker_map_size;
if (size <= old_size)
return 0;
mutex_lock(&memcg_shrinker_map_mutex);
if (!root_mem_cgroup)
goto unlock;
for_each_mem_cgroup(memcg) {
if (mem_cgroup_is_root(memcg))
continue;
ret = memcg_expand_one_shrinker_map(memcg, size, old_size);
if (ret) {
mem_cgroup_iter_break(NULL, memcg);
mm, memcg: assign memcg-aware shrinkers bitmap to memcg Imagine a big node with many cpus, memory cgroups and containers. Let we have 200 containers, every container has 10 mounts, and 10 cgroups. All container tasks don't touch foreign containers mounts. If there is intensive pages write, and global reclaim happens, a writing task has to iterate over all memcgs to shrink slab, before it's able to go to shrink_page_list(). Iteration over all the memcg slabs is very expensive: the task has to visit 200 * 10 = 2000 shrinkers for every memcg, and since there are 2000 memcgs, the total calls are 2000 * 2000 = 4000000. So, the shrinker makes 4 million do_shrink_slab() calls just to try to isolate SWAP_CLUSTER_MAX pages in one of the actively writing memcg via shrink_page_list(). I've observed a node spending almost 100% in kernel, making useless iteration over already shrinked slab. This patch adds bitmap of memcg-aware shrinkers to memcg. The size of the bitmap depends on bitmap_nr_ids, and during memcg life it's maintained to be enough to fit bitmap_nr_ids shrinkers. Every bit in the map is related to corresponding shrinker id. Next patches will maintain set bit only for really charged memcg. This will allow shrink_slab() to increase its performance in significant way. See the last patch for the numbers. [ktkhai@virtuozzo.com: v9] Link: http://lkml.kernel.org/r/153112549031.4097.3576147070498769979.stgit@localhost.localdomain [ktkhai@virtuozzo.com: add comment to mem_cgroup_css_online()] Link: http://lkml.kernel.org/r/521f9e5f-c436-b388-fe83-4dc870bfb489@virtuozzo.com Link: http://lkml.kernel.org/r/153063056619.1818.12550500883688681076.stgit@localhost.localdomain Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Tested-by: Shakeel Butt <shakeelb@google.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Guenter Roeck <linux@roeck-us.net> Cc: "Huang, Ying" <ying.huang@intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Josef Bacik <jbacik@fb.com> Cc: Li RongQing <lirongqing@baidu.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Matthias Kaehlcke <mka@chromium.org> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Michal Hocko <mhocko@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Philippe Ombredanne <pombredanne@nexb.com> Cc: Roman Gushchin <guro@fb.com> Cc: Sahitya Tummala <stummala@codeaurora.org> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Waiman Long <longman@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:37 +00:00
goto unlock;
}
mm, memcg: assign memcg-aware shrinkers bitmap to memcg Imagine a big node with many cpus, memory cgroups and containers. Let we have 200 containers, every container has 10 mounts, and 10 cgroups. All container tasks don't touch foreign containers mounts. If there is intensive pages write, and global reclaim happens, a writing task has to iterate over all memcgs to shrink slab, before it's able to go to shrink_page_list(). Iteration over all the memcg slabs is very expensive: the task has to visit 200 * 10 = 2000 shrinkers for every memcg, and since there are 2000 memcgs, the total calls are 2000 * 2000 = 4000000. So, the shrinker makes 4 million do_shrink_slab() calls just to try to isolate SWAP_CLUSTER_MAX pages in one of the actively writing memcg via shrink_page_list(). I've observed a node spending almost 100% in kernel, making useless iteration over already shrinked slab. This patch adds bitmap of memcg-aware shrinkers to memcg. The size of the bitmap depends on bitmap_nr_ids, and during memcg life it's maintained to be enough to fit bitmap_nr_ids shrinkers. Every bit in the map is related to corresponding shrinker id. Next patches will maintain set bit only for really charged memcg. This will allow shrink_slab() to increase its performance in significant way. See the last patch for the numbers. [ktkhai@virtuozzo.com: v9] Link: http://lkml.kernel.org/r/153112549031.4097.3576147070498769979.stgit@localhost.localdomain [ktkhai@virtuozzo.com: add comment to mem_cgroup_css_online()] Link: http://lkml.kernel.org/r/521f9e5f-c436-b388-fe83-4dc870bfb489@virtuozzo.com Link: http://lkml.kernel.org/r/153063056619.1818.12550500883688681076.stgit@localhost.localdomain Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Tested-by: Shakeel Butt <shakeelb@google.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Guenter Roeck <linux@roeck-us.net> Cc: "Huang, Ying" <ying.huang@intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Josef Bacik <jbacik@fb.com> Cc: Li RongQing <lirongqing@baidu.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Matthias Kaehlcke <mka@chromium.org> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Michal Hocko <mhocko@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Philippe Ombredanne <pombredanne@nexb.com> Cc: Roman Gushchin <guro@fb.com> Cc: Sahitya Tummala <stummala@codeaurora.org> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Waiman Long <longman@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:37 +00:00
}
unlock:
if (!ret)
memcg_shrinker_map_size = size;
mutex_unlock(&memcg_shrinker_map_mutex);
return ret;
}
mm/list_lru.c: set bit in memcg shrinker bitmap on first list_lru item appearance Introduce set_shrinker_bit() function to set shrinker-related bit in memcg shrinker bitmap, and set the bit after the first item is added and in case of reparenting destroyed memcg's items. This will allow next patch to make shrinkers be called only, in case of they have charged objects at the moment, and to improve shrink_slab() performance. [ktkhai@virtuozzo.com: v9] Link: http://lkml.kernel.org/r/153112557572.4097.17315791419810749985.stgit@localhost.localdomain Link: http://lkml.kernel.org/r/153063065671.1818.15914674956134687268.stgit@localhost.localdomain Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Tested-by: Shakeel Butt <shakeelb@google.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Guenter Roeck <linux@roeck-us.net> Cc: "Huang, Ying" <ying.huang@intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Josef Bacik <jbacik@fb.com> Cc: Li RongQing <lirongqing@baidu.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Matthias Kaehlcke <mka@chromium.org> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Michal Hocko <mhocko@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Philippe Ombredanne <pombredanne@nexb.com> Cc: Roman Gushchin <guro@fb.com> Cc: Sahitya Tummala <stummala@codeaurora.org> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Waiman Long <longman@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:48:10 +00:00
void memcg_set_shrinker_bit(struct mem_cgroup *memcg, int nid, int shrinker_id)
{
if (shrinker_id >= 0 && memcg && !mem_cgroup_is_root(memcg)) {
struct memcg_shrinker_map *map;
rcu_read_lock();
map = rcu_dereference(memcg->nodeinfo[nid]->shrinker_map);
mm/vmscan.c: clear shrinker bit if there are no objects related to memcg To avoid further unneed calls of do_shrink_slab() for shrinkers, which already do not have any charged objects in a memcg, their bits have to be cleared. This patch introduces a lockless mechanism to do that without races without parallel list lru add. After do_shrink_slab() returns SHRINK_EMPTY the first time, we clear the bit and call it once again. Then we restore the bit, if the new return value is different. Note, that single smp_mb__after_atomic() in shrink_slab_memcg() covers two situations: 1)list_lru_add() shrink_slab_memcg list_add_tail() for_each_set_bit() <--- read bit do_shrink_slab() <--- missed list update (no barrier) <MB> <MB> set_bit() do_shrink_slab() <--- seen list update This situation, when the first do_shrink_slab() sees set bit, but it doesn't see list update (i.e., race with the first element queueing), is rare. So we don't add <MB> before the first call of do_shrink_slab() instead of this to do not slow down generic case. Also, it's need the second call as seen in below in (2). 2)list_lru_add() shrink_slab_memcg() list_add_tail() ... set_bit() ... ... for_each_set_bit() do_shrink_slab() do_shrink_slab() clear_bit() ... ... ... list_lru_add() ... list_add_tail() clear_bit() <MB> <MB> set_bit() do_shrink_slab() The barriers guarantee that the second do_shrink_slab() in the right side task sees list update if really cleared the bit. This case is drawn in the code comment. [Results/performance of the patchset] After the whole patchset applied the below test shows signify increase of performance: $echo 1 > /sys/fs/cgroup/memory/memory.use_hierarchy $mkdir /sys/fs/cgroup/memory/ct $echo 4000M > /sys/fs/cgroup/memory/ct/memory.kmem.limit_in_bytes $for i in `seq 0 4000`; do mkdir /sys/fs/cgroup/memory/ct/$i; echo $$ > /sys/fs/cgroup/memory/ct/$i/cgroup.procs; mkdir -p s/$i; mount -t tmpfs $i s/$i; touch s/$i/file; done Then, 5 sequential calls of drop caches: $time echo 3 > /proc/sys/vm/drop_caches 1)Before: 0.00user 13.78system 0:13.78elapsed 99%CPU 0.00user 5.59system 0:05.60elapsed 99%CPU 0.00user 5.48system 0:05.48elapsed 99%CPU 0.00user 8.35system 0:08.35elapsed 99%CPU 0.00user 8.34system 0:08.35elapsed 99%CPU 2)After 0.00user 1.10system 0:01.10elapsed 99%CPU 0.00user 0.00system 0:00.01elapsed 64%CPU 0.00user 0.01system 0:00.01elapsed 82%CPU 0.00user 0.00system 0:00.01elapsed 64%CPU 0.00user 0.01system 0:00.01elapsed 82%CPU The results show the performance increases at least in 548 times. Shakeel Butt tested this patchset with fork-bomb on his configuration: > I created 255 memcgs, 255 ext4 mounts and made each memcg create a > file containing few KiBs on corresponding mount. Then in a separate > memcg of 200 MiB limit ran a fork-bomb. > > I ran the "perf record -ag -- sleep 60" and below are the results: > > Without the patch series: > Samples: 4M of event 'cycles', Event count (approx.): 3279403076005 > + 36.40% fb.sh [kernel.kallsyms] [k] shrink_slab > + 18.97% fb.sh [kernel.kallsyms] [k] list_lru_count_one > + 6.75% fb.sh [kernel.kallsyms] [k] super_cache_count > + 0.49% fb.sh [kernel.kallsyms] [k] down_read_trylock > + 0.44% fb.sh [kernel.kallsyms] [k] mem_cgroup_iter > + 0.27% fb.sh [kernel.kallsyms] [k] up_read > + 0.21% fb.sh [kernel.kallsyms] [k] osq_lock > + 0.13% fb.sh [kernel.kallsyms] [k] shmem_unused_huge_count > + 0.08% fb.sh [kernel.kallsyms] [k] shrink_node_memcg > + 0.08% fb.sh [kernel.kallsyms] [k] shrink_node > > With the patch series: > Samples: 4M of event 'cycles', Event count (approx.): 2756866824946 > + 47.49% fb.sh [kernel.kallsyms] [k] down_read_trylock > + 30.72% fb.sh [kernel.kallsyms] [k] up_read > + 9.51% fb.sh [kernel.kallsyms] [k] mem_cgroup_iter > + 1.69% fb.sh [kernel.kallsyms] [k] shrink_node_memcg > + 1.35% fb.sh [kernel.kallsyms] [k] mem_cgroup_protected > + 1.05% fb.sh [kernel.kallsyms] [k] queued_spin_lock_slowpath > + 0.85% fb.sh [kernel.kallsyms] [k] _raw_spin_lock > + 0.78% fb.sh [kernel.kallsyms] [k] lruvec_lru_size > + 0.57% fb.sh [kernel.kallsyms] [k] shrink_node > + 0.54% fb.sh [kernel.kallsyms] [k] queue_work_on > + 0.46% fb.sh [kernel.kallsyms] [k] shrink_slab_memcg [ktkhai@virtuozzo.com: v9] Link: http://lkml.kernel.org/r/153112561772.4097.11011071937553113003.stgit@localhost.localdomain Link: http://lkml.kernel.org/r/153063070859.1818.11870882950920963480.stgit@localhost.localdomain Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Tested-by: Shakeel Butt <shakeelb@google.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Guenter Roeck <linux@roeck-us.net> Cc: "Huang, Ying" <ying.huang@intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Josef Bacik <jbacik@fb.com> Cc: Li RongQing <lirongqing@baidu.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Matthias Kaehlcke <mka@chromium.org> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Michal Hocko <mhocko@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Philippe Ombredanne <pombredanne@nexb.com> Cc: Roman Gushchin <guro@fb.com> Cc: Sahitya Tummala <stummala@codeaurora.org> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Waiman Long <longman@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:48:25 +00:00
/* Pairs with smp mb in shrink_slab() */
smp_mb__before_atomic();
mm/list_lru.c: set bit in memcg shrinker bitmap on first list_lru item appearance Introduce set_shrinker_bit() function to set shrinker-related bit in memcg shrinker bitmap, and set the bit after the first item is added and in case of reparenting destroyed memcg's items. This will allow next patch to make shrinkers be called only, in case of they have charged objects at the moment, and to improve shrink_slab() performance. [ktkhai@virtuozzo.com: v9] Link: http://lkml.kernel.org/r/153112557572.4097.17315791419810749985.stgit@localhost.localdomain Link: http://lkml.kernel.org/r/153063065671.1818.15914674956134687268.stgit@localhost.localdomain Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Tested-by: Shakeel Butt <shakeelb@google.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Guenter Roeck <linux@roeck-us.net> Cc: "Huang, Ying" <ying.huang@intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Josef Bacik <jbacik@fb.com> Cc: Li RongQing <lirongqing@baidu.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Matthias Kaehlcke <mka@chromium.org> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Michal Hocko <mhocko@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Philippe Ombredanne <pombredanne@nexb.com> Cc: Roman Gushchin <guro@fb.com> Cc: Sahitya Tummala <stummala@codeaurora.org> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Waiman Long <longman@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:48:10 +00:00
set_bit(shrinker_id, map->map);
rcu_read_unlock();
}
}
/**
* mem_cgroup_css_from_page - css of the memcg associated with a page
* @page: page of interest
*
* If memcg is bound to the default hierarchy, css of the memcg associated
* with @page is returned. The returned css remains associated with @page
* until it is released.
*
* If memcg is bound to a traditional hierarchy, the css of root_mem_cgroup
* is returned.
*/
struct cgroup_subsys_state *mem_cgroup_css_from_page(struct page *page)
{
struct mem_cgroup *memcg;
memcg = page->mem_cgroup;
if (!memcg || !cgroup_subsys_on_dfl(memory_cgrp_subsys))
memcg = root_mem_cgroup;
return &memcg->css;
}
memcg: add page_cgroup_ino helper This patchset introduces a new user API for tracking user memory pages that have not been used for a given period of time. The purpose of this is to provide the userspace with the means of tracking a workload's working set, i.e. the set of pages that are actively used by the workload. Knowing the working set size can be useful for partitioning the system more efficiently, e.g. by tuning memory cgroup limits appropriately, or for job placement within a compute cluster. ==== USE CASES ==== The unified cgroup hierarchy has memory.low and memory.high knobs, which are defined as the low and high boundaries for the workload working set size. However, the working set size of a workload may be unknown or change in time. With this patch set, one can periodically estimate the amount of memory unused by each cgroup and tune their memory.low and memory.high parameters accordingly, therefore optimizing the overall memory utilization. Another use case is balancing workloads within a compute cluster. Knowing how much memory is not really used by a workload unit may help take a more optimal decision when considering migrating the unit to another node within the cluster. Also, as noted by Minchan, this would be useful for per-process reclaim (https://lwn.net/Articles/545668/). With idle tracking, we could reclaim idle pages only by smart user memory manager. ==== USER API ==== The user API consists of two new files: * /sys/kernel/mm/page_idle/bitmap. This file implements a bitmap where each bit corresponds to a page, indexed by PFN. When the bit is set, the corresponding page is idle. A page is considered idle if it has not been accessed since it was marked idle. To mark a page idle one should set the bit corresponding to the page by writing to the file. A value written to the file is OR-ed with the current bitmap value. Only user memory pages can be marked idle, for other page types input is silently ignored. Writing to this file beyond max PFN results in the ENXIO error. Only available when CONFIG_IDLE_PAGE_TRACKING is set. This file can be used to estimate the amount of pages that are not used by a particular workload as follows: 1. mark all pages of interest idle by setting corresponding bits in the /sys/kernel/mm/page_idle/bitmap 2. wait until the workload accesses its working set 3. read /sys/kernel/mm/page_idle/bitmap and count the number of bits set * /proc/kpagecgroup. This file contains a 64-bit inode number of the memory cgroup each page is charged to, indexed by PFN. Only available when CONFIG_MEMCG is set. This file can be used to find all pages (including unmapped file pages) accounted to a particular cgroup. Using /sys/kernel/mm/page_idle/bitmap, one can then estimate the cgroup working set size. For an example of using these files for estimating the amount of unused memory pages per each memory cgroup, please see the script attached below. ==== REASONING ==== The reason to introduce the new user API instead of using /proc/PID/{clear_refs,smaps} is that the latter has two serious drawbacks: - it does not count unmapped file pages - it affects the reclaimer logic The new API attempts to overcome them both. For more details on how it is achieved, please see the comment to patch 6. ==== PATCHSET STRUCTURE ==== The patch set is organized as follows: - patch 1 adds page_cgroup_ino() helper for the sake of /proc/kpagecgroup and patches 2-3 do related cleanup - patch 4 adds /proc/kpagecgroup, which reports cgroup ino each page is charged to - patch 5 introduces a new mmu notifier callback, clear_young, which is a lightweight version of clear_flush_young; it is used in patch 6 - patch 6 implements the idle page tracking feature, including the userspace API, /sys/kernel/mm/page_idle/bitmap - patch 7 exports idle flag via /proc/kpageflags ==== SIMILAR WORKS ==== Originally, the patch for tracking idle memory was proposed back in 2011 by Michel Lespinasse (see http://lwn.net/Articles/459269/). The main difference between Michel's patch and this one is that Michel implemented a kernel space daemon for estimating idle memory size per cgroup while this patch only provides the userspace with the minimal API for doing the job, leaving the rest up to the userspace. However, they both share the same idea of Idle/Young page flags to avoid affecting the reclaimer logic. ==== PERFORMANCE EVALUATION ==== SPECjvm2008 (https://www.spec.org/jvm2008/) was used to evaluate the performance impact introduced by this patch set. Three runs were carried out: - base: kernel without the patch - patched: patched kernel, the feature is not used - patched-active: patched kernel, 1 minute-period daemon is used for tracking idle memory For tracking idle memory, idlememstat utility was used: https://github.com/locker/idlememstat testcase base patched patched-active compiler 537.40 ( 0.00)% 532.26 (-0.96)% 538.31 ( 0.17)% compress 305.47 ( 0.00)% 301.08 (-1.44)% 300.71 (-1.56)% crypto 284.32 ( 0.00)% 282.21 (-0.74)% 284.87 ( 0.19)% derby 411.05 ( 0.00)% 413.44 ( 0.58)% 412.07 ( 0.25)% mpegaudio 189.96 ( 0.00)% 190.87 ( 0.48)% 189.42 (-0.28)% scimark.large 46.85 ( 0.00)% 46.41 (-0.94)% 47.83 ( 2.09)% scimark.small 412.91 ( 0.00)% 415.41 ( 0.61)% 421.17 ( 2.00)% serial 204.23 ( 0.00)% 213.46 ( 4.52)% 203.17 (-0.52)% startup 36.76 ( 0.00)% 35.49 (-3.45)% 35.64 (-3.05)% sunflow 115.34 ( 0.00)% 115.08 (-0.23)% 117.37 ( 1.76)% xml 620.55 ( 0.00)% 619.95 (-0.10)% 620.39 (-0.03)% composite 211.50 ( 0.00)% 211.15 (-0.17)% 211.67 ( 0.08)% time idlememstat: 17.20user 65.16system 2:15:23elapsed 1%CPU (0avgtext+0avgdata 8476maxresident)k 448inputs+40outputs (1major+36052minor)pagefaults 0swaps ==== SCRIPT FOR COUNTING IDLE PAGES PER CGROUP ==== #! /usr/bin/python # import os import stat import errno import struct CGROUP_MOUNT = "/sys/fs/cgroup/memory" BUFSIZE = 8 * 1024 # must be multiple of 8 def get_hugepage_size(): with open("/proc/meminfo", "r") as f: for s in f: k, v = s.split(":") if k == "Hugepagesize": return int(v.split()[0]) * 1024 PAGE_SIZE = os.sysconf("SC_PAGE_SIZE") HUGEPAGE_SIZE = get_hugepage_size() def set_idle(): f = open("/sys/kernel/mm/page_idle/bitmap", "wb", BUFSIZE) while True: try: f.write(struct.pack("Q", pow(2, 64) - 1)) except IOError as err: if err.errno == errno.ENXIO: break raise f.close() def count_idle(): f_flags = open("/proc/kpageflags", "rb", BUFSIZE) f_cgroup = open("/proc/kpagecgroup", "rb", BUFSIZE) with open("/sys/kernel/mm/page_idle/bitmap", "rb", BUFSIZE) as f: while f.read(BUFSIZE): pass # update idle flag idlememsz = {} while True: s1, s2 = f_flags.read(8), f_cgroup.read(8) if not s1 or not s2: break flags, = struct.unpack('Q', s1) cgino, = struct.unpack('Q', s2) unevictable = (flags >> 18) & 1 huge = (flags >> 22) & 1 idle = (flags >> 25) & 1 if idle and not unevictable: idlememsz[cgino] = idlememsz.get(cgino, 0) + \ (HUGEPAGE_SIZE if huge else PAGE_SIZE) f_flags.close() f_cgroup.close() return idlememsz if __name__ == "__main__": print "Setting the idle flag for each page..." set_idle() raw_input("Wait until the workload accesses its working set, " "then press Enter") print "Counting idle pages..." idlememsz = count_idle() for dir, subdirs, files in os.walk(CGROUP_MOUNT): ino = os.stat(dir)[stat.ST_INO] print dir + ": " + str(idlememsz.get(ino, 0) / 1024) + " kB" ==== END SCRIPT ==== This patch (of 8): Add page_cgroup_ino() helper to memcg. This function returns the inode number of the closest online ancestor of the memory cgroup a page is charged to. It is required for exporting information about which page is charged to which cgroup to userspace, which will be introduced by a following patch. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Reviewed-by: Andres Lagar-Cavilla <andreslc@google.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Raghavendra K T <raghavendra.kt@linux.vnet.ibm.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Greg Thelen <gthelen@google.com> Cc: Michel Lespinasse <walken@google.com> Cc: David Rientjes <rientjes@google.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: Cyrill Gorcunov <gorcunov@openvz.org> Cc: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-09-09 22:35:28 +00:00
/**
* page_cgroup_ino - return inode number of the memcg a page is charged to
* @page: the page
*
* Look up the closest online ancestor of the memory cgroup @page is charged to
* and return its inode number or 0 if @page is not charged to any cgroup. It
* is safe to call this function without holding a reference to @page.
*
* Note, this function is inherently racy, because there is nothing to prevent
* the cgroup inode from getting torn down and potentially reallocated a moment
* after page_cgroup_ino() returns, so it only should be used by callers that
* do not care (such as procfs interfaces).
*/
ino_t page_cgroup_ino(struct page *page)
{
struct mem_cgroup *memcg;
unsigned long ino = 0;
rcu_read_lock();
mm: slab: make page_cgroup_ino() to recognize non-compound slab pages properly page_cgroup_ino() doesn't return a valid memcg pointer for non-compound slab pages, because it depends on PgHead AND PgSlab flags to be set to determine the memory cgroup from the kmem_cache. It's correct for compound pages, but not for generic small pages. Those don't have PgHead set, so it ends up returning zero. Fix this by replacing the condition to PageSlab() && !PageTail(). Before this patch: [root@localhost ~]# ./page-types -c /sys/fs/cgroup/user.slice/user-0.slice/user@0.service/ | grep slab 0x0000000000000080 38 0 _______S___________________________________ slab After this patch: [root@localhost ~]# ./page-types -c /sys/fs/cgroup/user.slice/user-0.slice/user@0.service/ | grep slab 0x0000000000000080 147 0 _______S___________________________________ slab Also, hwpoison_filter_task() uses output of page_cgroup_ino() in order to filter error injection events based on memcg. So if page_cgroup_ino() fails to return memcg pointer, we just fail to inject memory error. Considering that hwpoison filter is for testing, affected users are limited and the impact should be marginal. [n-horiguchi@ah.jp.nec.com: changelog additions] Link: http://lkml.kernel.org/r/20191031012151.2722280-1-guro@fb.com Fixes: 4d96ba353075 ("mm: memcg/slab: stop setting page->mem_cgroup pointer for slab pages") Signed-off-by: Roman Gushchin <guro@fb.com> Reviewed-by: Shakeel Butt <shakeelb@google.com> Acked-by: David Rientjes <rientjes@google.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Daniel Jordan <daniel.m.jordan@oracle.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-11-06 05:17:03 +00:00
if (PageSlab(page) && !PageTail(page))
mm: memcg/slab: stop setting page->mem_cgroup pointer for slab pages Every slab page charged to a non-root memory cgroup has a pointer to the memory cgroup and holds a reference to it, which protects a non-empty memory cgroup from being released. At the same time the page has a pointer to the corresponding kmem_cache, and also hold a reference to the kmem_cache. And kmem_cache by itself holds a reference to the cgroup. So there is clearly some redundancy, which allows to stop setting the page->mem_cgroup pointer and rely on getting memcg pointer indirectly via kmem_cache. Further it will allow to change this pointer easier, without a need to go over all charged pages. So let's stop setting page->mem_cgroup pointer for slab pages, and stop using the css refcounter directly for protecting the memory cgroup from going away. Instead rely on kmem_cache as an intermediate object. Make sure that vmstats and shrinker lists are working as previously, as well as /proc/kpagecgroup interface. Link: http://lkml.kernel.org/r/20190611231813.3148843-10-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Reviewed-by: Shakeel Butt <shakeelb@google.com> Cc: Christoph Lameter <cl@linux.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Waiman Long <longman@redhat.com> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: Andrei Vagin <avagin@gmail.com> Cc: Qian Cai <cai@lca.pw> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:56:31 +00:00
memcg = memcg_from_slab_page(page);
else
memcg = READ_ONCE(page->mem_cgroup);
memcg: add page_cgroup_ino helper This patchset introduces a new user API for tracking user memory pages that have not been used for a given period of time. The purpose of this is to provide the userspace with the means of tracking a workload's working set, i.e. the set of pages that are actively used by the workload. Knowing the working set size can be useful for partitioning the system more efficiently, e.g. by tuning memory cgroup limits appropriately, or for job placement within a compute cluster. ==== USE CASES ==== The unified cgroup hierarchy has memory.low and memory.high knobs, which are defined as the low and high boundaries for the workload working set size. However, the working set size of a workload may be unknown or change in time. With this patch set, one can periodically estimate the amount of memory unused by each cgroup and tune their memory.low and memory.high parameters accordingly, therefore optimizing the overall memory utilization. Another use case is balancing workloads within a compute cluster. Knowing how much memory is not really used by a workload unit may help take a more optimal decision when considering migrating the unit to another node within the cluster. Also, as noted by Minchan, this would be useful for per-process reclaim (https://lwn.net/Articles/545668/). With idle tracking, we could reclaim idle pages only by smart user memory manager. ==== USER API ==== The user API consists of two new files: * /sys/kernel/mm/page_idle/bitmap. This file implements a bitmap where each bit corresponds to a page, indexed by PFN. When the bit is set, the corresponding page is idle. A page is considered idle if it has not been accessed since it was marked idle. To mark a page idle one should set the bit corresponding to the page by writing to the file. A value written to the file is OR-ed with the current bitmap value. Only user memory pages can be marked idle, for other page types input is silently ignored. Writing to this file beyond max PFN results in the ENXIO error. Only available when CONFIG_IDLE_PAGE_TRACKING is set. This file can be used to estimate the amount of pages that are not used by a particular workload as follows: 1. mark all pages of interest idle by setting corresponding bits in the /sys/kernel/mm/page_idle/bitmap 2. wait until the workload accesses its working set 3. read /sys/kernel/mm/page_idle/bitmap and count the number of bits set * /proc/kpagecgroup. This file contains a 64-bit inode number of the memory cgroup each page is charged to, indexed by PFN. Only available when CONFIG_MEMCG is set. This file can be used to find all pages (including unmapped file pages) accounted to a particular cgroup. Using /sys/kernel/mm/page_idle/bitmap, one can then estimate the cgroup working set size. For an example of using these files for estimating the amount of unused memory pages per each memory cgroup, please see the script attached below. ==== REASONING ==== The reason to introduce the new user API instead of using /proc/PID/{clear_refs,smaps} is that the latter has two serious drawbacks: - it does not count unmapped file pages - it affects the reclaimer logic The new API attempts to overcome them both. For more details on how it is achieved, please see the comment to patch 6. ==== PATCHSET STRUCTURE ==== The patch set is organized as follows: - patch 1 adds page_cgroup_ino() helper for the sake of /proc/kpagecgroup and patches 2-3 do related cleanup - patch 4 adds /proc/kpagecgroup, which reports cgroup ino each page is charged to - patch 5 introduces a new mmu notifier callback, clear_young, which is a lightweight version of clear_flush_young; it is used in patch 6 - patch 6 implements the idle page tracking feature, including the userspace API, /sys/kernel/mm/page_idle/bitmap - patch 7 exports idle flag via /proc/kpageflags ==== SIMILAR WORKS ==== Originally, the patch for tracking idle memory was proposed back in 2011 by Michel Lespinasse (see http://lwn.net/Articles/459269/). The main difference between Michel's patch and this one is that Michel implemented a kernel space daemon for estimating idle memory size per cgroup while this patch only provides the userspace with the minimal API for doing the job, leaving the rest up to the userspace. However, they both share the same idea of Idle/Young page flags to avoid affecting the reclaimer logic. ==== PERFORMANCE EVALUATION ==== SPECjvm2008 (https://www.spec.org/jvm2008/) was used to evaluate the performance impact introduced by this patch set. Three runs were carried out: - base: kernel without the patch - patched: patched kernel, the feature is not used - patched-active: patched kernel, 1 minute-period daemon is used for tracking idle memory For tracking idle memory, idlememstat utility was used: https://github.com/locker/idlememstat testcase base patched patched-active compiler 537.40 ( 0.00)% 532.26 (-0.96)% 538.31 ( 0.17)% compress 305.47 ( 0.00)% 301.08 (-1.44)% 300.71 (-1.56)% crypto 284.32 ( 0.00)% 282.21 (-0.74)% 284.87 ( 0.19)% derby 411.05 ( 0.00)% 413.44 ( 0.58)% 412.07 ( 0.25)% mpegaudio 189.96 ( 0.00)% 190.87 ( 0.48)% 189.42 (-0.28)% scimark.large 46.85 ( 0.00)% 46.41 (-0.94)% 47.83 ( 2.09)% scimark.small 412.91 ( 0.00)% 415.41 ( 0.61)% 421.17 ( 2.00)% serial 204.23 ( 0.00)% 213.46 ( 4.52)% 203.17 (-0.52)% startup 36.76 ( 0.00)% 35.49 (-3.45)% 35.64 (-3.05)% sunflow 115.34 ( 0.00)% 115.08 (-0.23)% 117.37 ( 1.76)% xml 620.55 ( 0.00)% 619.95 (-0.10)% 620.39 (-0.03)% composite 211.50 ( 0.00)% 211.15 (-0.17)% 211.67 ( 0.08)% time idlememstat: 17.20user 65.16system 2:15:23elapsed 1%CPU (0avgtext+0avgdata 8476maxresident)k 448inputs+40outputs (1major+36052minor)pagefaults 0swaps ==== SCRIPT FOR COUNTING IDLE PAGES PER CGROUP ==== #! /usr/bin/python # import os import stat import errno import struct CGROUP_MOUNT = "/sys/fs/cgroup/memory" BUFSIZE = 8 * 1024 # must be multiple of 8 def get_hugepage_size(): with open("/proc/meminfo", "r") as f: for s in f: k, v = s.split(":") if k == "Hugepagesize": return int(v.split()[0]) * 1024 PAGE_SIZE = os.sysconf("SC_PAGE_SIZE") HUGEPAGE_SIZE = get_hugepage_size() def set_idle(): f = open("/sys/kernel/mm/page_idle/bitmap", "wb", BUFSIZE) while True: try: f.write(struct.pack("Q", pow(2, 64) - 1)) except IOError as err: if err.errno == errno.ENXIO: break raise f.close() def count_idle(): f_flags = open("/proc/kpageflags", "rb", BUFSIZE) f_cgroup = open("/proc/kpagecgroup", "rb", BUFSIZE) with open("/sys/kernel/mm/page_idle/bitmap", "rb", BUFSIZE) as f: while f.read(BUFSIZE): pass # update idle flag idlememsz = {} while True: s1, s2 = f_flags.read(8), f_cgroup.read(8) if not s1 or not s2: break flags, = struct.unpack('Q', s1) cgino, = struct.unpack('Q', s2) unevictable = (flags >> 18) & 1 huge = (flags >> 22) & 1 idle = (flags >> 25) & 1 if idle and not unevictable: idlememsz[cgino] = idlememsz.get(cgino, 0) + \ (HUGEPAGE_SIZE if huge else PAGE_SIZE) f_flags.close() f_cgroup.close() return idlememsz if __name__ == "__main__": print "Setting the idle flag for each page..." set_idle() raw_input("Wait until the workload accesses its working set, " "then press Enter") print "Counting idle pages..." idlememsz = count_idle() for dir, subdirs, files in os.walk(CGROUP_MOUNT): ino = os.stat(dir)[stat.ST_INO] print dir + ": " + str(idlememsz.get(ino, 0) / 1024) + " kB" ==== END SCRIPT ==== This patch (of 8): Add page_cgroup_ino() helper to memcg. This function returns the inode number of the closest online ancestor of the memory cgroup a page is charged to. It is required for exporting information about which page is charged to which cgroup to userspace, which will be introduced by a following patch. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Reviewed-by: Andres Lagar-Cavilla <andreslc@google.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Raghavendra K T <raghavendra.kt@linux.vnet.ibm.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Greg Thelen <gthelen@google.com> Cc: Michel Lespinasse <walken@google.com> Cc: David Rientjes <rientjes@google.com> Cc: Pavel Emelyanov <xemul@parallels.com> Cc: Cyrill Gorcunov <gorcunov@openvz.org> Cc: Jonathan Corbet <corbet@lwn.net> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-09-09 22:35:28 +00:00
while (memcg && !(memcg->css.flags & CSS_ONLINE))
memcg = parent_mem_cgroup(memcg);
if (memcg)
ino = cgroup_ino(memcg->css.cgroup);
rcu_read_unlock();
return ino;
}
static struct mem_cgroup_per_node *
mem_cgroup_page_nodeinfo(struct mem_cgroup *memcg, struct page *page)
{
int nid = page_to_nid(page);
return memcg->nodeinfo[nid];
}
static struct mem_cgroup_tree_per_node *
soft_limit_tree_node(int nid)
{
return soft_limit_tree.rb_tree_per_node[nid];
}
static struct mem_cgroup_tree_per_node *
soft_limit_tree_from_page(struct page *page)
{
int nid = page_to_nid(page);
return soft_limit_tree.rb_tree_per_node[nid];
}
static void __mem_cgroup_insert_exceeded(struct mem_cgroup_per_node *mz,
struct mem_cgroup_tree_per_node *mctz,
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
unsigned long new_usage_in_excess)
{
struct rb_node **p = &mctz->rb_root.rb_node;
struct rb_node *parent = NULL;
struct mem_cgroup_per_node *mz_node;
bool rightmost = true;
if (mz->on_tree)
return;
mz->usage_in_excess = new_usage_in_excess;
if (!mz->usage_in_excess)
return;
while (*p) {
parent = *p;
mz_node = rb_entry(parent, struct mem_cgroup_per_node,
tree_node);
if (mz->usage_in_excess < mz_node->usage_in_excess) {
p = &(*p)->rb_left;
rightmost = false;
}
/*
* We can't avoid mem cgroups that are over their soft
* limit by the same amount
*/
else if (mz->usage_in_excess >= mz_node->usage_in_excess)
p = &(*p)->rb_right;
}
if (rightmost)
mctz->rb_rightmost = &mz->tree_node;
rb_link_node(&mz->tree_node, parent, p);
rb_insert_color(&mz->tree_node, &mctz->rb_root);
mz->on_tree = true;
}
static void __mem_cgroup_remove_exceeded(struct mem_cgroup_per_node *mz,
struct mem_cgroup_tree_per_node *mctz)
{
if (!mz->on_tree)
return;
if (&mz->tree_node == mctz->rb_rightmost)
mctz->rb_rightmost = rb_prev(&mz->tree_node);
rb_erase(&mz->tree_node, &mctz->rb_root);
mz->on_tree = false;
}
static void mem_cgroup_remove_exceeded(struct mem_cgroup_per_node *mz,
struct mem_cgroup_tree_per_node *mctz)
{
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
unsigned long flags;
spin_lock_irqsave(&mctz->lock, flags);
__mem_cgroup_remove_exceeded(mz, mctz);
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
spin_unlock_irqrestore(&mctz->lock, flags);
}
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
static unsigned long soft_limit_excess(struct mem_cgroup *memcg)
{
unsigned long nr_pages = page_counter_read(&memcg->memory);
unsigned long soft_limit = READ_ONCE(memcg->soft_limit);
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
unsigned long excess = 0;
if (nr_pages > soft_limit)
excess = nr_pages - soft_limit;
return excess;
}
static void mem_cgroup_update_tree(struct mem_cgroup *memcg, struct page *page)
{
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
unsigned long excess;
struct mem_cgroup_per_node *mz;
struct mem_cgroup_tree_per_node *mctz;
mctz = soft_limit_tree_from_page(page);
mm/cgroup: avoid panic when init with low memory The system may panic when initialisation is done when almost all the memory is assigned to the huge pages using the kernel command line parameter hugepage=xxxx. Panic may occur like this: Unable to handle kernel paging request for data at address 0x00000000 Faulting instruction address: 0xc000000000302b88 Oops: Kernel access of bad area, sig: 11 [#1] SMP NR_CPUS=2048 [ 0.082424] NUMA pSeries Modules linked in: CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.9.0-15-generic #16-Ubuntu task: c00000021ed01600 task.stack: c00000010d108000 NIP: c000000000302b88 LR: c000000000270e04 CTR: c00000000016cfd0 REGS: c00000010d10b2c0 TRAP: 0300 Not tainted (4.9.0-15-generic) MSR: 8000000002009033 <SF,VEC,EE,ME,IR,DR,RI,LE>[ 0.082770] CR: 28424422 XER: 00000000 CFAR: c0000000003d28b8 DAR: 0000000000000000 DSISR: 40000000 SOFTE: 1 GPR00: c000000000270e04 c00000010d10b540 c00000000141a300 c00000010fff6300 GPR04: 0000000000000000 00000000026012c0 c00000010d10b630 0000000487ab0000 GPR08: 000000010ee90000 c000000001454fd8 0000000000000000 0000000000000000 GPR12: 0000000000004400 c00000000fb80000 00000000026012c0 00000000026012c0 GPR16: 00000000026012c0 0000000000000000 0000000000000000 0000000000000002 GPR20: 000000000000000c 0000000000000000 0000000000000000 00000000024200c0 GPR24: c0000000016eef48 0000000000000000 c00000010fff7d00 00000000026012c0 GPR28: 0000000000000000 c00000010fff7d00 c00000010fff6300 c00000010d10b6d0 NIP mem_cgroup_soft_limit_reclaim+0xf8/0x4f0 LR do_try_to_free_pages+0x1b4/0x450 Call Trace: do_try_to_free_pages+0x1b4/0x450 try_to_free_pages+0xf8/0x270 __alloc_pages_nodemask+0x7a8/0xff0 new_slab+0x104/0x8e0 ___slab_alloc+0x620/0x700 __slab_alloc+0x34/0x60 kmem_cache_alloc_node_trace+0xdc/0x310 mem_cgroup_init+0x158/0x1c8 do_one_initcall+0x68/0x1d0 kernel_init_freeable+0x278/0x360 kernel_init+0x24/0x170 ret_from_kernel_thread+0x5c/0x74 Instruction dump: eb81ffe0 eba1ffe8 ebc1fff0 ebe1fff8 4e800020 3d230001 e9499a42 3d220004 3929acd8 794a1f24 7d295214 eac90100 <e9360000> 2fa90000 419eff74 3b200000 ---[ end trace 342f5208b00d01b6 ]--- This is a chicken and egg issue where the kernel try to get free memory when allocating per node data in mem_cgroup_init(), but in that path mem_cgroup_soft_limit_reclaim() is called which assumes that these data are allocated. As mem_cgroup_soft_limit_reclaim() is best effort, it should return when these data are not yet allocated. This patch also fixes potential null pointer access in mem_cgroup_remove_from_trees() and mem_cgroup_update_tree(). Link: http://lkml.kernel.org/r/1487856999-16581-2-git-send-email-ldufour@linux.vnet.ibm.com Signed-off-by: Laurent Dufour <ldufour@linux.vnet.ibm.com> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Balbir Singh <bsingharora@gmail.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-03-10 00:17:06 +00:00
if (!mctz)
return;
/*
* Necessary to update all ancestors when hierarchy is used.
* because their event counter is not touched.
*/
for (; memcg; memcg = parent_mem_cgroup(memcg)) {
mz = mem_cgroup_page_nodeinfo(memcg, page);
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
excess = soft_limit_excess(memcg);
/*
* We have to update the tree if mz is on RB-tree or
* mem is over its softlimit.
*/
if (excess || mz->on_tree) {
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
unsigned long flags;
spin_lock_irqsave(&mctz->lock, flags);
/* if on-tree, remove it */
if (mz->on_tree)
__mem_cgroup_remove_exceeded(mz, mctz);
/*
* Insert again. mz->usage_in_excess will be updated.
* If excess is 0, no tree ops.
*/
__mem_cgroup_insert_exceeded(mz, mctz, excess);
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
spin_unlock_irqrestore(&mctz->lock, flags);
}
}
}
static void mem_cgroup_remove_from_trees(struct mem_cgroup *memcg)
{
struct mem_cgroup_tree_per_node *mctz;
struct mem_cgroup_per_node *mz;
int nid;
for_each_node(nid) {
mz = mem_cgroup_nodeinfo(memcg, nid);
mctz = soft_limit_tree_node(nid);
mm/cgroup: avoid panic when init with low memory The system may panic when initialisation is done when almost all the memory is assigned to the huge pages using the kernel command line parameter hugepage=xxxx. Panic may occur like this: Unable to handle kernel paging request for data at address 0x00000000 Faulting instruction address: 0xc000000000302b88 Oops: Kernel access of bad area, sig: 11 [#1] SMP NR_CPUS=2048 [ 0.082424] NUMA pSeries Modules linked in: CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.9.0-15-generic #16-Ubuntu task: c00000021ed01600 task.stack: c00000010d108000 NIP: c000000000302b88 LR: c000000000270e04 CTR: c00000000016cfd0 REGS: c00000010d10b2c0 TRAP: 0300 Not tainted (4.9.0-15-generic) MSR: 8000000002009033 <SF,VEC,EE,ME,IR,DR,RI,LE>[ 0.082770] CR: 28424422 XER: 00000000 CFAR: c0000000003d28b8 DAR: 0000000000000000 DSISR: 40000000 SOFTE: 1 GPR00: c000000000270e04 c00000010d10b540 c00000000141a300 c00000010fff6300 GPR04: 0000000000000000 00000000026012c0 c00000010d10b630 0000000487ab0000 GPR08: 000000010ee90000 c000000001454fd8 0000000000000000 0000000000000000 GPR12: 0000000000004400 c00000000fb80000 00000000026012c0 00000000026012c0 GPR16: 00000000026012c0 0000000000000000 0000000000000000 0000000000000002 GPR20: 000000000000000c 0000000000000000 0000000000000000 00000000024200c0 GPR24: c0000000016eef48 0000000000000000 c00000010fff7d00 00000000026012c0 GPR28: 0000000000000000 c00000010fff7d00 c00000010fff6300 c00000010d10b6d0 NIP mem_cgroup_soft_limit_reclaim+0xf8/0x4f0 LR do_try_to_free_pages+0x1b4/0x450 Call Trace: do_try_to_free_pages+0x1b4/0x450 try_to_free_pages+0xf8/0x270 __alloc_pages_nodemask+0x7a8/0xff0 new_slab+0x104/0x8e0 ___slab_alloc+0x620/0x700 __slab_alloc+0x34/0x60 kmem_cache_alloc_node_trace+0xdc/0x310 mem_cgroup_init+0x158/0x1c8 do_one_initcall+0x68/0x1d0 kernel_init_freeable+0x278/0x360 kernel_init+0x24/0x170 ret_from_kernel_thread+0x5c/0x74 Instruction dump: eb81ffe0 eba1ffe8 ebc1fff0 ebe1fff8 4e800020 3d230001 e9499a42 3d220004 3929acd8 794a1f24 7d295214 eac90100 <e9360000> 2fa90000 419eff74 3b200000 ---[ end trace 342f5208b00d01b6 ]--- This is a chicken and egg issue where the kernel try to get free memory when allocating per node data in mem_cgroup_init(), but in that path mem_cgroup_soft_limit_reclaim() is called which assumes that these data are allocated. As mem_cgroup_soft_limit_reclaim() is best effort, it should return when these data are not yet allocated. This patch also fixes potential null pointer access in mem_cgroup_remove_from_trees() and mem_cgroup_update_tree(). Link: http://lkml.kernel.org/r/1487856999-16581-2-git-send-email-ldufour@linux.vnet.ibm.com Signed-off-by: Laurent Dufour <ldufour@linux.vnet.ibm.com> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Balbir Singh <bsingharora@gmail.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-03-10 00:17:06 +00:00
if (mctz)
mem_cgroup_remove_exceeded(mz, mctz);
}
}
static struct mem_cgroup_per_node *
__mem_cgroup_largest_soft_limit_node(struct mem_cgroup_tree_per_node *mctz)
{
struct mem_cgroup_per_node *mz;
retry:
mz = NULL;
if (!mctz->rb_rightmost)
goto done; /* Nothing to reclaim from */
mz = rb_entry(mctz->rb_rightmost,
struct mem_cgroup_per_node, tree_node);
/*
* Remove the node now but someone else can add it back,
* we will to add it back at the end of reclaim to its correct
* position in the tree.
*/
__mem_cgroup_remove_exceeded(mz, mctz);
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
if (!soft_limit_excess(mz->memcg) ||
memcg: css_tryget_online cleanups Currently multiple locations in memcg code, css_tryget_online() is being used. However it doesn't matter whether the cgroup is online for the callers. Online used to matter when we had reparenting on offlining and we needed a way to prevent new ones from showing up. The failure case for couple of these css_tryget_online usage is to fallback to root_mem_cgroup which kind of make bypassing the memcg limits possible for some workloads. For example creating an inotify group in a subcontainer and then deleting that container after moving the process to a different container will make all the event objects allocated for that group to the root_mem_cgroup. So, using css_tryget_online() is dangerous for such cases. Two locations still use the online version. The swapin of offlined memcg's pages and the memcg kmem cache creation. The kmem cache indeed needs the online version as the kernel does the reparenting of memcg kmem caches. For the swapin case, it has been left for later as the fallback is not really that concerning. With swap accounting enabled, if the memcg of the swapped out page is not online then the memcg extracted from the given 'mm' will be charged and if 'mm' is NULL then root memcg will be charged. However I could not find a code path where the given 'mm' will be NULL for swap-in case. Signed-off-by: Shakeel Butt <shakeelb@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Link: http://lkml.kernel.org/r/20200302203109.179417-1-shakeelb@google.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-04-02 04:07:10 +00:00
!css_tryget(&mz->memcg->css))
goto retry;
done:
return mz;
}
static struct mem_cgroup_per_node *
mem_cgroup_largest_soft_limit_node(struct mem_cgroup_tree_per_node *mctz)
{
struct mem_cgroup_per_node *mz;
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
spin_lock_irq(&mctz->lock);
mz = __mem_cgroup_largest_soft_limit_node(mctz);
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
spin_unlock_irq(&mctz->lock);
return mz;
}
/**
* __mod_memcg_state - update cgroup memory statistics
* @memcg: the memory cgroup
* @idx: the stat item - can be enum memcg_stat_item or enum node_stat_item
* @val: delta to add to the counter, can be negative
*/
void __mod_memcg_state(struct mem_cgroup *memcg, int idx, int val)
{
long x;
if (mem_cgroup_disabled())
return;
x = val + __this_cpu_read(memcg->vmstats_percpu->stat[idx]);
if (unlikely(abs(x) > MEMCG_CHARGE_BATCH)) {
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
struct mem_cgroup *mi;
/*
* Batch local counters to keep them in sync with
* the hierarchical ones.
*/
__this_cpu_add(memcg->vmstats_local->stat[idx], x);
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
for (mi = memcg; mi; mi = parent_mem_cgroup(mi))
atomic_long_add(x, &mi->vmstats[idx]);
x = 0;
}
__this_cpu_write(memcg->vmstats_percpu->stat[idx], x);
}
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
static struct mem_cgroup_per_node *
parent_nodeinfo(struct mem_cgroup_per_node *pn, int nid)
{
struct mem_cgroup *parent;
parent = parent_mem_cgroup(pn->memcg);
if (!parent)
return NULL;
return mem_cgroup_nodeinfo(parent, nid);
}
/**
* __mod_lruvec_state - update lruvec memory statistics
* @lruvec: the lruvec
* @idx: the stat item
* @val: delta to add to the counter, can be negative
*
* The lruvec is the intersection of the NUMA node and a cgroup. This
* function updates the all three counters that are affected by a
* change of state at this level: per-node, per-cgroup, per-lruvec.
*/
void __mod_lruvec_state(struct lruvec *lruvec, enum node_stat_item idx,
int val)
{
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
pg_data_t *pgdat = lruvec_pgdat(lruvec);
struct mem_cgroup_per_node *pn;
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
struct mem_cgroup *memcg;
long x;
/* Update node */
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
__mod_node_page_state(pgdat, idx, val);
if (mem_cgroup_disabled())
return;
pn = container_of(lruvec, struct mem_cgroup_per_node, lruvec);
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
memcg = pn->memcg;
/* Update memcg */
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
__mod_memcg_state(memcg, idx, val);
mm, memcg: partially revert "mm/memcontrol.c: keep local VM counters in sync with the hierarchical ones" Commit 766a4c19d880 ("mm/memcontrol.c: keep local VM counters in sync with the hierarchical ones") effectively decreased the precision of per-memcg vmstats_local and per-memcg-per-node lruvec percpu counters. That's good for displaying in memory.stat, but brings a serious regression into the reclaim process. One issue I've discovered and debugged is the following: lruvec_lru_size() can return 0 instead of the actual number of pages in the lru list, preventing the kernel to reclaim last remaining pages. Result is yet another dying memory cgroups flooding. The opposite is also happening: scanning an empty lru list is the waste of cpu time. Also, inactive_list_is_low() can return incorrect values, preventing the active lru from being scanned and freed. It can fail both because the size of active and inactive lists are inaccurate, and because the number of workingset refaults isn't precise. In other words, the result is pretty random. I'm not sure, if using the approximate number of slab pages in count_shadow_number() is acceptable, but issues described above are enough to partially revert the patch. Let's keep per-memcg vmstat_local batched (they are only used for displaying stats to the userspace), but keep lruvec stats precise. This change fixes the dead memcg flooding on my setup. Link: http://lkml.kernel.org/r/20190817004726.2530670-1-guro@fb.com Fixes: 766a4c19d880 ("mm/memcontrol.c: keep local VM counters in sync with the hierarchical ones") Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Yafang Shao <laoar.shao@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-08-30 23:04:39 +00:00
/* Update lruvec */
__this_cpu_add(pn->lruvec_stat_local->count[idx], val);
x = val + __this_cpu_read(pn->lruvec_stat_cpu->count[idx]);
if (unlikely(abs(x) > MEMCG_CHARGE_BATCH)) {
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
struct mem_cgroup_per_node *pi;
for (pi = pn; pi; pi = parent_nodeinfo(pi, pgdat->node_id))
atomic_long_add(x, &pi->lruvec_stat[idx]);
x = 0;
}
__this_cpu_write(pn->lruvec_stat_cpu->count[idx], x);
}
void __mod_lruvec_slab_state(void *p, enum node_stat_item idx, int val)
{
pg_data_t *pgdat = page_pgdat(virt_to_page(p));
struct mem_cgroup *memcg;
struct lruvec *lruvec;
rcu_read_lock();
memcg = mem_cgroup_from_obj(p);
/* Untracked pages have no memcg, no lruvec. Update only the node */
if (!memcg || memcg == root_mem_cgroup) {
__mod_node_page_state(pgdat, idx, val);
} else {
lruvec = mem_cgroup_lruvec(memcg, pgdat);
__mod_lruvec_state(lruvec, idx, val);
}
rcu_read_unlock();
}
mm: fork: fix kernel_stack memcg stats for various stack implementations Depending on CONFIG_VMAP_STACK and the THREAD_SIZE / PAGE_SIZE ratio the space for task stacks can be allocated using __vmalloc_node_range(), alloc_pages_node() and kmem_cache_alloc_node(). In the first and the second cases page->mem_cgroup pointer is set, but in the third it's not: memcg membership of a slab page should be determined using the memcg_from_slab_page() function, which looks at page->slab_cache->memcg_params.memcg . In this case, using mod_memcg_page_state() (as in account_kernel_stack()) is incorrect: page->mem_cgroup pointer is NULL even for pages charged to a non-root memory cgroup. It can lead to kernel_stack per-memcg counters permanently showing 0 on some architectures (depending on the configuration). In order to fix it, let's introduce a mod_memcg_obj_state() helper, which takes a pointer to a kernel object as a first argument, uses mem_cgroup_from_obj() to get a RCU-protected memcg pointer and calls mod_memcg_state(). It allows to handle all possible configurations (CONFIG_VMAP_STACK and various THREAD_SIZE/PAGE_SIZE values) without spilling any memcg/kmem specifics into fork.c . Note: This is a special version of the patch created for stable backports. It contains code from the following two patches: - mm: memcg/slab: introduce mem_cgroup_from_obj() - mm: fork: fix kernel_stack memcg stats for various stack implementations [guro@fb.com: introduce mem_cgroup_from_obj()] Link: http://lkml.kernel.org/r/20200324004221.GA36662@carbon.dhcp.thefacebook.com Fixes: 4d96ba353075 ("mm: memcg/slab: stop setting page->mem_cgroup pointer for slab pages") Signed-off-by: Roman Gushchin <guro@fb.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Bharata B Rao <bharata@linux.ibm.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: <stable@vger.kernel.org> Link: http://lkml.kernel.org/r/20200303233550.251375-1-guro@fb.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-03-29 02:17:25 +00:00
void mod_memcg_obj_state(void *p, int idx, int val)
{
struct mem_cgroup *memcg;
rcu_read_lock();
memcg = mem_cgroup_from_obj(p);
if (memcg)
mod_memcg_state(memcg, idx, val);
rcu_read_unlock();
}
/**
* __count_memcg_events - account VM events in a cgroup
* @memcg: the memory cgroup
* @idx: the event item
* @count: the number of events that occured
*/
void __count_memcg_events(struct mem_cgroup *memcg, enum vm_event_item idx,
unsigned long count)
{
unsigned long x;
if (mem_cgroup_disabled())
return;
x = count + __this_cpu_read(memcg->vmstats_percpu->events[idx]);
if (unlikely(x > MEMCG_CHARGE_BATCH)) {
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
struct mem_cgroup *mi;
/*
* Batch local counters to keep them in sync with
* the hierarchical ones.
*/
__this_cpu_add(memcg->vmstats_local->events[idx], x);
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
for (mi = memcg; mi; mi = parent_mem_cgroup(mi))
atomic_long_add(x, &mi->vmevents[idx]);
x = 0;
}
__this_cpu_write(memcg->vmstats_percpu->events[idx], x);
}
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
static unsigned long memcg_events(struct mem_cgroup *memcg, int event)
{
return atomic_long_read(&memcg->vmevents[event]);
}
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
static unsigned long memcg_events_local(struct mem_cgroup *memcg, int event)
{
mm: memcontrol: don't batch updates of local VM stats and events The kernel test robot noticed a 26% will-it-scale pagefault regression from commit 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty"). This appears to be caused by bouncing the additional cachelines from the new hierarchical statistics counters. We can fix this by getting rid of the batched local counters instead. Originally, there were *only* group-local counters, and they were fully maintained per cpu. A reader of a stats file high up in the cgroup tree would have to walk the entire subtree and collect each level's per-cpu counters to get the recursive view. This was prohibitively expensive, and so we switched to per-cpu batched updates of the local counters during a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting"), reducing the complexity from nr_subgroups * nr_cpus to nr_subgroups. With growing machines and cgroup trees, the tree walk itself became too expensive for monitoring top-level groups, and this is when the culprit patch added hierarchy counters on each cgroup level. When the per-cpu batch size would be reached, both the local and the hierarchy counters would get batch-updated from the per-cpu delta simultaneously. This makes local and hierarchical counter reads blazingly fast, but it unfortunately makes the write-side too cache line intense. Since local counter reads were never a problem - we only centralized them to accelerate the hierarchy walk - and use of the local counters are becoming rarer due to replacement with hierarchical views (ongoing rework in the page reclaim and workingset code), we can make those local counters unbatched per-cpu counters again. The scheme will then be as such: when a memcg statistic changes, the writer will: - update the local counter (per-cpu) - update the batch counter (per-cpu). If the batch is full: - spill the batch into the group's atomic_t - spill the batch into all ancestors' atomic_ts - empty out the batch counter (per-cpu) when a local memcg counter is read, the reader will: - collect the local counter from all cpus when a hiearchy memcg counter is read, the reader will: - read the atomic_t We might be able to simplify this further and make the recursive counters unbatched per-cpu counters as well (batch upward propagation, but leave per-cpu collection to the readers), but that will require a more in-depth analysis and testing of all the callsites. Deal with the immediate regression for now. Link: http://lkml.kernel.org/r/20190521151647.GB2870@cmpxchg.org Fixes: 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: kernel test robot <rong.a.chen@intel.com> Tested-by: kernel test robot <rong.a.chen@intel.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Shakeel Butt <shakeelb@google.com> Cc: Roman Gushchin <guro@fb.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-06-13 22:55:46 +00:00
long x = 0;
int cpu;
for_each_possible_cpu(cpu)
x += per_cpu(memcg->vmstats_local->events[event], cpu);
return x;
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
}
static void mem_cgroup_charge_statistics(struct mem_cgroup *memcg,
struct page *page,
mm: memcontrol: drop @compound parameter from memcg charging API The memcg charging API carries a boolean @compound parameter that tells whether the page we're dealing with is a hugepage. mem_cgroup_commit_charge() has another boolean @lrucare that indicates whether the page needs LRU locking or not while charging. The majority of callsites know those parameters at compile time, which results in a lot of naked "false, false" argument lists. This makes for cryptic code and is a breeding ground for subtle mistakes. Thankfully, the huge page state can be inferred from the page itself and doesn't need to be passed along. This is safe because charging completes before the page is published and somebody may split it. Simplify the callsites by removing @compound, and let memcg infer the state by using hpage_nr_pages() unconditionally. That function does PageTransHuge() to identify huge pages, which also helpfully asserts that nobody passes in tail pages by accident. The following patches will introduce a new charging API, best not to carry over unnecessary weight. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Alex Shi <alex.shi@linux.alibaba.com> Reviewed-by: Joonsoo Kim <iamjoonsoo.kim@lge.com> Reviewed-by: Shakeel Butt <shakeelb@google.com> Cc: Hugh Dickins <hughd@google.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.com> Cc: Roman Gushchin <guro@fb.com> Cc: Balbir Singh <bsingharora@gmail.com> Link: http://lkml.kernel.org/r/20200508183105.225460-4-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:01:31 +00:00
int nr_pages)
memory cgroup enhancements: add status accounting function for memory cgroup Add statistics account infrastructure for memory controller. All account information is stored per-cpu and caller will not have to take lock or use atomic ops. This will be used by memory.stat file later. CACHE includes swapcache now. I'd like to divide it to PAGECACHE and SWAPCACHE later. This patch adds 3 functions for accounting. * __mem_cgroup_stat_add() ... for usual routine. * __mem_cgroup_stat_add_safe ... for calling under irq_disabled section. * mem_cgroup_read_stat() ... for reading stat value. * renamed PAGECACHE to CACHE (because it may include swapcache *now*) [akpm@linux-foundation.org: coding-style fixes] [akpm@linux-foundation.org: fix smp_processor_id-in-preemptible] [akpm@linux-foundation.org: uninline things] [akpm@linux-foundation.org: remove dead code] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Signed-off-by: YAMAMOTO Takashi <yamamoto@valinux.co.jp> Cc: Balbir Singh <balbir@linux.vnet.ibm.com> Cc: Pavel Emelianov <xemul@openvz.org> Cc: Paul Menage <menage@google.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Cc: Kirill Korotaev <dev@sw.ru> Cc: Herbert Poetzl <herbert@13thfloor.at> Cc: David Rientjes <rientjes@google.com> Cc: Vaidyanathan Srinivasan <svaidy@linux.vnet.ibm.com> Cc: Kirill Korotaev <dev@sw.ru> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Cc: Paul Menage <menage@google.com> Cc: Pavel Emelianov <xemul@openvz.org> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: Vaidyanathan Srinivasan <svaidy@linux.vnet.ibm.com> Cc: YAMAMOTO Takashi <yamamoto@valinux.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-07 08:14:24 +00:00
{
/* pagein of a big page is an event. So, ignore page size */
if (nr_pages > 0)
__count_memcg_events(memcg, PGPGIN, 1);
else {
__count_memcg_events(memcg, PGPGOUT, 1);
nr_pages = -nr_pages; /* for event */
}
__this_cpu_add(memcg->vmstats_percpu->nr_page_events, nr_pages);
}
static bool mem_cgroup_event_ratelimit(struct mem_cgroup *memcg,
enum mem_cgroup_events_target target)
{
unsigned long val, next;
val = __this_cpu_read(memcg->vmstats_percpu->nr_page_events);
next = __this_cpu_read(memcg->vmstats_percpu->targets[target]);
/* from time_after() in jiffies.h */
if ((long)(next - val) < 0) {
switch (target) {
case MEM_CGROUP_TARGET_THRESH:
next = val + THRESHOLDS_EVENTS_TARGET;
break;
case MEM_CGROUP_TARGET_SOFTLIMIT:
next = val + SOFTLIMIT_EVENTS_TARGET;
break;
default:
break;
}
__this_cpu_write(memcg->vmstats_percpu->targets[target], next);
return true;
}
return false;
}
/*
* Check events in order.
*
*/
static void memcg_check_events(struct mem_cgroup *memcg, struct page *page)
{
/* threshold event is triggered in finer grain than soft limit */
if (unlikely(mem_cgroup_event_ratelimit(memcg,
MEM_CGROUP_TARGET_THRESH))) {
bool do_softlimit;
do_softlimit = mem_cgroup_event_ratelimit(memcg,
MEM_CGROUP_TARGET_SOFTLIMIT);
mem_cgroup_threshold(memcg);
if (unlikely(do_softlimit))
mem_cgroup_update_tree(memcg, page);
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
}
}
cgroups: add an owner to the mm_struct Remove the mem_cgroup member from mm_struct and instead adds an owner. This approach was suggested by Paul Menage. The advantage of this approach is that, once the mm->owner is known, using the subsystem id, the cgroup can be determined. It also allows several control groups that are virtually grouped by mm_struct, to exist independent of the memory controller i.e., without adding mem_cgroup's for each controller, to mm_struct. A new config option CONFIG_MM_OWNER is added and the memory resource controller selects this config option. This patch also adds cgroup callbacks to notify subsystems when mm->owner changes. The mm_cgroup_changed callback is called with the task_lock() of the new task held and is called just prior to changing the mm->owner. I am indebted to Paul Menage for the several reviews of this patchset and helping me make it lighter and simpler. This patch was tested on a powerpc box, it was compiled with both the MM_OWNER config turned on and off. After the thread group leader exits, it's moved to init_css_state by cgroup_exit(), thus all future charges from runnings threads would be redirected to the init_css_set's subsystem. Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com> Cc: Pavel Emelianov <xemul@openvz.org> Cc: Hugh Dickins <hugh@veritas.com> Cc: Sudhir Kumar <skumar@linux.vnet.ibm.com> Cc: YAMAMOTO Takashi <yamamoto@valinux.co.jp> Cc: Hirokazu Takahashi <taka@valinux.co.jp> Cc: David Rientjes <rientjes@google.com>, Cc: Balbir Singh <balbir@linux.vnet.ibm.com> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Pekka Enberg <penberg@cs.helsinki.fi> Reviewed-by: Paul Menage <menage@google.com> Cc: Oleg Nesterov <oleg@tv-sign.ru> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-04-29 08:00:16 +00:00
struct mem_cgroup *mem_cgroup_from_task(struct task_struct *p)
{
mm owner: fix race between swapoff and exit There's a race between mm->owner assignment and swapoff, more easily seen when task slab poisoning is turned on. The condition occurs when try_to_unuse() runs in parallel with an exiting task. A similar race can occur with callers of get_task_mm(), such as /proc/<pid>/<mmstats> or ptrace or page migration. CPU0 CPU1 try_to_unuse looks at mm = task0->mm increments mm->mm_users task 0 exits mm->owner needs to be updated, but no new owner is found (mm_users > 1, but no other task has task->mm = task0->mm) mm_update_next_owner() leaves mmput(mm) decrements mm->mm_users task0 freed dereferencing mm->owner fails The fix is to notify the subsystem via mm_owner_changed callback(), if no new owner is found, by specifying the new task as NULL. Jiri Slaby: mm->owner was set to NULL prior to calling cgroup_mm_owner_callbacks(), but must be set after that, so as not to pass NULL as old owner causing oops. Daisuke Nishimura: mm_update_next_owner() may set mm->owner to NULL, but mem_cgroup_from_task() and its callers need to take account of this situation to avoid oops. Hugh Dickins: Lockdep warning and hang below exec_mmap() when testing these patches. exit_mm() up_reads mmap_sem before calling mm_update_next_owner(), so exec_mmap() now needs to do the same. And with that repositioning, there's now no point in mm_need_new_owner() allowing for NULL mm. Reported-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com> Signed-off-by: Jiri Slaby <jirislaby@gmail.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Hugh Dickins <hugh@veritas.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Paul Menage <menage@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-09-28 22:09:31 +00:00
/*
* mm_update_next_owner() may clear mm->owner to NULL
* if it races with swapoff, page migration, etc.
* So this can be called with p == NULL.
*/
if (unlikely(!p))
return NULL;
cgroup: clean up cgroup_subsys names and initialization cgroup_subsys is a bit messier than it needs to be. * The name of a subsys can be different from its internal identifier defined in cgroup_subsys.h. Most subsystems use the matching name but three - cpu, memory and perf_event - use different ones. * cgroup_subsys_id enums are postfixed with _subsys_id and each cgroup_subsys is postfixed with _subsys. cgroup.h is widely included throughout various subsystems, it doesn't and shouldn't have claim on such generic names which don't have any qualifier indicating that they belong to cgroup. * cgroup_subsys->subsys_id should always equal the matching cgroup_subsys_id enum; however, we require each controller to initialize it and then BUG if they don't match, which is a bit silly. This patch cleans up cgroup_subsys names and initialization by doing the followings. * cgroup_subsys_id enums are now postfixed with _cgrp_id, and each cgroup_subsys with _cgrp_subsys. * With the above, renaming subsys identifiers to match the userland visible names doesn't cause any naming conflicts. All non-matching identifiers are renamed to match the official names. cpu_cgroup -> cpu mem_cgroup -> memory perf -> perf_event * controllers no longer need to initialize ->subsys_id and ->name. They're generated in cgroup core and set automatically during boot. * Redundant cgroup_subsys declarations removed. * While updating BUG_ON()s in cgroup_init_early(), convert them to WARN()s. BUGging that early during boot is stupid - the kernel can't print anything, even through serial console and the trap handler doesn't even link stack frame properly for back-tracing. This patch doesn't introduce any behavior changes. v2: Rebased on top of fe1217c4f3f7 ("net: net_cls: move cgroupfs classid handling into core"). Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Neil Horman <nhorman@tuxdriver.com> Acked-by: "David S. Miller" <davem@davemloft.net> Acked-by: "Rafael J. Wysocki" <rjw@rjwysocki.net> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Peter Zijlstra <peterz@infradead.org> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Ingo Molnar <mingo@redhat.com> Acked-by: Li Zefan <lizefan@huawei.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Serge E. Hallyn <serue@us.ibm.com> Cc: Vivek Goyal <vgoyal@redhat.com> Cc: Thomas Graf <tgraf@suug.ch>
2014-02-08 15:36:58 +00:00
return mem_cgroup_from_css(task_css(p, memory_cgrp_id));
}
EXPORT_SYMBOL(mem_cgroup_from_task);
fs: fsnotify: account fsnotify metadata to kmemcg Patch series "Directed kmem charging", v8. The Linux kernel's memory cgroup allows limiting the memory usage of the jobs running on the system to provide isolation between the jobs. All the kernel memory allocated in the context of the job and marked with __GFP_ACCOUNT will also be included in the memory usage and be limited by the job's limit. The kernel memory can only be charged to the memcg of the process in whose context kernel memory was allocated. However there are cases where the allocated kernel memory should be charged to the memcg different from the current processes's memcg. This patch series contains two such concrete use-cases i.e. fsnotify and buffer_head. The fsnotify event objects can consume a lot of system memory for large or unlimited queues if there is either no or slow listener. The events are allocated in the context of the event producer. However they should be charged to the event consumer. Similarly the buffer_head objects can be allocated in a memcg different from the memcg of the page for which buffer_head objects are being allocated. To solve this issue, this patch series introduces mechanism to charge kernel memory to a given memcg. In case of fsnotify events, the memcg of the consumer can be used for charging and for buffer_head, the memcg of the page can be charged. For directed charging, the caller can use the scope API memalloc_[un]use_memcg() to specify the memcg to charge for all the __GFP_ACCOUNT allocations within the scope. This patch (of 2): A lot of memory can be consumed by the events generated for the huge or unlimited queues if there is either no or slow listener. This can cause system level memory pressure or OOMs. So, it's better to account the fsnotify kmem caches to the memcg of the listener. However the listener can be in a different memcg than the memcg of the producer and these allocations happen in the context of the event producer. This patch introduces remote memcg charging API which the producer can use to charge the allocations to the memcg of the listener. There are seven fsnotify kmem caches and among them allocations from dnotify_struct_cache, dnotify_mark_cache, fanotify_mark_cache and inotify_inode_mark_cachep happens in the context of syscall from the listener. So, SLAB_ACCOUNT is enough for these caches. The objects from fsnotify_mark_connector_cachep are not accounted as they are small compared to the notification mark or events and it is unclear whom to account connector to since it is shared by all events attached to the inode. The allocations from the event caches happen in the context of the event producer. For such caches we will need to remote charge the allocations to the listener's memcg. Thus we save the memcg reference in the fsnotify_group structure of the listener. This patch has also moved the members of fsnotify_group to keep the size same, at least for 64 bit build, even with additional member by filling the holes. [shakeelb@google.com: use GFP_KERNEL_ACCOUNT rather than open-coding it] Link: http://lkml.kernel.org/r/20180702215439.211597-1-shakeelb@google.com Link: http://lkml.kernel.org/r/20180627191250.209150-2-shakeelb@google.com Signed-off-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Amir Goldstein <amir73il@gmail.com> Cc: Greg Thelen <gthelen@google.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:46:39 +00:00
/**
* get_mem_cgroup_from_mm: Obtain a reference on given mm_struct's memcg.
* @mm: mm from which memcg should be extracted. It can be NULL.
*
* Obtain a reference on mm->memcg and returns it if successful. Otherwise
* root_mem_cgroup is returned. However if mem_cgroup is disabled, NULL is
* returned.
*/
struct mem_cgroup *get_mem_cgroup_from_mm(struct mm_struct *mm)
{
fs: fsnotify: account fsnotify metadata to kmemcg Patch series "Directed kmem charging", v8. The Linux kernel's memory cgroup allows limiting the memory usage of the jobs running on the system to provide isolation between the jobs. All the kernel memory allocated in the context of the job and marked with __GFP_ACCOUNT will also be included in the memory usage and be limited by the job's limit. The kernel memory can only be charged to the memcg of the process in whose context kernel memory was allocated. However there are cases where the allocated kernel memory should be charged to the memcg different from the current processes's memcg. This patch series contains two such concrete use-cases i.e. fsnotify and buffer_head. The fsnotify event objects can consume a lot of system memory for large or unlimited queues if there is either no or slow listener. The events are allocated in the context of the event producer. However they should be charged to the event consumer. Similarly the buffer_head objects can be allocated in a memcg different from the memcg of the page for which buffer_head objects are being allocated. To solve this issue, this patch series introduces mechanism to charge kernel memory to a given memcg. In case of fsnotify events, the memcg of the consumer can be used for charging and for buffer_head, the memcg of the page can be charged. For directed charging, the caller can use the scope API memalloc_[un]use_memcg() to specify the memcg to charge for all the __GFP_ACCOUNT allocations within the scope. This patch (of 2): A lot of memory can be consumed by the events generated for the huge or unlimited queues if there is either no or slow listener. This can cause system level memory pressure or OOMs. So, it's better to account the fsnotify kmem caches to the memcg of the listener. However the listener can be in a different memcg than the memcg of the producer and these allocations happen in the context of the event producer. This patch introduces remote memcg charging API which the producer can use to charge the allocations to the memcg of the listener. There are seven fsnotify kmem caches and among them allocations from dnotify_struct_cache, dnotify_mark_cache, fanotify_mark_cache and inotify_inode_mark_cachep happens in the context of syscall from the listener. So, SLAB_ACCOUNT is enough for these caches. The objects from fsnotify_mark_connector_cachep are not accounted as they are small compared to the notification mark or events and it is unclear whom to account connector to since it is shared by all events attached to the inode. The allocations from the event caches happen in the context of the event producer. For such caches we will need to remote charge the allocations to the listener's memcg. Thus we save the memcg reference in the fsnotify_group structure of the listener. This patch has also moved the members of fsnotify_group to keep the size same, at least for 64 bit build, even with additional member by filling the holes. [shakeelb@google.com: use GFP_KERNEL_ACCOUNT rather than open-coding it] Link: http://lkml.kernel.org/r/20180702215439.211597-1-shakeelb@google.com Link: http://lkml.kernel.org/r/20180627191250.209150-2-shakeelb@google.com Signed-off-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Amir Goldstein <amir73il@gmail.com> Cc: Greg Thelen <gthelen@google.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:46:39 +00:00
struct mem_cgroup *memcg;
if (mem_cgroup_disabled())
return NULL;
rcu_read_lock();
do {
memcg: fix swapcache charge from kernel thread context Commit 284f39afeaa4 ("mm: memcg: push !mm handling out to page cache charge function") explicitly checks for page cache charges without any mm context (from kernel thread context[1]). This seemed to be the only possible case where memory could be charged without mm context so commit 03583f1a631c ("memcg: remove unnecessary !mm check from try_get_mem_cgroup_from_mm()") removed the mm check from get_mem_cgroup_from_mm(). This however caused another NULL ptr dereference during early boot when loopback kernel thread splices to tmpfs as reported by Stephan Kulow: BUG: unable to handle kernel NULL pointer dereference at 0000000000000360 IP: get_mem_cgroup_from_mm.isra.42+0x2b/0x60 Oops: 0000 [#1] SMP Modules linked in: btrfs dm_multipath dm_mod scsi_dh multipath raid10 raid456 async_raid6_recov async_memcpy async_pq raid6_pq async_xor xor async_tx raid1 raid0 md_mod parport_pc parport nls_utf8 isofs usb_storage iscsi_ibft iscsi_boot_sysfs arc4 ecb fan thermal nfs lockd fscache nls_iso8859_1 nls_cp437 sg st hid_generic usbhid af_packet sunrpc sr_mod cdrom ata_generic uhci_hcd virtio_net virtio_blk ehci_hcd usbcore ata_piix floppy processor button usb_common virtio_pci virtio_ring virtio edd squashfs loop ppa] CPU: 0 PID: 97 Comm: loop1 Not tainted 3.15.0-rc5-5-default #1 Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011 Call Trace: __mem_cgroup_try_charge_swapin+0x40/0xe0 mem_cgroup_charge_file+0x8b/0xd0 shmem_getpage_gfp+0x66b/0x7b0 shmem_file_splice_read+0x18f/0x430 splice_direct_to_actor+0xa2/0x1c0 do_lo_receive+0x5a/0x60 [loop] loop_thread+0x298/0x720 [loop] kthread+0xc6/0xe0 ret_from_fork+0x7c/0xb0 Also Branimir Maksimovic reported the following oops which is tiggered for the swapcache charge path from the accounting code for kernel threads: CPU: 1 PID: 160 Comm: kworker/u8:5 Tainted: P OE 3.15.0-rc5-core2-custom #159 Hardware name: System manufacturer System Product Name/MAXIMUSV GENE, BIOS 1903 08/19/2013 task: ffff880404e349b0 ti: ffff88040486a000 task.ti: ffff88040486a000 RIP: get_mem_cgroup_from_mm.isra.42+0x2b/0x60 Call Trace: __mem_cgroup_try_charge_swapin+0x45/0xf0 mem_cgroup_charge_file+0x9c/0xe0 shmem_getpage_gfp+0x62c/0x770 shmem_write_begin+0x38/0x40 generic_perform_write+0xc5/0x1c0 __generic_file_aio_write+0x1d1/0x3f0 generic_file_aio_write+0x4f/0xc0 do_sync_write+0x5a/0x90 do_acct_process+0x4b1/0x550 acct_process+0x6d/0xa0 do_exit+0x827/0xa70 kthread+0xc3/0xf0 This patch fixes the issue by reintroducing mm check into get_mem_cgroup_from_mm. We could do the same trick in __mem_cgroup_try_charge_swapin as we do for the regular page cache path but it is not worth troubles. The check is not that expensive and it is better to have get_mem_cgroup_from_mm more robust. [1] - http://marc.info/?l=linux-mm&m=139463617808941&w=2 Fixes: 03583f1a631c ("memcg: remove unnecessary !mm check from try_get_mem_cgroup_from_mm()") Reported-and-tested-by: Stephan Kulow <coolo@suse.com> Reported-by: Branimir Maksimovic <branimir.maksimovic@gmail.com> Signed-off-by: Michal Hocko <mhocko@suse.cz> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-05-22 18:54:19 +00:00
/*
* Page cache insertions can happen withou an
* actual mm context, e.g. during disk probing
* on boot, loopback IO, acct() writes etc.
*/
if (unlikely(!mm))
memcg = root_mem_cgroup;
memcg: fix swapcache charge from kernel thread context Commit 284f39afeaa4 ("mm: memcg: push !mm handling out to page cache charge function") explicitly checks for page cache charges without any mm context (from kernel thread context[1]). This seemed to be the only possible case where memory could be charged without mm context so commit 03583f1a631c ("memcg: remove unnecessary !mm check from try_get_mem_cgroup_from_mm()") removed the mm check from get_mem_cgroup_from_mm(). This however caused another NULL ptr dereference during early boot when loopback kernel thread splices to tmpfs as reported by Stephan Kulow: BUG: unable to handle kernel NULL pointer dereference at 0000000000000360 IP: get_mem_cgroup_from_mm.isra.42+0x2b/0x60 Oops: 0000 [#1] SMP Modules linked in: btrfs dm_multipath dm_mod scsi_dh multipath raid10 raid456 async_raid6_recov async_memcpy async_pq raid6_pq async_xor xor async_tx raid1 raid0 md_mod parport_pc parport nls_utf8 isofs usb_storage iscsi_ibft iscsi_boot_sysfs arc4 ecb fan thermal nfs lockd fscache nls_iso8859_1 nls_cp437 sg st hid_generic usbhid af_packet sunrpc sr_mod cdrom ata_generic uhci_hcd virtio_net virtio_blk ehci_hcd usbcore ata_piix floppy processor button usb_common virtio_pci virtio_ring virtio edd squashfs loop ppa] CPU: 0 PID: 97 Comm: loop1 Not tainted 3.15.0-rc5-5-default #1 Hardware name: Bochs Bochs, BIOS Bochs 01/01/2011 Call Trace: __mem_cgroup_try_charge_swapin+0x40/0xe0 mem_cgroup_charge_file+0x8b/0xd0 shmem_getpage_gfp+0x66b/0x7b0 shmem_file_splice_read+0x18f/0x430 splice_direct_to_actor+0xa2/0x1c0 do_lo_receive+0x5a/0x60 [loop] loop_thread+0x298/0x720 [loop] kthread+0xc6/0xe0 ret_from_fork+0x7c/0xb0 Also Branimir Maksimovic reported the following oops which is tiggered for the swapcache charge path from the accounting code for kernel threads: CPU: 1 PID: 160 Comm: kworker/u8:5 Tainted: P OE 3.15.0-rc5-core2-custom #159 Hardware name: System manufacturer System Product Name/MAXIMUSV GENE, BIOS 1903 08/19/2013 task: ffff880404e349b0 ti: ffff88040486a000 task.ti: ffff88040486a000 RIP: get_mem_cgroup_from_mm.isra.42+0x2b/0x60 Call Trace: __mem_cgroup_try_charge_swapin+0x45/0xf0 mem_cgroup_charge_file+0x9c/0xe0 shmem_getpage_gfp+0x62c/0x770 shmem_write_begin+0x38/0x40 generic_perform_write+0xc5/0x1c0 __generic_file_aio_write+0x1d1/0x3f0 generic_file_aio_write+0x4f/0xc0 do_sync_write+0x5a/0x90 do_acct_process+0x4b1/0x550 acct_process+0x6d/0xa0 do_exit+0x827/0xa70 kthread+0xc3/0xf0 This patch fixes the issue by reintroducing mm check into get_mem_cgroup_from_mm. We could do the same trick in __mem_cgroup_try_charge_swapin as we do for the regular page cache path but it is not worth troubles. The check is not that expensive and it is better to have get_mem_cgroup_from_mm more robust. [1] - http://marc.info/?l=linux-mm&m=139463617808941&w=2 Fixes: 03583f1a631c ("memcg: remove unnecessary !mm check from try_get_mem_cgroup_from_mm()") Reported-and-tested-by: Stephan Kulow <coolo@suse.com> Reported-by: Branimir Maksimovic <branimir.maksimovic@gmail.com> Signed-off-by: Michal Hocko <mhocko@suse.cz> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-05-22 18:54:19 +00:00
else {
memcg = mem_cgroup_from_task(rcu_dereference(mm->owner));
if (unlikely(!memcg))
memcg = root_mem_cgroup;
}
mm: memcg: switch to css_tryget() in get_mem_cgroup_from_mm() We've encountered a rcu stall in get_mem_cgroup_from_mm(): rcu: INFO: rcu_sched self-detected stall on CPU rcu: 33-....: (21000 ticks this GP) idle=6c6/1/0x4000000000000002 softirq=35441/35441 fqs=5017 (t=21031 jiffies g=324821 q=95837) NMI backtrace for cpu 33 <...> RIP: 0010:get_mem_cgroup_from_mm+0x2f/0x90 <...> __memcg_kmem_charge+0x55/0x140 __alloc_pages_nodemask+0x267/0x320 pipe_write+0x1ad/0x400 new_sync_write+0x127/0x1c0 __kernel_write+0x4f/0xf0 dump_emit+0x91/0xc0 writenote+0xa0/0xc0 elf_core_dump+0x11af/0x1430 do_coredump+0xc65/0xee0 get_signal+0x132/0x7c0 do_signal+0x36/0x640 exit_to_usermode_loop+0x61/0xd0 do_syscall_64+0xd4/0x100 entry_SYSCALL_64_after_hwframe+0x44/0xa9 The problem is caused by an exiting task which is associated with an offline memcg. We're iterating over and over in the do {} while (!css_tryget_online()) loop, but obviously the memcg won't become online and the exiting task won't be migrated to a live memcg. Let's fix it by switching from css_tryget_online() to css_tryget(). As css_tryget_online() cannot guarantee that the memcg won't go offline, the check is usually useless, except some rare cases when for example it determines if something should be presented to a user. A similar problem is described by commit 18fa84a2db0e ("cgroup: Use css_tryget() instead of css_tryget_online() in task_get_css()"). Johannes: : The bug aside, it doesn't matter whether the cgroup is online for the : callers. It used to matter when offlining needed to evacuate all charges : from the memcg, and so needed to prevent new ones from showing up, but we : don't care now. Link: http://lkml.kernel.org/r/20191106225131.3543616-1-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Tejun Heo <tj@kernel.org> Reviewed-by: Shakeel Butt <shakeeb@google.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Michal Koutn <mkoutny@suse.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-11-16 01:34:43 +00:00
} while (!css_tryget(&memcg->css));
rcu_read_unlock();
return memcg;
}
fs: fsnotify: account fsnotify metadata to kmemcg Patch series "Directed kmem charging", v8. The Linux kernel's memory cgroup allows limiting the memory usage of the jobs running on the system to provide isolation between the jobs. All the kernel memory allocated in the context of the job and marked with __GFP_ACCOUNT will also be included in the memory usage and be limited by the job's limit. The kernel memory can only be charged to the memcg of the process in whose context kernel memory was allocated. However there are cases where the allocated kernel memory should be charged to the memcg different from the current processes's memcg. This patch series contains two such concrete use-cases i.e. fsnotify and buffer_head. The fsnotify event objects can consume a lot of system memory for large or unlimited queues if there is either no or slow listener. The events are allocated in the context of the event producer. However they should be charged to the event consumer. Similarly the buffer_head objects can be allocated in a memcg different from the memcg of the page for which buffer_head objects are being allocated. To solve this issue, this patch series introduces mechanism to charge kernel memory to a given memcg. In case of fsnotify events, the memcg of the consumer can be used for charging and for buffer_head, the memcg of the page can be charged. For directed charging, the caller can use the scope API memalloc_[un]use_memcg() to specify the memcg to charge for all the __GFP_ACCOUNT allocations within the scope. This patch (of 2): A lot of memory can be consumed by the events generated for the huge or unlimited queues if there is either no or slow listener. This can cause system level memory pressure or OOMs. So, it's better to account the fsnotify kmem caches to the memcg of the listener. However the listener can be in a different memcg than the memcg of the producer and these allocations happen in the context of the event producer. This patch introduces remote memcg charging API which the producer can use to charge the allocations to the memcg of the listener. There are seven fsnotify kmem caches and among them allocations from dnotify_struct_cache, dnotify_mark_cache, fanotify_mark_cache and inotify_inode_mark_cachep happens in the context of syscall from the listener. So, SLAB_ACCOUNT is enough for these caches. The objects from fsnotify_mark_connector_cachep are not accounted as they are small compared to the notification mark or events and it is unclear whom to account connector to since it is shared by all events attached to the inode. The allocations from the event caches happen in the context of the event producer. For such caches we will need to remote charge the allocations to the listener's memcg. Thus we save the memcg reference in the fsnotify_group structure of the listener. This patch has also moved the members of fsnotify_group to keep the size same, at least for 64 bit build, even with additional member by filling the holes. [shakeelb@google.com: use GFP_KERNEL_ACCOUNT rather than open-coding it] Link: http://lkml.kernel.org/r/20180702215439.211597-1-shakeelb@google.com Link: http://lkml.kernel.org/r/20180627191250.209150-2-shakeelb@google.com Signed-off-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Amir Goldstein <amir73il@gmail.com> Cc: Greg Thelen <gthelen@google.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:46:39 +00:00
EXPORT_SYMBOL(get_mem_cgroup_from_mm);
fs, mm: account buffer_head to kmemcg The buffer_head can consume a significant amount of system memory and is directly related to the amount of page cache. In our production environment we have observed that a lot of machines are spending a significant amount of memory as buffer_head and can not be left as system memory overhead. Charging buffer_head is not as simple as adding __GFP_ACCOUNT to the allocation. The buffer_heads can be allocated in a memcg different from the memcg of the page for which buffer_heads are being allocated. One concrete example is memory reclaim. The reclaim can trigger I/O of pages of any memcg on the system. So, the right way to charge buffer_head is to extract the memcg from the page for which buffer_heads are being allocated and then use targeted memcg charging API. [shakeelb@google.com: use __GFP_ACCOUNT for directed memcg charging] Link: http://lkml.kernel.org/r/20180702220208.213380-1-shakeelb@google.com Link: http://lkml.kernel.org/r/20180627191250.209150-3-shakeelb@google.com Signed-off-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Amir Goldstein <amir73il@gmail.com> Cc: Greg Thelen <gthelen@google.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:46:44 +00:00
/**
* get_mem_cgroup_from_page: Obtain a reference on given page's memcg.
* @page: page from which memcg should be extracted.
*
* Obtain a reference on page->memcg and returns it if successful. Otherwise
* root_mem_cgroup is returned.
*/
struct mem_cgroup *get_mem_cgroup_from_page(struct page *page)
{
struct mem_cgroup *memcg = page->mem_cgroup;
if (mem_cgroup_disabled())
return NULL;
rcu_read_lock();
memcg: css_tryget_online cleanups Currently multiple locations in memcg code, css_tryget_online() is being used. However it doesn't matter whether the cgroup is online for the callers. Online used to matter when we had reparenting on offlining and we needed a way to prevent new ones from showing up. The failure case for couple of these css_tryget_online usage is to fallback to root_mem_cgroup which kind of make bypassing the memcg limits possible for some workloads. For example creating an inotify group in a subcontainer and then deleting that container after moving the process to a different container will make all the event objects allocated for that group to the root_mem_cgroup. So, using css_tryget_online() is dangerous for such cases. Two locations still use the online version. The swapin of offlined memcg's pages and the memcg kmem cache creation. The kmem cache indeed needs the online version as the kernel does the reparenting of memcg kmem caches. For the swapin case, it has been left for later as the fallback is not really that concerning. With swap accounting enabled, if the memcg of the swapped out page is not online then the memcg extracted from the given 'mm' will be charged and if 'mm' is NULL then root memcg will be charged. However I could not find a code path where the given 'mm' will be NULL for swap-in case. Signed-off-by: Shakeel Butt <shakeelb@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Link: http://lkml.kernel.org/r/20200302203109.179417-1-shakeelb@google.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-04-02 04:07:10 +00:00
/* Page should not get uncharged and freed memcg under us. */
if (!memcg || WARN_ON_ONCE(!css_tryget(&memcg->css)))
fs, mm: account buffer_head to kmemcg The buffer_head can consume a significant amount of system memory and is directly related to the amount of page cache. In our production environment we have observed that a lot of machines are spending a significant amount of memory as buffer_head and can not be left as system memory overhead. Charging buffer_head is not as simple as adding __GFP_ACCOUNT to the allocation. The buffer_heads can be allocated in a memcg different from the memcg of the page for which buffer_heads are being allocated. One concrete example is memory reclaim. The reclaim can trigger I/O of pages of any memcg on the system. So, the right way to charge buffer_head is to extract the memcg from the page for which buffer_heads are being allocated and then use targeted memcg charging API. [shakeelb@google.com: use __GFP_ACCOUNT for directed memcg charging] Link: http://lkml.kernel.org/r/20180702220208.213380-1-shakeelb@google.com Link: http://lkml.kernel.org/r/20180627191250.209150-3-shakeelb@google.com Signed-off-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Amir Goldstein <amir73il@gmail.com> Cc: Greg Thelen <gthelen@google.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:46:44 +00:00
memcg = root_mem_cgroup;
rcu_read_unlock();
return memcg;
}
EXPORT_SYMBOL(get_mem_cgroup_from_page);
fs: fsnotify: account fsnotify metadata to kmemcg Patch series "Directed kmem charging", v8. The Linux kernel's memory cgroup allows limiting the memory usage of the jobs running on the system to provide isolation between the jobs. All the kernel memory allocated in the context of the job and marked with __GFP_ACCOUNT will also be included in the memory usage and be limited by the job's limit. The kernel memory can only be charged to the memcg of the process in whose context kernel memory was allocated. However there are cases where the allocated kernel memory should be charged to the memcg different from the current processes's memcg. This patch series contains two such concrete use-cases i.e. fsnotify and buffer_head. The fsnotify event objects can consume a lot of system memory for large or unlimited queues if there is either no or slow listener. The events are allocated in the context of the event producer. However they should be charged to the event consumer. Similarly the buffer_head objects can be allocated in a memcg different from the memcg of the page for which buffer_head objects are being allocated. To solve this issue, this patch series introduces mechanism to charge kernel memory to a given memcg. In case of fsnotify events, the memcg of the consumer can be used for charging and for buffer_head, the memcg of the page can be charged. For directed charging, the caller can use the scope API memalloc_[un]use_memcg() to specify the memcg to charge for all the __GFP_ACCOUNT allocations within the scope. This patch (of 2): A lot of memory can be consumed by the events generated for the huge or unlimited queues if there is either no or slow listener. This can cause system level memory pressure or OOMs. So, it's better to account the fsnotify kmem caches to the memcg of the listener. However the listener can be in a different memcg than the memcg of the producer and these allocations happen in the context of the event producer. This patch introduces remote memcg charging API which the producer can use to charge the allocations to the memcg of the listener. There are seven fsnotify kmem caches and among them allocations from dnotify_struct_cache, dnotify_mark_cache, fanotify_mark_cache and inotify_inode_mark_cachep happens in the context of syscall from the listener. So, SLAB_ACCOUNT is enough for these caches. The objects from fsnotify_mark_connector_cachep are not accounted as they are small compared to the notification mark or events and it is unclear whom to account connector to since it is shared by all events attached to the inode. The allocations from the event caches happen in the context of the event producer. For such caches we will need to remote charge the allocations to the listener's memcg. Thus we save the memcg reference in the fsnotify_group structure of the listener. This patch has also moved the members of fsnotify_group to keep the size same, at least for 64 bit build, even with additional member by filling the holes. [shakeelb@google.com: use GFP_KERNEL_ACCOUNT rather than open-coding it] Link: http://lkml.kernel.org/r/20180702215439.211597-1-shakeelb@google.com Link: http://lkml.kernel.org/r/20180627191250.209150-2-shakeelb@google.com Signed-off-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Amir Goldstein <amir73il@gmail.com> Cc: Greg Thelen <gthelen@google.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:46:39 +00:00
/**
* If current->active_memcg is non-NULL, do not fallback to current->mm->memcg.
*/
static __always_inline struct mem_cgroup *get_mem_cgroup_from_current(void)
{
if (unlikely(current->active_memcg)) {
memcg: css_tryget_online cleanups Currently multiple locations in memcg code, css_tryget_online() is being used. However it doesn't matter whether the cgroup is online for the callers. Online used to matter when we had reparenting on offlining and we needed a way to prevent new ones from showing up. The failure case for couple of these css_tryget_online usage is to fallback to root_mem_cgroup which kind of make bypassing the memcg limits possible for some workloads. For example creating an inotify group in a subcontainer and then deleting that container after moving the process to a different container will make all the event objects allocated for that group to the root_mem_cgroup. So, using css_tryget_online() is dangerous for such cases. Two locations still use the online version. The swapin of offlined memcg's pages and the memcg kmem cache creation. The kmem cache indeed needs the online version as the kernel does the reparenting of memcg kmem caches. For the swapin case, it has been left for later as the fallback is not really that concerning. With swap accounting enabled, if the memcg of the swapped out page is not online then the memcg extracted from the given 'mm' will be charged and if 'mm' is NULL then root memcg will be charged. However I could not find a code path where the given 'mm' will be NULL for swap-in case. Signed-off-by: Shakeel Butt <shakeelb@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Link: http://lkml.kernel.org/r/20200302203109.179417-1-shakeelb@google.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-04-02 04:07:10 +00:00
struct mem_cgroup *memcg;
fs: fsnotify: account fsnotify metadata to kmemcg Patch series "Directed kmem charging", v8. The Linux kernel's memory cgroup allows limiting the memory usage of the jobs running on the system to provide isolation between the jobs. All the kernel memory allocated in the context of the job and marked with __GFP_ACCOUNT will also be included in the memory usage and be limited by the job's limit. The kernel memory can only be charged to the memcg of the process in whose context kernel memory was allocated. However there are cases where the allocated kernel memory should be charged to the memcg different from the current processes's memcg. This patch series contains two such concrete use-cases i.e. fsnotify and buffer_head. The fsnotify event objects can consume a lot of system memory for large or unlimited queues if there is either no or slow listener. The events are allocated in the context of the event producer. However they should be charged to the event consumer. Similarly the buffer_head objects can be allocated in a memcg different from the memcg of the page for which buffer_head objects are being allocated. To solve this issue, this patch series introduces mechanism to charge kernel memory to a given memcg. In case of fsnotify events, the memcg of the consumer can be used for charging and for buffer_head, the memcg of the page can be charged. For directed charging, the caller can use the scope API memalloc_[un]use_memcg() to specify the memcg to charge for all the __GFP_ACCOUNT allocations within the scope. This patch (of 2): A lot of memory can be consumed by the events generated for the huge or unlimited queues if there is either no or slow listener. This can cause system level memory pressure or OOMs. So, it's better to account the fsnotify kmem caches to the memcg of the listener. However the listener can be in a different memcg than the memcg of the producer and these allocations happen in the context of the event producer. This patch introduces remote memcg charging API which the producer can use to charge the allocations to the memcg of the listener. There are seven fsnotify kmem caches and among them allocations from dnotify_struct_cache, dnotify_mark_cache, fanotify_mark_cache and inotify_inode_mark_cachep happens in the context of syscall from the listener. So, SLAB_ACCOUNT is enough for these caches. The objects from fsnotify_mark_connector_cachep are not accounted as they are small compared to the notification mark or events and it is unclear whom to account connector to since it is shared by all events attached to the inode. The allocations from the event caches happen in the context of the event producer. For such caches we will need to remote charge the allocations to the listener's memcg. Thus we save the memcg reference in the fsnotify_group structure of the listener. This patch has also moved the members of fsnotify_group to keep the size same, at least for 64 bit build, even with additional member by filling the holes. [shakeelb@google.com: use GFP_KERNEL_ACCOUNT rather than open-coding it] Link: http://lkml.kernel.org/r/20180702215439.211597-1-shakeelb@google.com Link: http://lkml.kernel.org/r/20180627191250.209150-2-shakeelb@google.com Signed-off-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Amir Goldstein <amir73il@gmail.com> Cc: Greg Thelen <gthelen@google.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:46:39 +00:00
rcu_read_lock();
memcg: css_tryget_online cleanups Currently multiple locations in memcg code, css_tryget_online() is being used. However it doesn't matter whether the cgroup is online for the callers. Online used to matter when we had reparenting on offlining and we needed a way to prevent new ones from showing up. The failure case for couple of these css_tryget_online usage is to fallback to root_mem_cgroup which kind of make bypassing the memcg limits possible for some workloads. For example creating an inotify group in a subcontainer and then deleting that container after moving the process to a different container will make all the event objects allocated for that group to the root_mem_cgroup. So, using css_tryget_online() is dangerous for such cases. Two locations still use the online version. The swapin of offlined memcg's pages and the memcg kmem cache creation. The kmem cache indeed needs the online version as the kernel does the reparenting of memcg kmem caches. For the swapin case, it has been left for later as the fallback is not really that concerning. With swap accounting enabled, if the memcg of the swapped out page is not online then the memcg extracted from the given 'mm' will be charged and if 'mm' is NULL then root memcg will be charged. However I could not find a code path where the given 'mm' will be NULL for swap-in case. Signed-off-by: Shakeel Butt <shakeelb@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Link: http://lkml.kernel.org/r/20200302203109.179417-1-shakeelb@google.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-04-02 04:07:10 +00:00
/* current->active_memcg must hold a ref. */
if (WARN_ON_ONCE(!css_tryget(&current->active_memcg->css)))
memcg = root_mem_cgroup;
else
fs: fsnotify: account fsnotify metadata to kmemcg Patch series "Directed kmem charging", v8. The Linux kernel's memory cgroup allows limiting the memory usage of the jobs running on the system to provide isolation between the jobs. All the kernel memory allocated in the context of the job and marked with __GFP_ACCOUNT will also be included in the memory usage and be limited by the job's limit. The kernel memory can only be charged to the memcg of the process in whose context kernel memory was allocated. However there are cases where the allocated kernel memory should be charged to the memcg different from the current processes's memcg. This patch series contains two such concrete use-cases i.e. fsnotify and buffer_head. The fsnotify event objects can consume a lot of system memory for large or unlimited queues if there is either no or slow listener. The events are allocated in the context of the event producer. However they should be charged to the event consumer. Similarly the buffer_head objects can be allocated in a memcg different from the memcg of the page for which buffer_head objects are being allocated. To solve this issue, this patch series introduces mechanism to charge kernel memory to a given memcg. In case of fsnotify events, the memcg of the consumer can be used for charging and for buffer_head, the memcg of the page can be charged. For directed charging, the caller can use the scope API memalloc_[un]use_memcg() to specify the memcg to charge for all the __GFP_ACCOUNT allocations within the scope. This patch (of 2): A lot of memory can be consumed by the events generated for the huge or unlimited queues if there is either no or slow listener. This can cause system level memory pressure or OOMs. So, it's better to account the fsnotify kmem caches to the memcg of the listener. However the listener can be in a different memcg than the memcg of the producer and these allocations happen in the context of the event producer. This patch introduces remote memcg charging API which the producer can use to charge the allocations to the memcg of the listener. There are seven fsnotify kmem caches and among them allocations from dnotify_struct_cache, dnotify_mark_cache, fanotify_mark_cache and inotify_inode_mark_cachep happens in the context of syscall from the listener. So, SLAB_ACCOUNT is enough for these caches. The objects from fsnotify_mark_connector_cachep are not accounted as they are small compared to the notification mark or events and it is unclear whom to account connector to since it is shared by all events attached to the inode. The allocations from the event caches happen in the context of the event producer. For such caches we will need to remote charge the allocations to the listener's memcg. Thus we save the memcg reference in the fsnotify_group structure of the listener. This patch has also moved the members of fsnotify_group to keep the size same, at least for 64 bit build, even with additional member by filling the holes. [shakeelb@google.com: use GFP_KERNEL_ACCOUNT rather than open-coding it] Link: http://lkml.kernel.org/r/20180702215439.211597-1-shakeelb@google.com Link: http://lkml.kernel.org/r/20180627191250.209150-2-shakeelb@google.com Signed-off-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Amir Goldstein <amir73il@gmail.com> Cc: Greg Thelen <gthelen@google.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:46:39 +00:00
memcg = current->active_memcg;
rcu_read_unlock();
return memcg;
}
return get_mem_cgroup_from_mm(current->mm);
}
/**
* mem_cgroup_iter - iterate over memory cgroup hierarchy
* @root: hierarchy root
* @prev: previously returned memcg, NULL on first invocation
* @reclaim: cookie for shared reclaim walks, NULL for full walks
*
* Returns references to children of the hierarchy below @root, or
* @root itself, or %NULL after a full round-trip.
*
* Caller must pass the return value in @prev on subsequent
* invocations for reference counting, or use mem_cgroup_iter_break()
* to cancel a hierarchy walk before the round-trip is complete.
*
* Reclaimers can specify a node and a priority level in @reclaim to
* divide up the memcgs in the hierarchy among all concurrent
* reclaimers operating on the same node and priority.
*/
struct mem_cgroup *mem_cgroup_iter(struct mem_cgroup *root,
struct mem_cgroup *prev,
struct mem_cgroup_reclaim_cookie *reclaim)
{
struct mem_cgroup_reclaim_iter *uninitialized_var(iter);
struct cgroup_subsys_state *css = NULL;
mm: memcg: consolidate hierarchy iteration primitives The memcg naturalization series: Memory control groups are currently bolted onto the side of traditional memory management in places where better integration would be preferrable. To reclaim memory, for example, memory control groups maintain their own LRU list and reclaim strategy aside from the global per-zone LRU list reclaim. But an extra list head for each existing page frame is expensive and maintaining it requires additional code. This patchset disables the global per-zone LRU lists on memory cgroup configurations and converts all its users to operate on the per-memory cgroup lists instead. As LRU pages are then exclusively on one list, this saves two list pointers for each page frame in the system: page_cgroup array size with 4G physical memory vanilla: allocated 31457280 bytes of page_cgroup patched: allocated 15728640 bytes of page_cgroup At the same time, system performance for various workloads is unaffected: 100G sparse file cat, 4G physical memory, 10 runs, to test for code bloat in the traditional LRU handling and kswapd & direct reclaim paths, without/with the memory controller configured in vanilla: 71.603(0.207) seconds patched: 71.640(0.156) seconds vanilla: 79.558(0.288) seconds patched: 77.233(0.147) seconds 100G sparse file cat in 1G memory cgroup, 10 runs, to test for code bloat in the traditional memory cgroup LRU handling and reclaim path vanilla: 96.844(0.281) seconds patched: 94.454(0.311) seconds 4 unlimited memcgs running kbuild -j32 each, 4G physical memory, 500M swap on SSD, 10 runs, to test for regressions in kswapd & direct reclaim using per-memcg LRU lists with multiple memcgs and multiple allocators within each memcg vanilla: 717.722(1.440) seconds [ 69720.100(11600.835) majfaults ] patched: 714.106(2.313) seconds [ 71109.300(14886.186) majfaults ] 16 unlimited memcgs running kbuild, 1900M hierarchical limit, 500M swap on SSD, 10 runs, to test for regressions in hierarchical memcg setups vanilla: 2742.058(1.992) seconds [ 26479.600(1736.737) majfaults ] patched: 2743.267(1.214) seconds [ 27240.700(1076.063) majfaults ] This patch: There are currently two different implementations of iterating over a memory cgroup hierarchy tree. Consolidate them into one worker function and base the convenience looping-macros on top of it. Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Reviewed-by: Kirill A. Shutemov <kirill@shutemov.name> Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Ying Han <yinghan@google.com> Cc: Greg Thelen <gthelen@google.com> Cc: Michel Lespinasse <walken@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Minchan Kim <minchan.kim@gmail.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-13 01:17:48 +00:00
struct mem_cgroup *memcg = NULL;
struct mem_cgroup *pos = NULL;
if (mem_cgroup_disabled())
return NULL;
mm: memcg: consolidate hierarchy iteration primitives The memcg naturalization series: Memory control groups are currently bolted onto the side of traditional memory management in places where better integration would be preferrable. To reclaim memory, for example, memory control groups maintain their own LRU list and reclaim strategy aside from the global per-zone LRU list reclaim. But an extra list head for each existing page frame is expensive and maintaining it requires additional code. This patchset disables the global per-zone LRU lists on memory cgroup configurations and converts all its users to operate on the per-memory cgroup lists instead. As LRU pages are then exclusively on one list, this saves two list pointers for each page frame in the system: page_cgroup array size with 4G physical memory vanilla: allocated 31457280 bytes of page_cgroup patched: allocated 15728640 bytes of page_cgroup At the same time, system performance for various workloads is unaffected: 100G sparse file cat, 4G physical memory, 10 runs, to test for code bloat in the traditional LRU handling and kswapd & direct reclaim paths, without/with the memory controller configured in vanilla: 71.603(0.207) seconds patched: 71.640(0.156) seconds vanilla: 79.558(0.288) seconds patched: 77.233(0.147) seconds 100G sparse file cat in 1G memory cgroup, 10 runs, to test for code bloat in the traditional memory cgroup LRU handling and reclaim path vanilla: 96.844(0.281) seconds patched: 94.454(0.311) seconds 4 unlimited memcgs running kbuild -j32 each, 4G physical memory, 500M swap on SSD, 10 runs, to test for regressions in kswapd & direct reclaim using per-memcg LRU lists with multiple memcgs and multiple allocators within each memcg vanilla: 717.722(1.440) seconds [ 69720.100(11600.835) majfaults ] patched: 714.106(2.313) seconds [ 71109.300(14886.186) majfaults ] 16 unlimited memcgs running kbuild, 1900M hierarchical limit, 500M swap on SSD, 10 runs, to test for regressions in hierarchical memcg setups vanilla: 2742.058(1.992) seconds [ 26479.600(1736.737) majfaults ] patched: 2743.267(1.214) seconds [ 27240.700(1076.063) majfaults ] This patch: There are currently two different implementations of iterating over a memory cgroup hierarchy tree. Consolidate them into one worker function and base the convenience looping-macros on top of it. Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Reviewed-by: Kirill A. Shutemov <kirill@shutemov.name> Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Ying Han <yinghan@google.com> Cc: Greg Thelen <gthelen@google.com> Cc: Michel Lespinasse <walken@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Minchan Kim <minchan.kim@gmail.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-13 01:17:48 +00:00
if (!root)
root = root_mem_cgroup;
mm: memcg: consolidate hierarchy iteration primitives The memcg naturalization series: Memory control groups are currently bolted onto the side of traditional memory management in places where better integration would be preferrable. To reclaim memory, for example, memory control groups maintain their own LRU list and reclaim strategy aside from the global per-zone LRU list reclaim. But an extra list head for each existing page frame is expensive and maintaining it requires additional code. This patchset disables the global per-zone LRU lists on memory cgroup configurations and converts all its users to operate on the per-memory cgroup lists instead. As LRU pages are then exclusively on one list, this saves two list pointers for each page frame in the system: page_cgroup array size with 4G physical memory vanilla: allocated 31457280 bytes of page_cgroup patched: allocated 15728640 bytes of page_cgroup At the same time, system performance for various workloads is unaffected: 100G sparse file cat, 4G physical memory, 10 runs, to test for code bloat in the traditional LRU handling and kswapd & direct reclaim paths, without/with the memory controller configured in vanilla: 71.603(0.207) seconds patched: 71.640(0.156) seconds vanilla: 79.558(0.288) seconds patched: 77.233(0.147) seconds 100G sparse file cat in 1G memory cgroup, 10 runs, to test for code bloat in the traditional memory cgroup LRU handling and reclaim path vanilla: 96.844(0.281) seconds patched: 94.454(0.311) seconds 4 unlimited memcgs running kbuild -j32 each, 4G physical memory, 500M swap on SSD, 10 runs, to test for regressions in kswapd & direct reclaim using per-memcg LRU lists with multiple memcgs and multiple allocators within each memcg vanilla: 717.722(1.440) seconds [ 69720.100(11600.835) majfaults ] patched: 714.106(2.313) seconds [ 71109.300(14886.186) majfaults ] 16 unlimited memcgs running kbuild, 1900M hierarchical limit, 500M swap on SSD, 10 runs, to test for regressions in hierarchical memcg setups vanilla: 2742.058(1.992) seconds [ 26479.600(1736.737) majfaults ] patched: 2743.267(1.214) seconds [ 27240.700(1076.063) majfaults ] This patch: There are currently two different implementations of iterating over a memory cgroup hierarchy tree. Consolidate them into one worker function and base the convenience looping-macros on top of it. Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Reviewed-by: Kirill A. Shutemov <kirill@shutemov.name> Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Ying Han <yinghan@google.com> Cc: Greg Thelen <gthelen@google.com> Cc: Michel Lespinasse <walken@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Minchan Kim <minchan.kim@gmail.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-13 01:17:48 +00:00
if (prev && !reclaim)
pos = prev;
mm: memcg: consolidate hierarchy iteration primitives The memcg naturalization series: Memory control groups are currently bolted onto the side of traditional memory management in places where better integration would be preferrable. To reclaim memory, for example, memory control groups maintain their own LRU list and reclaim strategy aside from the global per-zone LRU list reclaim. But an extra list head for each existing page frame is expensive and maintaining it requires additional code. This patchset disables the global per-zone LRU lists on memory cgroup configurations and converts all its users to operate on the per-memory cgroup lists instead. As LRU pages are then exclusively on one list, this saves two list pointers for each page frame in the system: page_cgroup array size with 4G physical memory vanilla: allocated 31457280 bytes of page_cgroup patched: allocated 15728640 bytes of page_cgroup At the same time, system performance for various workloads is unaffected: 100G sparse file cat, 4G physical memory, 10 runs, to test for code bloat in the traditional LRU handling and kswapd & direct reclaim paths, without/with the memory controller configured in vanilla: 71.603(0.207) seconds patched: 71.640(0.156) seconds vanilla: 79.558(0.288) seconds patched: 77.233(0.147) seconds 100G sparse file cat in 1G memory cgroup, 10 runs, to test for code bloat in the traditional memory cgroup LRU handling and reclaim path vanilla: 96.844(0.281) seconds patched: 94.454(0.311) seconds 4 unlimited memcgs running kbuild -j32 each, 4G physical memory, 500M swap on SSD, 10 runs, to test for regressions in kswapd & direct reclaim using per-memcg LRU lists with multiple memcgs and multiple allocators within each memcg vanilla: 717.722(1.440) seconds [ 69720.100(11600.835) majfaults ] patched: 714.106(2.313) seconds [ 71109.300(14886.186) majfaults ] 16 unlimited memcgs running kbuild, 1900M hierarchical limit, 500M swap on SSD, 10 runs, to test for regressions in hierarchical memcg setups vanilla: 2742.058(1.992) seconds [ 26479.600(1736.737) majfaults ] patched: 2743.267(1.214) seconds [ 27240.700(1076.063) majfaults ] This patch: There are currently two different implementations of iterating over a memory cgroup hierarchy tree. Consolidate them into one worker function and base the convenience looping-macros on top of it. Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Reviewed-by: Kirill A. Shutemov <kirill@shutemov.name> Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Ying Han <yinghan@google.com> Cc: Greg Thelen <gthelen@google.com> Cc: Michel Lespinasse <walken@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Minchan Kim <minchan.kim@gmail.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-13 01:17:48 +00:00
if (!root->use_hierarchy && root != root_mem_cgroup) {
if (prev)
goto out;
return root;
mm: memcg: consolidate hierarchy iteration primitives The memcg naturalization series: Memory control groups are currently bolted onto the side of traditional memory management in places where better integration would be preferrable. To reclaim memory, for example, memory control groups maintain their own LRU list and reclaim strategy aside from the global per-zone LRU list reclaim. But an extra list head for each existing page frame is expensive and maintaining it requires additional code. This patchset disables the global per-zone LRU lists on memory cgroup configurations and converts all its users to operate on the per-memory cgroup lists instead. As LRU pages are then exclusively on one list, this saves two list pointers for each page frame in the system: page_cgroup array size with 4G physical memory vanilla: allocated 31457280 bytes of page_cgroup patched: allocated 15728640 bytes of page_cgroup At the same time, system performance for various workloads is unaffected: 100G sparse file cat, 4G physical memory, 10 runs, to test for code bloat in the traditional LRU handling and kswapd & direct reclaim paths, without/with the memory controller configured in vanilla: 71.603(0.207) seconds patched: 71.640(0.156) seconds vanilla: 79.558(0.288) seconds patched: 77.233(0.147) seconds 100G sparse file cat in 1G memory cgroup, 10 runs, to test for code bloat in the traditional memory cgroup LRU handling and reclaim path vanilla: 96.844(0.281) seconds patched: 94.454(0.311) seconds 4 unlimited memcgs running kbuild -j32 each, 4G physical memory, 500M swap on SSD, 10 runs, to test for regressions in kswapd & direct reclaim using per-memcg LRU lists with multiple memcgs and multiple allocators within each memcg vanilla: 717.722(1.440) seconds [ 69720.100(11600.835) majfaults ] patched: 714.106(2.313) seconds [ 71109.300(14886.186) majfaults ] 16 unlimited memcgs running kbuild, 1900M hierarchical limit, 500M swap on SSD, 10 runs, to test for regressions in hierarchical memcg setups vanilla: 2742.058(1.992) seconds [ 26479.600(1736.737) majfaults ] patched: 2743.267(1.214) seconds [ 27240.700(1076.063) majfaults ] This patch: There are currently two different implementations of iterating over a memory cgroup hierarchy tree. Consolidate them into one worker function and base the convenience looping-macros on top of it. Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Reviewed-by: Kirill A. Shutemov <kirill@shutemov.name> Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Ying Han <yinghan@google.com> Cc: Greg Thelen <gthelen@google.com> Cc: Michel Lespinasse <walken@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Minchan Kim <minchan.kim@gmail.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-13 01:17:48 +00:00
}
memcg: rework mem_cgroup_iter to use cgroup iterators mem_cgroup_iter curently relies on css->id when walking down a group hierarchy tree. This is really awkward because the tree walk depends on the groups creation ordering. The only guarantee is that a parent node is visited before its children. Example: 1) mkdir -p a a/d a/b/c 2) mkdir -a a/b/c a/d Will create the same trees but the tree walks will be different: 1) a, d, b, c 2) a, b, c, d Commit 574bd9f7c7c1 ("cgroup: implement generic child / descendant walk macros") has introduced generic cgroup tree walkers which provide either pre-order or post-order tree walk. This patch converts css->id based iteration to pre-order tree walk to keep the semantic with the original iterator where parent is always visited before its subtree. cgroup_for_each_descendant_pre suggests using post_create and pre_destroy for proper synchronization with groups addidition resp. removal. This implementation doesn't use those because a new memory cgroup is initialized sufficiently for iteration in mem_cgroup_css_alloc already and css reference counting enforces that the group is alive for both the last seen cgroup and the found one resp. it signals that the group is dead and it should be skipped. If the reclaim cookie is used we need to store the last visited group into the iterator so we have to be careful that it doesn't disappear in the mean time. Elevated reference count on the css keeps it alive even though the group have been removed (parked waiting for the last dput so that it can be freed). Per node-zone-prio iter_lock has been introduced to ensure that css_tryget and iter->last_visited is set atomically. Otherwise two racing walkers could both take a references and only one release it leading to a css leak (which pins cgroup dentry). Signed-off-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizefan@huawei.com> Cc: Ying Han <yinghan@google.com> Cc: Tejun Heo <htejun@gmail.com> Cc: Glauber Costa <glommer@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-04-29 22:07:15 +00:00
rcu_read_lock();
memcg: relax memcg iter caching Now that the per-node-zone-priority iterator caches memory cgroups rather than their css ids we have to be careful and remove them from the iterator when they are on the way out otherwise they might live for unbounded amount of time even though their group is already gone (until the global/targeted reclaim triggers the zone under priority to find out the group is dead and let it to find the final rest). We can fix this issue by relaxing rules for the last_visited memcg. Instead of taking a reference to the css before it is stored into iter->last_visited we can just store its pointer and track the number of removed groups from each memcg's subhierarchy. This number would be stored into iterator everytime when a memcg is cached. If the iter count doesn't match the curent walker root's one we will start from the root again. The group counter is incremented upwards the hierarchy every time a group is removed. The iter_lock can be dropped because racing iterators cannot leak the reference anymore as the reference count is not elevated for last_visited when it is cached. Locking rules got a bit complicated by this change though. The iterator primarily relies on rcu read lock which makes sure that once we see a valid last_visited pointer then it will be valid for the whole RCU walk. smp_rmb makes sure that dead_count is read before last_visited and last_dead_count while smp_wmb makes sure that last_visited is updated before last_dead_count so the up-to-date last_dead_count cannot point to an outdated last_visited. css_tryget then makes sure that the last_visited is still alive in case the iteration races with the cached group removal (css is invalidated before mem_cgroup_css_offline increments dead_count). In short: mem_cgroup_iter rcu_read_lock() dead_count = atomic_read(parent->dead_count) smp_rmb() if (dead_count != iter->last_dead_count) last_visited POSSIBLY INVALID -> last_visited = NULL if (!css_tryget(iter->last_visited)) last_visited DEAD -> last_visited = NULL next = find_next(last_visited) css_tryget(next) css_put(last_visited) // css would be invalidated and parent->dead_count // incremented if this was the last reference iter->last_visited = next smp_wmb() iter->last_dead_count = dead_count rcu_read_unlock() cgroup_rmdir cgroup_destroy_locked atomic_add(CSS_DEACT_BIAS, &css->refcnt) // subsequent css_tryget fail mem_cgroup_css_offline mem_cgroup_invalidate_reclaim_iterators while(parent = parent_mem_cgroup) atomic_inc(parent->dead_count) css_put(css) // last reference held by cgroup core Spotted by Ying Han. Original idea from Johannes Weiner. [akpm@linux-foundation.org: coding-style fixes] Signed-off-by: Michal Hocko <mhocko@suse.cz> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Ying Han <yinghan@google.com> Cc: Li Zefan <lizefan@huawei.com> Cc: Tejun Heo <htejun@gmail.com> Cc: Glauber Costa <glommer@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-04-29 22:07:17 +00:00
if (reclaim) {
struct mem_cgroup_per_node *mz;
mz = mem_cgroup_nodeinfo(root, reclaim->pgdat->node_id);
iter = &mz->iter;
if (prev && reclaim->generation != iter->generation)
goto out_unlock;
mm: memcontrol: fix possible memcg leak due to interrupted reclaim Memory cgroup reclaim can be interrupted with mem_cgroup_iter_break() once enough pages have been reclaimed, in which case, in contrast to a full round-trip over a cgroup sub-tree, the current position stored in mem_cgroup_reclaim_iter of the target cgroup does not get invalidated and so is left holding the reference to the last scanned cgroup. If the target cgroup does not get scanned again (we might have just reclaimed the last page or all processes might exit and free their memory voluntary), we will leak it, because there is nobody to put the reference held by the iterator. The problem is easy to reproduce by running the following command sequence in a loop: mkdir /sys/fs/cgroup/memory/test echo 100M > /sys/fs/cgroup/memory/test/memory.limit_in_bytes echo $$ > /sys/fs/cgroup/memory/test/cgroup.procs memhog 150M echo $$ > /sys/fs/cgroup/memory/cgroup.procs rmdir test The cgroups generated by it will never get freed. This patch fixes this issue by making mem_cgroup_iter avoid taking reference to the current position. In order not to hit use-after-free bug while running reclaim in parallel with cgroup deletion, we make use of ->css_released cgroup callback to clear references to the dying cgroup in all reclaim iterators that might refer to it. This callback is called right before scheduling rcu work which will free css, so if we access iter->position from rcu read section, we might be sure it won't go away under us. [hannes@cmpxchg.org: clean up css ref handling] Fixes: 5ac8fb31ad2e ("mm: memcontrol: convert reclaim iterator to simple css refcounting") Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-12-29 22:54:10 +00:00
while (1) {
pos = READ_ONCE(iter->position);
mm: memcontrol: fix possible memcg leak due to interrupted reclaim Memory cgroup reclaim can be interrupted with mem_cgroup_iter_break() once enough pages have been reclaimed, in which case, in contrast to a full round-trip over a cgroup sub-tree, the current position stored in mem_cgroup_reclaim_iter of the target cgroup does not get invalidated and so is left holding the reference to the last scanned cgroup. If the target cgroup does not get scanned again (we might have just reclaimed the last page or all processes might exit and free their memory voluntary), we will leak it, because there is nobody to put the reference held by the iterator. The problem is easy to reproduce by running the following command sequence in a loop: mkdir /sys/fs/cgroup/memory/test echo 100M > /sys/fs/cgroup/memory/test/memory.limit_in_bytes echo $$ > /sys/fs/cgroup/memory/test/cgroup.procs memhog 150M echo $$ > /sys/fs/cgroup/memory/cgroup.procs rmdir test The cgroups generated by it will never get freed. This patch fixes this issue by making mem_cgroup_iter avoid taking reference to the current position. In order not to hit use-after-free bug while running reclaim in parallel with cgroup deletion, we make use of ->css_released cgroup callback to clear references to the dying cgroup in all reclaim iterators that might refer to it. This callback is called right before scheduling rcu work which will free css, so if we access iter->position from rcu read section, we might be sure it won't go away under us. [hannes@cmpxchg.org: clean up css ref handling] Fixes: 5ac8fb31ad2e ("mm: memcontrol: convert reclaim iterator to simple css refcounting") Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-12-29 22:54:10 +00:00
if (!pos || css_tryget(&pos->css))
break;
/*
mm: memcontrol: fix possible memcg leak due to interrupted reclaim Memory cgroup reclaim can be interrupted with mem_cgroup_iter_break() once enough pages have been reclaimed, in which case, in contrast to a full round-trip over a cgroup sub-tree, the current position stored in mem_cgroup_reclaim_iter of the target cgroup does not get invalidated and so is left holding the reference to the last scanned cgroup. If the target cgroup does not get scanned again (we might have just reclaimed the last page or all processes might exit and free their memory voluntary), we will leak it, because there is nobody to put the reference held by the iterator. The problem is easy to reproduce by running the following command sequence in a loop: mkdir /sys/fs/cgroup/memory/test echo 100M > /sys/fs/cgroup/memory/test/memory.limit_in_bytes echo $$ > /sys/fs/cgroup/memory/test/cgroup.procs memhog 150M echo $$ > /sys/fs/cgroup/memory/cgroup.procs rmdir test The cgroups generated by it will never get freed. This patch fixes this issue by making mem_cgroup_iter avoid taking reference to the current position. In order not to hit use-after-free bug while running reclaim in parallel with cgroup deletion, we make use of ->css_released cgroup callback to clear references to the dying cgroup in all reclaim iterators that might refer to it. This callback is called right before scheduling rcu work which will free css, so if we access iter->position from rcu read section, we might be sure it won't go away under us. [hannes@cmpxchg.org: clean up css ref handling] Fixes: 5ac8fb31ad2e ("mm: memcontrol: convert reclaim iterator to simple css refcounting") Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-12-29 22:54:10 +00:00
* css reference reached zero, so iter->position will
* be cleared by ->css_released. However, we should not
* rely on this happening soon, because ->css_released
* is called from a work queue, and by busy-waiting we
* might block it. So we clear iter->position right
* away.
*/
mm: memcontrol: fix possible memcg leak due to interrupted reclaim Memory cgroup reclaim can be interrupted with mem_cgroup_iter_break() once enough pages have been reclaimed, in which case, in contrast to a full round-trip over a cgroup sub-tree, the current position stored in mem_cgroup_reclaim_iter of the target cgroup does not get invalidated and so is left holding the reference to the last scanned cgroup. If the target cgroup does not get scanned again (we might have just reclaimed the last page or all processes might exit and free their memory voluntary), we will leak it, because there is nobody to put the reference held by the iterator. The problem is easy to reproduce by running the following command sequence in a loop: mkdir /sys/fs/cgroup/memory/test echo 100M > /sys/fs/cgroup/memory/test/memory.limit_in_bytes echo $$ > /sys/fs/cgroup/memory/test/cgroup.procs memhog 150M echo $$ > /sys/fs/cgroup/memory/cgroup.procs rmdir test The cgroups generated by it will never get freed. This patch fixes this issue by making mem_cgroup_iter avoid taking reference to the current position. In order not to hit use-after-free bug while running reclaim in parallel with cgroup deletion, we make use of ->css_released cgroup callback to clear references to the dying cgroup in all reclaim iterators that might refer to it. This callback is called right before scheduling rcu work which will free css, so if we access iter->position from rcu read section, we might be sure it won't go away under us. [hannes@cmpxchg.org: clean up css ref handling] Fixes: 5ac8fb31ad2e ("mm: memcontrol: convert reclaim iterator to simple css refcounting") Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-12-29 22:54:10 +00:00
(void)cmpxchg(&iter->position, pos, NULL);
}
}
if (pos)
css = &pos->css;
for (;;) {
css = css_next_descendant_pre(css, &root->css);
if (!css) {
/*
* Reclaimers share the hierarchy walk, and a
* new one might jump in right at the end of
* the hierarchy - make sure they see at least
* one group and restart from the beginning.
*/
if (!prev)
continue;
break;
mm: memcg: per-priority per-zone hierarchy scan generations Memory cgroup limit reclaim currently picks one memory cgroup out of the target hierarchy, remembers it as the last scanned child, and reclaims all zones in it with decreasing priority levels. The new hierarchy reclaim code will pick memory cgroups from the same hierarchy concurrently from different zones and priority levels, it becomes necessary that hierarchy roots not only remember the last scanned child, but do so for each zone and priority level. Until now, we reclaimed memcgs like this: mem = mem_cgroup_iter(root) for each priority level: for each zone in zonelist: reclaim(mem, zone) But subsequent patches will move the memcg iteration inside the loop over the zones: for each priority level: for each zone in zonelist: mem = mem_cgroup_iter(root) reclaim(mem, zone) And to keep with the original scan order - memcg -> priority -> zone - the last scanned memcg has to be remembered per zone and per priority level. Furthermore, global reclaim will be switched to the hierarchy walk as well. Different from limit reclaim, which can just recheck the limit after some reclaim progress, its target is to scan all memcgs for the desired zone pages, proportional to the memcg size, and so reliably detecting a full hierarchy round-trip will become crucial. Currently, the code relies on one reclaimer encountering the same memcg twice, but that is error-prone with concurrent reclaimers. Instead, use a generation counter that is increased every time the child with the highest ID has been visited, so that reclaimers can stop when the generation changes. Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: Kirill A. Shutemov <kirill@shutemov.name> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Ying Han <yinghan@google.com> Cc: Greg Thelen <gthelen@google.com> Cc: Michel Lespinasse <walken@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Minchan Kim <minchan.kim@gmail.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-13 01:17:55 +00:00
}
/*
* Verify the css and acquire a reference. The root
* is provided by the caller, so we know it's alive
* and kicking, and don't take an extra reference.
*/
memcg = mem_cgroup_from_css(css);
if (css == &root->css)
break;
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
if (css_tryget(css))
break;
mm: memcg: consolidate hierarchy iteration primitives The memcg naturalization series: Memory control groups are currently bolted onto the side of traditional memory management in places where better integration would be preferrable. To reclaim memory, for example, memory control groups maintain their own LRU list and reclaim strategy aside from the global per-zone LRU list reclaim. But an extra list head for each existing page frame is expensive and maintaining it requires additional code. This patchset disables the global per-zone LRU lists on memory cgroup configurations and converts all its users to operate on the per-memory cgroup lists instead. As LRU pages are then exclusively on one list, this saves two list pointers for each page frame in the system: page_cgroup array size with 4G physical memory vanilla: allocated 31457280 bytes of page_cgroup patched: allocated 15728640 bytes of page_cgroup At the same time, system performance for various workloads is unaffected: 100G sparse file cat, 4G physical memory, 10 runs, to test for code bloat in the traditional LRU handling and kswapd & direct reclaim paths, without/with the memory controller configured in vanilla: 71.603(0.207) seconds patched: 71.640(0.156) seconds vanilla: 79.558(0.288) seconds patched: 77.233(0.147) seconds 100G sparse file cat in 1G memory cgroup, 10 runs, to test for code bloat in the traditional memory cgroup LRU handling and reclaim path vanilla: 96.844(0.281) seconds patched: 94.454(0.311) seconds 4 unlimited memcgs running kbuild -j32 each, 4G physical memory, 500M swap on SSD, 10 runs, to test for regressions in kswapd & direct reclaim using per-memcg LRU lists with multiple memcgs and multiple allocators within each memcg vanilla: 717.722(1.440) seconds [ 69720.100(11600.835) majfaults ] patched: 714.106(2.313) seconds [ 71109.300(14886.186) majfaults ] 16 unlimited memcgs running kbuild, 1900M hierarchical limit, 500M swap on SSD, 10 runs, to test for regressions in hierarchical memcg setups vanilla: 2742.058(1.992) seconds [ 26479.600(1736.737) majfaults ] patched: 2743.267(1.214) seconds [ 27240.700(1076.063) majfaults ] This patch: There are currently two different implementations of iterating over a memory cgroup hierarchy tree. Consolidate them into one worker function and base the convenience looping-macros on top of it. Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Reviewed-by: Kirill A. Shutemov <kirill@shutemov.name> Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Ying Han <yinghan@google.com> Cc: Greg Thelen <gthelen@google.com> Cc: Michel Lespinasse <walken@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Minchan Kim <minchan.kim@gmail.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-13 01:17:48 +00:00
memcg = NULL;
mm: memcg: consolidate hierarchy iteration primitives The memcg naturalization series: Memory control groups are currently bolted onto the side of traditional memory management in places where better integration would be preferrable. To reclaim memory, for example, memory control groups maintain their own LRU list and reclaim strategy aside from the global per-zone LRU list reclaim. But an extra list head for each existing page frame is expensive and maintaining it requires additional code. This patchset disables the global per-zone LRU lists on memory cgroup configurations and converts all its users to operate on the per-memory cgroup lists instead. As LRU pages are then exclusively on one list, this saves two list pointers for each page frame in the system: page_cgroup array size with 4G physical memory vanilla: allocated 31457280 bytes of page_cgroup patched: allocated 15728640 bytes of page_cgroup At the same time, system performance for various workloads is unaffected: 100G sparse file cat, 4G physical memory, 10 runs, to test for code bloat in the traditional LRU handling and kswapd & direct reclaim paths, without/with the memory controller configured in vanilla: 71.603(0.207) seconds patched: 71.640(0.156) seconds vanilla: 79.558(0.288) seconds patched: 77.233(0.147) seconds 100G sparse file cat in 1G memory cgroup, 10 runs, to test for code bloat in the traditional memory cgroup LRU handling and reclaim path vanilla: 96.844(0.281) seconds patched: 94.454(0.311) seconds 4 unlimited memcgs running kbuild -j32 each, 4G physical memory, 500M swap on SSD, 10 runs, to test for regressions in kswapd & direct reclaim using per-memcg LRU lists with multiple memcgs and multiple allocators within each memcg vanilla: 717.722(1.440) seconds [ 69720.100(11600.835) majfaults ] patched: 714.106(2.313) seconds [ 71109.300(14886.186) majfaults ] 16 unlimited memcgs running kbuild, 1900M hierarchical limit, 500M swap on SSD, 10 runs, to test for regressions in hierarchical memcg setups vanilla: 2742.058(1.992) seconds [ 26479.600(1736.737) majfaults ] patched: 2743.267(1.214) seconds [ 27240.700(1076.063) majfaults ] This patch: There are currently two different implementations of iterating over a memory cgroup hierarchy tree. Consolidate them into one worker function and base the convenience looping-macros on top of it. Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Reviewed-by: Kirill A. Shutemov <kirill@shutemov.name> Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Ying Han <yinghan@google.com> Cc: Greg Thelen <gthelen@google.com> Cc: Michel Lespinasse <walken@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Minchan Kim <minchan.kim@gmail.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-13 01:17:48 +00:00
}
if (reclaim) {
/*
mm: memcontrol: fix possible memcg leak due to interrupted reclaim Memory cgroup reclaim can be interrupted with mem_cgroup_iter_break() once enough pages have been reclaimed, in which case, in contrast to a full round-trip over a cgroup sub-tree, the current position stored in mem_cgroup_reclaim_iter of the target cgroup does not get invalidated and so is left holding the reference to the last scanned cgroup. If the target cgroup does not get scanned again (we might have just reclaimed the last page or all processes might exit and free their memory voluntary), we will leak it, because there is nobody to put the reference held by the iterator. The problem is easy to reproduce by running the following command sequence in a loop: mkdir /sys/fs/cgroup/memory/test echo 100M > /sys/fs/cgroup/memory/test/memory.limit_in_bytes echo $$ > /sys/fs/cgroup/memory/test/cgroup.procs memhog 150M echo $$ > /sys/fs/cgroup/memory/cgroup.procs rmdir test The cgroups generated by it will never get freed. This patch fixes this issue by making mem_cgroup_iter avoid taking reference to the current position. In order not to hit use-after-free bug while running reclaim in parallel with cgroup deletion, we make use of ->css_released cgroup callback to clear references to the dying cgroup in all reclaim iterators that might refer to it. This callback is called right before scheduling rcu work which will free css, so if we access iter->position from rcu read section, we might be sure it won't go away under us. [hannes@cmpxchg.org: clean up css ref handling] Fixes: 5ac8fb31ad2e ("mm: memcontrol: convert reclaim iterator to simple css refcounting") Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-12-29 22:54:10 +00:00
* The position could have already been updated by a competing
* thread, so check that the value hasn't changed since we read
* it to avoid reclaiming from the same cgroup twice.
*/
mm: memcontrol: fix possible memcg leak due to interrupted reclaim Memory cgroup reclaim can be interrupted with mem_cgroup_iter_break() once enough pages have been reclaimed, in which case, in contrast to a full round-trip over a cgroup sub-tree, the current position stored in mem_cgroup_reclaim_iter of the target cgroup does not get invalidated and so is left holding the reference to the last scanned cgroup. If the target cgroup does not get scanned again (we might have just reclaimed the last page or all processes might exit and free their memory voluntary), we will leak it, because there is nobody to put the reference held by the iterator. The problem is easy to reproduce by running the following command sequence in a loop: mkdir /sys/fs/cgroup/memory/test echo 100M > /sys/fs/cgroup/memory/test/memory.limit_in_bytes echo $$ > /sys/fs/cgroup/memory/test/cgroup.procs memhog 150M echo $$ > /sys/fs/cgroup/memory/cgroup.procs rmdir test The cgroups generated by it will never get freed. This patch fixes this issue by making mem_cgroup_iter avoid taking reference to the current position. In order not to hit use-after-free bug while running reclaim in parallel with cgroup deletion, we make use of ->css_released cgroup callback to clear references to the dying cgroup in all reclaim iterators that might refer to it. This callback is called right before scheduling rcu work which will free css, so if we access iter->position from rcu read section, we might be sure it won't go away under us. [hannes@cmpxchg.org: clean up css ref handling] Fixes: 5ac8fb31ad2e ("mm: memcontrol: convert reclaim iterator to simple css refcounting") Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-12-29 22:54:10 +00:00
(void)cmpxchg(&iter->position, pos, memcg);
if (pos)
css_put(&pos->css);
if (!memcg)
iter->generation++;
else if (!prev)
reclaim->generation = iter->generation;
mm: memcg: consolidate hierarchy iteration primitives The memcg naturalization series: Memory control groups are currently bolted onto the side of traditional memory management in places where better integration would be preferrable. To reclaim memory, for example, memory control groups maintain their own LRU list and reclaim strategy aside from the global per-zone LRU list reclaim. But an extra list head for each existing page frame is expensive and maintaining it requires additional code. This patchset disables the global per-zone LRU lists on memory cgroup configurations and converts all its users to operate on the per-memory cgroup lists instead. As LRU pages are then exclusively on one list, this saves two list pointers for each page frame in the system: page_cgroup array size with 4G physical memory vanilla: allocated 31457280 bytes of page_cgroup patched: allocated 15728640 bytes of page_cgroup At the same time, system performance for various workloads is unaffected: 100G sparse file cat, 4G physical memory, 10 runs, to test for code bloat in the traditional LRU handling and kswapd & direct reclaim paths, without/with the memory controller configured in vanilla: 71.603(0.207) seconds patched: 71.640(0.156) seconds vanilla: 79.558(0.288) seconds patched: 77.233(0.147) seconds 100G sparse file cat in 1G memory cgroup, 10 runs, to test for code bloat in the traditional memory cgroup LRU handling and reclaim path vanilla: 96.844(0.281) seconds patched: 94.454(0.311) seconds 4 unlimited memcgs running kbuild -j32 each, 4G physical memory, 500M swap on SSD, 10 runs, to test for regressions in kswapd & direct reclaim using per-memcg LRU lists with multiple memcgs and multiple allocators within each memcg vanilla: 717.722(1.440) seconds [ 69720.100(11600.835) majfaults ] patched: 714.106(2.313) seconds [ 71109.300(14886.186) majfaults ] 16 unlimited memcgs running kbuild, 1900M hierarchical limit, 500M swap on SSD, 10 runs, to test for regressions in hierarchical memcg setups vanilla: 2742.058(1.992) seconds [ 26479.600(1736.737) majfaults ] patched: 2743.267(1.214) seconds [ 27240.700(1076.063) majfaults ] This patch: There are currently two different implementations of iterating over a memory cgroup hierarchy tree. Consolidate them into one worker function and base the convenience looping-macros on top of it. Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Reviewed-by: Kirill A. Shutemov <kirill@shutemov.name> Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Ying Han <yinghan@google.com> Cc: Greg Thelen <gthelen@google.com> Cc: Michel Lespinasse <walken@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Minchan Kim <minchan.kim@gmail.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-13 01:17:48 +00:00
}
memcg: rework mem_cgroup_iter to use cgroup iterators mem_cgroup_iter curently relies on css->id when walking down a group hierarchy tree. This is really awkward because the tree walk depends on the groups creation ordering. The only guarantee is that a parent node is visited before its children. Example: 1) mkdir -p a a/d a/b/c 2) mkdir -a a/b/c a/d Will create the same trees but the tree walks will be different: 1) a, d, b, c 2) a, b, c, d Commit 574bd9f7c7c1 ("cgroup: implement generic child / descendant walk macros") has introduced generic cgroup tree walkers which provide either pre-order or post-order tree walk. This patch converts css->id based iteration to pre-order tree walk to keep the semantic with the original iterator where parent is always visited before its subtree. cgroup_for_each_descendant_pre suggests using post_create and pre_destroy for proper synchronization with groups addidition resp. removal. This implementation doesn't use those because a new memory cgroup is initialized sufficiently for iteration in mem_cgroup_css_alloc already and css reference counting enforces that the group is alive for both the last seen cgroup and the found one resp. it signals that the group is dead and it should be skipped. If the reclaim cookie is used we need to store the last visited group into the iterator so we have to be careful that it doesn't disappear in the mean time. Elevated reference count on the css keeps it alive even though the group have been removed (parked waiting for the last dput so that it can be freed). Per node-zone-prio iter_lock has been introduced to ensure that css_tryget and iter->last_visited is set atomically. Otherwise two racing walkers could both take a references and only one release it leading to a css leak (which pins cgroup dentry). Signed-off-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizefan@huawei.com> Cc: Ying Han <yinghan@google.com> Cc: Tejun Heo <htejun@gmail.com> Cc: Glauber Costa <glommer@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-04-29 22:07:15 +00:00
out_unlock:
rcu_read_unlock();
out:
memcg: keep prev's css alive for the whole mem_cgroup_iter The patchset tries to make mem_cgroup_iter saner in the way how it walks hierarchies. css->id based traversal is far from being ideal as it is not deterministic because it depends on the creation ordering. Additional to that css_id is considered a burden for cgroup maintainers because it is quite some code and memcg is the last user of it. After this series only the swap accounting uses css_id but that one will follow up later. Diffstat (if we exclude removed/added comments) looks quite promising. We got rid of some code: $ git diff mmotm... | grep -v "^[+-][[:space:]]*[/ ]\*" | diffstat b/include/linux/cgroup.h | 3 --- kernel/cgroup.c | 33 --------------------------------- mm/memcontrol.c | 4 +++- 3 files changed, 3 insertions(+), 37 deletions(-) The first patch is just preparatory and it changes when we release css of the previously returned memcg. Nothing controlversial. The second patch is the core of the patchset and it replaces css_get_next based on css_id by the generic cgroup pre-order. This brings some chalanges for the last visited group caching during the reclaim (mem_cgroup_per_zone::reclaim_iter). We have to use memcg pointers directly now which means that we have to keep a reference to those groups' css to keep them alive. I also folded iter_lock introduced by https://lkml.org/lkml/2013/1/3/295 in the previous version into this patch. Johannes felt the race I was describing should be mostly harmless and I haven't been able to trigger it so the lock doesn't deserve its own patch. It is still needed temporarily, though, because the reference counting on iter->last_visited depends on it. It will go away with the next patch. The next patch fixups an unbounded cgroup removal holdoff caused by the elevated css refcount. The issue has been observed by Ying Han. Johannes wasn't impressed by the previous version of the fix (https://lkml.org/lkml/2013/2/8/379) which cleaned up pending references during mem_cgroup_css_offline when a group is removed. He has suggested a different way when the iterator checks whether a cached memcg is still valid or no. More on that in the patch but the basic idea is that every memcg tracks the number removed subgroups and iterator records this number when a group is cached. These numbers are checked before iter->last_visited is about to be used and the iteration is restarted if it is invalid. The fourth and fifth patches are an attempt for simplification of the mem_cgroup_iter. css juggling is removed and the iteration logic is moved to a helper so that the reference counting and iteration are separated. The last patch just removes css_get_next as there is no user for it any longer. My testing looked as follows: A (use_hierarchy=1, limit_in_bytes=150M) /|\ 1 2 3 Children groups were created so that the number is never higher than 3 and their limits were random between 50-100M. Each group hosts a kernel build (starting with tar -xf so the tree is not shared and make -jNUM_CPUs/3) and terminated after random time - up to 5 minutes) and then it is removed. This should exercise both leaf and hierarchical reclaim as well as races with cgroup removals and debugging messages I added on top proved that. 100 groups were created during the test. This patch: css reference counting keeps the cgroup alive even though it has been already removed. mem_cgroup_iter relies on this fact and takes a reference to the returned group. The reference is then released on the next iteration or mem_cgroup_iter_break. mem_cgroup_iter currently releases the reference right after it gets the last css_id. This is correct because neither prev's memcg nor cgroup are accessed after then. This will change in the next patch so we need to hold the group alive a bit longer so let's move the css_put at the end of the function. Signed-off-by: Michal Hocko <mhocko@suse.cz> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Li Zefan <lizefan@huawei.com> Cc: Ying Han <yinghan@google.com> Cc: Tejun Heo <htejun@gmail.com> Cc: Glauber Costa <glommer@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-04-29 22:07:14 +00:00
if (prev && prev != root)
css_put(&prev->css);
mm: memcg: consolidate hierarchy iteration primitives The memcg naturalization series: Memory control groups are currently bolted onto the side of traditional memory management in places where better integration would be preferrable. To reclaim memory, for example, memory control groups maintain their own LRU list and reclaim strategy aside from the global per-zone LRU list reclaim. But an extra list head for each existing page frame is expensive and maintaining it requires additional code. This patchset disables the global per-zone LRU lists on memory cgroup configurations and converts all its users to operate on the per-memory cgroup lists instead. As LRU pages are then exclusively on one list, this saves two list pointers for each page frame in the system: page_cgroup array size with 4G physical memory vanilla: allocated 31457280 bytes of page_cgroup patched: allocated 15728640 bytes of page_cgroup At the same time, system performance for various workloads is unaffected: 100G sparse file cat, 4G physical memory, 10 runs, to test for code bloat in the traditional LRU handling and kswapd & direct reclaim paths, without/with the memory controller configured in vanilla: 71.603(0.207) seconds patched: 71.640(0.156) seconds vanilla: 79.558(0.288) seconds patched: 77.233(0.147) seconds 100G sparse file cat in 1G memory cgroup, 10 runs, to test for code bloat in the traditional memory cgroup LRU handling and reclaim path vanilla: 96.844(0.281) seconds patched: 94.454(0.311) seconds 4 unlimited memcgs running kbuild -j32 each, 4G physical memory, 500M swap on SSD, 10 runs, to test for regressions in kswapd & direct reclaim using per-memcg LRU lists with multiple memcgs and multiple allocators within each memcg vanilla: 717.722(1.440) seconds [ 69720.100(11600.835) majfaults ] patched: 714.106(2.313) seconds [ 71109.300(14886.186) majfaults ] 16 unlimited memcgs running kbuild, 1900M hierarchical limit, 500M swap on SSD, 10 runs, to test for regressions in hierarchical memcg setups vanilla: 2742.058(1.992) seconds [ 26479.600(1736.737) majfaults ] patched: 2743.267(1.214) seconds [ 27240.700(1076.063) majfaults ] This patch: There are currently two different implementations of iterating over a memory cgroup hierarchy tree. Consolidate them into one worker function and base the convenience looping-macros on top of it. Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Reviewed-by: Kirill A. Shutemov <kirill@shutemov.name> Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Ying Han <yinghan@google.com> Cc: Greg Thelen <gthelen@google.com> Cc: Michel Lespinasse <walken@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Minchan Kim <minchan.kim@gmail.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-13 01:17:48 +00:00
return memcg;
}
/**
* mem_cgroup_iter_break - abort a hierarchy walk prematurely
* @root: hierarchy root
* @prev: last visited hierarchy member as returned by mem_cgroup_iter()
*/
void mem_cgroup_iter_break(struct mem_cgroup *root,
struct mem_cgroup *prev)
mm: memcg: consolidate hierarchy iteration primitives The memcg naturalization series: Memory control groups are currently bolted onto the side of traditional memory management in places where better integration would be preferrable. To reclaim memory, for example, memory control groups maintain their own LRU list and reclaim strategy aside from the global per-zone LRU list reclaim. But an extra list head for each existing page frame is expensive and maintaining it requires additional code. This patchset disables the global per-zone LRU lists on memory cgroup configurations and converts all its users to operate on the per-memory cgroup lists instead. As LRU pages are then exclusively on one list, this saves two list pointers for each page frame in the system: page_cgroup array size with 4G physical memory vanilla: allocated 31457280 bytes of page_cgroup patched: allocated 15728640 bytes of page_cgroup At the same time, system performance for various workloads is unaffected: 100G sparse file cat, 4G physical memory, 10 runs, to test for code bloat in the traditional LRU handling and kswapd & direct reclaim paths, without/with the memory controller configured in vanilla: 71.603(0.207) seconds patched: 71.640(0.156) seconds vanilla: 79.558(0.288) seconds patched: 77.233(0.147) seconds 100G sparse file cat in 1G memory cgroup, 10 runs, to test for code bloat in the traditional memory cgroup LRU handling and reclaim path vanilla: 96.844(0.281) seconds patched: 94.454(0.311) seconds 4 unlimited memcgs running kbuild -j32 each, 4G physical memory, 500M swap on SSD, 10 runs, to test for regressions in kswapd & direct reclaim using per-memcg LRU lists with multiple memcgs and multiple allocators within each memcg vanilla: 717.722(1.440) seconds [ 69720.100(11600.835) majfaults ] patched: 714.106(2.313) seconds [ 71109.300(14886.186) majfaults ] 16 unlimited memcgs running kbuild, 1900M hierarchical limit, 500M swap on SSD, 10 runs, to test for regressions in hierarchical memcg setups vanilla: 2742.058(1.992) seconds [ 26479.600(1736.737) majfaults ] patched: 2743.267(1.214) seconds [ 27240.700(1076.063) majfaults ] This patch: There are currently two different implementations of iterating over a memory cgroup hierarchy tree. Consolidate them into one worker function and base the convenience looping-macros on top of it. Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Reviewed-by: Kirill A. Shutemov <kirill@shutemov.name> Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Ying Han <yinghan@google.com> Cc: Greg Thelen <gthelen@google.com> Cc: Michel Lespinasse <walken@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Minchan Kim <minchan.kim@gmail.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-13 01:17:48 +00:00
{
if (!root)
root = root_mem_cgroup;
if (prev && prev != root)
css_put(&prev->css);
}
mm/memcontrol.c: fix use after free in mem_cgroup_iter() This patch is sent to report an use after free in mem_cgroup_iter() after merging commit be2657752e9e ("mm: memcg: fix use after free in mem_cgroup_iter()"). I work with android kernel tree (4.9 & 4.14), and commit be2657752e9e ("mm: memcg: fix use after free in mem_cgroup_iter()") has been merged to the trees. However, I can still observe use after free issues addressed in the commit be2657752e9e. (on low-end devices, a few times this month) backtrace: css_tryget <- crash here mem_cgroup_iter shrink_node shrink_zones do_try_to_free_pages try_to_free_pages __perform_reclaim __alloc_pages_direct_reclaim __alloc_pages_slowpath __alloc_pages_nodemask To debug, I poisoned mem_cgroup before freeing it: static void __mem_cgroup_free(struct mem_cgroup *memcg) for_each_node(node) free_mem_cgroup_per_node_info(memcg, node); free_percpu(memcg->stat); + /* poison memcg before freeing it */ + memset(memcg, 0x78, sizeof(struct mem_cgroup)); kfree(memcg); } The coredump shows the position=0xdbbc2a00 is freed. (gdb) p/x ((struct mem_cgroup_per_node *)0xe5009e00)->iter[8] $13 = {position = 0xdbbc2a00, generation = 0x2efd} 0xdbbc2a00: 0xdbbc2e00 0x00000000 0xdbbc2800 0x00000100 0xdbbc2a10: 0x00000200 0x78787878 0x00026218 0x00000000 0xdbbc2a20: 0xdcad6000 0x00000001 0x78787800 0x00000000 0xdbbc2a30: 0x78780000 0x00000000 0x0068fb84 0x78787878 0xdbbc2a40: 0x78787878 0x78787878 0x78787878 0xe3fa5cc0 0xdbbc2a50: 0x78787878 0x78787878 0x00000000 0x00000000 0xdbbc2a60: 0x00000000 0x00000000 0x00000000 0x00000000 0xdbbc2a70: 0x00000000 0x00000000 0x00000000 0x00000000 0xdbbc2a80: 0x00000000 0x00000000 0x00000000 0x00000000 0xdbbc2a90: 0x00000001 0x00000000 0x00000000 0x00100000 0xdbbc2aa0: 0x00000001 0xdbbc2ac8 0x00000000 0x00000000 0xdbbc2ab0: 0x00000000 0x00000000 0x00000000 0x00000000 0xdbbc2ac0: 0x00000000 0x00000000 0xe5b02618 0x00001000 0xdbbc2ad0: 0x00000000 0x78787878 0x78787878 0x78787878 0xdbbc2ae0: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2af0: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b00: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b10: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b20: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b30: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b40: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b50: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b60: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b70: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b80: 0x78787878 0x78787878 0x00000000 0x78787878 0xdbbc2b90: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2ba0: 0x78787878 0x78787878 0x78787878 0x78787878 In the reclaim path, try_to_free_pages() does not setup sc.target_mem_cgroup and sc is passed to do_try_to_free_pages(), ..., shrink_node(). In mem_cgroup_iter(), root is set to root_mem_cgroup because sc->target_mem_cgroup is NULL. It is possible to assign a memcg to root_mem_cgroup.nodeinfo.iter in mem_cgroup_iter(). try_to_free_pages struct scan_control sc = {...}, target_mem_cgroup is 0x0; do_try_to_free_pages shrink_zones shrink_node mem_cgroup *root = sc->target_mem_cgroup; memcg = mem_cgroup_iter(root, NULL, &reclaim); mem_cgroup_iter() if (!root) root = root_mem_cgroup; ... css = css_next_descendant_pre(css, &root->css); memcg = mem_cgroup_from_css(css); cmpxchg(&iter->position, pos, memcg); My device uses memcg non-hierarchical mode. When we release a memcg: invalidate_reclaim_iterators() reaches only dead_memcg and its parents. If non-hierarchical mode is used, invalidate_reclaim_iterators() never reaches root_mem_cgroup. static void invalidate_reclaim_iterators(struct mem_cgroup *dead_memcg) { struct mem_cgroup *memcg = dead_memcg; for (; memcg; memcg = parent_mem_cgroup(memcg) ... } So the use after free scenario looks like: CPU1 CPU2 try_to_free_pages do_try_to_free_pages shrink_zones shrink_node mem_cgroup_iter() if (!root) root = root_mem_cgroup; ... css = css_next_descendant_pre(css, &root->css); memcg = mem_cgroup_from_css(css); cmpxchg(&iter->position, pos, memcg); invalidate_reclaim_iterators(memcg); ... __mem_cgroup_free() kfree(memcg); try_to_free_pages do_try_to_free_pages shrink_zones shrink_node mem_cgroup_iter() if (!root) root = root_mem_cgroup; ... mz = mem_cgroup_nodeinfo(root, reclaim->pgdat->node_id); iter = &mz->iter[reclaim->priority]; pos = READ_ONCE(iter->position); css_tryget(&pos->css) <- use after free To avoid this, we should also invalidate root_mem_cgroup.nodeinfo.iter in invalidate_reclaim_iterators(). [cai@lca.pw: fix -Wparentheses compilation warning] Link: http://lkml.kernel.org/r/1564580753-17531-1-git-send-email-cai@lca.pw Link: http://lkml.kernel.org/r/20190730015729.4406-1-miles.chen@mediatek.com Fixes: 5ac8fb31ad2e ("mm: memcontrol: convert reclaim iterator to simple css refcounting") Signed-off-by: Miles Chen <miles.chen@mediatek.com> Signed-off-by: Qian Cai <cai@lca.pw> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-08-13 22:37:28 +00:00
static void __invalidate_reclaim_iterators(struct mem_cgroup *from,
struct mem_cgroup *dead_memcg)
mm: memcontrol: fix possible memcg leak due to interrupted reclaim Memory cgroup reclaim can be interrupted with mem_cgroup_iter_break() once enough pages have been reclaimed, in which case, in contrast to a full round-trip over a cgroup sub-tree, the current position stored in mem_cgroup_reclaim_iter of the target cgroup does not get invalidated and so is left holding the reference to the last scanned cgroup. If the target cgroup does not get scanned again (we might have just reclaimed the last page or all processes might exit and free their memory voluntary), we will leak it, because there is nobody to put the reference held by the iterator. The problem is easy to reproduce by running the following command sequence in a loop: mkdir /sys/fs/cgroup/memory/test echo 100M > /sys/fs/cgroup/memory/test/memory.limit_in_bytes echo $$ > /sys/fs/cgroup/memory/test/cgroup.procs memhog 150M echo $$ > /sys/fs/cgroup/memory/cgroup.procs rmdir test The cgroups generated by it will never get freed. This patch fixes this issue by making mem_cgroup_iter avoid taking reference to the current position. In order not to hit use-after-free bug while running reclaim in parallel with cgroup deletion, we make use of ->css_released cgroup callback to clear references to the dying cgroup in all reclaim iterators that might refer to it. This callback is called right before scheduling rcu work which will free css, so if we access iter->position from rcu read section, we might be sure it won't go away under us. [hannes@cmpxchg.org: clean up css ref handling] Fixes: 5ac8fb31ad2e ("mm: memcontrol: convert reclaim iterator to simple css refcounting") Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-12-29 22:54:10 +00:00
{
struct mem_cgroup_reclaim_iter *iter;
struct mem_cgroup_per_node *mz;
int nid;
mm: memcontrol: fix possible memcg leak due to interrupted reclaim Memory cgroup reclaim can be interrupted with mem_cgroup_iter_break() once enough pages have been reclaimed, in which case, in contrast to a full round-trip over a cgroup sub-tree, the current position stored in mem_cgroup_reclaim_iter of the target cgroup does not get invalidated and so is left holding the reference to the last scanned cgroup. If the target cgroup does not get scanned again (we might have just reclaimed the last page or all processes might exit and free their memory voluntary), we will leak it, because there is nobody to put the reference held by the iterator. The problem is easy to reproduce by running the following command sequence in a loop: mkdir /sys/fs/cgroup/memory/test echo 100M > /sys/fs/cgroup/memory/test/memory.limit_in_bytes echo $$ > /sys/fs/cgroup/memory/test/cgroup.procs memhog 150M echo $$ > /sys/fs/cgroup/memory/cgroup.procs rmdir test The cgroups generated by it will never get freed. This patch fixes this issue by making mem_cgroup_iter avoid taking reference to the current position. In order not to hit use-after-free bug while running reclaim in parallel with cgroup deletion, we make use of ->css_released cgroup callback to clear references to the dying cgroup in all reclaim iterators that might refer to it. This callback is called right before scheduling rcu work which will free css, so if we access iter->position from rcu read section, we might be sure it won't go away under us. [hannes@cmpxchg.org: clean up css ref handling] Fixes: 5ac8fb31ad2e ("mm: memcontrol: convert reclaim iterator to simple css refcounting") Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-12-29 22:54:10 +00:00
mm/memcontrol.c: fix use after free in mem_cgroup_iter() This patch is sent to report an use after free in mem_cgroup_iter() after merging commit be2657752e9e ("mm: memcg: fix use after free in mem_cgroup_iter()"). I work with android kernel tree (4.9 & 4.14), and commit be2657752e9e ("mm: memcg: fix use after free in mem_cgroup_iter()") has been merged to the trees. However, I can still observe use after free issues addressed in the commit be2657752e9e. (on low-end devices, a few times this month) backtrace: css_tryget <- crash here mem_cgroup_iter shrink_node shrink_zones do_try_to_free_pages try_to_free_pages __perform_reclaim __alloc_pages_direct_reclaim __alloc_pages_slowpath __alloc_pages_nodemask To debug, I poisoned mem_cgroup before freeing it: static void __mem_cgroup_free(struct mem_cgroup *memcg) for_each_node(node) free_mem_cgroup_per_node_info(memcg, node); free_percpu(memcg->stat); + /* poison memcg before freeing it */ + memset(memcg, 0x78, sizeof(struct mem_cgroup)); kfree(memcg); } The coredump shows the position=0xdbbc2a00 is freed. (gdb) p/x ((struct mem_cgroup_per_node *)0xe5009e00)->iter[8] $13 = {position = 0xdbbc2a00, generation = 0x2efd} 0xdbbc2a00: 0xdbbc2e00 0x00000000 0xdbbc2800 0x00000100 0xdbbc2a10: 0x00000200 0x78787878 0x00026218 0x00000000 0xdbbc2a20: 0xdcad6000 0x00000001 0x78787800 0x00000000 0xdbbc2a30: 0x78780000 0x00000000 0x0068fb84 0x78787878 0xdbbc2a40: 0x78787878 0x78787878 0x78787878 0xe3fa5cc0 0xdbbc2a50: 0x78787878 0x78787878 0x00000000 0x00000000 0xdbbc2a60: 0x00000000 0x00000000 0x00000000 0x00000000 0xdbbc2a70: 0x00000000 0x00000000 0x00000000 0x00000000 0xdbbc2a80: 0x00000000 0x00000000 0x00000000 0x00000000 0xdbbc2a90: 0x00000001 0x00000000 0x00000000 0x00100000 0xdbbc2aa0: 0x00000001 0xdbbc2ac8 0x00000000 0x00000000 0xdbbc2ab0: 0x00000000 0x00000000 0x00000000 0x00000000 0xdbbc2ac0: 0x00000000 0x00000000 0xe5b02618 0x00001000 0xdbbc2ad0: 0x00000000 0x78787878 0x78787878 0x78787878 0xdbbc2ae0: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2af0: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b00: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b10: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b20: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b30: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b40: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b50: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b60: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b70: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b80: 0x78787878 0x78787878 0x00000000 0x78787878 0xdbbc2b90: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2ba0: 0x78787878 0x78787878 0x78787878 0x78787878 In the reclaim path, try_to_free_pages() does not setup sc.target_mem_cgroup and sc is passed to do_try_to_free_pages(), ..., shrink_node(). In mem_cgroup_iter(), root is set to root_mem_cgroup because sc->target_mem_cgroup is NULL. It is possible to assign a memcg to root_mem_cgroup.nodeinfo.iter in mem_cgroup_iter(). try_to_free_pages struct scan_control sc = {...}, target_mem_cgroup is 0x0; do_try_to_free_pages shrink_zones shrink_node mem_cgroup *root = sc->target_mem_cgroup; memcg = mem_cgroup_iter(root, NULL, &reclaim); mem_cgroup_iter() if (!root) root = root_mem_cgroup; ... css = css_next_descendant_pre(css, &root->css); memcg = mem_cgroup_from_css(css); cmpxchg(&iter->position, pos, memcg); My device uses memcg non-hierarchical mode. When we release a memcg: invalidate_reclaim_iterators() reaches only dead_memcg and its parents. If non-hierarchical mode is used, invalidate_reclaim_iterators() never reaches root_mem_cgroup. static void invalidate_reclaim_iterators(struct mem_cgroup *dead_memcg) { struct mem_cgroup *memcg = dead_memcg; for (; memcg; memcg = parent_mem_cgroup(memcg) ... } So the use after free scenario looks like: CPU1 CPU2 try_to_free_pages do_try_to_free_pages shrink_zones shrink_node mem_cgroup_iter() if (!root) root = root_mem_cgroup; ... css = css_next_descendant_pre(css, &root->css); memcg = mem_cgroup_from_css(css); cmpxchg(&iter->position, pos, memcg); invalidate_reclaim_iterators(memcg); ... __mem_cgroup_free() kfree(memcg); try_to_free_pages do_try_to_free_pages shrink_zones shrink_node mem_cgroup_iter() if (!root) root = root_mem_cgroup; ... mz = mem_cgroup_nodeinfo(root, reclaim->pgdat->node_id); iter = &mz->iter[reclaim->priority]; pos = READ_ONCE(iter->position); css_tryget(&pos->css) <- use after free To avoid this, we should also invalidate root_mem_cgroup.nodeinfo.iter in invalidate_reclaim_iterators(). [cai@lca.pw: fix -Wparentheses compilation warning] Link: http://lkml.kernel.org/r/1564580753-17531-1-git-send-email-cai@lca.pw Link: http://lkml.kernel.org/r/20190730015729.4406-1-miles.chen@mediatek.com Fixes: 5ac8fb31ad2e ("mm: memcontrol: convert reclaim iterator to simple css refcounting") Signed-off-by: Miles Chen <miles.chen@mediatek.com> Signed-off-by: Qian Cai <cai@lca.pw> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-08-13 22:37:28 +00:00
for_each_node(nid) {
mz = mem_cgroup_nodeinfo(from, nid);
iter = &mz->iter;
cmpxchg(&iter->position, dead_memcg, NULL);
mm: memcontrol: fix possible memcg leak due to interrupted reclaim Memory cgroup reclaim can be interrupted with mem_cgroup_iter_break() once enough pages have been reclaimed, in which case, in contrast to a full round-trip over a cgroup sub-tree, the current position stored in mem_cgroup_reclaim_iter of the target cgroup does not get invalidated and so is left holding the reference to the last scanned cgroup. If the target cgroup does not get scanned again (we might have just reclaimed the last page or all processes might exit and free their memory voluntary), we will leak it, because there is nobody to put the reference held by the iterator. The problem is easy to reproduce by running the following command sequence in a loop: mkdir /sys/fs/cgroup/memory/test echo 100M > /sys/fs/cgroup/memory/test/memory.limit_in_bytes echo $$ > /sys/fs/cgroup/memory/test/cgroup.procs memhog 150M echo $$ > /sys/fs/cgroup/memory/cgroup.procs rmdir test The cgroups generated by it will never get freed. This patch fixes this issue by making mem_cgroup_iter avoid taking reference to the current position. In order not to hit use-after-free bug while running reclaim in parallel with cgroup deletion, we make use of ->css_released cgroup callback to clear references to the dying cgroup in all reclaim iterators that might refer to it. This callback is called right before scheduling rcu work which will free css, so if we access iter->position from rcu read section, we might be sure it won't go away under us. [hannes@cmpxchg.org: clean up css ref handling] Fixes: 5ac8fb31ad2e ("mm: memcontrol: convert reclaim iterator to simple css refcounting") Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-12-29 22:54:10 +00:00
}
}
mm/memcontrol.c: fix use after free in mem_cgroup_iter() This patch is sent to report an use after free in mem_cgroup_iter() after merging commit be2657752e9e ("mm: memcg: fix use after free in mem_cgroup_iter()"). I work with android kernel tree (4.9 & 4.14), and commit be2657752e9e ("mm: memcg: fix use after free in mem_cgroup_iter()") has been merged to the trees. However, I can still observe use after free issues addressed in the commit be2657752e9e. (on low-end devices, a few times this month) backtrace: css_tryget <- crash here mem_cgroup_iter shrink_node shrink_zones do_try_to_free_pages try_to_free_pages __perform_reclaim __alloc_pages_direct_reclaim __alloc_pages_slowpath __alloc_pages_nodemask To debug, I poisoned mem_cgroup before freeing it: static void __mem_cgroup_free(struct mem_cgroup *memcg) for_each_node(node) free_mem_cgroup_per_node_info(memcg, node); free_percpu(memcg->stat); + /* poison memcg before freeing it */ + memset(memcg, 0x78, sizeof(struct mem_cgroup)); kfree(memcg); } The coredump shows the position=0xdbbc2a00 is freed. (gdb) p/x ((struct mem_cgroup_per_node *)0xe5009e00)->iter[8] $13 = {position = 0xdbbc2a00, generation = 0x2efd} 0xdbbc2a00: 0xdbbc2e00 0x00000000 0xdbbc2800 0x00000100 0xdbbc2a10: 0x00000200 0x78787878 0x00026218 0x00000000 0xdbbc2a20: 0xdcad6000 0x00000001 0x78787800 0x00000000 0xdbbc2a30: 0x78780000 0x00000000 0x0068fb84 0x78787878 0xdbbc2a40: 0x78787878 0x78787878 0x78787878 0xe3fa5cc0 0xdbbc2a50: 0x78787878 0x78787878 0x00000000 0x00000000 0xdbbc2a60: 0x00000000 0x00000000 0x00000000 0x00000000 0xdbbc2a70: 0x00000000 0x00000000 0x00000000 0x00000000 0xdbbc2a80: 0x00000000 0x00000000 0x00000000 0x00000000 0xdbbc2a90: 0x00000001 0x00000000 0x00000000 0x00100000 0xdbbc2aa0: 0x00000001 0xdbbc2ac8 0x00000000 0x00000000 0xdbbc2ab0: 0x00000000 0x00000000 0x00000000 0x00000000 0xdbbc2ac0: 0x00000000 0x00000000 0xe5b02618 0x00001000 0xdbbc2ad0: 0x00000000 0x78787878 0x78787878 0x78787878 0xdbbc2ae0: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2af0: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b00: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b10: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b20: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b30: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b40: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b50: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b60: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b70: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2b80: 0x78787878 0x78787878 0x00000000 0x78787878 0xdbbc2b90: 0x78787878 0x78787878 0x78787878 0x78787878 0xdbbc2ba0: 0x78787878 0x78787878 0x78787878 0x78787878 In the reclaim path, try_to_free_pages() does not setup sc.target_mem_cgroup and sc is passed to do_try_to_free_pages(), ..., shrink_node(). In mem_cgroup_iter(), root is set to root_mem_cgroup because sc->target_mem_cgroup is NULL. It is possible to assign a memcg to root_mem_cgroup.nodeinfo.iter in mem_cgroup_iter(). try_to_free_pages struct scan_control sc = {...}, target_mem_cgroup is 0x0; do_try_to_free_pages shrink_zones shrink_node mem_cgroup *root = sc->target_mem_cgroup; memcg = mem_cgroup_iter(root, NULL, &reclaim); mem_cgroup_iter() if (!root) root = root_mem_cgroup; ... css = css_next_descendant_pre(css, &root->css); memcg = mem_cgroup_from_css(css); cmpxchg(&iter->position, pos, memcg); My device uses memcg non-hierarchical mode. When we release a memcg: invalidate_reclaim_iterators() reaches only dead_memcg and its parents. If non-hierarchical mode is used, invalidate_reclaim_iterators() never reaches root_mem_cgroup. static void invalidate_reclaim_iterators(struct mem_cgroup *dead_memcg) { struct mem_cgroup *memcg = dead_memcg; for (; memcg; memcg = parent_mem_cgroup(memcg) ... } So the use after free scenario looks like: CPU1 CPU2 try_to_free_pages do_try_to_free_pages shrink_zones shrink_node mem_cgroup_iter() if (!root) root = root_mem_cgroup; ... css = css_next_descendant_pre(css, &root->css); memcg = mem_cgroup_from_css(css); cmpxchg(&iter->position, pos, memcg); invalidate_reclaim_iterators(memcg); ... __mem_cgroup_free() kfree(memcg); try_to_free_pages do_try_to_free_pages shrink_zones shrink_node mem_cgroup_iter() if (!root) root = root_mem_cgroup; ... mz = mem_cgroup_nodeinfo(root, reclaim->pgdat->node_id); iter = &mz->iter[reclaim->priority]; pos = READ_ONCE(iter->position); css_tryget(&pos->css) <- use after free To avoid this, we should also invalidate root_mem_cgroup.nodeinfo.iter in invalidate_reclaim_iterators(). [cai@lca.pw: fix -Wparentheses compilation warning] Link: http://lkml.kernel.org/r/1564580753-17531-1-git-send-email-cai@lca.pw Link: http://lkml.kernel.org/r/20190730015729.4406-1-miles.chen@mediatek.com Fixes: 5ac8fb31ad2e ("mm: memcontrol: convert reclaim iterator to simple css refcounting") Signed-off-by: Miles Chen <miles.chen@mediatek.com> Signed-off-by: Qian Cai <cai@lca.pw> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-08-13 22:37:28 +00:00
static void invalidate_reclaim_iterators(struct mem_cgroup *dead_memcg)
{
struct mem_cgroup *memcg = dead_memcg;
struct mem_cgroup *last;
do {
__invalidate_reclaim_iterators(memcg, dead_memcg);
last = memcg;
} while ((memcg = parent_mem_cgroup(memcg)));
/*
* When cgruop1 non-hierarchy mode is used,
* parent_mem_cgroup() does not walk all the way up to the
* cgroup root (root_mem_cgroup). So we have to handle
* dead_memcg from cgroup root separately.
*/
if (last != root_mem_cgroup)
__invalidate_reclaim_iterators(root_mem_cgroup,
dead_memcg);
}
/**
* mem_cgroup_scan_tasks - iterate over tasks of a memory cgroup hierarchy
* @memcg: hierarchy root
* @fn: function to call for each task
* @arg: argument passed to @fn
*
* This function iterates over tasks attached to @memcg or to any of its
* descendants and calls @fn for each task. If @fn returns a non-zero
* value, the function breaks the iteration loop and returns the value.
* Otherwise, it will iterate over all tasks and return 0.
*
* This function must not be called for the root memory cgroup.
*/
int mem_cgroup_scan_tasks(struct mem_cgroup *memcg,
int (*fn)(struct task_struct *, void *), void *arg)
{
struct mem_cgroup *iter;
int ret = 0;
BUG_ON(memcg == root_mem_cgroup);
for_each_mem_cgroup_tree(iter, memcg) {
struct css_task_iter it;
struct task_struct *task;
css_task_iter_start(&iter->css, CSS_TASK_ITER_PROCS, &it);
while (!ret && (task = css_task_iter_next(&it)))
ret = fn(task, arg);
css_task_iter_end(&it);
if (ret) {
mem_cgroup_iter_break(memcg, iter);
break;
}
}
return ret;
}
/**
* mem_cgroup_page_lruvec - return lruvec for isolating/putting an LRU page
* @page: the page
* @pgdat: pgdat of the page
*
* This function relies on page->mem_cgroup being stable - see the
* access rules in commit_charge().
*/
mm, vmscan: move LRU lists to node This moves the LRU lists from the zone to the node and related data such as counters, tracing, congestion tracking and writeback tracking. Unfortunately, due to reclaim and compaction retry logic, it is necessary to account for the number of LRU pages on both zone and node logic. Most reclaim logic is based on the node counters but the retry logic uses the zone counters which do not distinguish inactive and active sizes. It would be possible to leave the LRU counters on a per-zone basis but it's a heavier calculation across multiple cache lines that is much more frequent than the retry checks. Other than the LRU counters, this is mostly a mechanical patch but note that it introduces a number of anomalies. For example, the scans are per-zone but using per-node counters. We also mark a node as congested when a zone is congested. This causes weird problems that are fixed later but is easier to review. In the event that there is excessive overhead on 32-bit systems due to the nodes being on LRU then there are two potential solutions 1. Long-term isolation of highmem pages when reclaim is lowmem When pages are skipped, they are immediately added back onto the LRU list. If lowmem reclaim persisted for long periods of time, the same highmem pages get continually scanned. The idea would be that lowmem keeps those pages on a separate list until a reclaim for highmem pages arrives that splices the highmem pages back onto the LRU. It potentially could be implemented similar to the UNEVICTABLE list. That would reduce the skip rate with the potential corner case is that highmem pages have to be scanned and reclaimed to free lowmem slab pages. 2. Linear scan lowmem pages if the initial LRU shrink fails This will break LRU ordering but may be preferable and faster during memory pressure than skipping LRU pages. Link: http://lkml.kernel.org/r/1467970510-21195-4-git-send-email-mgorman@techsingularity.net Signed-off-by: Mel Gorman <mgorman@techsingularity.net> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vlastimil Babka <vbabka@suse.cz> Cc: Hillf Danton <hillf.zj@alibaba-inc.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Rik van Riel <riel@surriel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-28 22:45:31 +00:00
struct lruvec *mem_cgroup_page_lruvec(struct page *page, struct pglist_data *pgdat)
memcg: synchronized LRU A big patch for changing memcg's LRU semantics. Now, - page_cgroup is linked to mem_cgroup's its own LRU (per zone). - LRU of page_cgroup is not synchronous with global LRU. - page and page_cgroup is one-to-one and statically allocated. - To find page_cgroup is on what LRU, you have to check pc->mem_cgroup as - lru = page_cgroup_zoneinfo(pc, nid_of_pc, zid_of_pc); - SwapCache is handled. And, when we handle LRU list of page_cgroup, we do following. pc = lookup_page_cgroup(page); lock_page_cgroup(pc); .....................(1) mz = page_cgroup_zoneinfo(pc); spin_lock(&mz->lru_lock); .....add to LRU spin_unlock(&mz->lru_lock); unlock_page_cgroup(pc); But (1) is spin_lock and we have to be afraid of dead-lock with zone->lru_lock. So, trylock() is used at (1), now. Without (1), we can't trust "mz" is correct. This is a trial to remove this dirty nesting of locks. This patch changes mz->lru_lock to be zone->lru_lock. Then, above sequence will be written as spin_lock(&zone->lru_lock); # in vmscan.c or swap.c via global LRU mem_cgroup_add/remove/etc_lru() { pc = lookup_page_cgroup(page); mz = page_cgroup_zoneinfo(pc); if (PageCgroupUsed(pc)) { ....add to LRU } spin_lock(&zone->lru_lock); # in vmscan.c or swap.c via global LRU This is much simpler. (*) We're safe even if we don't take lock_page_cgroup(pc). Because.. 1. When pc->mem_cgroup can be modified. - at charge. - at account_move(). 2. at charge the PCG_USED bit is not set before pc->mem_cgroup is fixed. 3. at account_move() the page is isolated and not on LRU. Pros. - easy for maintenance. - memcg can make use of laziness of pagevec. - we don't have to duplicated LRU/Active/Unevictable bit in page_cgroup. - LRU status of memcg will be synchronized with global LRU's one. - # of locks are reduced. - account_move() is simplified very much. Cons. - may increase cost of LRU rotation. (no impact if memcg is not configured.) Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:08:01 +00:00
{
struct mem_cgroup_per_node *mz;
struct mem_cgroup *memcg;
memcg: fix hotplugged memory zone oops When MEMCG is configured on (even when it's disabled by boot option), when adding or removing a page to/from its lru list, the zone pointer used for stats updates is nowadays taken from the struct lruvec. (On many configurations, calculating zone from page is slower.) But we have no code to update all the lruvecs (per zone, per memcg) when a memory node is hotadded. Here's an extract from the oops which results when running numactl to bind a program to a newly onlined node: BUG: unable to handle kernel NULL pointer dereference at 0000000000000f60 IP: __mod_zone_page_state+0x9/0x60 Pid: 1219, comm: numactl Not tainted 3.6.0-rc5+ #180 Bochs Bochs Process numactl (pid: 1219, threadinfo ffff880039abc000, task ffff8800383c4ce0) Call Trace: __pagevec_lru_add_fn+0xdf/0x140 pagevec_lru_move_fn+0xb1/0x100 __pagevec_lru_add+0x1c/0x30 lru_add_drain_cpu+0xa3/0x130 lru_add_drain+0x2f/0x40 ... The natural solution might be to use a memcg callback whenever memory is hotadded; but that solution has not been scoped out, and it happens that we do have an easy location at which to update lruvec->zone. The lruvec pointer is discovered either by mem_cgroup_zone_lruvec() or by mem_cgroup_page_lruvec(), and both of those do know the right zone. So check and set lruvec->zone in those; and remove the inadequate attempt to set lruvec->zone from lruvec_init(), which is called before NODE_DATA(node) has been allocated in such cases. Ah, there was one exceptionr. For no particularly good reason, mem_cgroup_force_empty_list() has its own code for deciding lruvec. Change it to use the standard mem_cgroup_zone_lruvec() and mem_cgroup_get_lru_size() too. In fact it was already safe against such an oops (the lru lists in danger could only be empty), but we're better proofed against future changes this way. I've marked this for stable (3.6) since we introduced the problem in 3.5 (now closed to stable); but I have no idea if this is the only fix needed to get memory hotadd working with memcg in 3.6, and received no answer when I enquired twice before. Reported-by: Tang Chen <tangchen@cn.fujitsu.com> Signed-off-by: Hugh Dickins <hughd@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Konstantin Khlebnikov <khlebnikov@openvz.org> Cc: Wen Congyang <wency@cn.fujitsu.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-11-16 22:14:54 +00:00
struct lruvec *lruvec;
memcg: fix hotplugged memory zone oops When MEMCG is configured on (even when it's disabled by boot option), when adding or removing a page to/from its lru list, the zone pointer used for stats updates is nowadays taken from the struct lruvec. (On many configurations, calculating zone from page is slower.) But we have no code to update all the lruvecs (per zone, per memcg) when a memory node is hotadded. Here's an extract from the oops which results when running numactl to bind a program to a newly onlined node: BUG: unable to handle kernel NULL pointer dereference at 0000000000000f60 IP: __mod_zone_page_state+0x9/0x60 Pid: 1219, comm: numactl Not tainted 3.6.0-rc5+ #180 Bochs Bochs Process numactl (pid: 1219, threadinfo ffff880039abc000, task ffff8800383c4ce0) Call Trace: __pagevec_lru_add_fn+0xdf/0x140 pagevec_lru_move_fn+0xb1/0x100 __pagevec_lru_add+0x1c/0x30 lru_add_drain_cpu+0xa3/0x130 lru_add_drain+0x2f/0x40 ... The natural solution might be to use a memcg callback whenever memory is hotadded; but that solution has not been scoped out, and it happens that we do have an easy location at which to update lruvec->zone. The lruvec pointer is discovered either by mem_cgroup_zone_lruvec() or by mem_cgroup_page_lruvec(), and both of those do know the right zone. So check and set lruvec->zone in those; and remove the inadequate attempt to set lruvec->zone from lruvec_init(), which is called before NODE_DATA(node) has been allocated in such cases. Ah, there was one exceptionr. For no particularly good reason, mem_cgroup_force_empty_list() has its own code for deciding lruvec. Change it to use the standard mem_cgroup_zone_lruvec() and mem_cgroup_get_lru_size() too. In fact it was already safe against such an oops (the lru lists in danger could only be empty), but we're better proofed against future changes this way. I've marked this for stable (3.6) since we introduced the problem in 3.5 (now closed to stable); but I have no idea if this is the only fix needed to get memory hotadd working with memcg in 3.6, and received no answer when I enquired twice before. Reported-by: Tang Chen <tangchen@cn.fujitsu.com> Signed-off-by: Hugh Dickins <hughd@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Konstantin Khlebnikov <khlebnikov@openvz.org> Cc: Wen Congyang <wency@cn.fujitsu.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-11-16 22:14:54 +00:00
if (mem_cgroup_disabled()) {
lruvec = &pgdat->__lruvec;
memcg: fix hotplugged memory zone oops When MEMCG is configured on (even when it's disabled by boot option), when adding or removing a page to/from its lru list, the zone pointer used for stats updates is nowadays taken from the struct lruvec. (On many configurations, calculating zone from page is slower.) But we have no code to update all the lruvecs (per zone, per memcg) when a memory node is hotadded. Here's an extract from the oops which results when running numactl to bind a program to a newly onlined node: BUG: unable to handle kernel NULL pointer dereference at 0000000000000f60 IP: __mod_zone_page_state+0x9/0x60 Pid: 1219, comm: numactl Not tainted 3.6.0-rc5+ #180 Bochs Bochs Process numactl (pid: 1219, threadinfo ffff880039abc000, task ffff8800383c4ce0) Call Trace: __pagevec_lru_add_fn+0xdf/0x140 pagevec_lru_move_fn+0xb1/0x100 __pagevec_lru_add+0x1c/0x30 lru_add_drain_cpu+0xa3/0x130 lru_add_drain+0x2f/0x40 ... The natural solution might be to use a memcg callback whenever memory is hotadded; but that solution has not been scoped out, and it happens that we do have an easy location at which to update lruvec->zone. The lruvec pointer is discovered either by mem_cgroup_zone_lruvec() or by mem_cgroup_page_lruvec(), and both of those do know the right zone. So check and set lruvec->zone in those; and remove the inadequate attempt to set lruvec->zone from lruvec_init(), which is called before NODE_DATA(node) has been allocated in such cases. Ah, there was one exceptionr. For no particularly good reason, mem_cgroup_force_empty_list() has its own code for deciding lruvec. Change it to use the standard mem_cgroup_zone_lruvec() and mem_cgroup_get_lru_size() too. In fact it was already safe against such an oops (the lru lists in danger could only be empty), but we're better proofed against future changes this way. I've marked this for stable (3.6) since we introduced the problem in 3.5 (now closed to stable); but I have no idea if this is the only fix needed to get memory hotadd working with memcg in 3.6, and received no answer when I enquired twice before. Reported-by: Tang Chen <tangchen@cn.fujitsu.com> Signed-off-by: Hugh Dickins <hughd@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Konstantin Khlebnikov <khlebnikov@openvz.org> Cc: Wen Congyang <wency@cn.fujitsu.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-11-16 22:14:54 +00:00
goto out;
}
mm: embed the memcg pointer directly into struct page Memory cgroups used to have 5 per-page pointers. To allow users to disable that amount of overhead during runtime, those pointers were allocated in a separate array, with a translation layer between them and struct page. There is now only one page pointer remaining: the memcg pointer, that indicates which cgroup the page is associated with when charged. The complexity of runtime allocation and the runtime translation overhead is no longer justified to save that *potential* 0.19% of memory. With CONFIG_SLUB, page->mem_cgroup actually sits in the doubleword padding after the page->private member and doesn't even increase struct page, and then this patch actually saves space. Remaining users that care can still compile their kernels without CONFIG_MEMCG. text data bss dec hex filename 8828345 1725264 983040 11536649 b00909 vmlinux.old 8827425 1725264 966656 11519345 afc571 vmlinux.new [mhocko@suse.cz: update Documentation/cgroups/memory.txt] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: David S. Miller <davem@davemloft.net> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Acked-by: Konstantin Khlebnikov <koct9i@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:44:52 +00:00
memcg = page->mem_cgroup;
memcg: fix GPF when cgroup removal races with last exit When moving tasks from old memcg (with move_charge_at_immigrate on new memcg), followed by removal of old memcg, hit General Protection Fault in mem_cgroup_lru_del_list() (called from release_pages called from free_pages_and_swap_cache from tlb_flush_mmu from tlb_finish_mmu from exit_mmap from mmput from exit_mm from do_exit). Somewhat reproducible, takes a few hours: the old struct mem_cgroup has been freed and poisoned by SLAB_DEBUG, but mem_cgroup_lru_del_list() is still trying to update its stats, and take page off lru before freeing. A task, or a charge, or a page on lru: each secures a memcg against removal. In this case, the last task has been moved out of the old memcg, and it is exiting: anonymous pages are uncharged one by one from the memcg, as they are zapped from its pagetables, so the charge gets down to 0; but the pages themselves are queued in an mmu_gather for freeing. Most of those pages will be on lru (and force_empty is careful to lru_add_drain_all, to add pages from pagevec to lru first), but not necessarily all: perhaps some have been isolated for page reclaim, perhaps some isolated for other reasons. So, force_empty may find no task, no charge and no page on lru, and let the removal proceed. There would still be no problem if these pages were immediately freed; but typically (and the put_page_testzero protocol demands it) they have to be added back to lru before they are found freeable, then removed from lru and freed. We don't see the issue when adding, because the mem_cgroup_iter() loops keep their own reference to the memcg being scanned; but when it comes to mem_cgroup_lru_del_list(). I believe this was not an issue in v3.2: there, PageCgroupAcctLRU and PageCgroupUsed flags were used (like a trick with mirrors) to deflect view of pc->mem_cgroup to the stable root_mem_cgroup when neither set. 38c5d72f3ebe ("memcg: simplify LRU handling by new rule") mercifully removed those convolutions, but left this General Protection Fault. But it's surprisingly easy to restore the old behaviour: just check PageCgroupUsed in mem_cgroup_lru_add_list() (which decides on which lruvec to add), and reset pc to root_mem_cgroup if page is uncharged. A risky change? just going back to how it worked before; testing, and an audit of uses of pc->mem_cgroup, show no problem. And there's a nice bonus: with mem_cgroup_lru_add_list() itself making sure that an uncharged page goes to root lru, mem_cgroup_reset_owner() no longer has any purpose, and we can safely revert 4e5f01c2b9b9 ("memcg: clear pc->mem_cgroup if necessary"). Calling update_page_reclaim_stat() after add_page_to_lru_list() in swap.c is not strictly necessary: the lru_lock there, with RCU before memcg structures are freed, makes mem_cgroup_get_reclaim_stat_from_page safe without that; but it seems cleaner to rely on one dependency less. Signed-off-by: Hugh Dickins <hughd@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Konstantin Khlebnikov <khlebnikov@openvz.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-05 22:59:18 +00:00
/*
* Swapcache readahead pages are added to the LRU - and
* possibly migrated - before they are charged.
memcg: fix GPF when cgroup removal races with last exit When moving tasks from old memcg (with move_charge_at_immigrate on new memcg), followed by removal of old memcg, hit General Protection Fault in mem_cgroup_lru_del_list() (called from release_pages called from free_pages_and_swap_cache from tlb_flush_mmu from tlb_finish_mmu from exit_mmap from mmput from exit_mm from do_exit). Somewhat reproducible, takes a few hours: the old struct mem_cgroup has been freed and poisoned by SLAB_DEBUG, but mem_cgroup_lru_del_list() is still trying to update its stats, and take page off lru before freeing. A task, or a charge, or a page on lru: each secures a memcg against removal. In this case, the last task has been moved out of the old memcg, and it is exiting: anonymous pages are uncharged one by one from the memcg, as they are zapped from its pagetables, so the charge gets down to 0; but the pages themselves are queued in an mmu_gather for freeing. Most of those pages will be on lru (and force_empty is careful to lru_add_drain_all, to add pages from pagevec to lru first), but not necessarily all: perhaps some have been isolated for page reclaim, perhaps some isolated for other reasons. So, force_empty may find no task, no charge and no page on lru, and let the removal proceed. There would still be no problem if these pages were immediately freed; but typically (and the put_page_testzero protocol demands it) they have to be added back to lru before they are found freeable, then removed from lru and freed. We don't see the issue when adding, because the mem_cgroup_iter() loops keep their own reference to the memcg being scanned; but when it comes to mem_cgroup_lru_del_list(). I believe this was not an issue in v3.2: there, PageCgroupAcctLRU and PageCgroupUsed flags were used (like a trick with mirrors) to deflect view of pc->mem_cgroup to the stable root_mem_cgroup when neither set. 38c5d72f3ebe ("memcg: simplify LRU handling by new rule") mercifully removed those convolutions, but left this General Protection Fault. But it's surprisingly easy to restore the old behaviour: just check PageCgroupUsed in mem_cgroup_lru_add_list() (which decides on which lruvec to add), and reset pc to root_mem_cgroup if page is uncharged. A risky change? just going back to how it worked before; testing, and an audit of uses of pc->mem_cgroup, show no problem. And there's a nice bonus: with mem_cgroup_lru_add_list() itself making sure that an uncharged page goes to root lru, mem_cgroup_reset_owner() no longer has any purpose, and we can safely revert 4e5f01c2b9b9 ("memcg: clear pc->mem_cgroup if necessary"). Calling update_page_reclaim_stat() after add_page_to_lru_list() in swap.c is not strictly necessary: the lru_lock there, with RCU before memcg structures are freed, makes mem_cgroup_get_reclaim_stat_from_page safe without that; but it seems cleaner to rely on one dependency less. Signed-off-by: Hugh Dickins <hughd@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Konstantin Khlebnikov <khlebnikov@openvz.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-05 22:59:18 +00:00
*/
if (!memcg)
memcg = root_mem_cgroup;
memcg: fix GPF when cgroup removal races with last exit When moving tasks from old memcg (with move_charge_at_immigrate on new memcg), followed by removal of old memcg, hit General Protection Fault in mem_cgroup_lru_del_list() (called from release_pages called from free_pages_and_swap_cache from tlb_flush_mmu from tlb_finish_mmu from exit_mmap from mmput from exit_mm from do_exit). Somewhat reproducible, takes a few hours: the old struct mem_cgroup has been freed and poisoned by SLAB_DEBUG, but mem_cgroup_lru_del_list() is still trying to update its stats, and take page off lru before freeing. A task, or a charge, or a page on lru: each secures a memcg against removal. In this case, the last task has been moved out of the old memcg, and it is exiting: anonymous pages are uncharged one by one from the memcg, as they are zapped from its pagetables, so the charge gets down to 0; but the pages themselves are queued in an mmu_gather for freeing. Most of those pages will be on lru (and force_empty is careful to lru_add_drain_all, to add pages from pagevec to lru first), but not necessarily all: perhaps some have been isolated for page reclaim, perhaps some isolated for other reasons. So, force_empty may find no task, no charge and no page on lru, and let the removal proceed. There would still be no problem if these pages were immediately freed; but typically (and the put_page_testzero protocol demands it) they have to be added back to lru before they are found freeable, then removed from lru and freed. We don't see the issue when adding, because the mem_cgroup_iter() loops keep their own reference to the memcg being scanned; but when it comes to mem_cgroup_lru_del_list(). I believe this was not an issue in v3.2: there, PageCgroupAcctLRU and PageCgroupUsed flags were used (like a trick with mirrors) to deflect view of pc->mem_cgroup to the stable root_mem_cgroup when neither set. 38c5d72f3ebe ("memcg: simplify LRU handling by new rule") mercifully removed those convolutions, but left this General Protection Fault. But it's surprisingly easy to restore the old behaviour: just check PageCgroupUsed in mem_cgroup_lru_add_list() (which decides on which lruvec to add), and reset pc to root_mem_cgroup if page is uncharged. A risky change? just going back to how it worked before; testing, and an audit of uses of pc->mem_cgroup, show no problem. And there's a nice bonus: with mem_cgroup_lru_add_list() itself making sure that an uncharged page goes to root lru, mem_cgroup_reset_owner() no longer has any purpose, and we can safely revert 4e5f01c2b9b9 ("memcg: clear pc->mem_cgroup if necessary"). Calling update_page_reclaim_stat() after add_page_to_lru_list() in swap.c is not strictly necessary: the lru_lock there, with RCU before memcg structures are freed, makes mem_cgroup_get_reclaim_stat_from_page safe without that; but it seems cleaner to rely on one dependency less. Signed-off-by: Hugh Dickins <hughd@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Konstantin Khlebnikov <khlebnikov@openvz.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-05 22:59:18 +00:00
mz = mem_cgroup_page_nodeinfo(memcg, page);
memcg: fix hotplugged memory zone oops When MEMCG is configured on (even when it's disabled by boot option), when adding or removing a page to/from its lru list, the zone pointer used for stats updates is nowadays taken from the struct lruvec. (On many configurations, calculating zone from page is slower.) But we have no code to update all the lruvecs (per zone, per memcg) when a memory node is hotadded. Here's an extract from the oops which results when running numactl to bind a program to a newly onlined node: BUG: unable to handle kernel NULL pointer dereference at 0000000000000f60 IP: __mod_zone_page_state+0x9/0x60 Pid: 1219, comm: numactl Not tainted 3.6.0-rc5+ #180 Bochs Bochs Process numactl (pid: 1219, threadinfo ffff880039abc000, task ffff8800383c4ce0) Call Trace: __pagevec_lru_add_fn+0xdf/0x140 pagevec_lru_move_fn+0xb1/0x100 __pagevec_lru_add+0x1c/0x30 lru_add_drain_cpu+0xa3/0x130 lru_add_drain+0x2f/0x40 ... The natural solution might be to use a memcg callback whenever memory is hotadded; but that solution has not been scoped out, and it happens that we do have an easy location at which to update lruvec->zone. The lruvec pointer is discovered either by mem_cgroup_zone_lruvec() or by mem_cgroup_page_lruvec(), and both of those do know the right zone. So check and set lruvec->zone in those; and remove the inadequate attempt to set lruvec->zone from lruvec_init(), which is called before NODE_DATA(node) has been allocated in such cases. Ah, there was one exceptionr. For no particularly good reason, mem_cgroup_force_empty_list() has its own code for deciding lruvec. Change it to use the standard mem_cgroup_zone_lruvec() and mem_cgroup_get_lru_size() too. In fact it was already safe against such an oops (the lru lists in danger could only be empty), but we're better proofed against future changes this way. I've marked this for stable (3.6) since we introduced the problem in 3.5 (now closed to stable); but I have no idea if this is the only fix needed to get memory hotadd working with memcg in 3.6, and received no answer when I enquired twice before. Reported-by: Tang Chen <tangchen@cn.fujitsu.com> Signed-off-by: Hugh Dickins <hughd@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Konstantin Khlebnikov <khlebnikov@openvz.org> Cc: Wen Congyang <wency@cn.fujitsu.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-11-16 22:14:54 +00:00
lruvec = &mz->lruvec;
out:
/*
* Since a node can be onlined after the mem_cgroup was created,
* we have to be prepared to initialize lruvec->zone here;
* and if offlined then reonlined, we need to reinitialize it.
*/
mm, vmscan: move LRU lists to node This moves the LRU lists from the zone to the node and related data such as counters, tracing, congestion tracking and writeback tracking. Unfortunately, due to reclaim and compaction retry logic, it is necessary to account for the number of LRU pages on both zone and node logic. Most reclaim logic is based on the node counters but the retry logic uses the zone counters which do not distinguish inactive and active sizes. It would be possible to leave the LRU counters on a per-zone basis but it's a heavier calculation across multiple cache lines that is much more frequent than the retry checks. Other than the LRU counters, this is mostly a mechanical patch but note that it introduces a number of anomalies. For example, the scans are per-zone but using per-node counters. We also mark a node as congested when a zone is congested. This causes weird problems that are fixed later but is easier to review. In the event that there is excessive overhead on 32-bit systems due to the nodes being on LRU then there are two potential solutions 1. Long-term isolation of highmem pages when reclaim is lowmem When pages are skipped, they are immediately added back onto the LRU list. If lowmem reclaim persisted for long periods of time, the same highmem pages get continually scanned. The idea would be that lowmem keeps those pages on a separate list until a reclaim for highmem pages arrives that splices the highmem pages back onto the LRU. It potentially could be implemented similar to the UNEVICTABLE list. That would reduce the skip rate with the potential corner case is that highmem pages have to be scanned and reclaimed to free lowmem slab pages. 2. Linear scan lowmem pages if the initial LRU shrink fails This will break LRU ordering but may be preferable and faster during memory pressure than skipping LRU pages. Link: http://lkml.kernel.org/r/1467970510-21195-4-git-send-email-mgorman@techsingularity.net Signed-off-by: Mel Gorman <mgorman@techsingularity.net> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vlastimil Babka <vbabka@suse.cz> Cc: Hillf Danton <hillf.zj@alibaba-inc.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Rik van Riel <riel@surriel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-28 22:45:31 +00:00
if (unlikely(lruvec->pgdat != pgdat))
lruvec->pgdat = pgdat;
memcg: fix hotplugged memory zone oops When MEMCG is configured on (even when it's disabled by boot option), when adding or removing a page to/from its lru list, the zone pointer used for stats updates is nowadays taken from the struct lruvec. (On many configurations, calculating zone from page is slower.) But we have no code to update all the lruvecs (per zone, per memcg) when a memory node is hotadded. Here's an extract from the oops which results when running numactl to bind a program to a newly onlined node: BUG: unable to handle kernel NULL pointer dereference at 0000000000000f60 IP: __mod_zone_page_state+0x9/0x60 Pid: 1219, comm: numactl Not tainted 3.6.0-rc5+ #180 Bochs Bochs Process numactl (pid: 1219, threadinfo ffff880039abc000, task ffff8800383c4ce0) Call Trace: __pagevec_lru_add_fn+0xdf/0x140 pagevec_lru_move_fn+0xb1/0x100 __pagevec_lru_add+0x1c/0x30 lru_add_drain_cpu+0xa3/0x130 lru_add_drain+0x2f/0x40 ... The natural solution might be to use a memcg callback whenever memory is hotadded; but that solution has not been scoped out, and it happens that we do have an easy location at which to update lruvec->zone. The lruvec pointer is discovered either by mem_cgroup_zone_lruvec() or by mem_cgroup_page_lruvec(), and both of those do know the right zone. So check and set lruvec->zone in those; and remove the inadequate attempt to set lruvec->zone from lruvec_init(), which is called before NODE_DATA(node) has been allocated in such cases. Ah, there was one exceptionr. For no particularly good reason, mem_cgroup_force_empty_list() has its own code for deciding lruvec. Change it to use the standard mem_cgroup_zone_lruvec() and mem_cgroup_get_lru_size() too. In fact it was already safe against such an oops (the lru lists in danger could only be empty), but we're better proofed against future changes this way. I've marked this for stable (3.6) since we introduced the problem in 3.5 (now closed to stable); but I have no idea if this is the only fix needed to get memory hotadd working with memcg in 3.6, and received no answer when I enquired twice before. Reported-by: Tang Chen <tangchen@cn.fujitsu.com> Signed-off-by: Hugh Dickins <hughd@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Konstantin Khlebnikov <khlebnikov@openvz.org> Cc: Wen Congyang <wency@cn.fujitsu.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-11-16 22:14:54 +00:00
return lruvec;
memcg: synchronized LRU A big patch for changing memcg's LRU semantics. Now, - page_cgroup is linked to mem_cgroup's its own LRU (per zone). - LRU of page_cgroup is not synchronous with global LRU. - page and page_cgroup is one-to-one and statically allocated. - To find page_cgroup is on what LRU, you have to check pc->mem_cgroup as - lru = page_cgroup_zoneinfo(pc, nid_of_pc, zid_of_pc); - SwapCache is handled. And, when we handle LRU list of page_cgroup, we do following. pc = lookup_page_cgroup(page); lock_page_cgroup(pc); .....................(1) mz = page_cgroup_zoneinfo(pc); spin_lock(&mz->lru_lock); .....add to LRU spin_unlock(&mz->lru_lock); unlock_page_cgroup(pc); But (1) is spin_lock and we have to be afraid of dead-lock with zone->lru_lock. So, trylock() is used at (1), now. Without (1), we can't trust "mz" is correct. This is a trial to remove this dirty nesting of locks. This patch changes mz->lru_lock to be zone->lru_lock. Then, above sequence will be written as spin_lock(&zone->lru_lock); # in vmscan.c or swap.c via global LRU mem_cgroup_add/remove/etc_lru() { pc = lookup_page_cgroup(page); mz = page_cgroup_zoneinfo(pc); if (PageCgroupUsed(pc)) { ....add to LRU } spin_lock(&zone->lru_lock); # in vmscan.c or swap.c via global LRU This is much simpler. (*) We're safe even if we don't take lock_page_cgroup(pc). Because.. 1. When pc->mem_cgroup can be modified. - at charge. - at account_move(). 2. at charge the PCG_USED bit is not set before pc->mem_cgroup is fixed. 3. at account_move() the page is isolated and not on LRU. Pros. - easy for maintenance. - memcg can make use of laziness of pagevec. - we don't have to duplicated LRU/Active/Unevictable bit in page_cgroup. - LRU status of memcg will be synchronized with global LRU's one. - # of locks are reduced. - account_move() is simplified very much. Cons. - may increase cost of LRU rotation. (no impact if memcg is not configured.) Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:08:01 +00:00
}
/**
* mem_cgroup_update_lru_size - account for adding or removing an lru page
* @lruvec: mem_cgroup per zone lru vector
* @lru: index of lru list the page is sitting on
mm, memcg: fix the active list aging for lowmem requests when memcg is enabled Nils Holland and Klaus Ethgen have reported unexpected OOM killer invocations with 32b kernel starting with 4.8 kernels kworker/u4:5 invoked oom-killer: gfp_mask=0x2400840(GFP_NOFS|__GFP_NOFAIL), nodemask=0, order=0, oom_score_adj=0 kworker/u4:5 cpuset=/ mems_allowed=0 CPU: 1 PID: 2603 Comm: kworker/u4:5 Not tainted 4.9.0-gentoo #2 [...] Mem-Info: active_anon:58685 inactive_anon:90 isolated_anon:0 active_file:274324 inactive_file:281962 isolated_file:0 unevictable:0 dirty:649 writeback:0 unstable:0 slab_reclaimable:40662 slab_unreclaimable:17754 mapped:7382 shmem:202 pagetables:351 bounce:0 free:206736 free_pcp:332 free_cma:0 Node 0 active_anon:234740kB inactive_anon:360kB active_file:1097296kB inactive_file:1127848kB unevictable:0kB isolated(anon):0kB isolated(file):0kB mapped:29528kB dirty:2596kB writeback:0kB shmem:0kB shmem_thp: 0kB shmem_pmdmapped: 184320kB anon_thp: 808kB writeback_tmp:0kB unstable:0kB pages_scanned:0 all_unreclaimable? no DMA free:3952kB min:788kB low:984kB high:1180kB active_anon:0kB inactive_anon:0kB active_file:7316kB inactive_file:0kB unevictable:0kB writepending:96kB present:15992kB managed:15916kB mlocked:0kB slab_reclaimable:3200kB slab_unreclaimable:1408kB kernel_stack:0kB pagetables:0kB bounce:0kB free_pcp:0kB local_pcp:0kB free_cma:0kB lowmem_reserve[]: 0 813 3474 3474 Normal free:41332kB min:41368kB low:51708kB high:62048kB active_anon:0kB inactive_anon:0kB active_file:532748kB inactive_file:44kB unevictable:0kB writepending:24kB present:897016kB managed:836248kB mlocked:0kB slab_reclaimable:159448kB slab_unreclaimable:69608kB kernel_stack:1112kB pagetables:1404kB bounce:0kB free_pcp:528kB local_pcp:340kB free_cma:0kB lowmem_reserve[]: 0 0 21292 21292 HighMem free:781660kB min:512kB low:34356kB high:68200kB active_anon:234740kB inactive_anon:360kB active_file:557232kB inactive_file:1127804kB unevictable:0kB writepending:2592kB present:2725384kB managed:2725384kB mlocked:0kB slab_reclaimable:0kB slab_unreclaimable:0kB kernel_stack:0kB pagetables:0kB bounce:0kB free_pcp:800kB local_pcp:608kB free_cma:0kB the oom killer is clearly pre-mature because there there is still a lot of page cache in the zone Normal which should satisfy this lowmem request. Further debugging has shown that the reclaim cannot make any forward progress because the page cache is hidden in the active list which doesn't get rotated because inactive_list_is_low is not memcg aware. The code simply subtracts per-zone highmem counters from the respective memcg's lru sizes which doesn't make any sense. We can simply end up always seeing the resulting active and inactive counts 0 and return false. This issue is not limited to 32b kernels but in practice the effect on systems without CONFIG_HIGHMEM would be much harder to notice because we do not invoke the OOM killer for allocations requests targeting < ZONE_NORMAL. Fix the issue by tracking per zone lru page counts in mem_cgroup_per_node and subtract per-memcg highmem counts when memcg is enabled. Introduce helper lruvec_zone_lru_size which redirects to either zone counters or mem_cgroup_get_zone_lru_size when appropriate. We are losing empty LRU but non-zero lru size detection introduced by ca707239e8a7 ("mm: update_lru_size warn and reset bad lru_size") because of the inherent zone vs. node discrepancy. Fixes: f8d1a31163fc ("mm: consider whether to decivate based on eligible zones inactive ratio") Link: http://lkml.kernel.org/r/20170104100825.3729-1-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Reported-by: Nils Holland <nholland@tisys.org> Tested-by: Nils Holland <nholland@tisys.org> Reported-by: Klaus Ethgen <Klaus@Ethgen.de> Acked-by: Minchan Kim <minchan@kernel.org> Acked-by: Mel Gorman <mgorman@suse.de> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: <stable@vger.kernel.org> [4.8+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-01-11 00:58:04 +00:00
* @zid: zone id of the accounted pages
* @nr_pages: positive when adding or negative when removing
*
* This function must be called under lru_lock, just before a page is added
* to or just after a page is removed from an lru list (that ordering being
* so as to allow it to check that lru_size 0 is consistent with list_empty).
*/
void mem_cgroup_update_lru_size(struct lruvec *lruvec, enum lru_list lru,
mm, memcg: fix the active list aging for lowmem requests when memcg is enabled Nils Holland and Klaus Ethgen have reported unexpected OOM killer invocations with 32b kernel starting with 4.8 kernels kworker/u4:5 invoked oom-killer: gfp_mask=0x2400840(GFP_NOFS|__GFP_NOFAIL), nodemask=0, order=0, oom_score_adj=0 kworker/u4:5 cpuset=/ mems_allowed=0 CPU: 1 PID: 2603 Comm: kworker/u4:5 Not tainted 4.9.0-gentoo #2 [...] Mem-Info: active_anon:58685 inactive_anon:90 isolated_anon:0 active_file:274324 inactive_file:281962 isolated_file:0 unevictable:0 dirty:649 writeback:0 unstable:0 slab_reclaimable:40662 slab_unreclaimable:17754 mapped:7382 shmem:202 pagetables:351 bounce:0 free:206736 free_pcp:332 free_cma:0 Node 0 active_anon:234740kB inactive_anon:360kB active_file:1097296kB inactive_file:1127848kB unevictable:0kB isolated(anon):0kB isolated(file):0kB mapped:29528kB dirty:2596kB writeback:0kB shmem:0kB shmem_thp: 0kB shmem_pmdmapped: 184320kB anon_thp: 808kB writeback_tmp:0kB unstable:0kB pages_scanned:0 all_unreclaimable? no DMA free:3952kB min:788kB low:984kB high:1180kB active_anon:0kB inactive_anon:0kB active_file:7316kB inactive_file:0kB unevictable:0kB writepending:96kB present:15992kB managed:15916kB mlocked:0kB slab_reclaimable:3200kB slab_unreclaimable:1408kB kernel_stack:0kB pagetables:0kB bounce:0kB free_pcp:0kB local_pcp:0kB free_cma:0kB lowmem_reserve[]: 0 813 3474 3474 Normal free:41332kB min:41368kB low:51708kB high:62048kB active_anon:0kB inactive_anon:0kB active_file:532748kB inactive_file:44kB unevictable:0kB writepending:24kB present:897016kB managed:836248kB mlocked:0kB slab_reclaimable:159448kB slab_unreclaimable:69608kB kernel_stack:1112kB pagetables:1404kB bounce:0kB free_pcp:528kB local_pcp:340kB free_cma:0kB lowmem_reserve[]: 0 0 21292 21292 HighMem free:781660kB min:512kB low:34356kB high:68200kB active_anon:234740kB inactive_anon:360kB active_file:557232kB inactive_file:1127804kB unevictable:0kB writepending:2592kB present:2725384kB managed:2725384kB mlocked:0kB slab_reclaimable:0kB slab_unreclaimable:0kB kernel_stack:0kB pagetables:0kB bounce:0kB free_pcp:800kB local_pcp:608kB free_cma:0kB the oom killer is clearly pre-mature because there there is still a lot of page cache in the zone Normal which should satisfy this lowmem request. Further debugging has shown that the reclaim cannot make any forward progress because the page cache is hidden in the active list which doesn't get rotated because inactive_list_is_low is not memcg aware. The code simply subtracts per-zone highmem counters from the respective memcg's lru sizes which doesn't make any sense. We can simply end up always seeing the resulting active and inactive counts 0 and return false. This issue is not limited to 32b kernels but in practice the effect on systems without CONFIG_HIGHMEM would be much harder to notice because we do not invoke the OOM killer for allocations requests targeting < ZONE_NORMAL. Fix the issue by tracking per zone lru page counts in mem_cgroup_per_node and subtract per-memcg highmem counts when memcg is enabled. Introduce helper lruvec_zone_lru_size which redirects to either zone counters or mem_cgroup_get_zone_lru_size when appropriate. We are losing empty LRU but non-zero lru size detection introduced by ca707239e8a7 ("mm: update_lru_size warn and reset bad lru_size") because of the inherent zone vs. node discrepancy. Fixes: f8d1a31163fc ("mm: consider whether to decivate based on eligible zones inactive ratio") Link: http://lkml.kernel.org/r/20170104100825.3729-1-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Reported-by: Nils Holland <nholland@tisys.org> Tested-by: Nils Holland <nholland@tisys.org> Reported-by: Klaus Ethgen <Klaus@Ethgen.de> Acked-by: Minchan Kim <minchan@kernel.org> Acked-by: Mel Gorman <mgorman@suse.de> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: <stable@vger.kernel.org> [4.8+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-01-11 00:58:04 +00:00
int zid, int nr_pages)
{
struct mem_cgroup_per_node *mz;
unsigned long *lru_size;
long size;
if (mem_cgroup_disabled())
return;
mz = container_of(lruvec, struct mem_cgroup_per_node, lruvec);
mm, memcg: fix the active list aging for lowmem requests when memcg is enabled Nils Holland and Klaus Ethgen have reported unexpected OOM killer invocations with 32b kernel starting with 4.8 kernels kworker/u4:5 invoked oom-killer: gfp_mask=0x2400840(GFP_NOFS|__GFP_NOFAIL), nodemask=0, order=0, oom_score_adj=0 kworker/u4:5 cpuset=/ mems_allowed=0 CPU: 1 PID: 2603 Comm: kworker/u4:5 Not tainted 4.9.0-gentoo #2 [...] Mem-Info: active_anon:58685 inactive_anon:90 isolated_anon:0 active_file:274324 inactive_file:281962 isolated_file:0 unevictable:0 dirty:649 writeback:0 unstable:0 slab_reclaimable:40662 slab_unreclaimable:17754 mapped:7382 shmem:202 pagetables:351 bounce:0 free:206736 free_pcp:332 free_cma:0 Node 0 active_anon:234740kB inactive_anon:360kB active_file:1097296kB inactive_file:1127848kB unevictable:0kB isolated(anon):0kB isolated(file):0kB mapped:29528kB dirty:2596kB writeback:0kB shmem:0kB shmem_thp: 0kB shmem_pmdmapped: 184320kB anon_thp: 808kB writeback_tmp:0kB unstable:0kB pages_scanned:0 all_unreclaimable? no DMA free:3952kB min:788kB low:984kB high:1180kB active_anon:0kB inactive_anon:0kB active_file:7316kB inactive_file:0kB unevictable:0kB writepending:96kB present:15992kB managed:15916kB mlocked:0kB slab_reclaimable:3200kB slab_unreclaimable:1408kB kernel_stack:0kB pagetables:0kB bounce:0kB free_pcp:0kB local_pcp:0kB free_cma:0kB lowmem_reserve[]: 0 813 3474 3474 Normal free:41332kB min:41368kB low:51708kB high:62048kB active_anon:0kB inactive_anon:0kB active_file:532748kB inactive_file:44kB unevictable:0kB writepending:24kB present:897016kB managed:836248kB mlocked:0kB slab_reclaimable:159448kB slab_unreclaimable:69608kB kernel_stack:1112kB pagetables:1404kB bounce:0kB free_pcp:528kB local_pcp:340kB free_cma:0kB lowmem_reserve[]: 0 0 21292 21292 HighMem free:781660kB min:512kB low:34356kB high:68200kB active_anon:234740kB inactive_anon:360kB active_file:557232kB inactive_file:1127804kB unevictable:0kB writepending:2592kB present:2725384kB managed:2725384kB mlocked:0kB slab_reclaimable:0kB slab_unreclaimable:0kB kernel_stack:0kB pagetables:0kB bounce:0kB free_pcp:800kB local_pcp:608kB free_cma:0kB the oom killer is clearly pre-mature because there there is still a lot of page cache in the zone Normal which should satisfy this lowmem request. Further debugging has shown that the reclaim cannot make any forward progress because the page cache is hidden in the active list which doesn't get rotated because inactive_list_is_low is not memcg aware. The code simply subtracts per-zone highmem counters from the respective memcg's lru sizes which doesn't make any sense. We can simply end up always seeing the resulting active and inactive counts 0 and return false. This issue is not limited to 32b kernels but in practice the effect on systems without CONFIG_HIGHMEM would be much harder to notice because we do not invoke the OOM killer for allocations requests targeting < ZONE_NORMAL. Fix the issue by tracking per zone lru page counts in mem_cgroup_per_node and subtract per-memcg highmem counts when memcg is enabled. Introduce helper lruvec_zone_lru_size which redirects to either zone counters or mem_cgroup_get_zone_lru_size when appropriate. We are losing empty LRU but non-zero lru size detection introduced by ca707239e8a7 ("mm: update_lru_size warn and reset bad lru_size") because of the inherent zone vs. node discrepancy. Fixes: f8d1a31163fc ("mm: consider whether to decivate based on eligible zones inactive ratio") Link: http://lkml.kernel.org/r/20170104100825.3729-1-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Reported-by: Nils Holland <nholland@tisys.org> Tested-by: Nils Holland <nholland@tisys.org> Reported-by: Klaus Ethgen <Klaus@Ethgen.de> Acked-by: Minchan Kim <minchan@kernel.org> Acked-by: Mel Gorman <mgorman@suse.de> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: <stable@vger.kernel.org> [4.8+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-01-11 00:58:04 +00:00
lru_size = &mz->lru_zone_size[zid][lru];
if (nr_pages < 0)
*lru_size += nr_pages;
size = *lru_size;
mm, memcg: fix the active list aging for lowmem requests when memcg is enabled Nils Holland and Klaus Ethgen have reported unexpected OOM killer invocations with 32b kernel starting with 4.8 kernels kworker/u4:5 invoked oom-killer: gfp_mask=0x2400840(GFP_NOFS|__GFP_NOFAIL), nodemask=0, order=0, oom_score_adj=0 kworker/u4:5 cpuset=/ mems_allowed=0 CPU: 1 PID: 2603 Comm: kworker/u4:5 Not tainted 4.9.0-gentoo #2 [...] Mem-Info: active_anon:58685 inactive_anon:90 isolated_anon:0 active_file:274324 inactive_file:281962 isolated_file:0 unevictable:0 dirty:649 writeback:0 unstable:0 slab_reclaimable:40662 slab_unreclaimable:17754 mapped:7382 shmem:202 pagetables:351 bounce:0 free:206736 free_pcp:332 free_cma:0 Node 0 active_anon:234740kB inactive_anon:360kB active_file:1097296kB inactive_file:1127848kB unevictable:0kB isolated(anon):0kB isolated(file):0kB mapped:29528kB dirty:2596kB writeback:0kB shmem:0kB shmem_thp: 0kB shmem_pmdmapped: 184320kB anon_thp: 808kB writeback_tmp:0kB unstable:0kB pages_scanned:0 all_unreclaimable? no DMA free:3952kB min:788kB low:984kB high:1180kB active_anon:0kB inactive_anon:0kB active_file:7316kB inactive_file:0kB unevictable:0kB writepending:96kB present:15992kB managed:15916kB mlocked:0kB slab_reclaimable:3200kB slab_unreclaimable:1408kB kernel_stack:0kB pagetables:0kB bounce:0kB free_pcp:0kB local_pcp:0kB free_cma:0kB lowmem_reserve[]: 0 813 3474 3474 Normal free:41332kB min:41368kB low:51708kB high:62048kB active_anon:0kB inactive_anon:0kB active_file:532748kB inactive_file:44kB unevictable:0kB writepending:24kB present:897016kB managed:836248kB mlocked:0kB slab_reclaimable:159448kB slab_unreclaimable:69608kB kernel_stack:1112kB pagetables:1404kB bounce:0kB free_pcp:528kB local_pcp:340kB free_cma:0kB lowmem_reserve[]: 0 0 21292 21292 HighMem free:781660kB min:512kB low:34356kB high:68200kB active_anon:234740kB inactive_anon:360kB active_file:557232kB inactive_file:1127804kB unevictable:0kB writepending:2592kB present:2725384kB managed:2725384kB mlocked:0kB slab_reclaimable:0kB slab_unreclaimable:0kB kernel_stack:0kB pagetables:0kB bounce:0kB free_pcp:800kB local_pcp:608kB free_cma:0kB the oom killer is clearly pre-mature because there there is still a lot of page cache in the zone Normal which should satisfy this lowmem request. Further debugging has shown that the reclaim cannot make any forward progress because the page cache is hidden in the active list which doesn't get rotated because inactive_list_is_low is not memcg aware. The code simply subtracts per-zone highmem counters from the respective memcg's lru sizes which doesn't make any sense. We can simply end up always seeing the resulting active and inactive counts 0 and return false. This issue is not limited to 32b kernels but in practice the effect on systems without CONFIG_HIGHMEM would be much harder to notice because we do not invoke the OOM killer for allocations requests targeting < ZONE_NORMAL. Fix the issue by tracking per zone lru page counts in mem_cgroup_per_node and subtract per-memcg highmem counts when memcg is enabled. Introduce helper lruvec_zone_lru_size which redirects to either zone counters or mem_cgroup_get_zone_lru_size when appropriate. We are losing empty LRU but non-zero lru size detection introduced by ca707239e8a7 ("mm: update_lru_size warn and reset bad lru_size") because of the inherent zone vs. node discrepancy. Fixes: f8d1a31163fc ("mm: consider whether to decivate based on eligible zones inactive ratio") Link: http://lkml.kernel.org/r/20170104100825.3729-1-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Reported-by: Nils Holland <nholland@tisys.org> Tested-by: Nils Holland <nholland@tisys.org> Reported-by: Klaus Ethgen <Klaus@Ethgen.de> Acked-by: Minchan Kim <minchan@kernel.org> Acked-by: Mel Gorman <mgorman@suse.de> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: <stable@vger.kernel.org> [4.8+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-01-11 00:58:04 +00:00
if (WARN_ONCE(size < 0,
"%s(%p, %d, %d): lru_size %ld\n",
__func__, lruvec, lru, nr_pages, size)) {
VM_BUG_ON(1);
*lru_size = 0;
}
if (nr_pages > 0)
*lru_size += nr_pages;
memcg: synchronized LRU A big patch for changing memcg's LRU semantics. Now, - page_cgroup is linked to mem_cgroup's its own LRU (per zone). - LRU of page_cgroup is not synchronous with global LRU. - page and page_cgroup is one-to-one and statically allocated. - To find page_cgroup is on what LRU, you have to check pc->mem_cgroup as - lru = page_cgroup_zoneinfo(pc, nid_of_pc, zid_of_pc); - SwapCache is handled. And, when we handle LRU list of page_cgroup, we do following. pc = lookup_page_cgroup(page); lock_page_cgroup(pc); .....................(1) mz = page_cgroup_zoneinfo(pc); spin_lock(&mz->lru_lock); .....add to LRU spin_unlock(&mz->lru_lock); unlock_page_cgroup(pc); But (1) is spin_lock and we have to be afraid of dead-lock with zone->lru_lock. So, trylock() is used at (1), now. Without (1), we can't trust "mz" is correct. This is a trial to remove this dirty nesting of locks. This patch changes mz->lru_lock to be zone->lru_lock. Then, above sequence will be written as spin_lock(&zone->lru_lock); # in vmscan.c or swap.c via global LRU mem_cgroup_add/remove/etc_lru() { pc = lookup_page_cgroup(page); mz = page_cgroup_zoneinfo(pc); if (PageCgroupUsed(pc)) { ....add to LRU } spin_lock(&zone->lru_lock); # in vmscan.c or swap.c via global LRU This is much simpler. (*) We're safe even if we don't take lock_page_cgroup(pc). Because.. 1. When pc->mem_cgroup can be modified. - at charge. - at account_move(). 2. at charge the PCG_USED bit is not set before pc->mem_cgroup is fixed. 3. at account_move() the page is isolated and not on LRU. Pros. - easy for maintenance. - memcg can make use of laziness of pagevec. - we don't have to duplicated LRU/Active/Unevictable bit in page_cgroup. - LRU status of memcg will be synchronized with global LRU's one. - # of locks are reduced. - account_move() is simplified very much. Cons. - may increase cost of LRU rotation. (no impact if memcg is not configured.) Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:08:01 +00:00
}
/**
* mem_cgroup_margin - calculate chargeable space of a memory cgroup
* @memcg: the memory cgroup
*
* Returns the maximum amount of memory @mem can be charged with, in
* pages.
*/
static unsigned long mem_cgroup_margin(struct mem_cgroup *memcg)
{
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
unsigned long margin = 0;
unsigned long count;
unsigned long limit;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
count = page_counter_read(&memcg->memory);
limit = READ_ONCE(memcg->memory.max);
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
if (count < limit)
margin = limit - count;
if (do_memsw_account()) {
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
count = page_counter_read(&memcg->memsw);
limit = READ_ONCE(memcg->memsw.max);
if (count < limit)
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
margin = min(margin, limit - count);
else
margin = 0;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
}
return margin;
}
memcg: avoid lock in updating file_mapped (Was fix race in file_mapped accouting flag management At accounting file events per memory cgroup, we need to find memory cgroup via page_cgroup->mem_cgroup. Now, we use lock_page_cgroup() for guarantee pc->mem_cgroup is not overwritten while we make use of it. But, considering the context which page-cgroup for files are accessed, we can use alternative light-weight mutual execusion in the most case. At handling file-caches, the only race we have to take care of is "moving" account, IOW, overwriting page_cgroup->mem_cgroup. (See comment in the patch) Unlike charge/uncharge, "move" happens not so frequently. It happens only when rmdir() and task-moving (with a special settings.) This patch adds a race-checker for file-cache-status accounting v.s. account moving. The new per-cpu-per-memcg counter MEM_CGROUP_ON_MOVE is added. The routine for account move 1. Increment it before start moving 2. Call synchronize_rcu() 3. Decrement it after the end of moving. By this, file-status-counting routine can check it needs to call lock_page_cgroup(). In most case, I doesn't need to call it. Following is a perf data of a process which mmap()/munmap 32MB of file cache in a minute. Before patch: 28.25% mmap mmap [.] main 22.64% mmap [kernel.kallsyms] [k] page_fault 9.96% mmap [kernel.kallsyms] [k] mem_cgroup_update_file_mapped 3.67% mmap [kernel.kallsyms] [k] filemap_fault 3.50% mmap [kernel.kallsyms] [k] unmap_vmas 2.99% mmap [kernel.kallsyms] [k] __do_fault 2.76% mmap [kernel.kallsyms] [k] find_get_page After patch: 30.00% mmap mmap [.] main 23.78% mmap [kernel.kallsyms] [k] page_fault 5.52% mmap [kernel.kallsyms] [k] mem_cgroup_update_file_mapped 3.81% mmap [kernel.kallsyms] [k] unmap_vmas 3.26% mmap [kernel.kallsyms] [k] find_get_page 3.18% mmap [kernel.kallsyms] [k] __do_fault 3.03% mmap [kernel.kallsyms] [k] filemap_fault 2.40% mmap [kernel.kallsyms] [k] handle_mm_fault 2.40% mmap [kernel.kallsyms] [k] do_page_fault This patch reduces memcg's cost to some extent. (mem_cgroup_update_file_mapped is called by both of map/unmap) Note: It seems some more improvements are required..but no idea. maybe removing set/unset flag is required. Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-10-27 22:33:40 +00:00
/*
* A routine for checking "mem" is under move_account() or not.
memcg: avoid lock in updating file_mapped (Was fix race in file_mapped accouting flag management At accounting file events per memory cgroup, we need to find memory cgroup via page_cgroup->mem_cgroup. Now, we use lock_page_cgroup() for guarantee pc->mem_cgroup is not overwritten while we make use of it. But, considering the context which page-cgroup for files are accessed, we can use alternative light-weight mutual execusion in the most case. At handling file-caches, the only race we have to take care of is "moving" account, IOW, overwriting page_cgroup->mem_cgroup. (See comment in the patch) Unlike charge/uncharge, "move" happens not so frequently. It happens only when rmdir() and task-moving (with a special settings.) This patch adds a race-checker for file-cache-status accounting v.s. account moving. The new per-cpu-per-memcg counter MEM_CGROUP_ON_MOVE is added. The routine for account move 1. Increment it before start moving 2. Call synchronize_rcu() 3. Decrement it after the end of moving. By this, file-status-counting routine can check it needs to call lock_page_cgroup(). In most case, I doesn't need to call it. Following is a perf data of a process which mmap()/munmap 32MB of file cache in a minute. Before patch: 28.25% mmap mmap [.] main 22.64% mmap [kernel.kallsyms] [k] page_fault 9.96% mmap [kernel.kallsyms] [k] mem_cgroup_update_file_mapped 3.67% mmap [kernel.kallsyms] [k] filemap_fault 3.50% mmap [kernel.kallsyms] [k] unmap_vmas 2.99% mmap [kernel.kallsyms] [k] __do_fault 2.76% mmap [kernel.kallsyms] [k] find_get_page After patch: 30.00% mmap mmap [.] main 23.78% mmap [kernel.kallsyms] [k] page_fault 5.52% mmap [kernel.kallsyms] [k] mem_cgroup_update_file_mapped 3.81% mmap [kernel.kallsyms] [k] unmap_vmas 3.26% mmap [kernel.kallsyms] [k] find_get_page 3.18% mmap [kernel.kallsyms] [k] __do_fault 3.03% mmap [kernel.kallsyms] [k] filemap_fault 2.40% mmap [kernel.kallsyms] [k] handle_mm_fault 2.40% mmap [kernel.kallsyms] [k] do_page_fault This patch reduces memcg's cost to some extent. (mem_cgroup_update_file_mapped is called by both of map/unmap) Note: It seems some more improvements are required..but no idea. maybe removing set/unset flag is required. Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-10-27 22:33:40 +00:00
*
* Checking a cgroup is mc.from or mc.to or under hierarchy of
* moving cgroups. This is for waiting at high-memory pressure
* caused by "move".
memcg: avoid lock in updating file_mapped (Was fix race in file_mapped accouting flag management At accounting file events per memory cgroup, we need to find memory cgroup via page_cgroup->mem_cgroup. Now, we use lock_page_cgroup() for guarantee pc->mem_cgroup is not overwritten while we make use of it. But, considering the context which page-cgroup for files are accessed, we can use alternative light-weight mutual execusion in the most case. At handling file-caches, the only race we have to take care of is "moving" account, IOW, overwriting page_cgroup->mem_cgroup. (See comment in the patch) Unlike charge/uncharge, "move" happens not so frequently. It happens only when rmdir() and task-moving (with a special settings.) This patch adds a race-checker for file-cache-status accounting v.s. account moving. The new per-cpu-per-memcg counter MEM_CGROUP_ON_MOVE is added. The routine for account move 1. Increment it before start moving 2. Call synchronize_rcu() 3. Decrement it after the end of moving. By this, file-status-counting routine can check it needs to call lock_page_cgroup(). In most case, I doesn't need to call it. Following is a perf data of a process which mmap()/munmap 32MB of file cache in a minute. Before patch: 28.25% mmap mmap [.] main 22.64% mmap [kernel.kallsyms] [k] page_fault 9.96% mmap [kernel.kallsyms] [k] mem_cgroup_update_file_mapped 3.67% mmap [kernel.kallsyms] [k] filemap_fault 3.50% mmap [kernel.kallsyms] [k] unmap_vmas 2.99% mmap [kernel.kallsyms] [k] __do_fault 2.76% mmap [kernel.kallsyms] [k] find_get_page After patch: 30.00% mmap mmap [.] main 23.78% mmap [kernel.kallsyms] [k] page_fault 5.52% mmap [kernel.kallsyms] [k] mem_cgroup_update_file_mapped 3.81% mmap [kernel.kallsyms] [k] unmap_vmas 3.26% mmap [kernel.kallsyms] [k] find_get_page 3.18% mmap [kernel.kallsyms] [k] __do_fault 3.03% mmap [kernel.kallsyms] [k] filemap_fault 2.40% mmap [kernel.kallsyms] [k] handle_mm_fault 2.40% mmap [kernel.kallsyms] [k] do_page_fault This patch reduces memcg's cost to some extent. (mem_cgroup_update_file_mapped is called by both of map/unmap) Note: It seems some more improvements are required..but no idea. maybe removing set/unset flag is required. Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-10-27 22:33:40 +00:00
*/
static bool mem_cgroup_under_move(struct mem_cgroup *memcg)
{
struct mem_cgroup *from;
struct mem_cgroup *to;
bool ret = false;
/*
* Unlike task_move routines, we access mc.to, mc.from not under
* mutual exclusion by cgroup_mutex. Here, we take spinlock instead.
*/
spin_lock(&mc.lock);
from = mc.from;
to = mc.to;
if (!from)
goto unlock;
ret = mem_cgroup_is_descendant(from, memcg) ||
mem_cgroup_is_descendant(to, memcg);
unlock:
spin_unlock(&mc.lock);
return ret;
}
static bool mem_cgroup_wait_acct_move(struct mem_cgroup *memcg)
{
if (mc.moving_task && current != mc.moving_task) {
if (mem_cgroup_under_move(memcg)) {
DEFINE_WAIT(wait);
prepare_to_wait(&mc.waitq, &wait, TASK_INTERRUPTIBLE);
/* moving charge context might have finished. */
if (mc.moving_task)
schedule();
finish_wait(&mc.waitq, &wait);
return true;
}
}
return false;
}
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
static char *memory_stat_format(struct mem_cgroup *memcg)
{
struct seq_buf s;
int i;
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
seq_buf_init(&s, kmalloc(PAGE_SIZE, GFP_KERNEL), PAGE_SIZE);
if (!s.buffer)
return NULL;
/*
* Provide statistics on the state of the memory subsystem as
* well as cumulative event counters that show past behavior.
*
* This list is ordered following a combination of these gradients:
* 1) generic big picture -> specifics and details
* 2) reflecting userspace activity -> reflecting kernel heuristics
*
* Current memory state:
*/
seq_buf_printf(&s, "anon %llu\n",
mm: memcontrol: switch to native NR_ANON_MAPPED counter Memcg maintains a private MEMCG_RSS counter. This divergence from the generic VM accounting means unnecessary code overhead, and creates a dependency for memcg that page->mapping is set up at the time of charging, so that page types can be told apart. Convert the generic accounting sites to mod_lruvec_page_state and friends to maintain the per-cgroup vmstat counter of NR_ANON_MAPPED. We use lock_page_memcg() to stabilize page->mem_cgroup during rmap changes, the same way we do for NR_FILE_MAPPED. With the previous patch removing MEMCG_CACHE and the private NR_SHMEM counter, this patch finally eliminates the need to have page->mapping set up at charge time. However, we need to have page->mem_cgroup set up by the time rmap runs and does the accounting, so switch the commit and the rmap callbacks around. v2: fix temporary accounting bug by switching rmap<->commit (Joonsoo) Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Cc: Alex Shi <alex.shi@linux.alibaba.com> Cc: Hugh Dickins <hughd@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.com> Cc: Roman Gushchin <guro@fb.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Balbir Singh <bsingharora@gmail.com> Link: http://lkml.kernel.org/r/20200508183105.225460-11-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:01:57 +00:00
(u64)memcg_page_state(memcg, NR_ANON_MAPPED) *
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
PAGE_SIZE);
seq_buf_printf(&s, "file %llu\n",
(u64)memcg_page_state(memcg, NR_FILE_PAGES) *
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
PAGE_SIZE);
seq_buf_printf(&s, "kernel_stack %llu\n",
(u64)memcg_page_state(memcg, MEMCG_KERNEL_STACK_KB) *
1024);
seq_buf_printf(&s, "slab %llu\n",
(u64)(memcg_page_state(memcg, NR_SLAB_RECLAIMABLE) +
memcg_page_state(memcg, NR_SLAB_UNRECLAIMABLE)) *
PAGE_SIZE);
seq_buf_printf(&s, "sock %llu\n",
(u64)memcg_page_state(memcg, MEMCG_SOCK) *
PAGE_SIZE);
seq_buf_printf(&s, "shmem %llu\n",
(u64)memcg_page_state(memcg, NR_SHMEM) *
PAGE_SIZE);
seq_buf_printf(&s, "file_mapped %llu\n",
(u64)memcg_page_state(memcg, NR_FILE_MAPPED) *
PAGE_SIZE);
seq_buf_printf(&s, "file_dirty %llu\n",
(u64)memcg_page_state(memcg, NR_FILE_DIRTY) *
PAGE_SIZE);
seq_buf_printf(&s, "file_writeback %llu\n",
(u64)memcg_page_state(memcg, NR_WRITEBACK) *
PAGE_SIZE);
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
seq_buf_printf(&s, "anon_thp %llu\n",
(u64)memcg_page_state(memcg, NR_ANON_THPS) *
HPAGE_PMD_SIZE);
#endif
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
for (i = 0; i < NR_LRU_LISTS; i++)
seq_buf_printf(&s, "%s %llu\n", lru_list_name(i),
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
(u64)memcg_page_state(memcg, NR_LRU_BASE + i) *
PAGE_SIZE);
seq_buf_printf(&s, "slab_reclaimable %llu\n",
(u64)memcg_page_state(memcg, NR_SLAB_RECLAIMABLE) *
PAGE_SIZE);
seq_buf_printf(&s, "slab_unreclaimable %llu\n",
(u64)memcg_page_state(memcg, NR_SLAB_UNRECLAIMABLE) *
PAGE_SIZE);
/* Accumulated memory events */
seq_buf_printf(&s, "%s %lu\n", vm_event_name(PGFAULT),
memcg_events(memcg, PGFAULT));
seq_buf_printf(&s, "%s %lu\n", vm_event_name(PGMAJFAULT),
memcg_events(memcg, PGMAJFAULT));
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
seq_buf_printf(&s, "workingset_refault %lu\n",
memcg_page_state(memcg, WORKINGSET_REFAULT));
seq_buf_printf(&s, "workingset_activate %lu\n",
memcg_page_state(memcg, WORKINGSET_ACTIVATE));
mm, memcg: add workingset_restore in memory.stat There's a new workingset counter introduced in commit 1899ad18c607 ("mm: workingset: tell cache transitions from workingset thrashing"). With the help of this counter we can know the workingset is transitioning or thrashing. To leverage the benifit of this counter to memcg, we should introduce it into memory.stat. Then we could know the workingset of the workload inside a memcg better. Bellow is the verification of this new counter in memory.stat. Read a file into the memory and then read it again to make these pages be active. The size of this file is 1G. (memory.max is greater than file size) The counters in memory.stat will be inactive_file 0 active_file 1073639424 workingset_refault 0 workingset_activate 0 workingset_restore 0 workingset_nodereclaim 0 Trigger the memcg reclaim by setting a lower value to memory.high, and then some pages will be demoted into inactive list, and then some pages in the inactive list will be evicted into the storage. inactive_file 498094080 active_file 310063104 workingset_refault 0 workingset_activate 0 workingset_restore 0 workingset_nodereclaim 0 Then recover the memory.high and read the file into memory again. As a result of it, the transitioning will occur. Bellow is the result of this transitioning, inactive_file 498094080 active_file 575397888 workingset_refault 64746 workingset_activate 64746 workingset_restore 64746 workingset_nodereclaim 0 Signed-off-by: Yafang Shao <laoar.shao@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Chris Down <chris@chrisdown.name> Cc: Peter Zijlstra (Intel) <peterz@infradead.org> Cc: Suren Baghdasaryan <surenb@google.com> Cc: Shakeel Butt <shakeelb@google.com> Link: http://lkml.kernel.org/r/20200504153522.11553-1-laoar.shao@gmail.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:32 +00:00
seq_buf_printf(&s, "workingset_restore %lu\n",
memcg_page_state(memcg, WORKINGSET_RESTORE));
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
seq_buf_printf(&s, "workingset_nodereclaim %lu\n",
memcg_page_state(memcg, WORKINGSET_NODERECLAIM));
seq_buf_printf(&s, "%s %lu\n", vm_event_name(PGREFILL),
memcg_events(memcg, PGREFILL));
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
seq_buf_printf(&s, "pgscan %lu\n",
memcg_events(memcg, PGSCAN_KSWAPD) +
memcg_events(memcg, PGSCAN_DIRECT));
seq_buf_printf(&s, "pgsteal %lu\n",
memcg_events(memcg, PGSTEAL_KSWAPD) +
memcg_events(memcg, PGSTEAL_DIRECT));
seq_buf_printf(&s, "%s %lu\n", vm_event_name(PGACTIVATE),
memcg_events(memcg, PGACTIVATE));
seq_buf_printf(&s, "%s %lu\n", vm_event_name(PGDEACTIVATE),
memcg_events(memcg, PGDEACTIVATE));
seq_buf_printf(&s, "%s %lu\n", vm_event_name(PGLAZYFREE),
memcg_events(memcg, PGLAZYFREE));
seq_buf_printf(&s, "%s %lu\n", vm_event_name(PGLAZYFREED),
memcg_events(memcg, PGLAZYFREED));
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
seq_buf_printf(&s, "%s %lu\n", vm_event_name(THP_FAULT_ALLOC),
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
memcg_events(memcg, THP_FAULT_ALLOC));
seq_buf_printf(&s, "%s %lu\n", vm_event_name(THP_COLLAPSE_ALLOC),
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
memcg_events(memcg, THP_COLLAPSE_ALLOC));
#endif /* CONFIG_TRANSPARENT_HUGEPAGE */
/* The above should easily fit into one page */
WARN_ON_ONCE(seq_buf_has_overflowed(&s));
return s.buffer;
}
memcg, oom: provide more precise dump info while memcg oom happening Currently when a memcg oom is happening the oom dump messages is still global state and provides few useful info for users. This patch prints more pointed memcg page statistics for memcg-oom and take hierarchy into consideration: Based on Michal's advice, we take hierarchy into consideration: supppose we trigger an OOM on A's limit root_memcg | A (use_hierachy=1) / \ B C | D then the printed info will be: Memory cgroup stats for /A:... Memory cgroup stats for /A/B:... Memory cgroup stats for /A/C:... Memory cgroup stats for /A/B/D:... Following are samples of oom output: (1) Before change: mal-80 invoked oom-killer:gfp_mask=0xd0, order=0, oom_score_adj=0 mal-80 cpuset=/ mems_allowed=0 Pid: 2976, comm: mal-80 Not tainted 3.7.0+ #10 Call Trace: [<ffffffff8167fbfb>] dump_header+0x83/0x1ca ..... (call trace) [<ffffffff8168a818>] page_fault+0x28/0x30 <<<<<<<<<<<<<<<<<<<<< memcg specific information Task in /A/B/D killed as a result of limit of /A memory: usage 101376kB, limit 101376kB, failcnt 57 memory+swap: usage 101376kB, limit 101376kB, failcnt 0 kmem: usage 0kB, limit 9007199254740991kB, failcnt 0 <<<<<<<<<<<<<<<<<<<<< print per cpu pageset stat Mem-Info: Node 0 DMA per-cpu: CPU 0: hi: 0, btch: 1 usd: 0 ...... CPU 3: hi: 0, btch: 1 usd: 0 Node 0 DMA32 per-cpu: CPU 0: hi: 186, btch: 31 usd: 173 ...... CPU 3: hi: 186, btch: 31 usd: 130 <<<<<<<<<<<<<<<<<<<<< print global page state active_anon:92963 inactive_anon:40777 isolated_anon:0 active_file:33027 inactive_file:51718 isolated_file:0 unevictable:0 dirty:3 writeback:0 unstable:0 free:729995 slab_reclaimable:6897 slab_unreclaimable:6263 mapped:20278 shmem:35971 pagetables:5885 bounce:0 free_cma:0 <<<<<<<<<<<<<<<<<<<<< print per zone page state Node 0 DMA free:15836kB ... all_unreclaimable? no lowmem_reserve[]: 0 3175 3899 3899 Node 0 DMA32 free:2888564kB ... all_unrelaimable? no lowmem_reserve[]: 0 0 724 724 lowmem_reserve[]: 0 0 0 0 Node 0 DMA: 1*4kB (U) ... 3*4096kB (M) = 15836kB Node 0 DMA32: 41*4kB (UM) ... 702*4096kB (MR) = 2888316kB 120710 total pagecache pages 0 pages in swap cache <<<<<<<<<<<<<<<<<<<<< print global swap cache stat Swap cache stats: add 0, delete 0, find 0/0 Free swap = 499708kB Total swap = 499708kB 1040368 pages RAM 58678 pages reserved 169065 pages shared 173632 pages non-shared [ pid ] uid tgid total_vm rss nr_ptes swapents oom_score_adj name [ 2693] 0 2693 6005 1324 17 0 0 god [ 2754] 0 2754 6003 1320 16 0 0 god [ 2811] 0 2811 5992 1304 18 0 0 god [ 2874] 0 2874 6005 1323 18 0 0 god [ 2935] 0 2935 8720 7742 21 0 0 mal-30 [ 2976] 0 2976 21520 17577 42 0 0 mal-80 Memory cgroup out of memory: Kill process 2976 (mal-80) score 665 or sacrifice child Killed process 2976 (mal-80) total-vm:86080kB, anon-rss:69964kB, file-rss:344kB We can see that messages dumped by show_free_areas() are longsome and can provide so limited info for memcg that just happen oom. (2) After change mal-80 invoked oom-killer: gfp_mask=0xd0, order=0, oom_score_adj=0 mal-80 cpuset=/ mems_allowed=0 Pid: 2704, comm: mal-80 Not tainted 3.7.0+ #10 Call Trace: [<ffffffff8167fd0b>] dump_header+0x83/0x1d1 .......(call trace) [<ffffffff8168a918>] page_fault+0x28/0x30 Task in /A/B/D killed as a result of limit of /A <<<<<<<<<<<<<<<<<<<<< memcg specific information memory: usage 102400kB, limit 102400kB, failcnt 140 memory+swap: usage 102400kB, limit 102400kB, failcnt 0 kmem: usage 0kB, limit 9007199254740991kB, failcnt 0 Memory cgroup stats for /A: cache:32KB rss:30984KB mapped_file:0KB swap:0KB inactive_anon:6912KB active_anon:24072KB inactive_file:32KB active_file:0KB unevictable:0KB Memory cgroup stats for /A/B: cache:0KB rss:0KB mapped_file:0KB swap:0KB inactive_anon:0KB active_anon:0KB inactive_file:0KB active_file:0KB unevictable:0KB Memory cgroup stats for /A/C: cache:0KB rss:0KB mapped_file:0KB swap:0KB inactive_anon:0KB active_anon:0KB inactive_file:0KB active_file:0KB unevictable:0KB Memory cgroup stats for /A/B/D: cache:32KB rss:71352KB mapped_file:0KB swap:0KB inactive_anon:6656KB active_anon:64696KB inactive_file:16KB active_file:16KB unevictable:0KB [ pid ] uid tgid total_vm rss nr_ptes swapents oom_score_adj name [ 2260] 0 2260 6006 1325 18 0 0 god [ 2383] 0 2383 6003 1319 17 0 0 god [ 2503] 0 2503 6004 1321 18 0 0 god [ 2622] 0 2622 6004 1321 16 0 0 god [ 2695] 0 2695 8720 7741 22 0 0 mal-30 [ 2704] 0 2704 21520 17839 43 0 0 mal-80 Memory cgroup out of memory: Kill process 2704 (mal-80) score 669 or sacrifice child Killed process 2704 (mal-80) total-vm:86080kB, anon-rss:71016kB, file-rss:340kB This version provides more pointed info for memcg in "Memory cgroup stats for XXX" section. Signed-off-by: Sha Zhengju <handai.szj@taobao.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: David Rientjes <rientjes@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-23 00:32:05 +00:00
#define K(x) ((x) << (PAGE_SHIFT-10))
/**
mm, oom: add oom victim's memcg to the oom context information The current oom report doesn't display victim's memcg context during the global OOM situation. While this information is not strictly needed, it can be really helpful for containerized environments to locate which container has lost a process. Now that we have a single line for the oom context, we can trivially add both the oom memcg (this can be either global_oom or a specific memcg which hits its hard limits) and task_memcg which is the victim's memcg. Below is the single line output in the oom report after this patch. - global oom context information: oom-kill:constraint=<constraint>,nodemask=<nodemask>,cpuset=<cpuset>,mems_allowed=<mems_allowed>,global_oom,task_memcg=<memcg>,task=<comm>,pid=<pid>,uid=<uid> - memcg oom context information: oom-kill:constraint=<constraint>,nodemask=<nodemask>,cpuset=<cpuset>,mems_allowed=<mems_allowed>,oom_memcg=<memcg>,task_memcg=<memcg>,task=<comm>,pid=<pid>,uid=<uid> [penguin-kernel@I-love.SAKURA.ne.jp: use pr_cont() in mem_cgroup_print_oom_context()] Link: http://lkml.kernel.org/r/201812190723.wBJ7NdkN032628@www262.sakura.ne.jp Link: http://lkml.kernel.org/r/1542799799-36184-2-git-send-email-ufo19890607@gmail.com Signed-off-by: yuzhoujian <yuzhoujian@didichuxing.com> Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Acked-by: Michal Hocko <mhocko@suse.com> Cc: David Rientjes <rientjes@google.com> Cc: "Kirill A . Shutemov" <kirill.shutemov@linux.intel.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp> Cc: Roman Gushchin <guro@fb.com> Cc: Yang Shi <yang.s@alibaba-inc.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-12-28 08:36:10 +00:00
* mem_cgroup_print_oom_context: Print OOM information relevant to
* memory controller.
* @memcg: The memory cgroup that went over limit
* @p: Task that is going to be killed
*
* NOTE: @memcg and @p's mem_cgroup can be different when hierarchy is
* enabled
*/
mm, oom: add oom victim's memcg to the oom context information The current oom report doesn't display victim's memcg context during the global OOM situation. While this information is not strictly needed, it can be really helpful for containerized environments to locate which container has lost a process. Now that we have a single line for the oom context, we can trivially add both the oom memcg (this can be either global_oom or a specific memcg which hits its hard limits) and task_memcg which is the victim's memcg. Below is the single line output in the oom report after this patch. - global oom context information: oom-kill:constraint=<constraint>,nodemask=<nodemask>,cpuset=<cpuset>,mems_allowed=<mems_allowed>,global_oom,task_memcg=<memcg>,task=<comm>,pid=<pid>,uid=<uid> - memcg oom context information: oom-kill:constraint=<constraint>,nodemask=<nodemask>,cpuset=<cpuset>,mems_allowed=<mems_allowed>,oom_memcg=<memcg>,task_memcg=<memcg>,task=<comm>,pid=<pid>,uid=<uid> [penguin-kernel@I-love.SAKURA.ne.jp: use pr_cont() in mem_cgroup_print_oom_context()] Link: http://lkml.kernel.org/r/201812190723.wBJ7NdkN032628@www262.sakura.ne.jp Link: http://lkml.kernel.org/r/1542799799-36184-2-git-send-email-ufo19890607@gmail.com Signed-off-by: yuzhoujian <yuzhoujian@didichuxing.com> Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Acked-by: Michal Hocko <mhocko@suse.com> Cc: David Rientjes <rientjes@google.com> Cc: "Kirill A . Shutemov" <kirill.shutemov@linux.intel.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp> Cc: Roman Gushchin <guro@fb.com> Cc: Yang Shi <yang.s@alibaba-inc.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-12-28 08:36:10 +00:00
void mem_cgroup_print_oom_context(struct mem_cgroup *memcg, struct task_struct *p)
{
rcu_read_lock();
mm, oom: add oom victim's memcg to the oom context information The current oom report doesn't display victim's memcg context during the global OOM situation. While this information is not strictly needed, it can be really helpful for containerized environments to locate which container has lost a process. Now that we have a single line for the oom context, we can trivially add both the oom memcg (this can be either global_oom or a specific memcg which hits its hard limits) and task_memcg which is the victim's memcg. Below is the single line output in the oom report after this patch. - global oom context information: oom-kill:constraint=<constraint>,nodemask=<nodemask>,cpuset=<cpuset>,mems_allowed=<mems_allowed>,global_oom,task_memcg=<memcg>,task=<comm>,pid=<pid>,uid=<uid> - memcg oom context information: oom-kill:constraint=<constraint>,nodemask=<nodemask>,cpuset=<cpuset>,mems_allowed=<mems_allowed>,oom_memcg=<memcg>,task_memcg=<memcg>,task=<comm>,pid=<pid>,uid=<uid> [penguin-kernel@I-love.SAKURA.ne.jp: use pr_cont() in mem_cgroup_print_oom_context()] Link: http://lkml.kernel.org/r/201812190723.wBJ7NdkN032628@www262.sakura.ne.jp Link: http://lkml.kernel.org/r/1542799799-36184-2-git-send-email-ufo19890607@gmail.com Signed-off-by: yuzhoujian <yuzhoujian@didichuxing.com> Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Acked-by: Michal Hocko <mhocko@suse.com> Cc: David Rientjes <rientjes@google.com> Cc: "Kirill A . Shutemov" <kirill.shutemov@linux.intel.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp> Cc: Roman Gushchin <guro@fb.com> Cc: Yang Shi <yang.s@alibaba-inc.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-12-28 08:36:10 +00:00
if (memcg) {
pr_cont(",oom_memcg=");
pr_cont_cgroup_path(memcg->css.cgroup);
} else
pr_cont(",global_oom");
if (p) {
mm, oom: add oom victim's memcg to the oom context information The current oom report doesn't display victim's memcg context during the global OOM situation. While this information is not strictly needed, it can be really helpful for containerized environments to locate which container has lost a process. Now that we have a single line for the oom context, we can trivially add both the oom memcg (this can be either global_oom or a specific memcg which hits its hard limits) and task_memcg which is the victim's memcg. Below is the single line output in the oom report after this patch. - global oom context information: oom-kill:constraint=<constraint>,nodemask=<nodemask>,cpuset=<cpuset>,mems_allowed=<mems_allowed>,global_oom,task_memcg=<memcg>,task=<comm>,pid=<pid>,uid=<uid> - memcg oom context information: oom-kill:constraint=<constraint>,nodemask=<nodemask>,cpuset=<cpuset>,mems_allowed=<mems_allowed>,oom_memcg=<memcg>,task_memcg=<memcg>,task=<comm>,pid=<pid>,uid=<uid> [penguin-kernel@I-love.SAKURA.ne.jp: use pr_cont() in mem_cgroup_print_oom_context()] Link: http://lkml.kernel.org/r/201812190723.wBJ7NdkN032628@www262.sakura.ne.jp Link: http://lkml.kernel.org/r/1542799799-36184-2-git-send-email-ufo19890607@gmail.com Signed-off-by: yuzhoujian <yuzhoujian@didichuxing.com> Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Acked-by: Michal Hocko <mhocko@suse.com> Cc: David Rientjes <rientjes@google.com> Cc: "Kirill A . Shutemov" <kirill.shutemov@linux.intel.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp> Cc: Roman Gushchin <guro@fb.com> Cc: Yang Shi <yang.s@alibaba-inc.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-12-28 08:36:10 +00:00
pr_cont(",task_memcg=");
pr_cont_cgroup_path(task_cgroup(p, memory_cgrp_id));
}
rcu_read_unlock();
mm, oom: add oom victim's memcg to the oom context information The current oom report doesn't display victim's memcg context during the global OOM situation. While this information is not strictly needed, it can be really helpful for containerized environments to locate which container has lost a process. Now that we have a single line for the oom context, we can trivially add both the oom memcg (this can be either global_oom or a specific memcg which hits its hard limits) and task_memcg which is the victim's memcg. Below is the single line output in the oom report after this patch. - global oom context information: oom-kill:constraint=<constraint>,nodemask=<nodemask>,cpuset=<cpuset>,mems_allowed=<mems_allowed>,global_oom,task_memcg=<memcg>,task=<comm>,pid=<pid>,uid=<uid> - memcg oom context information: oom-kill:constraint=<constraint>,nodemask=<nodemask>,cpuset=<cpuset>,mems_allowed=<mems_allowed>,oom_memcg=<memcg>,task_memcg=<memcg>,task=<comm>,pid=<pid>,uid=<uid> [penguin-kernel@I-love.SAKURA.ne.jp: use pr_cont() in mem_cgroup_print_oom_context()] Link: http://lkml.kernel.org/r/201812190723.wBJ7NdkN032628@www262.sakura.ne.jp Link: http://lkml.kernel.org/r/1542799799-36184-2-git-send-email-ufo19890607@gmail.com Signed-off-by: yuzhoujian <yuzhoujian@didichuxing.com> Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Acked-by: Michal Hocko <mhocko@suse.com> Cc: David Rientjes <rientjes@google.com> Cc: "Kirill A . Shutemov" <kirill.shutemov@linux.intel.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp> Cc: Roman Gushchin <guro@fb.com> Cc: Yang Shi <yang.s@alibaba-inc.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-12-28 08:36:10 +00:00
}
/**
* mem_cgroup_print_oom_meminfo: Print OOM memory information relevant to
* memory controller.
* @memcg: The memory cgroup that went over limit
*/
void mem_cgroup_print_oom_meminfo(struct mem_cgroup *memcg)
{
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
char *buf;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
pr_info("memory: usage %llukB, limit %llukB, failcnt %lu\n",
K((u64)page_counter_read(&memcg->memory)),
K((u64)READ_ONCE(memcg->memory.max)), memcg->memory.failcnt);
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
if (cgroup_subsys_on_dfl(memory_cgrp_subsys))
pr_info("swap: usage %llukB, limit %llukB, failcnt %lu\n",
K((u64)page_counter_read(&memcg->swap)),
K((u64)READ_ONCE(memcg->swap.max)), memcg->swap.failcnt);
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
else {
pr_info("memory+swap: usage %llukB, limit %llukB, failcnt %lu\n",
K((u64)page_counter_read(&memcg->memsw)),
K((u64)memcg->memsw.max), memcg->memsw.failcnt);
pr_info("kmem: usage %llukB, limit %llukB, failcnt %lu\n",
K((u64)page_counter_read(&memcg->kmem)),
K((u64)memcg->kmem.max), memcg->kmem.failcnt);
memcg, oom: provide more precise dump info while memcg oom happening Currently when a memcg oom is happening the oom dump messages is still global state and provides few useful info for users. This patch prints more pointed memcg page statistics for memcg-oom and take hierarchy into consideration: Based on Michal's advice, we take hierarchy into consideration: supppose we trigger an OOM on A's limit root_memcg | A (use_hierachy=1) / \ B C | D then the printed info will be: Memory cgroup stats for /A:... Memory cgroup stats for /A/B:... Memory cgroup stats for /A/C:... Memory cgroup stats for /A/B/D:... Following are samples of oom output: (1) Before change: mal-80 invoked oom-killer:gfp_mask=0xd0, order=0, oom_score_adj=0 mal-80 cpuset=/ mems_allowed=0 Pid: 2976, comm: mal-80 Not tainted 3.7.0+ #10 Call Trace: [<ffffffff8167fbfb>] dump_header+0x83/0x1ca ..... (call trace) [<ffffffff8168a818>] page_fault+0x28/0x30 <<<<<<<<<<<<<<<<<<<<< memcg specific information Task in /A/B/D killed as a result of limit of /A memory: usage 101376kB, limit 101376kB, failcnt 57 memory+swap: usage 101376kB, limit 101376kB, failcnt 0 kmem: usage 0kB, limit 9007199254740991kB, failcnt 0 <<<<<<<<<<<<<<<<<<<<< print per cpu pageset stat Mem-Info: Node 0 DMA per-cpu: CPU 0: hi: 0, btch: 1 usd: 0 ...... CPU 3: hi: 0, btch: 1 usd: 0 Node 0 DMA32 per-cpu: CPU 0: hi: 186, btch: 31 usd: 173 ...... CPU 3: hi: 186, btch: 31 usd: 130 <<<<<<<<<<<<<<<<<<<<< print global page state active_anon:92963 inactive_anon:40777 isolated_anon:0 active_file:33027 inactive_file:51718 isolated_file:0 unevictable:0 dirty:3 writeback:0 unstable:0 free:729995 slab_reclaimable:6897 slab_unreclaimable:6263 mapped:20278 shmem:35971 pagetables:5885 bounce:0 free_cma:0 <<<<<<<<<<<<<<<<<<<<< print per zone page state Node 0 DMA free:15836kB ... all_unreclaimable? no lowmem_reserve[]: 0 3175 3899 3899 Node 0 DMA32 free:2888564kB ... all_unrelaimable? no lowmem_reserve[]: 0 0 724 724 lowmem_reserve[]: 0 0 0 0 Node 0 DMA: 1*4kB (U) ... 3*4096kB (M) = 15836kB Node 0 DMA32: 41*4kB (UM) ... 702*4096kB (MR) = 2888316kB 120710 total pagecache pages 0 pages in swap cache <<<<<<<<<<<<<<<<<<<<< print global swap cache stat Swap cache stats: add 0, delete 0, find 0/0 Free swap = 499708kB Total swap = 499708kB 1040368 pages RAM 58678 pages reserved 169065 pages shared 173632 pages non-shared [ pid ] uid tgid total_vm rss nr_ptes swapents oom_score_adj name [ 2693] 0 2693 6005 1324 17 0 0 god [ 2754] 0 2754 6003 1320 16 0 0 god [ 2811] 0 2811 5992 1304 18 0 0 god [ 2874] 0 2874 6005 1323 18 0 0 god [ 2935] 0 2935 8720 7742 21 0 0 mal-30 [ 2976] 0 2976 21520 17577 42 0 0 mal-80 Memory cgroup out of memory: Kill process 2976 (mal-80) score 665 or sacrifice child Killed process 2976 (mal-80) total-vm:86080kB, anon-rss:69964kB, file-rss:344kB We can see that messages dumped by show_free_areas() are longsome and can provide so limited info for memcg that just happen oom. (2) After change mal-80 invoked oom-killer: gfp_mask=0xd0, order=0, oom_score_adj=0 mal-80 cpuset=/ mems_allowed=0 Pid: 2704, comm: mal-80 Not tainted 3.7.0+ #10 Call Trace: [<ffffffff8167fd0b>] dump_header+0x83/0x1d1 .......(call trace) [<ffffffff8168a918>] page_fault+0x28/0x30 Task in /A/B/D killed as a result of limit of /A <<<<<<<<<<<<<<<<<<<<< memcg specific information memory: usage 102400kB, limit 102400kB, failcnt 140 memory+swap: usage 102400kB, limit 102400kB, failcnt 0 kmem: usage 0kB, limit 9007199254740991kB, failcnt 0 Memory cgroup stats for /A: cache:32KB rss:30984KB mapped_file:0KB swap:0KB inactive_anon:6912KB active_anon:24072KB inactive_file:32KB active_file:0KB unevictable:0KB Memory cgroup stats for /A/B: cache:0KB rss:0KB mapped_file:0KB swap:0KB inactive_anon:0KB active_anon:0KB inactive_file:0KB active_file:0KB unevictable:0KB Memory cgroup stats for /A/C: cache:0KB rss:0KB mapped_file:0KB swap:0KB inactive_anon:0KB active_anon:0KB inactive_file:0KB active_file:0KB unevictable:0KB Memory cgroup stats for /A/B/D: cache:32KB rss:71352KB mapped_file:0KB swap:0KB inactive_anon:6656KB active_anon:64696KB inactive_file:16KB active_file:16KB unevictable:0KB [ pid ] uid tgid total_vm rss nr_ptes swapents oom_score_adj name [ 2260] 0 2260 6006 1325 18 0 0 god [ 2383] 0 2383 6003 1319 17 0 0 god [ 2503] 0 2503 6004 1321 18 0 0 god [ 2622] 0 2622 6004 1321 16 0 0 god [ 2695] 0 2695 8720 7741 22 0 0 mal-30 [ 2704] 0 2704 21520 17839 43 0 0 mal-80 Memory cgroup out of memory: Kill process 2704 (mal-80) score 669 or sacrifice child Killed process 2704 (mal-80) total-vm:86080kB, anon-rss:71016kB, file-rss:340kB This version provides more pointed info for memcg in "Memory cgroup stats for XXX" section. Signed-off-by: Sha Zhengju <handai.szj@taobao.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: David Rientjes <rientjes@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-23 00:32:05 +00:00
}
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
pr_info("Memory cgroup stats for ");
pr_cont_cgroup_path(memcg->css.cgroup);
pr_cont(":");
buf = memory_stat_format(memcg);
if (!buf)
return;
pr_info("%s", buf);
kfree(buf);
}
oom: badness heuristic rewrite This a complete rewrite of the oom killer's badness() heuristic which is used to determine which task to kill in oom conditions. The goal is to make it as simple and predictable as possible so the results are better understood and we end up killing the task which will lead to the most memory freeing while still respecting the fine-tuning from userspace. Instead of basing the heuristic on mm->total_vm for each task, the task's rss and swap space is used instead. This is a better indication of the amount of memory that will be freeable if the oom killed task is chosen and subsequently exits. This helps specifically in cases where KDE or GNOME is chosen for oom kill on desktop systems instead of a memory hogging task. The baseline for the heuristic is a proportion of memory that each task is currently using in memory plus swap compared to the amount of "allowable" memory. "Allowable," in this sense, means the system-wide resources for unconstrained oom conditions, the set of mempolicy nodes, the mems attached to current's cpuset, or a memory controller's limit. The proportion is given on a scale of 0 (never kill) to 1000 (always kill), roughly meaning that if a task has a badness() score of 500 that the task consumes approximately 50% of allowable memory resident in RAM or in swap space. The proportion is always relative to the amount of "allowable" memory and not the total amount of RAM systemwide so that mempolicies and cpusets may operate in isolation; they shall not need to know the true size of the machine on which they are running if they are bound to a specific set of nodes or mems, respectively. Root tasks are given 3% extra memory just like __vm_enough_memory() provides in LSMs. In the event of two tasks consuming similar amounts of memory, it is generally better to save root's task. Because of the change in the badness() heuristic's baseline, it is also necessary to introduce a new user interface to tune it. It's not possible to redefine the meaning of /proc/pid/oom_adj with a new scale since the ABI cannot be changed for backward compatability. Instead, a new tunable, /proc/pid/oom_score_adj, is added that ranges from -1000 to +1000. It may be used to polarize the heuristic such that certain tasks are never considered for oom kill while others may always be considered. The value is added directly into the badness() score so a value of -500, for example, means to discount 50% of its memory consumption in comparison to other tasks either on the system, bound to the mempolicy, in the cpuset, or sharing the same memory controller. /proc/pid/oom_adj is changed so that its meaning is rescaled into the units used by /proc/pid/oom_score_adj, and vice versa. Changing one of these per-task tunables will rescale the value of the other to an equivalent meaning. Although /proc/pid/oom_adj was originally defined as a bitshift on the badness score, it now shares the same linear growth as /proc/pid/oom_score_adj but with different granularity. This is required so the ABI is not broken with userspace applications and allows oom_adj to be deprecated for future removal. Signed-off-by: David Rientjes <rientjes@google.com> Cc: Nick Piggin <npiggin@suse.de> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-08-10 00:19:46 +00:00
/*
* Return the memory (and swap, if configured) limit for a memcg.
*/
unsigned long mem_cgroup_get_max(struct mem_cgroup *memcg)
oom: badness heuristic rewrite This a complete rewrite of the oom killer's badness() heuristic which is used to determine which task to kill in oom conditions. The goal is to make it as simple and predictable as possible so the results are better understood and we end up killing the task which will lead to the most memory freeing while still respecting the fine-tuning from userspace. Instead of basing the heuristic on mm->total_vm for each task, the task's rss and swap space is used instead. This is a better indication of the amount of memory that will be freeable if the oom killed task is chosen and subsequently exits. This helps specifically in cases where KDE or GNOME is chosen for oom kill on desktop systems instead of a memory hogging task. The baseline for the heuristic is a proportion of memory that each task is currently using in memory plus swap compared to the amount of "allowable" memory. "Allowable," in this sense, means the system-wide resources for unconstrained oom conditions, the set of mempolicy nodes, the mems attached to current's cpuset, or a memory controller's limit. The proportion is given on a scale of 0 (never kill) to 1000 (always kill), roughly meaning that if a task has a badness() score of 500 that the task consumes approximately 50% of allowable memory resident in RAM or in swap space. The proportion is always relative to the amount of "allowable" memory and not the total amount of RAM systemwide so that mempolicies and cpusets may operate in isolation; they shall not need to know the true size of the machine on which they are running if they are bound to a specific set of nodes or mems, respectively. Root tasks are given 3% extra memory just like __vm_enough_memory() provides in LSMs. In the event of two tasks consuming similar amounts of memory, it is generally better to save root's task. Because of the change in the badness() heuristic's baseline, it is also necessary to introduce a new user interface to tune it. It's not possible to redefine the meaning of /proc/pid/oom_adj with a new scale since the ABI cannot be changed for backward compatability. Instead, a new tunable, /proc/pid/oom_score_adj, is added that ranges from -1000 to +1000. It may be used to polarize the heuristic such that certain tasks are never considered for oom kill while others may always be considered. The value is added directly into the badness() score so a value of -500, for example, means to discount 50% of its memory consumption in comparison to other tasks either on the system, bound to the mempolicy, in the cpuset, or sharing the same memory controller. /proc/pid/oom_adj is changed so that its meaning is rescaled into the units used by /proc/pid/oom_score_adj, and vice versa. Changing one of these per-task tunables will rescale the value of the other to an equivalent meaning. Although /proc/pid/oom_adj was originally defined as a bitshift on the badness score, it now shares the same linear growth as /proc/pid/oom_score_adj but with different granularity. This is required so the ABI is not broken with userspace applications and allows oom_adj to be deprecated for future removal. Signed-off-by: David Rientjes <rientjes@google.com> Cc: Nick Piggin <npiggin@suse.de> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-08-10 00:19:46 +00:00
{
unsigned long max;
max = READ_ONCE(memcg->memory.max);
if (mem_cgroup_swappiness(memcg)) {
unsigned long memsw_max;
unsigned long swap_max;
memsw_max = memcg->memsw.max;
swap_max = READ_ONCE(memcg->swap.max);
swap_max = min(swap_max, (unsigned long)total_swap_pages);
max = min(max + swap_max, memsw_max);
}
return max;
oom: badness heuristic rewrite This a complete rewrite of the oom killer's badness() heuristic which is used to determine which task to kill in oom conditions. The goal is to make it as simple and predictable as possible so the results are better understood and we end up killing the task which will lead to the most memory freeing while still respecting the fine-tuning from userspace. Instead of basing the heuristic on mm->total_vm for each task, the task's rss and swap space is used instead. This is a better indication of the amount of memory that will be freeable if the oom killed task is chosen and subsequently exits. This helps specifically in cases where KDE or GNOME is chosen for oom kill on desktop systems instead of a memory hogging task. The baseline for the heuristic is a proportion of memory that each task is currently using in memory plus swap compared to the amount of "allowable" memory. "Allowable," in this sense, means the system-wide resources for unconstrained oom conditions, the set of mempolicy nodes, the mems attached to current's cpuset, or a memory controller's limit. The proportion is given on a scale of 0 (never kill) to 1000 (always kill), roughly meaning that if a task has a badness() score of 500 that the task consumes approximately 50% of allowable memory resident in RAM or in swap space. The proportion is always relative to the amount of "allowable" memory and not the total amount of RAM systemwide so that mempolicies and cpusets may operate in isolation; they shall not need to know the true size of the machine on which they are running if they are bound to a specific set of nodes or mems, respectively. Root tasks are given 3% extra memory just like __vm_enough_memory() provides in LSMs. In the event of two tasks consuming similar amounts of memory, it is generally better to save root's task. Because of the change in the badness() heuristic's baseline, it is also necessary to introduce a new user interface to tune it. It's not possible to redefine the meaning of /proc/pid/oom_adj with a new scale since the ABI cannot be changed for backward compatability. Instead, a new tunable, /proc/pid/oom_score_adj, is added that ranges from -1000 to +1000. It may be used to polarize the heuristic such that certain tasks are never considered for oom kill while others may always be considered. The value is added directly into the badness() score so a value of -500, for example, means to discount 50% of its memory consumption in comparison to other tasks either on the system, bound to the mempolicy, in the cpuset, or sharing the same memory controller. /proc/pid/oom_adj is changed so that its meaning is rescaled into the units used by /proc/pid/oom_score_adj, and vice versa. Changing one of these per-task tunables will rescale the value of the other to an equivalent meaning. Although /proc/pid/oom_adj was originally defined as a bitshift on the badness score, it now shares the same linear growth as /proc/pid/oom_score_adj but with different granularity. This is required so the ABI is not broken with userspace applications and allows oom_adj to be deprecated for future removal. Signed-off-by: David Rientjes <rientjes@google.com> Cc: Nick Piggin <npiggin@suse.de> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-08-10 00:19:46 +00:00
}
mm, memcg: proportional memory.{low,min} reclaim cgroup v2 introduces two memory protection thresholds: memory.low (best-effort) and memory.min (hard protection). While they generally do what they say on the tin, there is a limitation in their implementation that makes them difficult to use effectively: that cliff behaviour often manifests when they become eligible for reclaim. This patch implements more intuitive and usable behaviour, where we gradually mount more reclaim pressure as cgroups further and further exceed their protection thresholds. This cliff edge behaviour happens because we only choose whether or not to reclaim based on whether the memcg is within its protection limits (see the use of mem_cgroup_protected in shrink_node), but we don't vary our reclaim behaviour based on this information. Imagine the following timeline, with the numbers the lruvec size in this zone: 1. memory.low=1000000, memory.current=999999. 0 pages may be scanned. 2. memory.low=1000000, memory.current=1000000. 0 pages may be scanned. 3. memory.low=1000000, memory.current=1000001. 1000001* pages may be scanned. (?!) * Of course, we won't usually scan all available pages in the zone even without this patch because of scan control priority, over-reclaim protection, etc. However, as shown by the tests at the end, these techniques don't sufficiently throttle such an extreme change in input, so cliff-like behaviour isn't really averted by their existence alone. Here's an example of how this plays out in practice. At Facebook, we are trying to protect various workloads from "system" software, like configuration management tools, metric collectors, etc (see this[0] case study). In order to find a suitable memory.low value, we start by determining the expected memory range within which the workload will be comfortable operating. This isn't an exact science -- memory usage deemed "comfortable" will vary over time due to user behaviour, differences in composition of work, etc, etc. As such we need to ballpark memory.low, but doing this is currently problematic: 1. If we end up setting it too low for the workload, it won't have *any* effect (see discussion above). The group will receive the full weight of reclaim and won't have any priority while competing with the less important system software, as if we had no memory.low configured at all. 2. Because of this behaviour, we end up erring on the side of setting it too high, such that the comfort range is reliably covered. However, protected memory is completely unavailable to the rest of the system, so we might cause undue memory and IO pressure there when we *know* we have some elasticity in the workload. 3. Even if we get the value totally right, smack in the middle of the comfort zone, we get extreme jumps between no pressure and full pressure that cause unpredictable pressure spikes in the workload due to the current binary reclaim behaviour. With this patch, we can set it to our ballpark estimation without too much worry. Any undesirable behaviour, such as too much or too little reclaim pressure on the workload or system will be proportional to how far our estimation is off. This means we can set memory.low much more conservatively and thus waste less resources *without* the risk of the workload falling off a cliff if we overshoot. As a more abstract technical description, this unintuitive behaviour results in having to give high-priority workloads a large protection buffer on top of their expected usage to function reliably, as otherwise we have abrupt periods of dramatically increased memory pressure which hamper performance. Having to set these thresholds so high wastes resources and generally works against the principle of work conservation. In addition, having proportional memory reclaim behaviour has other benefits. Most notably, before this patch it's basically mandatory to set memory.low to a higher than desirable value because otherwise as soon as you exceed memory.low, all protection is lost, and all pages are eligible to scan again. By contrast, having a gradual ramp in reclaim pressure means that you now still get some protection when thresholds are exceeded, which means that one can now be more comfortable setting memory.low to lower values without worrying that all protection will be lost. This is important because workingset size is really hard to know exactly, especially with variable workloads, so at least getting *some* protection if your workingset size grows larger than you expect increases user confidence in setting memory.low without a huge buffer on top being needed. Thanks a lot to Johannes Weiner and Tejun Heo for their advice and assistance in thinking about how to make this work better. In testing these changes, I intended to verify that: 1. Changes in page scanning become gradual and proportional instead of binary. To test this, I experimented stepping further and further down memory.low protection on a workload that floats around 19G workingset when under memory.low protection, watching page scan rates for the workload cgroup: +------------+-----------------+--------------------+--------------+ | memory.low | test (pgscan/s) | control (pgscan/s) | % of control | +------------+-----------------+--------------------+--------------+ | 21G | 0 | 0 | N/A | | 17G | 867 | 3799 | 23% | | 12G | 1203 | 3543 | 34% | | 8G | 2534 | 3979 | 64% | | 4G | 3980 | 4147 | 96% | | 0 | 3799 | 3980 | 95% | +------------+-----------------+--------------------+--------------+ As you can see, the test kernel (with a kernel containing this patch) ramps up page scanning significantly more gradually than the control kernel (without this patch). 2. More gradual ramp up in reclaim aggression doesn't result in premature OOMs. To test this, I wrote a script that slowly increments the number of pages held by stress(1)'s --vm-keep mode until a production system entered severe overall memory contention. This script runs in a highly protected slice taking up the majority of available system memory. Watching vmstat revealed that page scanning continued essentially nominally between test and control, without causing forward reclaim progress to become arrested. [0]: https://facebookmicrosites.github.io/cgroup2/docs/overview.html#case-study-the-fbtax2-project [akpm@linux-foundation.org: reflow block comments to fit in 80 cols] [chris@chrisdown.name: handle cgroup_disable=memory when getting memcg protection] Link: http://lkml.kernel.org/r/20190201045711.GA18302@chrisdown.name Link: http://lkml.kernel.org/r/20190124014455.GA6396@chrisdown.name Signed-off-by: Chris Down <chris@chrisdown.name> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Cc: Dennis Zhou <dennis@kernel.org> Cc: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-10-07 00:58:32 +00:00
unsigned long mem_cgroup_size(struct mem_cgroup *memcg)
{
return page_counter_read(&memcg->memory);
}
static bool mem_cgroup_out_of_memory(struct mem_cgroup *memcg, gfp_t gfp_mask,
int order)
mm, memcg: introduce own oom handler to iterate only over its own threads The global oom killer is serialized by the per-zonelist try_set_zonelist_oom() which is used in the page allocator. Concurrent oom kills are thus a rare event and only occur in systems using mempolicies and with a large number of nodes. Memory controller oom kills, however, can frequently be concurrent since there is no serialization once the oom killer is called for oom conditions in several different memcgs in parallel. This creates a massive contention on tasklist_lock since the oom killer requires the readside for the tasklist iteration. If several memcgs are calling the oom killer, this lock can be held for a substantial amount of time, especially if threads continue to enter it as other threads are exiting. Since the exit path grabs the writeside of the lock with irqs disabled in a few different places, this can cause a soft lockup on cpus as a result of tasklist_lock starvation. The kernel lacks unfair writelocks, and successful calls to the oom killer usually result in at least one thread entering the exit path, so an alternative solution is needed. This patch introduces a seperate oom handler for memcgs so that they do not require tasklist_lock for as much time. Instead, it iterates only over the threads attached to the oom memcg and grabs a reference to the selected thread before calling oom_kill_process() to ensure it doesn't prematurely exit. This still requires tasklist_lock for the tasklist dump, iterating children of the selected process, and killing all other threads on the system sharing the same memory as the selected victim. So while this isn't a complete solution to tasklist_lock starvation, it significantly reduces the amount of time that it is held. Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Michal Hocko <mhocko@suse.cz> Signed-off-by: David Rientjes <rientjes@google.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Reviewed-by: Sha Zhengju <handai.szj@taobao.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-07-31 23:43:44 +00:00
{
struct oom_control oc = {
.zonelist = NULL,
.nodemask = NULL,
.memcg = memcg,
.gfp_mask = gfp_mask,
.order = order,
};
bool ret;
mm, memcg: introduce own oom handler to iterate only over its own threads The global oom killer is serialized by the per-zonelist try_set_zonelist_oom() which is used in the page allocator. Concurrent oom kills are thus a rare event and only occur in systems using mempolicies and with a large number of nodes. Memory controller oom kills, however, can frequently be concurrent since there is no serialization once the oom killer is called for oom conditions in several different memcgs in parallel. This creates a massive contention on tasklist_lock since the oom killer requires the readside for the tasklist iteration. If several memcgs are calling the oom killer, this lock can be held for a substantial amount of time, especially if threads continue to enter it as other threads are exiting. Since the exit path grabs the writeside of the lock with irqs disabled in a few different places, this can cause a soft lockup on cpus as a result of tasklist_lock starvation. The kernel lacks unfair writelocks, and successful calls to the oom killer usually result in at least one thread entering the exit path, so an alternative solution is needed. This patch introduces a seperate oom handler for memcgs so that they do not require tasklist_lock for as much time. Instead, it iterates only over the threads attached to the oom memcg and grabs a reference to the selected thread before calling oom_kill_process() to ensure it doesn't prematurely exit. This still requires tasklist_lock for the tasklist dump, iterating children of the selected process, and killing all other threads on the system sharing the same memory as the selected victim. So while this isn't a complete solution to tasklist_lock starvation, it significantly reduces the amount of time that it is held. Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Michal Hocko <mhocko@suse.cz> Signed-off-by: David Rientjes <rientjes@google.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Reviewed-by: Sha Zhengju <handai.szj@taobao.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-07-31 23:43:44 +00:00
memcg: killed threads should not invoke memcg OOM killer If a memory cgroup contains a single process with many threads (including different process group sharing the mm) then it is possible to trigger a race when the oom killer complains that there are no oom elible tasks and complain into the log which is both annoying and confusing because there is no actual problem. The race looks as follows: P1 oom_reaper P2 try_charge try_charge mem_cgroup_out_of_memory mutex_lock(oom_lock) out_of_memory oom_kill_process(P1,P2) wake_oom_reaper mutex_unlock(oom_lock) oom_reap_task mutex_lock(oom_lock) select_bad_process # no victim The problem is more visible with many threads. Fix this by checking for fatal_signal_pending from mem_cgroup_out_of_memory when the oom_lock is already held. The oom bypass is safe because we do the same early in the try_charge path already. The situation migh have changed in the mean time. It should be safe to check for fatal_signal_pending and tsk_is_oom_victim but for a better code readability abstract the current charge bypass condition into should_force_charge and reuse it from that path. " Link: http://lkml.kernel.org/r/01370f70-e1f6-ebe4-b95e-0df21a0bc15e@i-love.sakura.ne.jp Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: David Rientjes <rientjes@google.com> Cc: Kirill Tkhai <ktkhai@virtuozzo.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-03-05 23:46:47 +00:00
if (mutex_lock_killable(&oom_lock))
return true;
/*
* A few threads which were not waiting at mutex_lock_killable() can
* fail to bail out. Therefore, check again after holding oom_lock.
*/
ret = should_force_charge() || out_of_memory(&oc);
mutex_unlock(&oom_lock);
return ret;
mm, memcg: introduce own oom handler to iterate only over its own threads The global oom killer is serialized by the per-zonelist try_set_zonelist_oom() which is used in the page allocator. Concurrent oom kills are thus a rare event and only occur in systems using mempolicies and with a large number of nodes. Memory controller oom kills, however, can frequently be concurrent since there is no serialization once the oom killer is called for oom conditions in several different memcgs in parallel. This creates a massive contention on tasklist_lock since the oom killer requires the readside for the tasklist iteration. If several memcgs are calling the oom killer, this lock can be held for a substantial amount of time, especially if threads continue to enter it as other threads are exiting. Since the exit path grabs the writeside of the lock with irqs disabled in a few different places, this can cause a soft lockup on cpus as a result of tasklist_lock starvation. The kernel lacks unfair writelocks, and successful calls to the oom killer usually result in at least one thread entering the exit path, so an alternative solution is needed. This patch introduces a seperate oom handler for memcgs so that they do not require tasklist_lock for as much time. Instead, it iterates only over the threads attached to the oom memcg and grabs a reference to the selected thread before calling oom_kill_process() to ensure it doesn't prematurely exit. This still requires tasklist_lock for the tasklist dump, iterating children of the selected process, and killing all other threads on the system sharing the same memory as the selected victim. So while this isn't a complete solution to tasklist_lock starvation, it significantly reduces the amount of time that it is held. Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Michal Hocko <mhocko@suse.cz> Signed-off-by: David Rientjes <rientjes@google.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Reviewed-by: Sha Zhengju <handai.szj@taobao.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-07-31 23:43:44 +00:00
}
static int mem_cgroup_soft_reclaim(struct mem_cgroup *root_memcg,
pg_data_t *pgdat,
gfp_t gfp_mask,
unsigned long *total_scanned)
{
struct mem_cgroup *victim = NULL;
int total = 0;
int loop = 0;
unsigned long excess;
unsigned long nr_scanned;
struct mem_cgroup_reclaim_cookie reclaim = {
.pgdat = pgdat,
};
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
excess = soft_limit_excess(root_memcg);
while (1) {
victim = mem_cgroup_iter(root_memcg, victim, &reclaim);
if (!victim) {
loop++;
if (loop >= 2) {
/*
* If we have not been able to reclaim
* anything, it might because there are
* no reclaimable pages under this hierarchy
*/
if (!total)
break;
/*
* We want to do more targeted reclaim.
* excess >> 2 is not to excessive so as to
* reclaim too much, nor too less that we keep
* coming back to reclaim from this cgroup
*/
if (total >= (excess >> 2) ||
(loop > MEM_CGROUP_MAX_RECLAIM_LOOPS))
break;
}
continue;
}
total += mem_cgroup_shrink_node(victim, gfp_mask, false,
pgdat, &nr_scanned);
*total_scanned += nr_scanned;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
if (!soft_limit_excess(root_memcg))
break;
}
mem_cgroup_iter_break(root_memcg, victim);
return total;
}
#ifdef CONFIG_LOCKDEP
static struct lockdep_map memcg_oom_lock_dep_map = {
.name = "memcg_oom_lock",
};
#endif
mm: memcg: rework and document OOM waiting and wakeup The memcg OOM handler open-codes a sleeping lock for OOM serialization (trylock, wait, repeat) because the required locking is so specific to memcg hierarchies. However, it would be nice if this construct would be clearly recognizable and not be as obfuscated as it is right now. Clean up as follows: 1. Remove the return value of mem_cgroup_oom_unlock() 2. Rename mem_cgroup_oom_lock() to mem_cgroup_oom_trylock(). 3. Pull the prepare_to_wait() out of the memcg_oom_lock scope. This makes it more obvious that the task has to be on the waitqueue before attempting to OOM-trylock the hierarchy, to not miss any wakeups before going to sleep. It just didn't matter until now because it was all lumped together into the global memcg_oom_lock spinlock section. 4. Pull the mem_cgroup_oom_notify() out of the memcg_oom_lock scope. It is proctected by the hierarchical OOM-lock. 5. The memcg_oom_lock spinlock is only required to propagate the OOM lock in any given hierarchy atomically. Restrict its scope to mem_cgroup_oom_(trylock|unlock). 6. Do not wake up the waitqueue unconditionally at the end of the function. Only the lockholder has to wake up the next in line after releasing the lock. Note that the lockholder kicks off the OOM-killer, which in turn leads to wakeups from the uncharges of the exiting task. But a contender is not guaranteed to see them if it enters the OOM path after the OOM kills but before the lockholder releases the lock. Thus there has to be an explicit wakeup after releasing the lock. 7. Put the OOM task on the waitqueue before marking the hierarchy as under OOM as that is the point where we start to receive wakeups. No point in listening before being on the waitqueue. 8. Likewise, unmark the hierarchy before finishing the sleep, for symmetry. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: azurIt <azurit@pobox.sk> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:43 +00:00
static DEFINE_SPINLOCK(memcg_oom_lock);
memcg: fix oom kill behavior In current page-fault code, handle_mm_fault() -> ... -> mem_cgroup_charge() -> map page or handle error. -> check return code. If page fault's return code is VM_FAULT_OOM, page_fault_out_of_memory() is called. But if it's caused by memcg, OOM should have been already invoked. Then, I added a patch: a636b327f731143ccc544b966cfd8de6cb6d72c6. That patch records last_oom_jiffies for memcg's sub-hierarchy and prevents page_fault_out_of_memory from being invoked in near future. But Nishimura-san reported that check by jiffies is not enough when the system is terribly heavy. This patch changes memcg's oom logic as. * If memcg causes OOM-kill, continue to retry. * remove jiffies check which is used now. * add memcg-oom-lock which works like perzone oom lock. * If current is killed(as a process), bypass charge. Something more sophisticated can be added but this pactch does fundamental things. TODO: - add oom notifier - add permemcg disable-oom-kill flag and freezer at oom. - more chances for wake up oom waiter (when changing memory limit etc..) Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Tested-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-03-10 23:22:39 +00:00
/*
* Check OOM-Killer is already running under our hierarchy.
* If someone is running, return false.
*/
mm: memcg: rework and document OOM waiting and wakeup The memcg OOM handler open-codes a sleeping lock for OOM serialization (trylock, wait, repeat) because the required locking is so specific to memcg hierarchies. However, it would be nice if this construct would be clearly recognizable and not be as obfuscated as it is right now. Clean up as follows: 1. Remove the return value of mem_cgroup_oom_unlock() 2. Rename mem_cgroup_oom_lock() to mem_cgroup_oom_trylock(). 3. Pull the prepare_to_wait() out of the memcg_oom_lock scope. This makes it more obvious that the task has to be on the waitqueue before attempting to OOM-trylock the hierarchy, to not miss any wakeups before going to sleep. It just didn't matter until now because it was all lumped together into the global memcg_oom_lock spinlock section. 4. Pull the mem_cgroup_oom_notify() out of the memcg_oom_lock scope. It is proctected by the hierarchical OOM-lock. 5. The memcg_oom_lock spinlock is only required to propagate the OOM lock in any given hierarchy atomically. Restrict its scope to mem_cgroup_oom_(trylock|unlock). 6. Do not wake up the waitqueue unconditionally at the end of the function. Only the lockholder has to wake up the next in line after releasing the lock. Note that the lockholder kicks off the OOM-killer, which in turn leads to wakeups from the uncharges of the exiting task. But a contender is not guaranteed to see them if it enters the OOM path after the OOM kills but before the lockholder releases the lock. Thus there has to be an explicit wakeup after releasing the lock. 7. Put the OOM task on the waitqueue before marking the hierarchy as under OOM as that is the point where we start to receive wakeups. No point in listening before being on the waitqueue. 8. Likewise, unmark the hierarchy before finishing the sleep, for symmetry. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: azurIt <azurit@pobox.sk> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:43 +00:00
static bool mem_cgroup_oom_trylock(struct mem_cgroup *memcg)
memcg: fix oom kill behavior In current page-fault code, handle_mm_fault() -> ... -> mem_cgroup_charge() -> map page or handle error. -> check return code. If page fault's return code is VM_FAULT_OOM, page_fault_out_of_memory() is called. But if it's caused by memcg, OOM should have been already invoked. Then, I added a patch: a636b327f731143ccc544b966cfd8de6cb6d72c6. That patch records last_oom_jiffies for memcg's sub-hierarchy and prevents page_fault_out_of_memory from being invoked in near future. But Nishimura-san reported that check by jiffies is not enough when the system is terribly heavy. This patch changes memcg's oom logic as. * If memcg causes OOM-kill, continue to retry. * remove jiffies check which is used now. * add memcg-oom-lock which works like perzone oom lock. * If current is killed(as a process), bypass charge. Something more sophisticated can be added but this pactch does fundamental things. TODO: - add oom notifier - add permemcg disable-oom-kill flag and freezer at oom. - more chances for wake up oom waiter (when changing memory limit etc..) Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Tested-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-03-10 23:22:39 +00:00
{
memcg: make oom_lock 0 and 1 based rather than counter Commit 867578cb ("memcg: fix oom kill behavior") introduced a oom_lock counter which is incremented by mem_cgroup_oom_lock when we are about to handle memcg OOM situation. mem_cgroup_handle_oom falls back to a sleep if oom_lock > 1 to prevent from multiple oom kills at the same time. The counter is then decremented by mem_cgroup_oom_unlock called from the same function. This works correctly but it can lead to serious starvations when we have many processes triggering OOM and many CPUs available for them (I have tested with 16 CPUs). Consider a process (call it A) which gets the oom_lock (the first one that got to mem_cgroup_handle_oom and grabbed memcg_oom_mutex) and other processes that are blocked on the mutex. While A releases the mutex and calls mem_cgroup_out_of_memory others will wake up (one after another) and increase the counter and fall into sleep (memcg_oom_waitq). Once A finishes mem_cgroup_out_of_memory it takes the mutex again and decreases oom_lock and wakes other tasks (if releasing memory by somebody else - e.g. killed process - hasn't done it yet). A testcase would look like: Assume malloc XXX is a program allocating XXX Megabytes of memory which touches all allocated pages in a tight loop # swapoff SWAP_DEVICE # cgcreate -g memory:A # cgset -r memory.oom_control=0 A # cgset -r memory.limit_in_bytes= 200M # for i in `seq 100` # do # cgexec -g memory:A malloc 10 & # done The main problem here is that all processes still race for the mutex and there is no guarantee that we will get counter back to 0 for those that got back to mem_cgroup_handle_oom. In the end the whole convoy in/decreases the counter but we do not get to 1 that would enable killing so nothing useful can be done. The time is basically unbounded because it highly depends on scheduling and ordering on mutex (I have seen this taking hours...). This patch replaces the counter by a simple {un}lock semantic. As mem_cgroup_oom_{un}lock works on the a subtree of a hierarchy we have to make sure that nobody else races with us which is guaranteed by the memcg_oom_mutex. We have to be careful while locking subtrees because we can encounter a subtree which is already locked: hierarchy: A / \ B \ /\ \ C D E B - C - D tree might be already locked. While we want to enable locking E subtree because OOM situations cannot influence each other we definitely do not want to allow locking A. Therefore we have to refuse lock if any subtree is already locked and clear up the lock for all nodes that have been set up to the failure point. On the other hand we have to make sure that the rest of the world will recognize that a group is under OOM even though it doesn't have a lock. Therefore we have to introduce under_oom variable which is incremented and decremented for the whole subtree when we enter resp. leave mem_cgroup_handle_oom. under_oom, unlike oom_lock, doesn't need be updated under memcg_oom_mutex because its users only check a single group and they use atomic operations for that. This can be checked easily by the following test case: # cgcreate -g memory:A # cgset -r memory.use_hierarchy=1 A # cgset -r memory.oom_control=1 A # cgset -r memory.limit_in_bytes= 100M # cgset -r memory.memsw.limit_in_bytes= 100M # cgcreate -g memory:A/B # cgset -r memory.oom_control=1 A/B # cgset -r memory.limit_in_bytes=20M # cgset -r memory.memsw.limit_in_bytes=20M # cgexec -g memory:A/B malloc 30 & #->this will be blocked by OOM of group B # cgexec -g memory:A malloc 80 & #->this will be blocked by OOM of group A While B gets oom_lock A will not get it. Both of them go into sleep and wait for an external action. We can make the limit higher for A to enforce waking it up # cgset -r memory.memsw.limit_in_bytes=300M A # cgset -r memory.limit_in_bytes=300M A malloc in A has to wake up even though it doesn't have oom_lock. Finally, the unlock path is very easy because we always unlock only the subtree we have locked previously while we always decrement under_oom. Signed-off-by: Michal Hocko <mhocko@suse.cz> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <bsingharora@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-07-26 23:08:23 +00:00
struct mem_cgroup *iter, *failed = NULL;
mm: memcg: rework and document OOM waiting and wakeup The memcg OOM handler open-codes a sleeping lock for OOM serialization (trylock, wait, repeat) because the required locking is so specific to memcg hierarchies. However, it would be nice if this construct would be clearly recognizable and not be as obfuscated as it is right now. Clean up as follows: 1. Remove the return value of mem_cgroup_oom_unlock() 2. Rename mem_cgroup_oom_lock() to mem_cgroup_oom_trylock(). 3. Pull the prepare_to_wait() out of the memcg_oom_lock scope. This makes it more obvious that the task has to be on the waitqueue before attempting to OOM-trylock the hierarchy, to not miss any wakeups before going to sleep. It just didn't matter until now because it was all lumped together into the global memcg_oom_lock spinlock section. 4. Pull the mem_cgroup_oom_notify() out of the memcg_oom_lock scope. It is proctected by the hierarchical OOM-lock. 5. The memcg_oom_lock spinlock is only required to propagate the OOM lock in any given hierarchy atomically. Restrict its scope to mem_cgroup_oom_(trylock|unlock). 6. Do not wake up the waitqueue unconditionally at the end of the function. Only the lockholder has to wake up the next in line after releasing the lock. Note that the lockholder kicks off the OOM-killer, which in turn leads to wakeups from the uncharges of the exiting task. But a contender is not guaranteed to see them if it enters the OOM path after the OOM kills but before the lockholder releases the lock. Thus there has to be an explicit wakeup after releasing the lock. 7. Put the OOM task on the waitqueue before marking the hierarchy as under OOM as that is the point where we start to receive wakeups. No point in listening before being on the waitqueue. 8. Likewise, unmark the hierarchy before finishing the sleep, for symmetry. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: azurIt <azurit@pobox.sk> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:43 +00:00
spin_lock(&memcg_oom_lock);
mm: memcg: consolidate hierarchy iteration primitives The memcg naturalization series: Memory control groups are currently bolted onto the side of traditional memory management in places where better integration would be preferrable. To reclaim memory, for example, memory control groups maintain their own LRU list and reclaim strategy aside from the global per-zone LRU list reclaim. But an extra list head for each existing page frame is expensive and maintaining it requires additional code. This patchset disables the global per-zone LRU lists on memory cgroup configurations and converts all its users to operate on the per-memory cgroup lists instead. As LRU pages are then exclusively on one list, this saves two list pointers for each page frame in the system: page_cgroup array size with 4G physical memory vanilla: allocated 31457280 bytes of page_cgroup patched: allocated 15728640 bytes of page_cgroup At the same time, system performance for various workloads is unaffected: 100G sparse file cat, 4G physical memory, 10 runs, to test for code bloat in the traditional LRU handling and kswapd & direct reclaim paths, without/with the memory controller configured in vanilla: 71.603(0.207) seconds patched: 71.640(0.156) seconds vanilla: 79.558(0.288) seconds patched: 77.233(0.147) seconds 100G sparse file cat in 1G memory cgroup, 10 runs, to test for code bloat in the traditional memory cgroup LRU handling and reclaim path vanilla: 96.844(0.281) seconds patched: 94.454(0.311) seconds 4 unlimited memcgs running kbuild -j32 each, 4G physical memory, 500M swap on SSD, 10 runs, to test for regressions in kswapd & direct reclaim using per-memcg LRU lists with multiple memcgs and multiple allocators within each memcg vanilla: 717.722(1.440) seconds [ 69720.100(11600.835) majfaults ] patched: 714.106(2.313) seconds [ 71109.300(14886.186) majfaults ] 16 unlimited memcgs running kbuild, 1900M hierarchical limit, 500M swap on SSD, 10 runs, to test for regressions in hierarchical memcg setups vanilla: 2742.058(1.992) seconds [ 26479.600(1736.737) majfaults ] patched: 2743.267(1.214) seconds [ 27240.700(1076.063) majfaults ] This patch: There are currently two different implementations of iterating over a memory cgroup hierarchy tree. Consolidate them into one worker function and base the convenience looping-macros on top of it. Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Reviewed-by: Kirill A. Shutemov <kirill@shutemov.name> Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Ying Han <yinghan@google.com> Cc: Greg Thelen <gthelen@google.com> Cc: Michel Lespinasse <walken@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Minchan Kim <minchan.kim@gmail.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-13 01:17:48 +00:00
for_each_mem_cgroup_tree(iter, memcg) {
if (iter->oom_lock) {
memcg: make oom_lock 0 and 1 based rather than counter Commit 867578cb ("memcg: fix oom kill behavior") introduced a oom_lock counter which is incremented by mem_cgroup_oom_lock when we are about to handle memcg OOM situation. mem_cgroup_handle_oom falls back to a sleep if oom_lock > 1 to prevent from multiple oom kills at the same time. The counter is then decremented by mem_cgroup_oom_unlock called from the same function. This works correctly but it can lead to serious starvations when we have many processes triggering OOM and many CPUs available for them (I have tested with 16 CPUs). Consider a process (call it A) which gets the oom_lock (the first one that got to mem_cgroup_handle_oom and grabbed memcg_oom_mutex) and other processes that are blocked on the mutex. While A releases the mutex and calls mem_cgroup_out_of_memory others will wake up (one after another) and increase the counter and fall into sleep (memcg_oom_waitq). Once A finishes mem_cgroup_out_of_memory it takes the mutex again and decreases oom_lock and wakes other tasks (if releasing memory by somebody else - e.g. killed process - hasn't done it yet). A testcase would look like: Assume malloc XXX is a program allocating XXX Megabytes of memory which touches all allocated pages in a tight loop # swapoff SWAP_DEVICE # cgcreate -g memory:A # cgset -r memory.oom_control=0 A # cgset -r memory.limit_in_bytes= 200M # for i in `seq 100` # do # cgexec -g memory:A malloc 10 & # done The main problem here is that all processes still race for the mutex and there is no guarantee that we will get counter back to 0 for those that got back to mem_cgroup_handle_oom. In the end the whole convoy in/decreases the counter but we do not get to 1 that would enable killing so nothing useful can be done. The time is basically unbounded because it highly depends on scheduling and ordering on mutex (I have seen this taking hours...). This patch replaces the counter by a simple {un}lock semantic. As mem_cgroup_oom_{un}lock works on the a subtree of a hierarchy we have to make sure that nobody else races with us which is guaranteed by the memcg_oom_mutex. We have to be careful while locking subtrees because we can encounter a subtree which is already locked: hierarchy: A / \ B \ /\ \ C D E B - C - D tree might be already locked. While we want to enable locking E subtree because OOM situations cannot influence each other we definitely do not want to allow locking A. Therefore we have to refuse lock if any subtree is already locked and clear up the lock for all nodes that have been set up to the failure point. On the other hand we have to make sure that the rest of the world will recognize that a group is under OOM even though it doesn't have a lock. Therefore we have to introduce under_oom variable which is incremented and decremented for the whole subtree when we enter resp. leave mem_cgroup_handle_oom. under_oom, unlike oom_lock, doesn't need be updated under memcg_oom_mutex because its users only check a single group and they use atomic operations for that. This can be checked easily by the following test case: # cgcreate -g memory:A # cgset -r memory.use_hierarchy=1 A # cgset -r memory.oom_control=1 A # cgset -r memory.limit_in_bytes= 100M # cgset -r memory.memsw.limit_in_bytes= 100M # cgcreate -g memory:A/B # cgset -r memory.oom_control=1 A/B # cgset -r memory.limit_in_bytes=20M # cgset -r memory.memsw.limit_in_bytes=20M # cgexec -g memory:A/B malloc 30 & #->this will be blocked by OOM of group B # cgexec -g memory:A malloc 80 & #->this will be blocked by OOM of group A While B gets oom_lock A will not get it. Both of them go into sleep and wait for an external action. We can make the limit higher for A to enforce waking it up # cgset -r memory.memsw.limit_in_bytes=300M A # cgset -r memory.limit_in_bytes=300M A malloc in A has to wake up even though it doesn't have oom_lock. Finally, the unlock path is very easy because we always unlock only the subtree we have locked previously while we always decrement under_oom. Signed-off-by: Michal Hocko <mhocko@suse.cz> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <bsingharora@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-07-26 23:08:23 +00:00
/*
* this subtree of our hierarchy is already locked
* so we cannot give a lock.
*/
failed = iter;
mm: memcg: consolidate hierarchy iteration primitives The memcg naturalization series: Memory control groups are currently bolted onto the side of traditional memory management in places where better integration would be preferrable. To reclaim memory, for example, memory control groups maintain their own LRU list and reclaim strategy aside from the global per-zone LRU list reclaim. But an extra list head for each existing page frame is expensive and maintaining it requires additional code. This patchset disables the global per-zone LRU lists on memory cgroup configurations and converts all its users to operate on the per-memory cgroup lists instead. As LRU pages are then exclusively on one list, this saves two list pointers for each page frame in the system: page_cgroup array size with 4G physical memory vanilla: allocated 31457280 bytes of page_cgroup patched: allocated 15728640 bytes of page_cgroup At the same time, system performance for various workloads is unaffected: 100G sparse file cat, 4G physical memory, 10 runs, to test for code bloat in the traditional LRU handling and kswapd & direct reclaim paths, without/with the memory controller configured in vanilla: 71.603(0.207) seconds patched: 71.640(0.156) seconds vanilla: 79.558(0.288) seconds patched: 77.233(0.147) seconds 100G sparse file cat in 1G memory cgroup, 10 runs, to test for code bloat in the traditional memory cgroup LRU handling and reclaim path vanilla: 96.844(0.281) seconds patched: 94.454(0.311) seconds 4 unlimited memcgs running kbuild -j32 each, 4G physical memory, 500M swap on SSD, 10 runs, to test for regressions in kswapd & direct reclaim using per-memcg LRU lists with multiple memcgs and multiple allocators within each memcg vanilla: 717.722(1.440) seconds [ 69720.100(11600.835) majfaults ] patched: 714.106(2.313) seconds [ 71109.300(14886.186) majfaults ] 16 unlimited memcgs running kbuild, 1900M hierarchical limit, 500M swap on SSD, 10 runs, to test for regressions in hierarchical memcg setups vanilla: 2742.058(1.992) seconds [ 26479.600(1736.737) majfaults ] patched: 2743.267(1.214) seconds [ 27240.700(1076.063) majfaults ] This patch: There are currently two different implementations of iterating over a memory cgroup hierarchy tree. Consolidate them into one worker function and base the convenience looping-macros on top of it. Signed-off-by: Johannes Weiner <jweiner@redhat.com> Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Michal Hocko <mhocko@suse.cz> Reviewed-by: Kirill A. Shutemov <kirill@shutemov.name> Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Ying Han <yinghan@google.com> Cc: Greg Thelen <gthelen@google.com> Cc: Michel Lespinasse <walken@google.com> Cc: Rik van Riel <riel@redhat.com> Cc: Minchan Kim <minchan.kim@gmail.com> Cc: Christoph Hellwig <hch@infradead.org> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-01-13 01:17:48 +00:00
mem_cgroup_iter_break(memcg, iter);
break;
} else
iter->oom_lock = true;
}
memcg: fix oom kill behavior In current page-fault code, handle_mm_fault() -> ... -> mem_cgroup_charge() -> map page or handle error. -> check return code. If page fault's return code is VM_FAULT_OOM, page_fault_out_of_memory() is called. But if it's caused by memcg, OOM should have been already invoked. Then, I added a patch: a636b327f731143ccc544b966cfd8de6cb6d72c6. That patch records last_oom_jiffies for memcg's sub-hierarchy and prevents page_fault_out_of_memory from being invoked in near future. But Nishimura-san reported that check by jiffies is not enough when the system is terribly heavy. This patch changes memcg's oom logic as. * If memcg causes OOM-kill, continue to retry. * remove jiffies check which is used now. * add memcg-oom-lock which works like perzone oom lock. * If current is killed(as a process), bypass charge. Something more sophisticated can be added but this pactch does fundamental things. TODO: - add oom notifier - add permemcg disable-oom-kill flag and freezer at oom. - more chances for wake up oom waiter (when changing memory limit etc..) Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Tested-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-03-10 23:22:39 +00:00
mm: memcg: rework and document OOM waiting and wakeup The memcg OOM handler open-codes a sleeping lock for OOM serialization (trylock, wait, repeat) because the required locking is so specific to memcg hierarchies. However, it would be nice if this construct would be clearly recognizable and not be as obfuscated as it is right now. Clean up as follows: 1. Remove the return value of mem_cgroup_oom_unlock() 2. Rename mem_cgroup_oom_lock() to mem_cgroup_oom_trylock(). 3. Pull the prepare_to_wait() out of the memcg_oom_lock scope. This makes it more obvious that the task has to be on the waitqueue before attempting to OOM-trylock the hierarchy, to not miss any wakeups before going to sleep. It just didn't matter until now because it was all lumped together into the global memcg_oom_lock spinlock section. 4. Pull the mem_cgroup_oom_notify() out of the memcg_oom_lock scope. It is proctected by the hierarchical OOM-lock. 5. The memcg_oom_lock spinlock is only required to propagate the OOM lock in any given hierarchy atomically. Restrict its scope to mem_cgroup_oom_(trylock|unlock). 6. Do not wake up the waitqueue unconditionally at the end of the function. Only the lockholder has to wake up the next in line after releasing the lock. Note that the lockholder kicks off the OOM-killer, which in turn leads to wakeups from the uncharges of the exiting task. But a contender is not guaranteed to see them if it enters the OOM path after the OOM kills but before the lockholder releases the lock. Thus there has to be an explicit wakeup after releasing the lock. 7. Put the OOM task on the waitqueue before marking the hierarchy as under OOM as that is the point where we start to receive wakeups. No point in listening before being on the waitqueue. 8. Likewise, unmark the hierarchy before finishing the sleep, for symmetry. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: azurIt <azurit@pobox.sk> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:43 +00:00
if (failed) {
/*
* OK, we failed to lock the whole subtree so we have
* to clean up what we set up to the failing subtree
*/
for_each_mem_cgroup_tree(iter, memcg) {
if (iter == failed) {
mem_cgroup_iter_break(memcg, iter);
break;
}
iter->oom_lock = false;
memcg: make oom_lock 0 and 1 based rather than counter Commit 867578cb ("memcg: fix oom kill behavior") introduced a oom_lock counter which is incremented by mem_cgroup_oom_lock when we are about to handle memcg OOM situation. mem_cgroup_handle_oom falls back to a sleep if oom_lock > 1 to prevent from multiple oom kills at the same time. The counter is then decremented by mem_cgroup_oom_unlock called from the same function. This works correctly but it can lead to serious starvations when we have many processes triggering OOM and many CPUs available for them (I have tested with 16 CPUs). Consider a process (call it A) which gets the oom_lock (the first one that got to mem_cgroup_handle_oom and grabbed memcg_oom_mutex) and other processes that are blocked on the mutex. While A releases the mutex and calls mem_cgroup_out_of_memory others will wake up (one after another) and increase the counter and fall into sleep (memcg_oom_waitq). Once A finishes mem_cgroup_out_of_memory it takes the mutex again and decreases oom_lock and wakes other tasks (if releasing memory by somebody else - e.g. killed process - hasn't done it yet). A testcase would look like: Assume malloc XXX is a program allocating XXX Megabytes of memory which touches all allocated pages in a tight loop # swapoff SWAP_DEVICE # cgcreate -g memory:A # cgset -r memory.oom_control=0 A # cgset -r memory.limit_in_bytes= 200M # for i in `seq 100` # do # cgexec -g memory:A malloc 10 & # done The main problem here is that all processes still race for the mutex and there is no guarantee that we will get counter back to 0 for those that got back to mem_cgroup_handle_oom. In the end the whole convoy in/decreases the counter but we do not get to 1 that would enable killing so nothing useful can be done. The time is basically unbounded because it highly depends on scheduling and ordering on mutex (I have seen this taking hours...). This patch replaces the counter by a simple {un}lock semantic. As mem_cgroup_oom_{un}lock works on the a subtree of a hierarchy we have to make sure that nobody else races with us which is guaranteed by the memcg_oom_mutex. We have to be careful while locking subtrees because we can encounter a subtree which is already locked: hierarchy: A / \ B \ /\ \ C D E B - C - D tree might be already locked. While we want to enable locking E subtree because OOM situations cannot influence each other we definitely do not want to allow locking A. Therefore we have to refuse lock if any subtree is already locked and clear up the lock for all nodes that have been set up to the failure point. On the other hand we have to make sure that the rest of the world will recognize that a group is under OOM even though it doesn't have a lock. Therefore we have to introduce under_oom variable which is incremented and decremented for the whole subtree when we enter resp. leave mem_cgroup_handle_oom. under_oom, unlike oom_lock, doesn't need be updated under memcg_oom_mutex because its users only check a single group and they use atomic operations for that. This can be checked easily by the following test case: # cgcreate -g memory:A # cgset -r memory.use_hierarchy=1 A # cgset -r memory.oom_control=1 A # cgset -r memory.limit_in_bytes= 100M # cgset -r memory.memsw.limit_in_bytes= 100M # cgcreate -g memory:A/B # cgset -r memory.oom_control=1 A/B # cgset -r memory.limit_in_bytes=20M # cgset -r memory.memsw.limit_in_bytes=20M # cgexec -g memory:A/B malloc 30 & #->this will be blocked by OOM of group B # cgexec -g memory:A malloc 80 & #->this will be blocked by OOM of group A While B gets oom_lock A will not get it. Both of them go into sleep and wait for an external action. We can make the limit higher for A to enforce waking it up # cgset -r memory.memsw.limit_in_bytes=300M A # cgset -r memory.limit_in_bytes=300M A malloc in A has to wake up even though it doesn't have oom_lock. Finally, the unlock path is very easy because we always unlock only the subtree we have locked previously while we always decrement under_oom. Signed-off-by: Michal Hocko <mhocko@suse.cz> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <bsingharora@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-07-26 23:08:23 +00:00
}
} else
mutex_acquire(&memcg_oom_lock_dep_map, 0, 1, _RET_IP_);
mm: memcg: rework and document OOM waiting and wakeup The memcg OOM handler open-codes a sleeping lock for OOM serialization (trylock, wait, repeat) because the required locking is so specific to memcg hierarchies. However, it would be nice if this construct would be clearly recognizable and not be as obfuscated as it is right now. Clean up as follows: 1. Remove the return value of mem_cgroup_oom_unlock() 2. Rename mem_cgroup_oom_lock() to mem_cgroup_oom_trylock(). 3. Pull the prepare_to_wait() out of the memcg_oom_lock scope. This makes it more obvious that the task has to be on the waitqueue before attempting to OOM-trylock the hierarchy, to not miss any wakeups before going to sleep. It just didn't matter until now because it was all lumped together into the global memcg_oom_lock spinlock section. 4. Pull the mem_cgroup_oom_notify() out of the memcg_oom_lock scope. It is proctected by the hierarchical OOM-lock. 5. The memcg_oom_lock spinlock is only required to propagate the OOM lock in any given hierarchy atomically. Restrict its scope to mem_cgroup_oom_(trylock|unlock). 6. Do not wake up the waitqueue unconditionally at the end of the function. Only the lockholder has to wake up the next in line after releasing the lock. Note that the lockholder kicks off the OOM-killer, which in turn leads to wakeups from the uncharges of the exiting task. But a contender is not guaranteed to see them if it enters the OOM path after the OOM kills but before the lockholder releases the lock. Thus there has to be an explicit wakeup after releasing the lock. 7. Put the OOM task on the waitqueue before marking the hierarchy as under OOM as that is the point where we start to receive wakeups. No point in listening before being on the waitqueue. 8. Likewise, unmark the hierarchy before finishing the sleep, for symmetry. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: azurIt <azurit@pobox.sk> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:43 +00:00
spin_unlock(&memcg_oom_lock);
return !failed;
}
mm: memcg: rework and document OOM waiting and wakeup The memcg OOM handler open-codes a sleeping lock for OOM serialization (trylock, wait, repeat) because the required locking is so specific to memcg hierarchies. However, it would be nice if this construct would be clearly recognizable and not be as obfuscated as it is right now. Clean up as follows: 1. Remove the return value of mem_cgroup_oom_unlock() 2. Rename mem_cgroup_oom_lock() to mem_cgroup_oom_trylock(). 3. Pull the prepare_to_wait() out of the memcg_oom_lock scope. This makes it more obvious that the task has to be on the waitqueue before attempting to OOM-trylock the hierarchy, to not miss any wakeups before going to sleep. It just didn't matter until now because it was all lumped together into the global memcg_oom_lock spinlock section. 4. Pull the mem_cgroup_oom_notify() out of the memcg_oom_lock scope. It is proctected by the hierarchical OOM-lock. 5. The memcg_oom_lock spinlock is only required to propagate the OOM lock in any given hierarchy atomically. Restrict its scope to mem_cgroup_oom_(trylock|unlock). 6. Do not wake up the waitqueue unconditionally at the end of the function. Only the lockholder has to wake up the next in line after releasing the lock. Note that the lockholder kicks off the OOM-killer, which in turn leads to wakeups from the uncharges of the exiting task. But a contender is not guaranteed to see them if it enters the OOM path after the OOM kills but before the lockholder releases the lock. Thus there has to be an explicit wakeup after releasing the lock. 7. Put the OOM task on the waitqueue before marking the hierarchy as under OOM as that is the point where we start to receive wakeups. No point in listening before being on the waitqueue. 8. Likewise, unmark the hierarchy before finishing the sleep, for symmetry. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: azurIt <azurit@pobox.sk> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:43 +00:00
static void mem_cgroup_oom_unlock(struct mem_cgroup *memcg)
{
struct mem_cgroup *iter;
mm: memcg: rework and document OOM waiting and wakeup The memcg OOM handler open-codes a sleeping lock for OOM serialization (trylock, wait, repeat) because the required locking is so specific to memcg hierarchies. However, it would be nice if this construct would be clearly recognizable and not be as obfuscated as it is right now. Clean up as follows: 1. Remove the return value of mem_cgroup_oom_unlock() 2. Rename mem_cgroup_oom_lock() to mem_cgroup_oom_trylock(). 3. Pull the prepare_to_wait() out of the memcg_oom_lock scope. This makes it more obvious that the task has to be on the waitqueue before attempting to OOM-trylock the hierarchy, to not miss any wakeups before going to sleep. It just didn't matter until now because it was all lumped together into the global memcg_oom_lock spinlock section. 4. Pull the mem_cgroup_oom_notify() out of the memcg_oom_lock scope. It is proctected by the hierarchical OOM-lock. 5. The memcg_oom_lock spinlock is only required to propagate the OOM lock in any given hierarchy atomically. Restrict its scope to mem_cgroup_oom_(trylock|unlock). 6. Do not wake up the waitqueue unconditionally at the end of the function. Only the lockholder has to wake up the next in line after releasing the lock. Note that the lockholder kicks off the OOM-killer, which in turn leads to wakeups from the uncharges of the exiting task. But a contender is not guaranteed to see them if it enters the OOM path after the OOM kills but before the lockholder releases the lock. Thus there has to be an explicit wakeup after releasing the lock. 7. Put the OOM task on the waitqueue before marking the hierarchy as under OOM as that is the point where we start to receive wakeups. No point in listening before being on the waitqueue. 8. Likewise, unmark the hierarchy before finishing the sleep, for symmetry. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: azurIt <azurit@pobox.sk> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:43 +00:00
spin_lock(&memcg_oom_lock);
2019-09-19 16:09:40 +00:00
mutex_release(&memcg_oom_lock_dep_map, _RET_IP_);
for_each_mem_cgroup_tree(iter, memcg)
memcg: make oom_lock 0 and 1 based rather than counter Commit 867578cb ("memcg: fix oom kill behavior") introduced a oom_lock counter which is incremented by mem_cgroup_oom_lock when we are about to handle memcg OOM situation. mem_cgroup_handle_oom falls back to a sleep if oom_lock > 1 to prevent from multiple oom kills at the same time. The counter is then decremented by mem_cgroup_oom_unlock called from the same function. This works correctly but it can lead to serious starvations when we have many processes triggering OOM and many CPUs available for them (I have tested with 16 CPUs). Consider a process (call it A) which gets the oom_lock (the first one that got to mem_cgroup_handle_oom and grabbed memcg_oom_mutex) and other processes that are blocked on the mutex. While A releases the mutex and calls mem_cgroup_out_of_memory others will wake up (one after another) and increase the counter and fall into sleep (memcg_oom_waitq). Once A finishes mem_cgroup_out_of_memory it takes the mutex again and decreases oom_lock and wakes other tasks (if releasing memory by somebody else - e.g. killed process - hasn't done it yet). A testcase would look like: Assume malloc XXX is a program allocating XXX Megabytes of memory which touches all allocated pages in a tight loop # swapoff SWAP_DEVICE # cgcreate -g memory:A # cgset -r memory.oom_control=0 A # cgset -r memory.limit_in_bytes= 200M # for i in `seq 100` # do # cgexec -g memory:A malloc 10 & # done The main problem here is that all processes still race for the mutex and there is no guarantee that we will get counter back to 0 for those that got back to mem_cgroup_handle_oom. In the end the whole convoy in/decreases the counter but we do not get to 1 that would enable killing so nothing useful can be done. The time is basically unbounded because it highly depends on scheduling and ordering on mutex (I have seen this taking hours...). This patch replaces the counter by a simple {un}lock semantic. As mem_cgroup_oom_{un}lock works on the a subtree of a hierarchy we have to make sure that nobody else races with us which is guaranteed by the memcg_oom_mutex. We have to be careful while locking subtrees because we can encounter a subtree which is already locked: hierarchy: A / \ B \ /\ \ C D E B - C - D tree might be already locked. While we want to enable locking E subtree because OOM situations cannot influence each other we definitely do not want to allow locking A. Therefore we have to refuse lock if any subtree is already locked and clear up the lock for all nodes that have been set up to the failure point. On the other hand we have to make sure that the rest of the world will recognize that a group is under OOM even though it doesn't have a lock. Therefore we have to introduce under_oom variable which is incremented and decremented for the whole subtree when we enter resp. leave mem_cgroup_handle_oom. under_oom, unlike oom_lock, doesn't need be updated under memcg_oom_mutex because its users only check a single group and they use atomic operations for that. This can be checked easily by the following test case: # cgcreate -g memory:A # cgset -r memory.use_hierarchy=1 A # cgset -r memory.oom_control=1 A # cgset -r memory.limit_in_bytes= 100M # cgset -r memory.memsw.limit_in_bytes= 100M # cgcreate -g memory:A/B # cgset -r memory.oom_control=1 A/B # cgset -r memory.limit_in_bytes=20M # cgset -r memory.memsw.limit_in_bytes=20M # cgexec -g memory:A/B malloc 30 & #->this will be blocked by OOM of group B # cgexec -g memory:A malloc 80 & #->this will be blocked by OOM of group A While B gets oom_lock A will not get it. Both of them go into sleep and wait for an external action. We can make the limit higher for A to enforce waking it up # cgset -r memory.memsw.limit_in_bytes=300M A # cgset -r memory.limit_in_bytes=300M A malloc in A has to wake up even though it doesn't have oom_lock. Finally, the unlock path is very easy because we always unlock only the subtree we have locked previously while we always decrement under_oom. Signed-off-by: Michal Hocko <mhocko@suse.cz> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <bsingharora@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-07-26 23:08:23 +00:00
iter->oom_lock = false;
mm: memcg: rework and document OOM waiting and wakeup The memcg OOM handler open-codes a sleeping lock for OOM serialization (trylock, wait, repeat) because the required locking is so specific to memcg hierarchies. However, it would be nice if this construct would be clearly recognizable and not be as obfuscated as it is right now. Clean up as follows: 1. Remove the return value of mem_cgroup_oom_unlock() 2. Rename mem_cgroup_oom_lock() to mem_cgroup_oom_trylock(). 3. Pull the prepare_to_wait() out of the memcg_oom_lock scope. This makes it more obvious that the task has to be on the waitqueue before attempting to OOM-trylock the hierarchy, to not miss any wakeups before going to sleep. It just didn't matter until now because it was all lumped together into the global memcg_oom_lock spinlock section. 4. Pull the mem_cgroup_oom_notify() out of the memcg_oom_lock scope. It is proctected by the hierarchical OOM-lock. 5. The memcg_oom_lock spinlock is only required to propagate the OOM lock in any given hierarchy atomically. Restrict its scope to mem_cgroup_oom_(trylock|unlock). 6. Do not wake up the waitqueue unconditionally at the end of the function. Only the lockholder has to wake up the next in line after releasing the lock. Note that the lockholder kicks off the OOM-killer, which in turn leads to wakeups from the uncharges of the exiting task. But a contender is not guaranteed to see them if it enters the OOM path after the OOM kills but before the lockholder releases the lock. Thus there has to be an explicit wakeup after releasing the lock. 7. Put the OOM task on the waitqueue before marking the hierarchy as under OOM as that is the point where we start to receive wakeups. No point in listening before being on the waitqueue. 8. Likewise, unmark the hierarchy before finishing the sleep, for symmetry. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: azurIt <azurit@pobox.sk> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:43 +00:00
spin_unlock(&memcg_oom_lock);
memcg: make oom_lock 0 and 1 based rather than counter Commit 867578cb ("memcg: fix oom kill behavior") introduced a oom_lock counter which is incremented by mem_cgroup_oom_lock when we are about to handle memcg OOM situation. mem_cgroup_handle_oom falls back to a sleep if oom_lock > 1 to prevent from multiple oom kills at the same time. The counter is then decremented by mem_cgroup_oom_unlock called from the same function. This works correctly but it can lead to serious starvations when we have many processes triggering OOM and many CPUs available for them (I have tested with 16 CPUs). Consider a process (call it A) which gets the oom_lock (the first one that got to mem_cgroup_handle_oom and grabbed memcg_oom_mutex) and other processes that are blocked on the mutex. While A releases the mutex and calls mem_cgroup_out_of_memory others will wake up (one after another) and increase the counter and fall into sleep (memcg_oom_waitq). Once A finishes mem_cgroup_out_of_memory it takes the mutex again and decreases oom_lock and wakes other tasks (if releasing memory by somebody else - e.g. killed process - hasn't done it yet). A testcase would look like: Assume malloc XXX is a program allocating XXX Megabytes of memory which touches all allocated pages in a tight loop # swapoff SWAP_DEVICE # cgcreate -g memory:A # cgset -r memory.oom_control=0 A # cgset -r memory.limit_in_bytes= 200M # for i in `seq 100` # do # cgexec -g memory:A malloc 10 & # done The main problem here is that all processes still race for the mutex and there is no guarantee that we will get counter back to 0 for those that got back to mem_cgroup_handle_oom. In the end the whole convoy in/decreases the counter but we do not get to 1 that would enable killing so nothing useful can be done. The time is basically unbounded because it highly depends on scheduling and ordering on mutex (I have seen this taking hours...). This patch replaces the counter by a simple {un}lock semantic. As mem_cgroup_oom_{un}lock works on the a subtree of a hierarchy we have to make sure that nobody else races with us which is guaranteed by the memcg_oom_mutex. We have to be careful while locking subtrees because we can encounter a subtree which is already locked: hierarchy: A / \ B \ /\ \ C D E B - C - D tree might be already locked. While we want to enable locking E subtree because OOM situations cannot influence each other we definitely do not want to allow locking A. Therefore we have to refuse lock if any subtree is already locked and clear up the lock for all nodes that have been set up to the failure point. On the other hand we have to make sure that the rest of the world will recognize that a group is under OOM even though it doesn't have a lock. Therefore we have to introduce under_oom variable which is incremented and decremented for the whole subtree when we enter resp. leave mem_cgroup_handle_oom. under_oom, unlike oom_lock, doesn't need be updated under memcg_oom_mutex because its users only check a single group and they use atomic operations for that. This can be checked easily by the following test case: # cgcreate -g memory:A # cgset -r memory.use_hierarchy=1 A # cgset -r memory.oom_control=1 A # cgset -r memory.limit_in_bytes= 100M # cgset -r memory.memsw.limit_in_bytes= 100M # cgcreate -g memory:A/B # cgset -r memory.oom_control=1 A/B # cgset -r memory.limit_in_bytes=20M # cgset -r memory.memsw.limit_in_bytes=20M # cgexec -g memory:A/B malloc 30 & #->this will be blocked by OOM of group B # cgexec -g memory:A malloc 80 & #->this will be blocked by OOM of group A While B gets oom_lock A will not get it. Both of them go into sleep and wait for an external action. We can make the limit higher for A to enforce waking it up # cgset -r memory.memsw.limit_in_bytes=300M A # cgset -r memory.limit_in_bytes=300M A malloc in A has to wake up even though it doesn't have oom_lock. Finally, the unlock path is very easy because we always unlock only the subtree we have locked previously while we always decrement under_oom. Signed-off-by: Michal Hocko <mhocko@suse.cz> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <bsingharora@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-07-26 23:08:23 +00:00
}
static void mem_cgroup_mark_under_oom(struct mem_cgroup *memcg)
memcg: make oom_lock 0 and 1 based rather than counter Commit 867578cb ("memcg: fix oom kill behavior") introduced a oom_lock counter which is incremented by mem_cgroup_oom_lock when we are about to handle memcg OOM situation. mem_cgroup_handle_oom falls back to a sleep if oom_lock > 1 to prevent from multiple oom kills at the same time. The counter is then decremented by mem_cgroup_oom_unlock called from the same function. This works correctly but it can lead to serious starvations when we have many processes triggering OOM and many CPUs available for them (I have tested with 16 CPUs). Consider a process (call it A) which gets the oom_lock (the first one that got to mem_cgroup_handle_oom and grabbed memcg_oom_mutex) and other processes that are blocked on the mutex. While A releases the mutex and calls mem_cgroup_out_of_memory others will wake up (one after another) and increase the counter and fall into sleep (memcg_oom_waitq). Once A finishes mem_cgroup_out_of_memory it takes the mutex again and decreases oom_lock and wakes other tasks (if releasing memory by somebody else - e.g. killed process - hasn't done it yet). A testcase would look like: Assume malloc XXX is a program allocating XXX Megabytes of memory which touches all allocated pages in a tight loop # swapoff SWAP_DEVICE # cgcreate -g memory:A # cgset -r memory.oom_control=0 A # cgset -r memory.limit_in_bytes= 200M # for i in `seq 100` # do # cgexec -g memory:A malloc 10 & # done The main problem here is that all processes still race for the mutex and there is no guarantee that we will get counter back to 0 for those that got back to mem_cgroup_handle_oom. In the end the whole convoy in/decreases the counter but we do not get to 1 that would enable killing so nothing useful can be done. The time is basically unbounded because it highly depends on scheduling and ordering on mutex (I have seen this taking hours...). This patch replaces the counter by a simple {un}lock semantic. As mem_cgroup_oom_{un}lock works on the a subtree of a hierarchy we have to make sure that nobody else races with us which is guaranteed by the memcg_oom_mutex. We have to be careful while locking subtrees because we can encounter a subtree which is already locked: hierarchy: A / \ B \ /\ \ C D E B - C - D tree might be already locked. While we want to enable locking E subtree because OOM situations cannot influence each other we definitely do not want to allow locking A. Therefore we have to refuse lock if any subtree is already locked and clear up the lock for all nodes that have been set up to the failure point. On the other hand we have to make sure that the rest of the world will recognize that a group is under OOM even though it doesn't have a lock. Therefore we have to introduce under_oom variable which is incremented and decremented for the whole subtree when we enter resp. leave mem_cgroup_handle_oom. under_oom, unlike oom_lock, doesn't need be updated under memcg_oom_mutex because its users only check a single group and they use atomic operations for that. This can be checked easily by the following test case: # cgcreate -g memory:A # cgset -r memory.use_hierarchy=1 A # cgset -r memory.oom_control=1 A # cgset -r memory.limit_in_bytes= 100M # cgset -r memory.memsw.limit_in_bytes= 100M # cgcreate -g memory:A/B # cgset -r memory.oom_control=1 A/B # cgset -r memory.limit_in_bytes=20M # cgset -r memory.memsw.limit_in_bytes=20M # cgexec -g memory:A/B malloc 30 & #->this will be blocked by OOM of group B # cgexec -g memory:A malloc 80 & #->this will be blocked by OOM of group A While B gets oom_lock A will not get it. Both of them go into sleep and wait for an external action. We can make the limit higher for A to enforce waking it up # cgset -r memory.memsw.limit_in_bytes=300M A # cgset -r memory.limit_in_bytes=300M A malloc in A has to wake up even though it doesn't have oom_lock. Finally, the unlock path is very easy because we always unlock only the subtree we have locked previously while we always decrement under_oom. Signed-off-by: Michal Hocko <mhocko@suse.cz> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <bsingharora@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-07-26 23:08:23 +00:00
{
struct mem_cgroup *iter;
spin_lock(&memcg_oom_lock);
for_each_mem_cgroup_tree(iter, memcg)
iter->under_oom++;
spin_unlock(&memcg_oom_lock);
memcg: make oom_lock 0 and 1 based rather than counter Commit 867578cb ("memcg: fix oom kill behavior") introduced a oom_lock counter which is incremented by mem_cgroup_oom_lock when we are about to handle memcg OOM situation. mem_cgroup_handle_oom falls back to a sleep if oom_lock > 1 to prevent from multiple oom kills at the same time. The counter is then decremented by mem_cgroup_oom_unlock called from the same function. This works correctly but it can lead to serious starvations when we have many processes triggering OOM and many CPUs available for them (I have tested with 16 CPUs). Consider a process (call it A) which gets the oom_lock (the first one that got to mem_cgroup_handle_oom and grabbed memcg_oom_mutex) and other processes that are blocked on the mutex. While A releases the mutex and calls mem_cgroup_out_of_memory others will wake up (one after another) and increase the counter and fall into sleep (memcg_oom_waitq). Once A finishes mem_cgroup_out_of_memory it takes the mutex again and decreases oom_lock and wakes other tasks (if releasing memory by somebody else - e.g. killed process - hasn't done it yet). A testcase would look like: Assume malloc XXX is a program allocating XXX Megabytes of memory which touches all allocated pages in a tight loop # swapoff SWAP_DEVICE # cgcreate -g memory:A # cgset -r memory.oom_control=0 A # cgset -r memory.limit_in_bytes= 200M # for i in `seq 100` # do # cgexec -g memory:A malloc 10 & # done The main problem here is that all processes still race for the mutex and there is no guarantee that we will get counter back to 0 for those that got back to mem_cgroup_handle_oom. In the end the whole convoy in/decreases the counter but we do not get to 1 that would enable killing so nothing useful can be done. The time is basically unbounded because it highly depends on scheduling and ordering on mutex (I have seen this taking hours...). This patch replaces the counter by a simple {un}lock semantic. As mem_cgroup_oom_{un}lock works on the a subtree of a hierarchy we have to make sure that nobody else races with us which is guaranteed by the memcg_oom_mutex. We have to be careful while locking subtrees because we can encounter a subtree which is already locked: hierarchy: A / \ B \ /\ \ C D E B - C - D tree might be already locked. While we want to enable locking E subtree because OOM situations cannot influence each other we definitely do not want to allow locking A. Therefore we have to refuse lock if any subtree is already locked and clear up the lock for all nodes that have been set up to the failure point. On the other hand we have to make sure that the rest of the world will recognize that a group is under OOM even though it doesn't have a lock. Therefore we have to introduce under_oom variable which is incremented and decremented for the whole subtree when we enter resp. leave mem_cgroup_handle_oom. under_oom, unlike oom_lock, doesn't need be updated under memcg_oom_mutex because its users only check a single group and they use atomic operations for that. This can be checked easily by the following test case: # cgcreate -g memory:A # cgset -r memory.use_hierarchy=1 A # cgset -r memory.oom_control=1 A # cgset -r memory.limit_in_bytes= 100M # cgset -r memory.memsw.limit_in_bytes= 100M # cgcreate -g memory:A/B # cgset -r memory.oom_control=1 A/B # cgset -r memory.limit_in_bytes=20M # cgset -r memory.memsw.limit_in_bytes=20M # cgexec -g memory:A/B malloc 30 & #->this will be blocked by OOM of group B # cgexec -g memory:A malloc 80 & #->this will be blocked by OOM of group A While B gets oom_lock A will not get it. Both of them go into sleep and wait for an external action. We can make the limit higher for A to enforce waking it up # cgset -r memory.memsw.limit_in_bytes=300M A # cgset -r memory.limit_in_bytes=300M A malloc in A has to wake up even though it doesn't have oom_lock. Finally, the unlock path is very easy because we always unlock only the subtree we have locked previously while we always decrement under_oom. Signed-off-by: Michal Hocko <mhocko@suse.cz> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <bsingharora@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-07-26 23:08:23 +00:00
}
static void mem_cgroup_unmark_under_oom(struct mem_cgroup *memcg)
memcg: make oom_lock 0 and 1 based rather than counter Commit 867578cb ("memcg: fix oom kill behavior") introduced a oom_lock counter which is incremented by mem_cgroup_oom_lock when we are about to handle memcg OOM situation. mem_cgroup_handle_oom falls back to a sleep if oom_lock > 1 to prevent from multiple oom kills at the same time. The counter is then decremented by mem_cgroup_oom_unlock called from the same function. This works correctly but it can lead to serious starvations when we have many processes triggering OOM and many CPUs available for them (I have tested with 16 CPUs). Consider a process (call it A) which gets the oom_lock (the first one that got to mem_cgroup_handle_oom and grabbed memcg_oom_mutex) and other processes that are blocked on the mutex. While A releases the mutex and calls mem_cgroup_out_of_memory others will wake up (one after another) and increase the counter and fall into sleep (memcg_oom_waitq). Once A finishes mem_cgroup_out_of_memory it takes the mutex again and decreases oom_lock and wakes other tasks (if releasing memory by somebody else - e.g. killed process - hasn't done it yet). A testcase would look like: Assume malloc XXX is a program allocating XXX Megabytes of memory which touches all allocated pages in a tight loop # swapoff SWAP_DEVICE # cgcreate -g memory:A # cgset -r memory.oom_control=0 A # cgset -r memory.limit_in_bytes= 200M # for i in `seq 100` # do # cgexec -g memory:A malloc 10 & # done The main problem here is that all processes still race for the mutex and there is no guarantee that we will get counter back to 0 for those that got back to mem_cgroup_handle_oom. In the end the whole convoy in/decreases the counter but we do not get to 1 that would enable killing so nothing useful can be done. The time is basically unbounded because it highly depends on scheduling and ordering on mutex (I have seen this taking hours...). This patch replaces the counter by a simple {un}lock semantic. As mem_cgroup_oom_{un}lock works on the a subtree of a hierarchy we have to make sure that nobody else races with us which is guaranteed by the memcg_oom_mutex. We have to be careful while locking subtrees because we can encounter a subtree which is already locked: hierarchy: A / \ B \ /\ \ C D E B - C - D tree might be already locked. While we want to enable locking E subtree because OOM situations cannot influence each other we definitely do not want to allow locking A. Therefore we have to refuse lock if any subtree is already locked and clear up the lock for all nodes that have been set up to the failure point. On the other hand we have to make sure that the rest of the world will recognize that a group is under OOM even though it doesn't have a lock. Therefore we have to introduce under_oom variable which is incremented and decremented for the whole subtree when we enter resp. leave mem_cgroup_handle_oom. under_oom, unlike oom_lock, doesn't need be updated under memcg_oom_mutex because its users only check a single group and they use atomic operations for that. This can be checked easily by the following test case: # cgcreate -g memory:A # cgset -r memory.use_hierarchy=1 A # cgset -r memory.oom_control=1 A # cgset -r memory.limit_in_bytes= 100M # cgset -r memory.memsw.limit_in_bytes= 100M # cgcreate -g memory:A/B # cgset -r memory.oom_control=1 A/B # cgset -r memory.limit_in_bytes=20M # cgset -r memory.memsw.limit_in_bytes=20M # cgexec -g memory:A/B malloc 30 & #->this will be blocked by OOM of group B # cgexec -g memory:A malloc 80 & #->this will be blocked by OOM of group A While B gets oom_lock A will not get it. Both of them go into sleep and wait for an external action. We can make the limit higher for A to enforce waking it up # cgset -r memory.memsw.limit_in_bytes=300M A # cgset -r memory.limit_in_bytes=300M A malloc in A has to wake up even though it doesn't have oom_lock. Finally, the unlock path is very easy because we always unlock only the subtree we have locked previously while we always decrement under_oom. Signed-off-by: Michal Hocko <mhocko@suse.cz> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <bsingharora@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-07-26 23:08:23 +00:00
{
struct mem_cgroup *iter;
memcg: fix oom kill behavior In current page-fault code, handle_mm_fault() -> ... -> mem_cgroup_charge() -> map page or handle error. -> check return code. If page fault's return code is VM_FAULT_OOM, page_fault_out_of_memory() is called. But if it's caused by memcg, OOM should have been already invoked. Then, I added a patch: a636b327f731143ccc544b966cfd8de6cb6d72c6. That patch records last_oom_jiffies for memcg's sub-hierarchy and prevents page_fault_out_of_memory from being invoked in near future. But Nishimura-san reported that check by jiffies is not enough when the system is terribly heavy. This patch changes memcg's oom logic as. * If memcg causes OOM-kill, continue to retry. * remove jiffies check which is used now. * add memcg-oom-lock which works like perzone oom lock. * If current is killed(as a process), bypass charge. Something more sophisticated can be added but this pactch does fundamental things. TODO: - add oom notifier - add permemcg disable-oom-kill flag and freezer at oom. - more chances for wake up oom waiter (when changing memory limit etc..) Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Tested-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-03-10 23:22:39 +00:00
/*
* When a new child is created while the hierarchy is under oom,
* mem_cgroup_oom_lock() may not be called. Watch for underflow.
memcg: fix oom kill behavior In current page-fault code, handle_mm_fault() -> ... -> mem_cgroup_charge() -> map page or handle error. -> check return code. If page fault's return code is VM_FAULT_OOM, page_fault_out_of_memory() is called. But if it's caused by memcg, OOM should have been already invoked. Then, I added a patch: a636b327f731143ccc544b966cfd8de6cb6d72c6. That patch records last_oom_jiffies for memcg's sub-hierarchy and prevents page_fault_out_of_memory from being invoked in near future. But Nishimura-san reported that check by jiffies is not enough when the system is terribly heavy. This patch changes memcg's oom logic as. * If memcg causes OOM-kill, continue to retry. * remove jiffies check which is used now. * add memcg-oom-lock which works like perzone oom lock. * If current is killed(as a process), bypass charge. Something more sophisticated can be added but this pactch does fundamental things. TODO: - add oom notifier - add permemcg disable-oom-kill flag and freezer at oom. - more chances for wake up oom waiter (when changing memory limit etc..) Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Tested-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-03-10 23:22:39 +00:00
*/
spin_lock(&memcg_oom_lock);
for_each_mem_cgroup_tree(iter, memcg)
if (iter->under_oom > 0)
iter->under_oom--;
spin_unlock(&memcg_oom_lock);
}
memcg: fix oom kill behavior In current page-fault code, handle_mm_fault() -> ... -> mem_cgroup_charge() -> map page or handle error. -> check return code. If page fault's return code is VM_FAULT_OOM, page_fault_out_of_memory() is called. But if it's caused by memcg, OOM should have been already invoked. Then, I added a patch: a636b327f731143ccc544b966cfd8de6cb6d72c6. That patch records last_oom_jiffies for memcg's sub-hierarchy and prevents page_fault_out_of_memory from being invoked in near future. But Nishimura-san reported that check by jiffies is not enough when the system is terribly heavy. This patch changes memcg's oom logic as. * If memcg causes OOM-kill, continue to retry. * remove jiffies check which is used now. * add memcg-oom-lock which works like perzone oom lock. * If current is killed(as a process), bypass charge. Something more sophisticated can be added but this pactch does fundamental things. TODO: - add oom notifier - add permemcg disable-oom-kill flag and freezer at oom. - more chances for wake up oom waiter (when changing memory limit etc..) Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Tested-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-03-10 23:22:39 +00:00
static DECLARE_WAIT_QUEUE_HEAD(memcg_oom_waitq);
struct oom_wait_info {
struct mem_cgroup *memcg;
wait_queue_entry_t wait;
};
static int memcg_oom_wake_function(wait_queue_entry_t *wait,
unsigned mode, int sync, void *arg)
{
struct mem_cgroup *wake_memcg = (struct mem_cgroup *)arg;
struct mem_cgroup *oom_wait_memcg;
struct oom_wait_info *oom_wait_info;
oom_wait_info = container_of(wait, struct oom_wait_info, wait);
oom_wait_memcg = oom_wait_info->memcg;
if (!mem_cgroup_is_descendant(wake_memcg, oom_wait_memcg) &&
!mem_cgroup_is_descendant(oom_wait_memcg, wake_memcg))
return 0;
return autoremove_wake_function(wait, mode, sync, arg);
}
static void memcg_oom_recover(struct mem_cgroup *memcg)
{
/*
* For the following lockless ->under_oom test, the only required
* guarantee is that it must see the state asserted by an OOM when
* this function is called as a result of userland actions
* triggered by the notification of the OOM. This is trivially
* achieved by invoking mem_cgroup_mark_under_oom() before
* triggering notification.
*/
if (memcg && memcg->under_oom)
__wake_up(&memcg_oom_waitq, TASK_NORMAL, 0, memcg);
}
memcg, oom: move out_of_memory back to the charge path Commit 3812c8c8f395 ("mm: memcg: do not trap chargers with full callstack on OOM") has changed the ENOMEM semantic of memcg charges. Rather than invoking the oom killer from the charging context it delays the oom killer to the page fault path (pagefault_out_of_memory). This in turn means that many users (e.g. slab or g-u-p) will get ENOMEM when the corresponding memcg hits the hard limit and the memcg is is OOM. This is behavior is inconsistent with !memcg case where the oom killer is invoked from the allocation context and the allocator keeps retrying until it succeeds. The difference in the behavior is user visible. mmap(MAP_POPULATE) might result in not fully populated ranges while the mmap return code doesn't tell that to the userspace. Random syscalls might fail with ENOMEM etc. The primary motivation of the different memcg oom semantic was the deadlock avoidance. Things have changed since then, though. We have an async oom teardown by the oom reaper now and so we do not have to rely on the victim to tear down its memory anymore. Therefore we can return to the original semantic as long as the memcg oom killer is not handed over to the users space. There is still one thing to be careful about here though. If the oom killer is not able to make any forward progress - e.g. because there is no eligible task to kill - then we have to bail out of the charge path to prevent from same class of deadlocks. We have basically two options here. Either we fail the charge with ENOMEM or force the charge and allow overcharge. The first option has been considered more harmful than useful because rare inconsistencies in the ENOMEM behavior is hard to test for and error prone. Basically the same reason why the page allocator doesn't fail allocations under such conditions. The later might allow runaways but those should be really unlikely unless somebody misconfigures the system. E.g. allowing to migrate tasks away from the memcg to a different unlimited memcg with move_charge_at_immigrate disabled. Link: http://lkml.kernel.org/r/20180628151101.25307-1-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Acked-by: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Shakeel Butt <shakeelb@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:11 +00:00
enum oom_status {
OOM_SUCCESS,
OOM_FAILED,
OOM_ASYNC,
OOM_SKIPPED
};
static enum oom_status mem_cgroup_oom(struct mem_cgroup *memcg, gfp_t mask, int order)
{
enum oom_status ret;
bool locked;
memcg, oom: move out_of_memory back to the charge path Commit 3812c8c8f395 ("mm: memcg: do not trap chargers with full callstack on OOM") has changed the ENOMEM semantic of memcg charges. Rather than invoking the oom killer from the charging context it delays the oom killer to the page fault path (pagefault_out_of_memory). This in turn means that many users (e.g. slab or g-u-p) will get ENOMEM when the corresponding memcg hits the hard limit and the memcg is is OOM. This is behavior is inconsistent with !memcg case where the oom killer is invoked from the allocation context and the allocator keeps retrying until it succeeds. The difference in the behavior is user visible. mmap(MAP_POPULATE) might result in not fully populated ranges while the mmap return code doesn't tell that to the userspace. Random syscalls might fail with ENOMEM etc. The primary motivation of the different memcg oom semantic was the deadlock avoidance. Things have changed since then, though. We have an async oom teardown by the oom reaper now and so we do not have to rely on the victim to tear down its memory anymore. Therefore we can return to the original semantic as long as the memcg oom killer is not handed over to the users space. There is still one thing to be careful about here though. If the oom killer is not able to make any forward progress - e.g. because there is no eligible task to kill - then we have to bail out of the charge path to prevent from same class of deadlocks. We have basically two options here. Either we fail the charge with ENOMEM or force the charge and allow overcharge. The first option has been considered more harmful than useful because rare inconsistencies in the ENOMEM behavior is hard to test for and error prone. Basically the same reason why the page allocator doesn't fail allocations under such conditions. The later might allow runaways but those should be really unlikely unless somebody misconfigures the system. E.g. allowing to migrate tasks away from the memcg to a different unlimited memcg with move_charge_at_immigrate disabled. Link: http://lkml.kernel.org/r/20180628151101.25307-1-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Acked-by: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Shakeel Butt <shakeelb@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:11 +00:00
if (order > PAGE_ALLOC_COSTLY_ORDER)
return OOM_SKIPPED;
mm: don't raise MEMCG_OOM event due to failed high-order allocation It was reported that on some of our machines containers were restarted with OOM symptoms without an obvious reason. Despite there were almost no memory pressure and plenty of page cache, MEMCG_OOM event was raised occasionally, causing the container management software to think, that OOM has happened. However, no tasks have been killed. The following investigation showed that the problem is caused by a failing attempt to charge a high-order page. In such case, the OOM killer is never invoked. As shown below, it can happen under conditions, which are very far from a real OOM: e.g. there is plenty of clean page cache and no memory pressure. There is no sense in raising an OOM event in this case, as it might confuse a user and lead to wrong and excessive actions (e.g. restart the workload, as in my case). Let's look at the charging path in try_charge(). If the memory usage is about memory.max, which is absolutely natural for most memory cgroups, we try to reclaim some pages. Even if we were able to reclaim enough memory for the allocation, the following check can fail due to a race with another concurrent allocation: if (mem_cgroup_margin(mem_over_limit) >= nr_pages) goto retry; For regular pages the following condition will save us from triggering the OOM: if (nr_reclaimed && nr_pages <= (1 << PAGE_ALLOC_COSTLY_ORDER)) goto retry; But for high-order allocation this condition will intentionally fail. The reason behind is that we'll likely fall to regular pages anyway, so it's ok and even preferred to return ENOMEM. In this case the idea of raising MEMCG_OOM looks dubious. Fix this by moving MEMCG_OOM raising to mem_cgroup_oom() after allocation order check, so that the event won't be raised for high order allocations. This change doesn't affect regular pages allocation and charging. Link: http://lkml.kernel.org/r/20181004214050.7417-1-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: David Rientjes <rientjes@google.com> Acked-by: Michal Hocko <mhocko@kernel.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-10-26 22:09:48 +00:00
memcg_memory_event(memcg, MEMCG_OOM);
memcg: fix oom kill behavior In current page-fault code, handle_mm_fault() -> ... -> mem_cgroup_charge() -> map page or handle error. -> check return code. If page fault's return code is VM_FAULT_OOM, page_fault_out_of_memory() is called. But if it's caused by memcg, OOM should have been already invoked. Then, I added a patch: a636b327f731143ccc544b966cfd8de6cb6d72c6. That patch records last_oom_jiffies for memcg's sub-hierarchy and prevents page_fault_out_of_memory from being invoked in near future. But Nishimura-san reported that check by jiffies is not enough when the system is terribly heavy. This patch changes memcg's oom logic as. * If memcg causes OOM-kill, continue to retry. * remove jiffies check which is used now. * add memcg-oom-lock which works like perzone oom lock. * If current is killed(as a process), bypass charge. Something more sophisticated can be added but this pactch does fundamental things. TODO: - add oom notifier - add permemcg disable-oom-kill flag and freezer at oom. - more chances for wake up oom waiter (when changing memory limit etc..) Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Tested-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-03-10 23:22:39 +00:00
/*
* We are in the middle of the charge context here, so we
* don't want to block when potentially sitting on a callstack
* that holds all kinds of filesystem and mm locks.
*
memcg, oom: move out_of_memory back to the charge path Commit 3812c8c8f395 ("mm: memcg: do not trap chargers with full callstack on OOM") has changed the ENOMEM semantic of memcg charges. Rather than invoking the oom killer from the charging context it delays the oom killer to the page fault path (pagefault_out_of_memory). This in turn means that many users (e.g. slab or g-u-p) will get ENOMEM when the corresponding memcg hits the hard limit and the memcg is is OOM. This is behavior is inconsistent with !memcg case where the oom killer is invoked from the allocation context and the allocator keeps retrying until it succeeds. The difference in the behavior is user visible. mmap(MAP_POPULATE) might result in not fully populated ranges while the mmap return code doesn't tell that to the userspace. Random syscalls might fail with ENOMEM etc. The primary motivation of the different memcg oom semantic was the deadlock avoidance. Things have changed since then, though. We have an async oom teardown by the oom reaper now and so we do not have to rely on the victim to tear down its memory anymore. Therefore we can return to the original semantic as long as the memcg oom killer is not handed over to the users space. There is still one thing to be careful about here though. If the oom killer is not able to make any forward progress - e.g. because there is no eligible task to kill - then we have to bail out of the charge path to prevent from same class of deadlocks. We have basically two options here. Either we fail the charge with ENOMEM or force the charge and allow overcharge. The first option has been considered more harmful than useful because rare inconsistencies in the ENOMEM behavior is hard to test for and error prone. Basically the same reason why the page allocator doesn't fail allocations under such conditions. The later might allow runaways but those should be really unlikely unless somebody misconfigures the system. E.g. allowing to migrate tasks away from the memcg to a different unlimited memcg with move_charge_at_immigrate disabled. Link: http://lkml.kernel.org/r/20180628151101.25307-1-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Acked-by: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Shakeel Butt <shakeelb@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:11 +00:00
* cgroup1 allows disabling the OOM killer and waiting for outside
* handling until the charge can succeed; remember the context and put
* the task to sleep at the end of the page fault when all locks are
* released.
*
memcg, oom: move out_of_memory back to the charge path Commit 3812c8c8f395 ("mm: memcg: do not trap chargers with full callstack on OOM") has changed the ENOMEM semantic of memcg charges. Rather than invoking the oom killer from the charging context it delays the oom killer to the page fault path (pagefault_out_of_memory). This in turn means that many users (e.g. slab or g-u-p) will get ENOMEM when the corresponding memcg hits the hard limit and the memcg is is OOM. This is behavior is inconsistent with !memcg case where the oom killer is invoked from the allocation context and the allocator keeps retrying until it succeeds. The difference in the behavior is user visible. mmap(MAP_POPULATE) might result in not fully populated ranges while the mmap return code doesn't tell that to the userspace. Random syscalls might fail with ENOMEM etc. The primary motivation of the different memcg oom semantic was the deadlock avoidance. Things have changed since then, though. We have an async oom teardown by the oom reaper now and so we do not have to rely on the victim to tear down its memory anymore. Therefore we can return to the original semantic as long as the memcg oom killer is not handed over to the users space. There is still one thing to be careful about here though. If the oom killer is not able to make any forward progress - e.g. because there is no eligible task to kill - then we have to bail out of the charge path to prevent from same class of deadlocks. We have basically two options here. Either we fail the charge with ENOMEM or force the charge and allow overcharge. The first option has been considered more harmful than useful because rare inconsistencies in the ENOMEM behavior is hard to test for and error prone. Basically the same reason why the page allocator doesn't fail allocations under such conditions. The later might allow runaways but those should be really unlikely unless somebody misconfigures the system. E.g. allowing to migrate tasks away from the memcg to a different unlimited memcg with move_charge_at_immigrate disabled. Link: http://lkml.kernel.org/r/20180628151101.25307-1-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Acked-by: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Shakeel Butt <shakeelb@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:11 +00:00
* On the other hand, in-kernel OOM killer allows for an async victim
* memory reclaim (oom_reaper) and that means that we are not solely
* relying on the oom victim to make a forward progress and we can
* invoke the oom killer here.
*
* Please note that mem_cgroup_out_of_memory might fail to find a
* victim and then we have to bail out from the charge path.
memcg: fix oom kill behavior In current page-fault code, handle_mm_fault() -> ... -> mem_cgroup_charge() -> map page or handle error. -> check return code. If page fault's return code is VM_FAULT_OOM, page_fault_out_of_memory() is called. But if it's caused by memcg, OOM should have been already invoked. Then, I added a patch: a636b327f731143ccc544b966cfd8de6cb6d72c6. That patch records last_oom_jiffies for memcg's sub-hierarchy and prevents page_fault_out_of_memory from being invoked in near future. But Nishimura-san reported that check by jiffies is not enough when the system is terribly heavy. This patch changes memcg's oom logic as. * If memcg causes OOM-kill, continue to retry. * remove jiffies check which is used now. * add memcg-oom-lock which works like perzone oom lock. * If current is killed(as a process), bypass charge. Something more sophisticated can be added but this pactch does fundamental things. TODO: - add oom notifier - add permemcg disable-oom-kill flag and freezer at oom. - more chances for wake up oom waiter (when changing memory limit etc..) Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Tested-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-03-10 23:22:39 +00:00
*/
memcg, oom: move out_of_memory back to the charge path Commit 3812c8c8f395 ("mm: memcg: do not trap chargers with full callstack on OOM") has changed the ENOMEM semantic of memcg charges. Rather than invoking the oom killer from the charging context it delays the oom killer to the page fault path (pagefault_out_of_memory). This in turn means that many users (e.g. slab or g-u-p) will get ENOMEM when the corresponding memcg hits the hard limit and the memcg is is OOM. This is behavior is inconsistent with !memcg case where the oom killer is invoked from the allocation context and the allocator keeps retrying until it succeeds. The difference in the behavior is user visible. mmap(MAP_POPULATE) might result in not fully populated ranges while the mmap return code doesn't tell that to the userspace. Random syscalls might fail with ENOMEM etc. The primary motivation of the different memcg oom semantic was the deadlock avoidance. Things have changed since then, though. We have an async oom teardown by the oom reaper now and so we do not have to rely on the victim to tear down its memory anymore. Therefore we can return to the original semantic as long as the memcg oom killer is not handed over to the users space. There is still one thing to be careful about here though. If the oom killer is not able to make any forward progress - e.g. because there is no eligible task to kill - then we have to bail out of the charge path to prevent from same class of deadlocks. We have basically two options here. Either we fail the charge with ENOMEM or force the charge and allow overcharge. The first option has been considered more harmful than useful because rare inconsistencies in the ENOMEM behavior is hard to test for and error prone. Basically the same reason why the page allocator doesn't fail allocations under such conditions. The later might allow runaways but those should be really unlikely unless somebody misconfigures the system. E.g. allowing to migrate tasks away from the memcg to a different unlimited memcg with move_charge_at_immigrate disabled. Link: http://lkml.kernel.org/r/20180628151101.25307-1-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Acked-by: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Shakeel Butt <shakeelb@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:11 +00:00
if (memcg->oom_kill_disable) {
if (!current->in_user_fault)
return OOM_SKIPPED;
css_get(&memcg->css);
current->memcg_in_oom = memcg;
current->memcg_oom_gfp_mask = mask;
current->memcg_oom_order = order;
return OOM_ASYNC;
}
mem_cgroup_mark_under_oom(memcg);
locked = mem_cgroup_oom_trylock(memcg);
if (locked)
mem_cgroup_oom_notify(memcg);
mem_cgroup_unmark_under_oom(memcg);
memcg, oom: move out_of_memory back to the charge path Commit 3812c8c8f395 ("mm: memcg: do not trap chargers with full callstack on OOM") has changed the ENOMEM semantic of memcg charges. Rather than invoking the oom killer from the charging context it delays the oom killer to the page fault path (pagefault_out_of_memory). This in turn means that many users (e.g. slab or g-u-p) will get ENOMEM when the corresponding memcg hits the hard limit and the memcg is is OOM. This is behavior is inconsistent with !memcg case where the oom killer is invoked from the allocation context and the allocator keeps retrying until it succeeds. The difference in the behavior is user visible. mmap(MAP_POPULATE) might result in not fully populated ranges while the mmap return code doesn't tell that to the userspace. Random syscalls might fail with ENOMEM etc. The primary motivation of the different memcg oom semantic was the deadlock avoidance. Things have changed since then, though. We have an async oom teardown by the oom reaper now and so we do not have to rely on the victim to tear down its memory anymore. Therefore we can return to the original semantic as long as the memcg oom killer is not handed over to the users space. There is still one thing to be careful about here though. If the oom killer is not able to make any forward progress - e.g. because there is no eligible task to kill - then we have to bail out of the charge path to prevent from same class of deadlocks. We have basically two options here. Either we fail the charge with ENOMEM or force the charge and allow overcharge. The first option has been considered more harmful than useful because rare inconsistencies in the ENOMEM behavior is hard to test for and error prone. Basically the same reason why the page allocator doesn't fail allocations under such conditions. The later might allow runaways but those should be really unlikely unless somebody misconfigures the system. E.g. allowing to migrate tasks away from the memcg to a different unlimited memcg with move_charge_at_immigrate disabled. Link: http://lkml.kernel.org/r/20180628151101.25307-1-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Acked-by: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Shakeel Butt <shakeelb@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:11 +00:00
if (mem_cgroup_out_of_memory(memcg, mask, order))
ret = OOM_SUCCESS;
else
ret = OOM_FAILED;
if (locked)
mem_cgroup_oom_unlock(memcg);
memcg, oom: move out_of_memory back to the charge path Commit 3812c8c8f395 ("mm: memcg: do not trap chargers with full callstack on OOM") has changed the ENOMEM semantic of memcg charges. Rather than invoking the oom killer from the charging context it delays the oom killer to the page fault path (pagefault_out_of_memory). This in turn means that many users (e.g. slab or g-u-p) will get ENOMEM when the corresponding memcg hits the hard limit and the memcg is is OOM. This is behavior is inconsistent with !memcg case where the oom killer is invoked from the allocation context and the allocator keeps retrying until it succeeds. The difference in the behavior is user visible. mmap(MAP_POPULATE) might result in not fully populated ranges while the mmap return code doesn't tell that to the userspace. Random syscalls might fail with ENOMEM etc. The primary motivation of the different memcg oom semantic was the deadlock avoidance. Things have changed since then, though. We have an async oom teardown by the oom reaper now and so we do not have to rely on the victim to tear down its memory anymore. Therefore we can return to the original semantic as long as the memcg oom killer is not handed over to the users space. There is still one thing to be careful about here though. If the oom killer is not able to make any forward progress - e.g. because there is no eligible task to kill - then we have to bail out of the charge path to prevent from same class of deadlocks. We have basically two options here. Either we fail the charge with ENOMEM or force the charge and allow overcharge. The first option has been considered more harmful than useful because rare inconsistencies in the ENOMEM behavior is hard to test for and error prone. Basically the same reason why the page allocator doesn't fail allocations under such conditions. The later might allow runaways but those should be really unlikely unless somebody misconfigures the system. E.g. allowing to migrate tasks away from the memcg to a different unlimited memcg with move_charge_at_immigrate disabled. Link: http://lkml.kernel.org/r/20180628151101.25307-1-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Acked-by: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Shakeel Butt <shakeelb@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:11 +00:00
return ret;
mm: memcg: do not trap chargers with full callstack on OOM The memcg OOM handling is incredibly fragile and can deadlock. When a task fails to charge memory, it invokes the OOM killer and loops right there in the charge code until it succeeds. Comparably, any other task that enters the charge path at this point will go to a waitqueue right then and there and sleep until the OOM situation is resolved. The problem is that these tasks may hold filesystem locks and the mmap_sem; locks that the selected OOM victim may need to exit. For example, in one reported case, the task invoking the OOM killer was about to charge a page cache page during a write(), which holds the i_mutex. The OOM killer selected a task that was just entering truncate() and trying to acquire the i_mutex: OOM invoking task: mem_cgroup_handle_oom+0x241/0x3b0 mem_cgroup_cache_charge+0xbe/0xe0 add_to_page_cache_locked+0x4c/0x140 add_to_page_cache_lru+0x22/0x50 grab_cache_page_write_begin+0x8b/0xe0 ext3_write_begin+0x88/0x270 generic_file_buffered_write+0x116/0x290 __generic_file_aio_write+0x27c/0x480 generic_file_aio_write+0x76/0xf0 # takes ->i_mutex do_sync_write+0xea/0x130 vfs_write+0xf3/0x1f0 sys_write+0x51/0x90 system_call_fastpath+0x18/0x1d OOM kill victim: do_truncate+0x58/0xa0 # takes i_mutex do_last+0x250/0xa30 path_openat+0xd7/0x440 do_filp_open+0x49/0xa0 do_sys_open+0x106/0x240 sys_open+0x20/0x30 system_call_fastpath+0x18/0x1d The OOM handling task will retry the charge indefinitely while the OOM killed task is not releasing any resources. A similar scenario can happen when the kernel OOM killer for a memcg is disabled and a userspace task is in charge of resolving OOM situations. In this case, ALL tasks that enter the OOM path will be made to sleep on the OOM waitqueue and wait for userspace to free resources or increase the group's limit. But a userspace OOM handler is prone to deadlock itself on the locks held by the waiting tasks. For example one of the sleeping tasks may be stuck in a brk() call with the mmap_sem held for writing but the userspace handler, in order to pick an optimal victim, may need to read files from /proc/<pid>, which tries to acquire the same mmap_sem for reading and deadlocks. This patch changes the way tasks behave after detecting a memcg OOM and makes sure nobody loops or sleeps with locks held: 1. When OOMing in a user fault, invoke the OOM killer and restart the fault instead of looping on the charge attempt. This way, the OOM victim can not get stuck on locks the looping task may hold. 2. When OOMing in a user fault but somebody else is handling it (either the kernel OOM killer or a userspace handler), don't go to sleep in the charge context. Instead, remember the OOMing memcg in the task struct and then fully unwind the page fault stack with -ENOMEM. pagefault_out_of_memory() will then call back into the memcg code to check if the -ENOMEM came from the memcg, and then either put the task to sleep on the memcg's OOM waitqueue or just restart the fault. The OOM victim can no longer get stuck on any lock a sleeping task may hold. Debugged by Michal Hocko. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: azurIt <azurit@pobox.sk> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:44 +00:00
}
/**
* mem_cgroup_oom_synchronize - complete memcg OOM handling
* @handle: actually kill/wait or just clean up the OOM state
mm: memcg: do not trap chargers with full callstack on OOM The memcg OOM handling is incredibly fragile and can deadlock. When a task fails to charge memory, it invokes the OOM killer and loops right there in the charge code until it succeeds. Comparably, any other task that enters the charge path at this point will go to a waitqueue right then and there and sleep until the OOM situation is resolved. The problem is that these tasks may hold filesystem locks and the mmap_sem; locks that the selected OOM victim may need to exit. For example, in one reported case, the task invoking the OOM killer was about to charge a page cache page during a write(), which holds the i_mutex. The OOM killer selected a task that was just entering truncate() and trying to acquire the i_mutex: OOM invoking task: mem_cgroup_handle_oom+0x241/0x3b0 mem_cgroup_cache_charge+0xbe/0xe0 add_to_page_cache_locked+0x4c/0x140 add_to_page_cache_lru+0x22/0x50 grab_cache_page_write_begin+0x8b/0xe0 ext3_write_begin+0x88/0x270 generic_file_buffered_write+0x116/0x290 __generic_file_aio_write+0x27c/0x480 generic_file_aio_write+0x76/0xf0 # takes ->i_mutex do_sync_write+0xea/0x130 vfs_write+0xf3/0x1f0 sys_write+0x51/0x90 system_call_fastpath+0x18/0x1d OOM kill victim: do_truncate+0x58/0xa0 # takes i_mutex do_last+0x250/0xa30 path_openat+0xd7/0x440 do_filp_open+0x49/0xa0 do_sys_open+0x106/0x240 sys_open+0x20/0x30 system_call_fastpath+0x18/0x1d The OOM handling task will retry the charge indefinitely while the OOM killed task is not releasing any resources. A similar scenario can happen when the kernel OOM killer for a memcg is disabled and a userspace task is in charge of resolving OOM situations. In this case, ALL tasks that enter the OOM path will be made to sleep on the OOM waitqueue and wait for userspace to free resources or increase the group's limit. But a userspace OOM handler is prone to deadlock itself on the locks held by the waiting tasks. For example one of the sleeping tasks may be stuck in a brk() call with the mmap_sem held for writing but the userspace handler, in order to pick an optimal victim, may need to read files from /proc/<pid>, which tries to acquire the same mmap_sem for reading and deadlocks. This patch changes the way tasks behave after detecting a memcg OOM and makes sure nobody loops or sleeps with locks held: 1. When OOMing in a user fault, invoke the OOM killer and restart the fault instead of looping on the charge attempt. This way, the OOM victim can not get stuck on locks the looping task may hold. 2. When OOMing in a user fault but somebody else is handling it (either the kernel OOM killer or a userspace handler), don't go to sleep in the charge context. Instead, remember the OOMing memcg in the task struct and then fully unwind the page fault stack with -ENOMEM. pagefault_out_of_memory() will then call back into the memcg code to check if the -ENOMEM came from the memcg, and then either put the task to sleep on the memcg's OOM waitqueue or just restart the fault. The OOM victim can no longer get stuck on any lock a sleeping task may hold. Debugged by Michal Hocko. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: azurIt <azurit@pobox.sk> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:44 +00:00
*
* This has to be called at the end of a page fault if the memcg OOM
* handler was enabled.
mm: memcg: do not trap chargers with full callstack on OOM The memcg OOM handling is incredibly fragile and can deadlock. When a task fails to charge memory, it invokes the OOM killer and loops right there in the charge code until it succeeds. Comparably, any other task that enters the charge path at this point will go to a waitqueue right then and there and sleep until the OOM situation is resolved. The problem is that these tasks may hold filesystem locks and the mmap_sem; locks that the selected OOM victim may need to exit. For example, in one reported case, the task invoking the OOM killer was about to charge a page cache page during a write(), which holds the i_mutex. The OOM killer selected a task that was just entering truncate() and trying to acquire the i_mutex: OOM invoking task: mem_cgroup_handle_oom+0x241/0x3b0 mem_cgroup_cache_charge+0xbe/0xe0 add_to_page_cache_locked+0x4c/0x140 add_to_page_cache_lru+0x22/0x50 grab_cache_page_write_begin+0x8b/0xe0 ext3_write_begin+0x88/0x270 generic_file_buffered_write+0x116/0x290 __generic_file_aio_write+0x27c/0x480 generic_file_aio_write+0x76/0xf0 # takes ->i_mutex do_sync_write+0xea/0x130 vfs_write+0xf3/0x1f0 sys_write+0x51/0x90 system_call_fastpath+0x18/0x1d OOM kill victim: do_truncate+0x58/0xa0 # takes i_mutex do_last+0x250/0xa30 path_openat+0xd7/0x440 do_filp_open+0x49/0xa0 do_sys_open+0x106/0x240 sys_open+0x20/0x30 system_call_fastpath+0x18/0x1d The OOM handling task will retry the charge indefinitely while the OOM killed task is not releasing any resources. A similar scenario can happen when the kernel OOM killer for a memcg is disabled and a userspace task is in charge of resolving OOM situations. In this case, ALL tasks that enter the OOM path will be made to sleep on the OOM waitqueue and wait for userspace to free resources or increase the group's limit. But a userspace OOM handler is prone to deadlock itself on the locks held by the waiting tasks. For example one of the sleeping tasks may be stuck in a brk() call with the mmap_sem held for writing but the userspace handler, in order to pick an optimal victim, may need to read files from /proc/<pid>, which tries to acquire the same mmap_sem for reading and deadlocks. This patch changes the way tasks behave after detecting a memcg OOM and makes sure nobody loops or sleeps with locks held: 1. When OOMing in a user fault, invoke the OOM killer and restart the fault instead of looping on the charge attempt. This way, the OOM victim can not get stuck on locks the looping task may hold. 2. When OOMing in a user fault but somebody else is handling it (either the kernel OOM killer or a userspace handler), don't go to sleep in the charge context. Instead, remember the OOMing memcg in the task struct and then fully unwind the page fault stack with -ENOMEM. pagefault_out_of_memory() will then call back into the memcg code to check if the -ENOMEM came from the memcg, and then either put the task to sleep on the memcg's OOM waitqueue or just restart the fault. The OOM victim can no longer get stuck on any lock a sleeping task may hold. Debugged by Michal Hocko. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: azurIt <azurit@pobox.sk> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:44 +00:00
*
* Memcg supports userspace OOM handling where failed allocations must
mm: memcg: do not trap chargers with full callstack on OOM The memcg OOM handling is incredibly fragile and can deadlock. When a task fails to charge memory, it invokes the OOM killer and loops right there in the charge code until it succeeds. Comparably, any other task that enters the charge path at this point will go to a waitqueue right then and there and sleep until the OOM situation is resolved. The problem is that these tasks may hold filesystem locks and the mmap_sem; locks that the selected OOM victim may need to exit. For example, in one reported case, the task invoking the OOM killer was about to charge a page cache page during a write(), which holds the i_mutex. The OOM killer selected a task that was just entering truncate() and trying to acquire the i_mutex: OOM invoking task: mem_cgroup_handle_oom+0x241/0x3b0 mem_cgroup_cache_charge+0xbe/0xe0 add_to_page_cache_locked+0x4c/0x140 add_to_page_cache_lru+0x22/0x50 grab_cache_page_write_begin+0x8b/0xe0 ext3_write_begin+0x88/0x270 generic_file_buffered_write+0x116/0x290 __generic_file_aio_write+0x27c/0x480 generic_file_aio_write+0x76/0xf0 # takes ->i_mutex do_sync_write+0xea/0x130 vfs_write+0xf3/0x1f0 sys_write+0x51/0x90 system_call_fastpath+0x18/0x1d OOM kill victim: do_truncate+0x58/0xa0 # takes i_mutex do_last+0x250/0xa30 path_openat+0xd7/0x440 do_filp_open+0x49/0xa0 do_sys_open+0x106/0x240 sys_open+0x20/0x30 system_call_fastpath+0x18/0x1d The OOM handling task will retry the charge indefinitely while the OOM killed task is not releasing any resources. A similar scenario can happen when the kernel OOM killer for a memcg is disabled and a userspace task is in charge of resolving OOM situations. In this case, ALL tasks that enter the OOM path will be made to sleep on the OOM waitqueue and wait for userspace to free resources or increase the group's limit. But a userspace OOM handler is prone to deadlock itself on the locks held by the waiting tasks. For example one of the sleeping tasks may be stuck in a brk() call with the mmap_sem held for writing but the userspace handler, in order to pick an optimal victim, may need to read files from /proc/<pid>, which tries to acquire the same mmap_sem for reading and deadlocks. This patch changes the way tasks behave after detecting a memcg OOM and makes sure nobody loops or sleeps with locks held: 1. When OOMing in a user fault, invoke the OOM killer and restart the fault instead of looping on the charge attempt. This way, the OOM victim can not get stuck on locks the looping task may hold. 2. When OOMing in a user fault but somebody else is handling it (either the kernel OOM killer or a userspace handler), don't go to sleep in the charge context. Instead, remember the OOMing memcg in the task struct and then fully unwind the page fault stack with -ENOMEM. pagefault_out_of_memory() will then call back into the memcg code to check if the -ENOMEM came from the memcg, and then either put the task to sleep on the memcg's OOM waitqueue or just restart the fault. The OOM victim can no longer get stuck on any lock a sleeping task may hold. Debugged by Michal Hocko. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: azurIt <azurit@pobox.sk> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:44 +00:00
* sleep on a waitqueue until the userspace task resolves the
* situation. Sleeping directly in the charge context with all kinds
* of locks held is not a good idea, instead we remember an OOM state
* in the task and mem_cgroup_oom_synchronize() has to be called at
* the end of the page fault to complete the OOM handling.
mm: memcg: do not trap chargers with full callstack on OOM The memcg OOM handling is incredibly fragile and can deadlock. When a task fails to charge memory, it invokes the OOM killer and loops right there in the charge code until it succeeds. Comparably, any other task that enters the charge path at this point will go to a waitqueue right then and there and sleep until the OOM situation is resolved. The problem is that these tasks may hold filesystem locks and the mmap_sem; locks that the selected OOM victim may need to exit. For example, in one reported case, the task invoking the OOM killer was about to charge a page cache page during a write(), which holds the i_mutex. The OOM killer selected a task that was just entering truncate() and trying to acquire the i_mutex: OOM invoking task: mem_cgroup_handle_oom+0x241/0x3b0 mem_cgroup_cache_charge+0xbe/0xe0 add_to_page_cache_locked+0x4c/0x140 add_to_page_cache_lru+0x22/0x50 grab_cache_page_write_begin+0x8b/0xe0 ext3_write_begin+0x88/0x270 generic_file_buffered_write+0x116/0x290 __generic_file_aio_write+0x27c/0x480 generic_file_aio_write+0x76/0xf0 # takes ->i_mutex do_sync_write+0xea/0x130 vfs_write+0xf3/0x1f0 sys_write+0x51/0x90 system_call_fastpath+0x18/0x1d OOM kill victim: do_truncate+0x58/0xa0 # takes i_mutex do_last+0x250/0xa30 path_openat+0xd7/0x440 do_filp_open+0x49/0xa0 do_sys_open+0x106/0x240 sys_open+0x20/0x30 system_call_fastpath+0x18/0x1d The OOM handling task will retry the charge indefinitely while the OOM killed task is not releasing any resources. A similar scenario can happen when the kernel OOM killer for a memcg is disabled and a userspace task is in charge of resolving OOM situations. In this case, ALL tasks that enter the OOM path will be made to sleep on the OOM waitqueue and wait for userspace to free resources or increase the group's limit. But a userspace OOM handler is prone to deadlock itself on the locks held by the waiting tasks. For example one of the sleeping tasks may be stuck in a brk() call with the mmap_sem held for writing but the userspace handler, in order to pick an optimal victim, may need to read files from /proc/<pid>, which tries to acquire the same mmap_sem for reading and deadlocks. This patch changes the way tasks behave after detecting a memcg OOM and makes sure nobody loops or sleeps with locks held: 1. When OOMing in a user fault, invoke the OOM killer and restart the fault instead of looping on the charge attempt. This way, the OOM victim can not get stuck on locks the looping task may hold. 2. When OOMing in a user fault but somebody else is handling it (either the kernel OOM killer or a userspace handler), don't go to sleep in the charge context. Instead, remember the OOMing memcg in the task struct and then fully unwind the page fault stack with -ENOMEM. pagefault_out_of_memory() will then call back into the memcg code to check if the -ENOMEM came from the memcg, and then either put the task to sleep on the memcg's OOM waitqueue or just restart the fault. The OOM victim can no longer get stuck on any lock a sleeping task may hold. Debugged by Michal Hocko. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: azurIt <azurit@pobox.sk> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:44 +00:00
*
* Returns %true if an ongoing memcg OOM situation was detected and
* completed, %false otherwise.
mm: memcg: do not trap chargers with full callstack on OOM The memcg OOM handling is incredibly fragile and can deadlock. When a task fails to charge memory, it invokes the OOM killer and loops right there in the charge code until it succeeds. Comparably, any other task that enters the charge path at this point will go to a waitqueue right then and there and sleep until the OOM situation is resolved. The problem is that these tasks may hold filesystem locks and the mmap_sem; locks that the selected OOM victim may need to exit. For example, in one reported case, the task invoking the OOM killer was about to charge a page cache page during a write(), which holds the i_mutex. The OOM killer selected a task that was just entering truncate() and trying to acquire the i_mutex: OOM invoking task: mem_cgroup_handle_oom+0x241/0x3b0 mem_cgroup_cache_charge+0xbe/0xe0 add_to_page_cache_locked+0x4c/0x140 add_to_page_cache_lru+0x22/0x50 grab_cache_page_write_begin+0x8b/0xe0 ext3_write_begin+0x88/0x270 generic_file_buffered_write+0x116/0x290 __generic_file_aio_write+0x27c/0x480 generic_file_aio_write+0x76/0xf0 # takes ->i_mutex do_sync_write+0xea/0x130 vfs_write+0xf3/0x1f0 sys_write+0x51/0x90 system_call_fastpath+0x18/0x1d OOM kill victim: do_truncate+0x58/0xa0 # takes i_mutex do_last+0x250/0xa30 path_openat+0xd7/0x440 do_filp_open+0x49/0xa0 do_sys_open+0x106/0x240 sys_open+0x20/0x30 system_call_fastpath+0x18/0x1d The OOM handling task will retry the charge indefinitely while the OOM killed task is not releasing any resources. A similar scenario can happen when the kernel OOM killer for a memcg is disabled and a userspace task is in charge of resolving OOM situations. In this case, ALL tasks that enter the OOM path will be made to sleep on the OOM waitqueue and wait for userspace to free resources or increase the group's limit. But a userspace OOM handler is prone to deadlock itself on the locks held by the waiting tasks. For example one of the sleeping tasks may be stuck in a brk() call with the mmap_sem held for writing but the userspace handler, in order to pick an optimal victim, may need to read files from /proc/<pid>, which tries to acquire the same mmap_sem for reading and deadlocks. This patch changes the way tasks behave after detecting a memcg OOM and makes sure nobody loops or sleeps with locks held: 1. When OOMing in a user fault, invoke the OOM killer and restart the fault instead of looping on the charge attempt. This way, the OOM victim can not get stuck on locks the looping task may hold. 2. When OOMing in a user fault but somebody else is handling it (either the kernel OOM killer or a userspace handler), don't go to sleep in the charge context. Instead, remember the OOMing memcg in the task struct and then fully unwind the page fault stack with -ENOMEM. pagefault_out_of_memory() will then call back into the memcg code to check if the -ENOMEM came from the memcg, and then either put the task to sleep on the memcg's OOM waitqueue or just restart the fault. The OOM victim can no longer get stuck on any lock a sleeping task may hold. Debugged by Michal Hocko. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: azurIt <azurit@pobox.sk> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:44 +00:00
*/
bool mem_cgroup_oom_synchronize(bool handle)
mm: memcg: do not trap chargers with full callstack on OOM The memcg OOM handling is incredibly fragile and can deadlock. When a task fails to charge memory, it invokes the OOM killer and loops right there in the charge code until it succeeds. Comparably, any other task that enters the charge path at this point will go to a waitqueue right then and there and sleep until the OOM situation is resolved. The problem is that these tasks may hold filesystem locks and the mmap_sem; locks that the selected OOM victim may need to exit. For example, in one reported case, the task invoking the OOM killer was about to charge a page cache page during a write(), which holds the i_mutex. The OOM killer selected a task that was just entering truncate() and trying to acquire the i_mutex: OOM invoking task: mem_cgroup_handle_oom+0x241/0x3b0 mem_cgroup_cache_charge+0xbe/0xe0 add_to_page_cache_locked+0x4c/0x140 add_to_page_cache_lru+0x22/0x50 grab_cache_page_write_begin+0x8b/0xe0 ext3_write_begin+0x88/0x270 generic_file_buffered_write+0x116/0x290 __generic_file_aio_write+0x27c/0x480 generic_file_aio_write+0x76/0xf0 # takes ->i_mutex do_sync_write+0xea/0x130 vfs_write+0xf3/0x1f0 sys_write+0x51/0x90 system_call_fastpath+0x18/0x1d OOM kill victim: do_truncate+0x58/0xa0 # takes i_mutex do_last+0x250/0xa30 path_openat+0xd7/0x440 do_filp_open+0x49/0xa0 do_sys_open+0x106/0x240 sys_open+0x20/0x30 system_call_fastpath+0x18/0x1d The OOM handling task will retry the charge indefinitely while the OOM killed task is not releasing any resources. A similar scenario can happen when the kernel OOM killer for a memcg is disabled and a userspace task is in charge of resolving OOM situations. In this case, ALL tasks that enter the OOM path will be made to sleep on the OOM waitqueue and wait for userspace to free resources or increase the group's limit. But a userspace OOM handler is prone to deadlock itself on the locks held by the waiting tasks. For example one of the sleeping tasks may be stuck in a brk() call with the mmap_sem held for writing but the userspace handler, in order to pick an optimal victim, may need to read files from /proc/<pid>, which tries to acquire the same mmap_sem for reading and deadlocks. This patch changes the way tasks behave after detecting a memcg OOM and makes sure nobody loops or sleeps with locks held: 1. When OOMing in a user fault, invoke the OOM killer and restart the fault instead of looping on the charge attempt. This way, the OOM victim can not get stuck on locks the looping task may hold. 2. When OOMing in a user fault but somebody else is handling it (either the kernel OOM killer or a userspace handler), don't go to sleep in the charge context. Instead, remember the OOMing memcg in the task struct and then fully unwind the page fault stack with -ENOMEM. pagefault_out_of_memory() will then call back into the memcg code to check if the -ENOMEM came from the memcg, and then either put the task to sleep on the memcg's OOM waitqueue or just restart the fault. The OOM victim can no longer get stuck on any lock a sleeping task may hold. Debugged by Michal Hocko. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: azurIt <azurit@pobox.sk> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:44 +00:00
{
struct mem_cgroup *memcg = current->memcg_in_oom;
mm: memcg: do not trap chargers with full callstack on OOM The memcg OOM handling is incredibly fragile and can deadlock. When a task fails to charge memory, it invokes the OOM killer and loops right there in the charge code until it succeeds. Comparably, any other task that enters the charge path at this point will go to a waitqueue right then and there and sleep until the OOM situation is resolved. The problem is that these tasks may hold filesystem locks and the mmap_sem; locks that the selected OOM victim may need to exit. For example, in one reported case, the task invoking the OOM killer was about to charge a page cache page during a write(), which holds the i_mutex. The OOM killer selected a task that was just entering truncate() and trying to acquire the i_mutex: OOM invoking task: mem_cgroup_handle_oom+0x241/0x3b0 mem_cgroup_cache_charge+0xbe/0xe0 add_to_page_cache_locked+0x4c/0x140 add_to_page_cache_lru+0x22/0x50 grab_cache_page_write_begin+0x8b/0xe0 ext3_write_begin+0x88/0x270 generic_file_buffered_write+0x116/0x290 __generic_file_aio_write+0x27c/0x480 generic_file_aio_write+0x76/0xf0 # takes ->i_mutex do_sync_write+0xea/0x130 vfs_write+0xf3/0x1f0 sys_write+0x51/0x90 system_call_fastpath+0x18/0x1d OOM kill victim: do_truncate+0x58/0xa0 # takes i_mutex do_last+0x250/0xa30 path_openat+0xd7/0x440 do_filp_open+0x49/0xa0 do_sys_open+0x106/0x240 sys_open+0x20/0x30 system_call_fastpath+0x18/0x1d The OOM handling task will retry the charge indefinitely while the OOM killed task is not releasing any resources. A similar scenario can happen when the kernel OOM killer for a memcg is disabled and a userspace task is in charge of resolving OOM situations. In this case, ALL tasks that enter the OOM path will be made to sleep on the OOM waitqueue and wait for userspace to free resources or increase the group's limit. But a userspace OOM handler is prone to deadlock itself on the locks held by the waiting tasks. For example one of the sleeping tasks may be stuck in a brk() call with the mmap_sem held for writing but the userspace handler, in order to pick an optimal victim, may need to read files from /proc/<pid>, which tries to acquire the same mmap_sem for reading and deadlocks. This patch changes the way tasks behave after detecting a memcg OOM and makes sure nobody loops or sleeps with locks held: 1. When OOMing in a user fault, invoke the OOM killer and restart the fault instead of looping on the charge attempt. This way, the OOM victim can not get stuck on locks the looping task may hold. 2. When OOMing in a user fault but somebody else is handling it (either the kernel OOM killer or a userspace handler), don't go to sleep in the charge context. Instead, remember the OOMing memcg in the task struct and then fully unwind the page fault stack with -ENOMEM. pagefault_out_of_memory() will then call back into the memcg code to check if the -ENOMEM came from the memcg, and then either put the task to sleep on the memcg's OOM waitqueue or just restart the fault. The OOM victim can no longer get stuck on any lock a sleeping task may hold. Debugged by Michal Hocko. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: azurIt <azurit@pobox.sk> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:44 +00:00
struct oom_wait_info owait;
bool locked;
mm: memcg: do not trap chargers with full callstack on OOM The memcg OOM handling is incredibly fragile and can deadlock. When a task fails to charge memory, it invokes the OOM killer and loops right there in the charge code until it succeeds. Comparably, any other task that enters the charge path at this point will go to a waitqueue right then and there and sleep until the OOM situation is resolved. The problem is that these tasks may hold filesystem locks and the mmap_sem; locks that the selected OOM victim may need to exit. For example, in one reported case, the task invoking the OOM killer was about to charge a page cache page during a write(), which holds the i_mutex. The OOM killer selected a task that was just entering truncate() and trying to acquire the i_mutex: OOM invoking task: mem_cgroup_handle_oom+0x241/0x3b0 mem_cgroup_cache_charge+0xbe/0xe0 add_to_page_cache_locked+0x4c/0x140 add_to_page_cache_lru+0x22/0x50 grab_cache_page_write_begin+0x8b/0xe0 ext3_write_begin+0x88/0x270 generic_file_buffered_write+0x116/0x290 __generic_file_aio_write+0x27c/0x480 generic_file_aio_write+0x76/0xf0 # takes ->i_mutex do_sync_write+0xea/0x130 vfs_write+0xf3/0x1f0 sys_write+0x51/0x90 system_call_fastpath+0x18/0x1d OOM kill victim: do_truncate+0x58/0xa0 # takes i_mutex do_last+0x250/0xa30 path_openat+0xd7/0x440 do_filp_open+0x49/0xa0 do_sys_open+0x106/0x240 sys_open+0x20/0x30 system_call_fastpath+0x18/0x1d The OOM handling task will retry the charge indefinitely while the OOM killed task is not releasing any resources. A similar scenario can happen when the kernel OOM killer for a memcg is disabled and a userspace task is in charge of resolving OOM situations. In this case, ALL tasks that enter the OOM path will be made to sleep on the OOM waitqueue and wait for userspace to free resources or increase the group's limit. But a userspace OOM handler is prone to deadlock itself on the locks held by the waiting tasks. For example one of the sleeping tasks may be stuck in a brk() call with the mmap_sem held for writing but the userspace handler, in order to pick an optimal victim, may need to read files from /proc/<pid>, which tries to acquire the same mmap_sem for reading and deadlocks. This patch changes the way tasks behave after detecting a memcg OOM and makes sure nobody loops or sleeps with locks held: 1. When OOMing in a user fault, invoke the OOM killer and restart the fault instead of looping on the charge attempt. This way, the OOM victim can not get stuck on locks the looping task may hold. 2. When OOMing in a user fault but somebody else is handling it (either the kernel OOM killer or a userspace handler), don't go to sleep in the charge context. Instead, remember the OOMing memcg in the task struct and then fully unwind the page fault stack with -ENOMEM. pagefault_out_of_memory() will then call back into the memcg code to check if the -ENOMEM came from the memcg, and then either put the task to sleep on the memcg's OOM waitqueue or just restart the fault. The OOM victim can no longer get stuck on any lock a sleeping task may hold. Debugged by Michal Hocko. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: azurIt <azurit@pobox.sk> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:44 +00:00
/* OOM is global, do not handle */
if (!memcg)
return false;
mm: memcg: do not trap chargers with full callstack on OOM The memcg OOM handling is incredibly fragile and can deadlock. When a task fails to charge memory, it invokes the OOM killer and loops right there in the charge code until it succeeds. Comparably, any other task that enters the charge path at this point will go to a waitqueue right then and there and sleep until the OOM situation is resolved. The problem is that these tasks may hold filesystem locks and the mmap_sem; locks that the selected OOM victim may need to exit. For example, in one reported case, the task invoking the OOM killer was about to charge a page cache page during a write(), which holds the i_mutex. The OOM killer selected a task that was just entering truncate() and trying to acquire the i_mutex: OOM invoking task: mem_cgroup_handle_oom+0x241/0x3b0 mem_cgroup_cache_charge+0xbe/0xe0 add_to_page_cache_locked+0x4c/0x140 add_to_page_cache_lru+0x22/0x50 grab_cache_page_write_begin+0x8b/0xe0 ext3_write_begin+0x88/0x270 generic_file_buffered_write+0x116/0x290 __generic_file_aio_write+0x27c/0x480 generic_file_aio_write+0x76/0xf0 # takes ->i_mutex do_sync_write+0xea/0x130 vfs_write+0xf3/0x1f0 sys_write+0x51/0x90 system_call_fastpath+0x18/0x1d OOM kill victim: do_truncate+0x58/0xa0 # takes i_mutex do_last+0x250/0xa30 path_openat+0xd7/0x440 do_filp_open+0x49/0xa0 do_sys_open+0x106/0x240 sys_open+0x20/0x30 system_call_fastpath+0x18/0x1d The OOM handling task will retry the charge indefinitely while the OOM killed task is not releasing any resources. A similar scenario can happen when the kernel OOM killer for a memcg is disabled and a userspace task is in charge of resolving OOM situations. In this case, ALL tasks that enter the OOM path will be made to sleep on the OOM waitqueue and wait for userspace to free resources or increase the group's limit. But a userspace OOM handler is prone to deadlock itself on the locks held by the waiting tasks. For example one of the sleeping tasks may be stuck in a brk() call with the mmap_sem held for writing but the userspace handler, in order to pick an optimal victim, may need to read files from /proc/<pid>, which tries to acquire the same mmap_sem for reading and deadlocks. This patch changes the way tasks behave after detecting a memcg OOM and makes sure nobody loops or sleeps with locks held: 1. When OOMing in a user fault, invoke the OOM killer and restart the fault instead of looping on the charge attempt. This way, the OOM victim can not get stuck on locks the looping task may hold. 2. When OOMing in a user fault but somebody else is handling it (either the kernel OOM killer or a userspace handler), don't go to sleep in the charge context. Instead, remember the OOMing memcg in the task struct and then fully unwind the page fault stack with -ENOMEM. pagefault_out_of_memory() will then call back into the memcg code to check if the -ENOMEM came from the memcg, and then either put the task to sleep on the memcg's OOM waitqueue or just restart the fault. The OOM victim can no longer get stuck on any lock a sleeping task may hold. Debugged by Michal Hocko. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: azurIt <azurit@pobox.sk> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:44 +00:00
if (!handle)
goto cleanup;
mm: memcg: do not trap chargers with full callstack on OOM The memcg OOM handling is incredibly fragile and can deadlock. When a task fails to charge memory, it invokes the OOM killer and loops right there in the charge code until it succeeds. Comparably, any other task that enters the charge path at this point will go to a waitqueue right then and there and sleep until the OOM situation is resolved. The problem is that these tasks may hold filesystem locks and the mmap_sem; locks that the selected OOM victim may need to exit. For example, in one reported case, the task invoking the OOM killer was about to charge a page cache page during a write(), which holds the i_mutex. The OOM killer selected a task that was just entering truncate() and trying to acquire the i_mutex: OOM invoking task: mem_cgroup_handle_oom+0x241/0x3b0 mem_cgroup_cache_charge+0xbe/0xe0 add_to_page_cache_locked+0x4c/0x140 add_to_page_cache_lru+0x22/0x50 grab_cache_page_write_begin+0x8b/0xe0 ext3_write_begin+0x88/0x270 generic_file_buffered_write+0x116/0x290 __generic_file_aio_write+0x27c/0x480 generic_file_aio_write+0x76/0xf0 # takes ->i_mutex do_sync_write+0xea/0x130 vfs_write+0xf3/0x1f0 sys_write+0x51/0x90 system_call_fastpath+0x18/0x1d OOM kill victim: do_truncate+0x58/0xa0 # takes i_mutex do_last+0x250/0xa30 path_openat+0xd7/0x440 do_filp_open+0x49/0xa0 do_sys_open+0x106/0x240 sys_open+0x20/0x30 system_call_fastpath+0x18/0x1d The OOM handling task will retry the charge indefinitely while the OOM killed task is not releasing any resources. A similar scenario can happen when the kernel OOM killer for a memcg is disabled and a userspace task is in charge of resolving OOM situations. In this case, ALL tasks that enter the OOM path will be made to sleep on the OOM waitqueue and wait for userspace to free resources or increase the group's limit. But a userspace OOM handler is prone to deadlock itself on the locks held by the waiting tasks. For example one of the sleeping tasks may be stuck in a brk() call with the mmap_sem held for writing but the userspace handler, in order to pick an optimal victim, may need to read files from /proc/<pid>, which tries to acquire the same mmap_sem for reading and deadlocks. This patch changes the way tasks behave after detecting a memcg OOM and makes sure nobody loops or sleeps with locks held: 1. When OOMing in a user fault, invoke the OOM killer and restart the fault instead of looping on the charge attempt. This way, the OOM victim can not get stuck on locks the looping task may hold. 2. When OOMing in a user fault but somebody else is handling it (either the kernel OOM killer or a userspace handler), don't go to sleep in the charge context. Instead, remember the OOMing memcg in the task struct and then fully unwind the page fault stack with -ENOMEM. pagefault_out_of_memory() will then call back into the memcg code to check if the -ENOMEM came from the memcg, and then either put the task to sleep on the memcg's OOM waitqueue or just restart the fault. The OOM victim can no longer get stuck on any lock a sleeping task may hold. Debugged by Michal Hocko. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: azurIt <azurit@pobox.sk> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:44 +00:00
owait.memcg = memcg;
owait.wait.flags = 0;
owait.wait.func = memcg_oom_wake_function;
owait.wait.private = current;
sched/wait: Disambiguate wq_entry->task_list and wq_head->task_list naming So I've noticed a number of instances where it was not obvious from the code whether ->task_list was for a wait-queue head or a wait-queue entry. Furthermore, there's a number of wait-queue users where the lists are not for 'tasks' but other entities (poll tables, etc.), in which case the 'task_list' name is actively confusing. To clear this all up, name the wait-queue head and entry list structure fields unambiguously: struct wait_queue_head::task_list => ::head struct wait_queue_entry::task_list => ::entry For example, this code: rqw->wait.task_list.next != &wait->task_list ... is was pretty unclear (to me) what it's doing, while now it's written this way: rqw->wait.head.next != &wait->entry ... which makes it pretty clear that we are iterating a list until we see the head. Other examples are: list_for_each_entry_safe(pos, next, &x->task_list, task_list) { list_for_each_entry(wq, &fence->wait.task_list, task_list) { ... where it's unclear (to me) what we are iterating, and during review it's hard to tell whether it's trying to walk a wait-queue entry (which would be a bug), while now it's written as: list_for_each_entry_safe(pos, next, &x->head, entry) { list_for_each_entry(wq, &fence->wait.head, entry) { Cc: Linus Torvalds <torvalds@linux-foundation.org> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: linux-kernel@vger.kernel.org Signed-off-by: Ingo Molnar <mingo@kernel.org>
2017-06-20 10:06:46 +00:00
INIT_LIST_HEAD(&owait.wait.entry);
memcg: fix oom kill behavior In current page-fault code, handle_mm_fault() -> ... -> mem_cgroup_charge() -> map page or handle error. -> check return code. If page fault's return code is VM_FAULT_OOM, page_fault_out_of_memory() is called. But if it's caused by memcg, OOM should have been already invoked. Then, I added a patch: a636b327f731143ccc544b966cfd8de6cb6d72c6. That patch records last_oom_jiffies for memcg's sub-hierarchy and prevents page_fault_out_of_memory from being invoked in near future. But Nishimura-san reported that check by jiffies is not enough when the system is terribly heavy. This patch changes memcg's oom logic as. * If memcg causes OOM-kill, continue to retry. * remove jiffies check which is used now. * add memcg-oom-lock which works like perzone oom lock. * If current is killed(as a process), bypass charge. Something more sophisticated can be added but this pactch does fundamental things. TODO: - add oom notifier - add permemcg disable-oom-kill flag and freezer at oom. - more chances for wake up oom waiter (when changing memory limit etc..) Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Tested-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-03-10 23:22:39 +00:00
mm: memcg: do not trap chargers with full callstack on OOM The memcg OOM handling is incredibly fragile and can deadlock. When a task fails to charge memory, it invokes the OOM killer and loops right there in the charge code until it succeeds. Comparably, any other task that enters the charge path at this point will go to a waitqueue right then and there and sleep until the OOM situation is resolved. The problem is that these tasks may hold filesystem locks and the mmap_sem; locks that the selected OOM victim may need to exit. For example, in one reported case, the task invoking the OOM killer was about to charge a page cache page during a write(), which holds the i_mutex. The OOM killer selected a task that was just entering truncate() and trying to acquire the i_mutex: OOM invoking task: mem_cgroup_handle_oom+0x241/0x3b0 mem_cgroup_cache_charge+0xbe/0xe0 add_to_page_cache_locked+0x4c/0x140 add_to_page_cache_lru+0x22/0x50 grab_cache_page_write_begin+0x8b/0xe0 ext3_write_begin+0x88/0x270 generic_file_buffered_write+0x116/0x290 __generic_file_aio_write+0x27c/0x480 generic_file_aio_write+0x76/0xf0 # takes ->i_mutex do_sync_write+0xea/0x130 vfs_write+0xf3/0x1f0 sys_write+0x51/0x90 system_call_fastpath+0x18/0x1d OOM kill victim: do_truncate+0x58/0xa0 # takes i_mutex do_last+0x250/0xa30 path_openat+0xd7/0x440 do_filp_open+0x49/0xa0 do_sys_open+0x106/0x240 sys_open+0x20/0x30 system_call_fastpath+0x18/0x1d The OOM handling task will retry the charge indefinitely while the OOM killed task is not releasing any resources. A similar scenario can happen when the kernel OOM killer for a memcg is disabled and a userspace task is in charge of resolving OOM situations. In this case, ALL tasks that enter the OOM path will be made to sleep on the OOM waitqueue and wait for userspace to free resources or increase the group's limit. But a userspace OOM handler is prone to deadlock itself on the locks held by the waiting tasks. For example one of the sleeping tasks may be stuck in a brk() call with the mmap_sem held for writing but the userspace handler, in order to pick an optimal victim, may need to read files from /proc/<pid>, which tries to acquire the same mmap_sem for reading and deadlocks. This patch changes the way tasks behave after detecting a memcg OOM and makes sure nobody loops or sleeps with locks held: 1. When OOMing in a user fault, invoke the OOM killer and restart the fault instead of looping on the charge attempt. This way, the OOM victim can not get stuck on locks the looping task may hold. 2. When OOMing in a user fault but somebody else is handling it (either the kernel OOM killer or a userspace handler), don't go to sleep in the charge context. Instead, remember the OOMing memcg in the task struct and then fully unwind the page fault stack with -ENOMEM. pagefault_out_of_memory() will then call back into the memcg code to check if the -ENOMEM came from the memcg, and then either put the task to sleep on the memcg's OOM waitqueue or just restart the fault. The OOM victim can no longer get stuck on any lock a sleeping task may hold. Debugged by Michal Hocko. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: azurIt <azurit@pobox.sk> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:44 +00:00
prepare_to_wait(&memcg_oom_waitq, &owait.wait, TASK_KILLABLE);
mem_cgroup_mark_under_oom(memcg);
locked = mem_cgroup_oom_trylock(memcg);
if (locked)
mem_cgroup_oom_notify(memcg);
if (locked && !memcg->oom_kill_disable) {
mem_cgroup_unmark_under_oom(memcg);
finish_wait(&memcg_oom_waitq, &owait.wait);
mem_cgroup_out_of_memory(memcg, current->memcg_oom_gfp_mask,
current->memcg_oom_order);
} else {
mm: memcg: do not trap chargers with full callstack on OOM The memcg OOM handling is incredibly fragile and can deadlock. When a task fails to charge memory, it invokes the OOM killer and loops right there in the charge code until it succeeds. Comparably, any other task that enters the charge path at this point will go to a waitqueue right then and there and sleep until the OOM situation is resolved. The problem is that these tasks may hold filesystem locks and the mmap_sem; locks that the selected OOM victim may need to exit. For example, in one reported case, the task invoking the OOM killer was about to charge a page cache page during a write(), which holds the i_mutex. The OOM killer selected a task that was just entering truncate() and trying to acquire the i_mutex: OOM invoking task: mem_cgroup_handle_oom+0x241/0x3b0 mem_cgroup_cache_charge+0xbe/0xe0 add_to_page_cache_locked+0x4c/0x140 add_to_page_cache_lru+0x22/0x50 grab_cache_page_write_begin+0x8b/0xe0 ext3_write_begin+0x88/0x270 generic_file_buffered_write+0x116/0x290 __generic_file_aio_write+0x27c/0x480 generic_file_aio_write+0x76/0xf0 # takes ->i_mutex do_sync_write+0xea/0x130 vfs_write+0xf3/0x1f0 sys_write+0x51/0x90 system_call_fastpath+0x18/0x1d OOM kill victim: do_truncate+0x58/0xa0 # takes i_mutex do_last+0x250/0xa30 path_openat+0xd7/0x440 do_filp_open+0x49/0xa0 do_sys_open+0x106/0x240 sys_open+0x20/0x30 system_call_fastpath+0x18/0x1d The OOM handling task will retry the charge indefinitely while the OOM killed task is not releasing any resources. A similar scenario can happen when the kernel OOM killer for a memcg is disabled and a userspace task is in charge of resolving OOM situations. In this case, ALL tasks that enter the OOM path will be made to sleep on the OOM waitqueue and wait for userspace to free resources or increase the group's limit. But a userspace OOM handler is prone to deadlock itself on the locks held by the waiting tasks. For example one of the sleeping tasks may be stuck in a brk() call with the mmap_sem held for writing but the userspace handler, in order to pick an optimal victim, may need to read files from /proc/<pid>, which tries to acquire the same mmap_sem for reading and deadlocks. This patch changes the way tasks behave after detecting a memcg OOM and makes sure nobody loops or sleeps with locks held: 1. When OOMing in a user fault, invoke the OOM killer and restart the fault instead of looping on the charge attempt. This way, the OOM victim can not get stuck on locks the looping task may hold. 2. When OOMing in a user fault but somebody else is handling it (either the kernel OOM killer or a userspace handler), don't go to sleep in the charge context. Instead, remember the OOMing memcg in the task struct and then fully unwind the page fault stack with -ENOMEM. pagefault_out_of_memory() will then call back into the memcg code to check if the -ENOMEM came from the memcg, and then either put the task to sleep on the memcg's OOM waitqueue or just restart the fault. The OOM victim can no longer get stuck on any lock a sleeping task may hold. Debugged by Michal Hocko. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: azurIt <azurit@pobox.sk> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:44 +00:00
schedule();
mem_cgroup_unmark_under_oom(memcg);
finish_wait(&memcg_oom_waitq, &owait.wait);
}
if (locked) {
mm: memcg: rework and document OOM waiting and wakeup The memcg OOM handler open-codes a sleeping lock for OOM serialization (trylock, wait, repeat) because the required locking is so specific to memcg hierarchies. However, it would be nice if this construct would be clearly recognizable and not be as obfuscated as it is right now. Clean up as follows: 1. Remove the return value of mem_cgroup_oom_unlock() 2. Rename mem_cgroup_oom_lock() to mem_cgroup_oom_trylock(). 3. Pull the prepare_to_wait() out of the memcg_oom_lock scope. This makes it more obvious that the task has to be on the waitqueue before attempting to OOM-trylock the hierarchy, to not miss any wakeups before going to sleep. It just didn't matter until now because it was all lumped together into the global memcg_oom_lock spinlock section. 4. Pull the mem_cgroup_oom_notify() out of the memcg_oom_lock scope. It is proctected by the hierarchical OOM-lock. 5. The memcg_oom_lock spinlock is only required to propagate the OOM lock in any given hierarchy atomically. Restrict its scope to mem_cgroup_oom_(trylock|unlock). 6. Do not wake up the waitqueue unconditionally at the end of the function. Only the lockholder has to wake up the next in line after releasing the lock. Note that the lockholder kicks off the OOM-killer, which in turn leads to wakeups from the uncharges of the exiting task. But a contender is not guaranteed to see them if it enters the OOM path after the OOM kills but before the lockholder releases the lock. Thus there has to be an explicit wakeup after releasing the lock. 7. Put the OOM task on the waitqueue before marking the hierarchy as under OOM as that is the point where we start to receive wakeups. No point in listening before being on the waitqueue. 8. Likewise, unmark the hierarchy before finishing the sleep, for symmetry. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: azurIt <azurit@pobox.sk> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:43 +00:00
mem_cgroup_oom_unlock(memcg);
/*
* There is no guarantee that an OOM-lock contender
* sees the wakeups triggered by the OOM kill
* uncharges. Wake any sleepers explicitely.
*/
memcg_oom_recover(memcg);
}
cleanup:
current->memcg_in_oom = NULL;
mm: memcg: do not trap chargers with full callstack on OOM The memcg OOM handling is incredibly fragile and can deadlock. When a task fails to charge memory, it invokes the OOM killer and loops right there in the charge code until it succeeds. Comparably, any other task that enters the charge path at this point will go to a waitqueue right then and there and sleep until the OOM situation is resolved. The problem is that these tasks may hold filesystem locks and the mmap_sem; locks that the selected OOM victim may need to exit. For example, in one reported case, the task invoking the OOM killer was about to charge a page cache page during a write(), which holds the i_mutex. The OOM killer selected a task that was just entering truncate() and trying to acquire the i_mutex: OOM invoking task: mem_cgroup_handle_oom+0x241/0x3b0 mem_cgroup_cache_charge+0xbe/0xe0 add_to_page_cache_locked+0x4c/0x140 add_to_page_cache_lru+0x22/0x50 grab_cache_page_write_begin+0x8b/0xe0 ext3_write_begin+0x88/0x270 generic_file_buffered_write+0x116/0x290 __generic_file_aio_write+0x27c/0x480 generic_file_aio_write+0x76/0xf0 # takes ->i_mutex do_sync_write+0xea/0x130 vfs_write+0xf3/0x1f0 sys_write+0x51/0x90 system_call_fastpath+0x18/0x1d OOM kill victim: do_truncate+0x58/0xa0 # takes i_mutex do_last+0x250/0xa30 path_openat+0xd7/0x440 do_filp_open+0x49/0xa0 do_sys_open+0x106/0x240 sys_open+0x20/0x30 system_call_fastpath+0x18/0x1d The OOM handling task will retry the charge indefinitely while the OOM killed task is not releasing any resources. A similar scenario can happen when the kernel OOM killer for a memcg is disabled and a userspace task is in charge of resolving OOM situations. In this case, ALL tasks that enter the OOM path will be made to sleep on the OOM waitqueue and wait for userspace to free resources or increase the group's limit. But a userspace OOM handler is prone to deadlock itself on the locks held by the waiting tasks. For example one of the sleeping tasks may be stuck in a brk() call with the mmap_sem held for writing but the userspace handler, in order to pick an optimal victim, may need to read files from /proc/<pid>, which tries to acquire the same mmap_sem for reading and deadlocks. This patch changes the way tasks behave after detecting a memcg OOM and makes sure nobody loops or sleeps with locks held: 1. When OOMing in a user fault, invoke the OOM killer and restart the fault instead of looping on the charge attempt. This way, the OOM victim can not get stuck on locks the looping task may hold. 2. When OOMing in a user fault but somebody else is handling it (either the kernel OOM killer or a userspace handler), don't go to sleep in the charge context. Instead, remember the OOMing memcg in the task struct and then fully unwind the page fault stack with -ENOMEM. pagefault_out_of_memory() will then call back into the memcg code to check if the -ENOMEM came from the memcg, and then either put the task to sleep on the memcg's OOM waitqueue or just restart the fault. The OOM victim can no longer get stuck on any lock a sleeping task may hold. Debugged by Michal Hocko. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: azurIt <azurit@pobox.sk> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: David Rientjes <rientjes@google.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-12 22:13:44 +00:00
css_put(&memcg->css);
memcg: fix oom kill behavior In current page-fault code, handle_mm_fault() -> ... -> mem_cgroup_charge() -> map page or handle error. -> check return code. If page fault's return code is VM_FAULT_OOM, page_fault_out_of_memory() is called. But if it's caused by memcg, OOM should have been already invoked. Then, I added a patch: a636b327f731143ccc544b966cfd8de6cb6d72c6. That patch records last_oom_jiffies for memcg's sub-hierarchy and prevents page_fault_out_of_memory from being invoked in near future. But Nishimura-san reported that check by jiffies is not enough when the system is terribly heavy. This patch changes memcg's oom logic as. * If memcg causes OOM-kill, continue to retry. * remove jiffies check which is used now. * add memcg-oom-lock which works like perzone oom lock. * If current is killed(as a process), bypass charge. Something more sophisticated can be added but this pactch does fundamental things. TODO: - add oom notifier - add permemcg disable-oom-kill flag and freezer at oom. - more chances for wake up oom waiter (when changing memory limit etc..) Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Tested-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-03-10 23:22:39 +00:00
return true;
}
mm, oom: introduce memory.oom.group For some workloads an intervention from the OOM killer can be painful. Killing a random task can bring the workload into an inconsistent state. Historically, there are two common solutions for this problem: 1) enabling panic_on_oom, 2) using a userspace daemon to monitor OOMs and kill all outstanding processes. Both approaches have their downsides: rebooting on each OOM is an obvious waste of capacity, and handling all in userspace is tricky and requires a userspace agent, which will monitor all cgroups for OOMs. In most cases an in-kernel after-OOM cleaning-up mechanism can eliminate the necessity of enabling panic_on_oom. Also, it can simplify the cgroup management for userspace applications. This commit introduces a new knob for cgroup v2 memory controller: memory.oom.group. The knob determines whether the cgroup should be treated as an indivisible workload by the OOM killer. If set, all tasks belonging to the cgroup or to its descendants (if the memory cgroup is not a leaf cgroup) are killed together or not at all. To determine which cgroup has to be killed, we do traverse the cgroup hierarchy from the victim task's cgroup up to the OOMing cgroup (or root) and looking for the highest-level cgroup with memory.oom.group set. Tasks with the OOM protection (oom_score_adj set to -1000) are treated as an exception and are never killed. This patch doesn't change the OOM victim selection algorithm. Link: http://lkml.kernel.org/r/20180802003201.817-4-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: David Rientjes <rientjes@google.com> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-22 04:53:54 +00:00
/**
* mem_cgroup_get_oom_group - get a memory cgroup to clean up after OOM
* @victim: task to be killed by the OOM killer
* @oom_domain: memcg in case of memcg OOM, NULL in case of system-wide OOM
*
* Returns a pointer to a memory cgroup, which has to be cleaned up
* by killing all belonging OOM-killable tasks.
*
* Caller has to call mem_cgroup_put() on the returned non-NULL memcg.
*/
struct mem_cgroup *mem_cgroup_get_oom_group(struct task_struct *victim,
struct mem_cgroup *oom_domain)
{
struct mem_cgroup *oom_group = NULL;
struct mem_cgroup *memcg;
if (!cgroup_subsys_on_dfl(memory_cgrp_subsys))
return NULL;
if (!oom_domain)
oom_domain = root_mem_cgroup;
rcu_read_lock();
memcg = mem_cgroup_from_task(victim);
if (memcg == root_mem_cgroup)
goto out;
/*
* If the victim task has been asynchronously moved to a different
* memory cgroup, we might end up killing tasks outside oom_domain.
* In this case it's better to ignore memory.group.oom.
*/
if (unlikely(!mem_cgroup_is_descendant(memcg, oom_domain)))
goto out;
mm, oom: introduce memory.oom.group For some workloads an intervention from the OOM killer can be painful. Killing a random task can bring the workload into an inconsistent state. Historically, there are two common solutions for this problem: 1) enabling panic_on_oom, 2) using a userspace daemon to monitor OOMs and kill all outstanding processes. Both approaches have their downsides: rebooting on each OOM is an obvious waste of capacity, and handling all in userspace is tricky and requires a userspace agent, which will monitor all cgroups for OOMs. In most cases an in-kernel after-OOM cleaning-up mechanism can eliminate the necessity of enabling panic_on_oom. Also, it can simplify the cgroup management for userspace applications. This commit introduces a new knob for cgroup v2 memory controller: memory.oom.group. The knob determines whether the cgroup should be treated as an indivisible workload by the OOM killer. If set, all tasks belonging to the cgroup or to its descendants (if the memory cgroup is not a leaf cgroup) are killed together or not at all. To determine which cgroup has to be killed, we do traverse the cgroup hierarchy from the victim task's cgroup up to the OOMing cgroup (or root) and looking for the highest-level cgroup with memory.oom.group set. Tasks with the OOM protection (oom_score_adj set to -1000) are treated as an exception and are never killed. This patch doesn't change the OOM victim selection algorithm. Link: http://lkml.kernel.org/r/20180802003201.817-4-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: David Rientjes <rientjes@google.com> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-22 04:53:54 +00:00
/*
* Traverse the memory cgroup hierarchy from the victim task's
* cgroup up to the OOMing cgroup (or root) to find the
* highest-level memory cgroup with oom.group set.
*/
for (; memcg; memcg = parent_mem_cgroup(memcg)) {
if (memcg->oom_group)
oom_group = memcg;
if (memcg == oom_domain)
break;
}
if (oom_group)
css_get(&oom_group->css);
out:
rcu_read_unlock();
return oom_group;
}
void mem_cgroup_print_oom_group(struct mem_cgroup *memcg)
{
pr_info("Tasks in ");
pr_cont_cgroup_path(memcg->css.cgroup);
pr_cont(" are going to be killed due to memory.oom.group set\n");
}
mm: memcontrol: fix missed end-writeback page accounting Commit 0a31bc97c80c ("mm: memcontrol: rewrite uncharge API") changed page migration to uncharge the old page right away. The page is locked, unmapped, truncated, and off the LRU, but it could race with writeback ending, which then doesn't unaccount the page properly: test_clear_page_writeback() migration wait_on_page_writeback() TestClearPageWriteback() mem_cgroup_migrate() clear PCG_USED mem_cgroup_update_page_stat() if (PageCgroupUsed(pc)) decrease memcg pages under writeback release pc->mem_cgroup->move_lock The per-page statistics interface is heavily optimized to avoid a function call and a lookup_page_cgroup() in the file unmap fast path, which means it doesn't verify whether a page is still charged before clearing PageWriteback() and it has to do it in the stat update later. Rework it so that it looks up the page's memcg once at the beginning of the transaction and then uses it throughout. The charge will be verified before clearing PageWriteback() and migration can't uncharge the page as long as that is still set. The RCU lock will protect the memcg past uncharge. As far as losing the optimization goes, the following test results are from a microbenchmark that maps, faults, and unmaps a 4GB sparse file three times in a nested fashion, so that there are two negative passes that don't account but still go through the new transaction overhead. There is no actual difference: old: 33.195102545 seconds time elapsed ( +- 0.01% ) new: 33.199231369 seconds time elapsed ( +- 0.03% ) The time spent in page_remove_rmap()'s callees still adds up to the same, but the time spent in the function itself seems reduced: # Children Self Command Shared Object Symbol old: 0.12% 0.11% filemapstress [kernel.kallsyms] [k] page_remove_rmap new: 0.12% 0.08% filemapstress [kernel.kallsyms] [k] page_remove_rmap Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: <stable@vger.kernel.org> [3.17.x] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-29 21:50:48 +00:00
/**
* lock_page_memcg - lock a page->mem_cgroup binding
* @page: the page
memcg: avoid lock in updating file_mapped (Was fix race in file_mapped accouting flag management At accounting file events per memory cgroup, we need to find memory cgroup via page_cgroup->mem_cgroup. Now, we use lock_page_cgroup() for guarantee pc->mem_cgroup is not overwritten while we make use of it. But, considering the context which page-cgroup for files are accessed, we can use alternative light-weight mutual execusion in the most case. At handling file-caches, the only race we have to take care of is "moving" account, IOW, overwriting page_cgroup->mem_cgroup. (See comment in the patch) Unlike charge/uncharge, "move" happens not so frequently. It happens only when rmdir() and task-moving (with a special settings.) This patch adds a race-checker for file-cache-status accounting v.s. account moving. The new per-cpu-per-memcg counter MEM_CGROUP_ON_MOVE is added. The routine for account move 1. Increment it before start moving 2. Call synchronize_rcu() 3. Decrement it after the end of moving. By this, file-status-counting routine can check it needs to call lock_page_cgroup(). In most case, I doesn't need to call it. Following is a perf data of a process which mmap()/munmap 32MB of file cache in a minute. Before patch: 28.25% mmap mmap [.] main 22.64% mmap [kernel.kallsyms] [k] page_fault 9.96% mmap [kernel.kallsyms] [k] mem_cgroup_update_file_mapped 3.67% mmap [kernel.kallsyms] [k] filemap_fault 3.50% mmap [kernel.kallsyms] [k] unmap_vmas 2.99% mmap [kernel.kallsyms] [k] __do_fault 2.76% mmap [kernel.kallsyms] [k] find_get_page After patch: 30.00% mmap mmap [.] main 23.78% mmap [kernel.kallsyms] [k] page_fault 5.52% mmap [kernel.kallsyms] [k] mem_cgroup_update_file_mapped 3.81% mmap [kernel.kallsyms] [k] unmap_vmas 3.26% mmap [kernel.kallsyms] [k] find_get_page 3.18% mmap [kernel.kallsyms] [k] __do_fault 3.03% mmap [kernel.kallsyms] [k] filemap_fault 2.40% mmap [kernel.kallsyms] [k] handle_mm_fault 2.40% mmap [kernel.kallsyms] [k] do_page_fault This patch reduces memcg's cost to some extent. (mem_cgroup_update_file_mapped is called by both of map/unmap) Note: It seems some more improvements are required..but no idea. maybe removing set/unset flag is required. Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2010-10-27 22:33:40 +00:00
*
* This function protects unlocked LRU pages from being moved to
mm: memcontrol: fix NULL pointer crash in test_clear_page_writeback() Jaegeuk and Brad report a NULL pointer crash when writeback ending tries to update the memcg stats: BUG: unable to handle kernel NULL pointer dereference at 00000000000003b0 IP: test_clear_page_writeback+0x12e/0x2c0 [...] RIP: 0010:test_clear_page_writeback+0x12e/0x2c0 Call Trace: <IRQ> end_page_writeback+0x47/0x70 f2fs_write_end_io+0x76/0x180 [f2fs] bio_endio+0x9f/0x120 blk_update_request+0xa8/0x2f0 scsi_end_request+0x39/0x1d0 scsi_io_completion+0x211/0x690 scsi_finish_command+0xd9/0x120 scsi_softirq_done+0x127/0x150 __blk_mq_complete_request_remote+0x13/0x20 flush_smp_call_function_queue+0x56/0x110 generic_smp_call_function_single_interrupt+0x13/0x30 smp_call_function_single_interrupt+0x27/0x40 call_function_single_interrupt+0x89/0x90 RIP: 0010:native_safe_halt+0x6/0x10 (gdb) l *(test_clear_page_writeback+0x12e) 0xffffffff811bae3e is in test_clear_page_writeback (./include/linux/memcontrol.h:619). 614 mod_node_page_state(page_pgdat(page), idx, val); 615 if (mem_cgroup_disabled() || !page->mem_cgroup) 616 return; 617 mod_memcg_state(page->mem_cgroup, idx, val); 618 pn = page->mem_cgroup->nodeinfo[page_to_nid(page)]; 619 this_cpu_add(pn->lruvec_stat->count[idx], val); 620 } 621 622 unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order, 623 gfp_t gfp_mask, The issue is that writeback doesn't hold a page reference and the page might get freed after PG_writeback is cleared (and the mapping is unlocked) in test_clear_page_writeback(). The stat functions looking up the page's node or zone are safe, as those attributes are static across allocation and free cycles. But page->mem_cgroup is not, and it will get cleared if we race with truncation or migration. It appears this race window has been around for a while, but less likely to trigger when the memcg stats were updated first thing after PG_writeback is cleared. Recent changes reshuffled this code to update the global node stats before the memcg ones, though, stretching the race window out to an extent where people can reproduce the problem. Update test_clear_page_writeback() to look up and pin page->mem_cgroup before clearing PG_writeback, then not use that pointer afterward. It is a partial revert of 62cccb8c8e7a ("mm: simplify lock_page_memcg()") but leaves the pageref-holding callsites that aren't affected alone. Link: http://lkml.kernel.org/r/20170809183825.GA26387@cmpxchg.org Fixes: 62cccb8c8e7a ("mm: simplify lock_page_memcg()") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Jaegeuk Kim <jaegeuk@kernel.org> Tested-by: Jaegeuk Kim <jaegeuk@kernel.org> Reported-by: Bradley Bolen <bradleybolen@gmail.com> Tested-by: Brad Bolen <bradleybolen@gmail.com> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> [4.6+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-08-18 22:15:48 +00:00
* another cgroup.
*
* It ensures lifetime of the returned memcg. Caller is responsible
* for the lifetime of the page; __unlock_page_memcg() is available
* when @page might get freed inside the locked section.
*/
mm: memcontrol: fix NULL pointer crash in test_clear_page_writeback() Jaegeuk and Brad report a NULL pointer crash when writeback ending tries to update the memcg stats: BUG: unable to handle kernel NULL pointer dereference at 00000000000003b0 IP: test_clear_page_writeback+0x12e/0x2c0 [...] RIP: 0010:test_clear_page_writeback+0x12e/0x2c0 Call Trace: <IRQ> end_page_writeback+0x47/0x70 f2fs_write_end_io+0x76/0x180 [f2fs] bio_endio+0x9f/0x120 blk_update_request+0xa8/0x2f0 scsi_end_request+0x39/0x1d0 scsi_io_completion+0x211/0x690 scsi_finish_command+0xd9/0x120 scsi_softirq_done+0x127/0x150 __blk_mq_complete_request_remote+0x13/0x20 flush_smp_call_function_queue+0x56/0x110 generic_smp_call_function_single_interrupt+0x13/0x30 smp_call_function_single_interrupt+0x27/0x40 call_function_single_interrupt+0x89/0x90 RIP: 0010:native_safe_halt+0x6/0x10 (gdb) l *(test_clear_page_writeback+0x12e) 0xffffffff811bae3e is in test_clear_page_writeback (./include/linux/memcontrol.h:619). 614 mod_node_page_state(page_pgdat(page), idx, val); 615 if (mem_cgroup_disabled() || !page->mem_cgroup) 616 return; 617 mod_memcg_state(page->mem_cgroup, idx, val); 618 pn = page->mem_cgroup->nodeinfo[page_to_nid(page)]; 619 this_cpu_add(pn->lruvec_stat->count[idx], val); 620 } 621 622 unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order, 623 gfp_t gfp_mask, The issue is that writeback doesn't hold a page reference and the page might get freed after PG_writeback is cleared (and the mapping is unlocked) in test_clear_page_writeback(). The stat functions looking up the page's node or zone are safe, as those attributes are static across allocation and free cycles. But page->mem_cgroup is not, and it will get cleared if we race with truncation or migration. It appears this race window has been around for a while, but less likely to trigger when the memcg stats were updated first thing after PG_writeback is cleared. Recent changes reshuffled this code to update the global node stats before the memcg ones, though, stretching the race window out to an extent where people can reproduce the problem. Update test_clear_page_writeback() to look up and pin page->mem_cgroup before clearing PG_writeback, then not use that pointer afterward. It is a partial revert of 62cccb8c8e7a ("mm: simplify lock_page_memcg()") but leaves the pageref-holding callsites that aren't affected alone. Link: http://lkml.kernel.org/r/20170809183825.GA26387@cmpxchg.org Fixes: 62cccb8c8e7a ("mm: simplify lock_page_memcg()") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Jaegeuk Kim <jaegeuk@kernel.org> Tested-by: Jaegeuk Kim <jaegeuk@kernel.org> Reported-by: Bradley Bolen <bradleybolen@gmail.com> Tested-by: Brad Bolen <bradleybolen@gmail.com> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> [4.6+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-08-18 22:15:48 +00:00
struct mem_cgroup *lock_page_memcg(struct page *page)
memcg: use new logic for page stat accounting Now, page-stat-per-memcg is recorded into per page_cgroup flag by duplicating page's status into the flag. The reason is that memcg has a feature to move a page from a group to another group and we have race between "move" and "page stat accounting", Under current logic, assume CPU-A and CPU-B. CPU-A does "move" and CPU-B does "page stat accounting". When CPU-A goes 1st, CPU-A CPU-B update "struct page" info. move_lock_mem_cgroup(memcg) see pc->flags copy page stat to new group overwrite pc->mem_cgroup. move_unlock_mem_cgroup(memcg) move_lock_mem_cgroup(mem) set pc->flags update page stat accounting move_unlock_mem_cgroup(mem) stat accounting is guarded by move_lock_mem_cgroup() and "move" logic (CPU-A) doesn't see changes in "struct page" information. But it's costly to have the same information both in 'struct page' and 'struct page_cgroup'. And, there is a potential problem. For example, assume we have PG_dirty accounting in memcg. PG_..is a flag for struct page. PCG_ is a flag for struct page_cgroup. (This is just an example. The same problem can be found in any kind of page stat accounting.) CPU-A CPU-B TestSet PG_dirty (delay) TestClear PG_dirty if (TestClear(PCG_dirty)) memcg->nr_dirty-- if (TestSet(PCG_dirty)) memcg->nr_dirty++ Here, memcg->nr_dirty = +1, this is wrong. This race was reported by Greg Thelen <gthelen@google.com>. Now, only FILE_MAPPED is supported but fortunately, it's serialized by page table lock and this is not real bug, _now_, If this potential problem is caused by having duplicated information in struct page and struct page_cgroup, we may be able to fix this by using original 'struct page' information. But we'll have a problem in "move account" Assume we use only PG_dirty. CPU-A CPU-B TestSet PG_dirty (delay) move_lock_mem_cgroup() if (PageDirty(page)) new_memcg->nr_dirty++ pc->mem_cgroup = new_memcg; move_unlock_mem_cgroup() move_lock_mem_cgroup() memcg = pc->mem_cgroup new_memcg->nr_dirty++ accounting information may be double-counted. This was original reason to have PCG_xxx flags but it seems PCG_xxx has another problem. I think we need a bigger lock as move_lock_mem_cgroup(page) TestSetPageDirty(page) update page stats (without any checks) move_unlock_mem_cgroup(page) This fixes both of problems and we don't have to duplicate page flag into page_cgroup. Please note: move_lock_mem_cgroup() is held only when there are possibility of "account move" under the system. So, in most path, status update will go without atomic locks. This patch introduces mem_cgroup_begin_update_page_stat() and mem_cgroup_end_update_page_stat() both should be called at modifying 'struct page' information if memcg takes care of it. as mem_cgroup_begin_update_page_stat() modify page information mem_cgroup_update_page_stat() => never check any 'struct page' info, just update counters. mem_cgroup_end_update_page_stat(). This patch is slow because we need to call begin_update_page_stat()/ end_update_page_stat() regardless of accounted will be changed or not. A following patch adds an easy optimization and reduces the cost. [akpm@linux-foundation.org: s/lock/locked/] [hughd@google.com: fix deadlock by avoiding stat lock when anon] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Greg Thelen <gthelen@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Ying Han <yinghan@google.com> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-21 23:34:25 +00:00
{
struct page *head = compound_head(page); /* rmap on tail pages */
memcg: use new logic for page stat accounting Now, page-stat-per-memcg is recorded into per page_cgroup flag by duplicating page's status into the flag. The reason is that memcg has a feature to move a page from a group to another group and we have race between "move" and "page stat accounting", Under current logic, assume CPU-A and CPU-B. CPU-A does "move" and CPU-B does "page stat accounting". When CPU-A goes 1st, CPU-A CPU-B update "struct page" info. move_lock_mem_cgroup(memcg) see pc->flags copy page stat to new group overwrite pc->mem_cgroup. move_unlock_mem_cgroup(memcg) move_lock_mem_cgroup(mem) set pc->flags update page stat accounting move_unlock_mem_cgroup(mem) stat accounting is guarded by move_lock_mem_cgroup() and "move" logic (CPU-A) doesn't see changes in "struct page" information. But it's costly to have the same information both in 'struct page' and 'struct page_cgroup'. And, there is a potential problem. For example, assume we have PG_dirty accounting in memcg. PG_..is a flag for struct page. PCG_ is a flag for struct page_cgroup. (This is just an example. The same problem can be found in any kind of page stat accounting.) CPU-A CPU-B TestSet PG_dirty (delay) TestClear PG_dirty if (TestClear(PCG_dirty)) memcg->nr_dirty-- if (TestSet(PCG_dirty)) memcg->nr_dirty++ Here, memcg->nr_dirty = +1, this is wrong. This race was reported by Greg Thelen <gthelen@google.com>. Now, only FILE_MAPPED is supported but fortunately, it's serialized by page table lock and this is not real bug, _now_, If this potential problem is caused by having duplicated information in struct page and struct page_cgroup, we may be able to fix this by using original 'struct page' information. But we'll have a problem in "move account" Assume we use only PG_dirty. CPU-A CPU-B TestSet PG_dirty (delay) move_lock_mem_cgroup() if (PageDirty(page)) new_memcg->nr_dirty++ pc->mem_cgroup = new_memcg; move_unlock_mem_cgroup() move_lock_mem_cgroup() memcg = pc->mem_cgroup new_memcg->nr_dirty++ accounting information may be double-counted. This was original reason to have PCG_xxx flags but it seems PCG_xxx has another problem. I think we need a bigger lock as move_lock_mem_cgroup(page) TestSetPageDirty(page) update page stats (without any checks) move_unlock_mem_cgroup(page) This fixes both of problems and we don't have to duplicate page flag into page_cgroup. Please note: move_lock_mem_cgroup() is held only when there are possibility of "account move" under the system. So, in most path, status update will go without atomic locks. This patch introduces mem_cgroup_begin_update_page_stat() and mem_cgroup_end_update_page_stat() both should be called at modifying 'struct page' information if memcg takes care of it. as mem_cgroup_begin_update_page_stat() modify page information mem_cgroup_update_page_stat() => never check any 'struct page' info, just update counters. mem_cgroup_end_update_page_stat(). This patch is slow because we need to call begin_update_page_stat()/ end_update_page_stat() regardless of accounted will be changed or not. A following patch adds an easy optimization and reduces the cost. [akpm@linux-foundation.org: s/lock/locked/] [hughd@google.com: fix deadlock by avoiding stat lock when anon] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Greg Thelen <gthelen@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Ying Han <yinghan@google.com> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-21 23:34:25 +00:00
struct mem_cgroup *memcg;
unsigned long flags;
memcg: use new logic for page stat accounting Now, page-stat-per-memcg is recorded into per page_cgroup flag by duplicating page's status into the flag. The reason is that memcg has a feature to move a page from a group to another group and we have race between "move" and "page stat accounting", Under current logic, assume CPU-A and CPU-B. CPU-A does "move" and CPU-B does "page stat accounting". When CPU-A goes 1st, CPU-A CPU-B update "struct page" info. move_lock_mem_cgroup(memcg) see pc->flags copy page stat to new group overwrite pc->mem_cgroup. move_unlock_mem_cgroup(memcg) move_lock_mem_cgroup(mem) set pc->flags update page stat accounting move_unlock_mem_cgroup(mem) stat accounting is guarded by move_lock_mem_cgroup() and "move" logic (CPU-A) doesn't see changes in "struct page" information. But it's costly to have the same information both in 'struct page' and 'struct page_cgroup'. And, there is a potential problem. For example, assume we have PG_dirty accounting in memcg. PG_..is a flag for struct page. PCG_ is a flag for struct page_cgroup. (This is just an example. The same problem can be found in any kind of page stat accounting.) CPU-A CPU-B TestSet PG_dirty (delay) TestClear PG_dirty if (TestClear(PCG_dirty)) memcg->nr_dirty-- if (TestSet(PCG_dirty)) memcg->nr_dirty++ Here, memcg->nr_dirty = +1, this is wrong. This race was reported by Greg Thelen <gthelen@google.com>. Now, only FILE_MAPPED is supported but fortunately, it's serialized by page table lock and this is not real bug, _now_, If this potential problem is caused by having duplicated information in struct page and struct page_cgroup, we may be able to fix this by using original 'struct page' information. But we'll have a problem in "move account" Assume we use only PG_dirty. CPU-A CPU-B TestSet PG_dirty (delay) move_lock_mem_cgroup() if (PageDirty(page)) new_memcg->nr_dirty++ pc->mem_cgroup = new_memcg; move_unlock_mem_cgroup() move_lock_mem_cgroup() memcg = pc->mem_cgroup new_memcg->nr_dirty++ accounting information may be double-counted. This was original reason to have PCG_xxx flags but it seems PCG_xxx has another problem. I think we need a bigger lock as move_lock_mem_cgroup(page) TestSetPageDirty(page) update page stats (without any checks) move_unlock_mem_cgroup(page) This fixes both of problems and we don't have to duplicate page flag into page_cgroup. Please note: move_lock_mem_cgroup() is held only when there are possibility of "account move" under the system. So, in most path, status update will go without atomic locks. This patch introduces mem_cgroup_begin_update_page_stat() and mem_cgroup_end_update_page_stat() both should be called at modifying 'struct page' information if memcg takes care of it. as mem_cgroup_begin_update_page_stat() modify page information mem_cgroup_update_page_stat() => never check any 'struct page' info, just update counters. mem_cgroup_end_update_page_stat(). This patch is slow because we need to call begin_update_page_stat()/ end_update_page_stat() regardless of accounted will be changed or not. A following patch adds an easy optimization and reduces the cost. [akpm@linux-foundation.org: s/lock/locked/] [hughd@google.com: fix deadlock by avoiding stat lock when anon] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Greg Thelen <gthelen@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Ying Han <yinghan@google.com> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-21 23:34:25 +00:00
/*
* The RCU lock is held throughout the transaction. The fast
* path can get away without acquiring the memcg->move_lock
* because page moving starts with an RCU grace period.
mm: memcontrol: fix NULL pointer crash in test_clear_page_writeback() Jaegeuk and Brad report a NULL pointer crash when writeback ending tries to update the memcg stats: BUG: unable to handle kernel NULL pointer dereference at 00000000000003b0 IP: test_clear_page_writeback+0x12e/0x2c0 [...] RIP: 0010:test_clear_page_writeback+0x12e/0x2c0 Call Trace: <IRQ> end_page_writeback+0x47/0x70 f2fs_write_end_io+0x76/0x180 [f2fs] bio_endio+0x9f/0x120 blk_update_request+0xa8/0x2f0 scsi_end_request+0x39/0x1d0 scsi_io_completion+0x211/0x690 scsi_finish_command+0xd9/0x120 scsi_softirq_done+0x127/0x150 __blk_mq_complete_request_remote+0x13/0x20 flush_smp_call_function_queue+0x56/0x110 generic_smp_call_function_single_interrupt+0x13/0x30 smp_call_function_single_interrupt+0x27/0x40 call_function_single_interrupt+0x89/0x90 RIP: 0010:native_safe_halt+0x6/0x10 (gdb) l *(test_clear_page_writeback+0x12e) 0xffffffff811bae3e is in test_clear_page_writeback (./include/linux/memcontrol.h:619). 614 mod_node_page_state(page_pgdat(page), idx, val); 615 if (mem_cgroup_disabled() || !page->mem_cgroup) 616 return; 617 mod_memcg_state(page->mem_cgroup, idx, val); 618 pn = page->mem_cgroup->nodeinfo[page_to_nid(page)]; 619 this_cpu_add(pn->lruvec_stat->count[idx], val); 620 } 621 622 unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order, 623 gfp_t gfp_mask, The issue is that writeback doesn't hold a page reference and the page might get freed after PG_writeback is cleared (and the mapping is unlocked) in test_clear_page_writeback(). The stat functions looking up the page's node or zone are safe, as those attributes are static across allocation and free cycles. But page->mem_cgroup is not, and it will get cleared if we race with truncation or migration. It appears this race window has been around for a while, but less likely to trigger when the memcg stats were updated first thing after PG_writeback is cleared. Recent changes reshuffled this code to update the global node stats before the memcg ones, though, stretching the race window out to an extent where people can reproduce the problem. Update test_clear_page_writeback() to look up and pin page->mem_cgroup before clearing PG_writeback, then not use that pointer afterward. It is a partial revert of 62cccb8c8e7a ("mm: simplify lock_page_memcg()") but leaves the pageref-holding callsites that aren't affected alone. Link: http://lkml.kernel.org/r/20170809183825.GA26387@cmpxchg.org Fixes: 62cccb8c8e7a ("mm: simplify lock_page_memcg()") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Jaegeuk Kim <jaegeuk@kernel.org> Tested-by: Jaegeuk Kim <jaegeuk@kernel.org> Reported-by: Bradley Bolen <bradleybolen@gmail.com> Tested-by: Brad Bolen <bradleybolen@gmail.com> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> [4.6+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-08-18 22:15:48 +00:00
*
* The RCU lock also protects the memcg from being freed when
* the page state that is going to change is the only thing
* preventing the page itself from being freed. E.g. writeback
* doesn't hold a page reference and relies on PG_writeback to
* keep off truncation, migration and so forth.
*/
mm: memcontrol: fix missed end-writeback page accounting Commit 0a31bc97c80c ("mm: memcontrol: rewrite uncharge API") changed page migration to uncharge the old page right away. The page is locked, unmapped, truncated, and off the LRU, but it could race with writeback ending, which then doesn't unaccount the page properly: test_clear_page_writeback() migration wait_on_page_writeback() TestClearPageWriteback() mem_cgroup_migrate() clear PCG_USED mem_cgroup_update_page_stat() if (PageCgroupUsed(pc)) decrease memcg pages under writeback release pc->mem_cgroup->move_lock The per-page statistics interface is heavily optimized to avoid a function call and a lookup_page_cgroup() in the file unmap fast path, which means it doesn't verify whether a page is still charged before clearing PageWriteback() and it has to do it in the stat update later. Rework it so that it looks up the page's memcg once at the beginning of the transaction and then uses it throughout. The charge will be verified before clearing PageWriteback() and migration can't uncharge the page as long as that is still set. The RCU lock will protect the memcg past uncharge. As far as losing the optimization goes, the following test results are from a microbenchmark that maps, faults, and unmaps a 4GB sparse file three times in a nested fashion, so that there are two negative passes that don't account but still go through the new transaction overhead. There is no actual difference: old: 33.195102545 seconds time elapsed ( +- 0.01% ) new: 33.199231369 seconds time elapsed ( +- 0.03% ) The time spent in page_remove_rmap()'s callees still adds up to the same, but the time spent in the function itself seems reduced: # Children Self Command Shared Object Symbol old: 0.12% 0.11% filemapstress [kernel.kallsyms] [k] page_remove_rmap new: 0.12% 0.08% filemapstress [kernel.kallsyms] [k] page_remove_rmap Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: <stable@vger.kernel.org> [3.17.x] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-29 21:50:48 +00:00
rcu_read_lock();
if (mem_cgroup_disabled())
mm: memcontrol: fix NULL pointer crash in test_clear_page_writeback() Jaegeuk and Brad report a NULL pointer crash when writeback ending tries to update the memcg stats: BUG: unable to handle kernel NULL pointer dereference at 00000000000003b0 IP: test_clear_page_writeback+0x12e/0x2c0 [...] RIP: 0010:test_clear_page_writeback+0x12e/0x2c0 Call Trace: <IRQ> end_page_writeback+0x47/0x70 f2fs_write_end_io+0x76/0x180 [f2fs] bio_endio+0x9f/0x120 blk_update_request+0xa8/0x2f0 scsi_end_request+0x39/0x1d0 scsi_io_completion+0x211/0x690 scsi_finish_command+0xd9/0x120 scsi_softirq_done+0x127/0x150 __blk_mq_complete_request_remote+0x13/0x20 flush_smp_call_function_queue+0x56/0x110 generic_smp_call_function_single_interrupt+0x13/0x30 smp_call_function_single_interrupt+0x27/0x40 call_function_single_interrupt+0x89/0x90 RIP: 0010:native_safe_halt+0x6/0x10 (gdb) l *(test_clear_page_writeback+0x12e) 0xffffffff811bae3e is in test_clear_page_writeback (./include/linux/memcontrol.h:619). 614 mod_node_page_state(page_pgdat(page), idx, val); 615 if (mem_cgroup_disabled() || !page->mem_cgroup) 616 return; 617 mod_memcg_state(page->mem_cgroup, idx, val); 618 pn = page->mem_cgroup->nodeinfo[page_to_nid(page)]; 619 this_cpu_add(pn->lruvec_stat->count[idx], val); 620 } 621 622 unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order, 623 gfp_t gfp_mask, The issue is that writeback doesn't hold a page reference and the page might get freed after PG_writeback is cleared (and the mapping is unlocked) in test_clear_page_writeback(). The stat functions looking up the page's node or zone are safe, as those attributes are static across allocation and free cycles. But page->mem_cgroup is not, and it will get cleared if we race with truncation or migration. It appears this race window has been around for a while, but less likely to trigger when the memcg stats were updated first thing after PG_writeback is cleared. Recent changes reshuffled this code to update the global node stats before the memcg ones, though, stretching the race window out to an extent where people can reproduce the problem. Update test_clear_page_writeback() to look up and pin page->mem_cgroup before clearing PG_writeback, then not use that pointer afterward. It is a partial revert of 62cccb8c8e7a ("mm: simplify lock_page_memcg()") but leaves the pageref-holding callsites that aren't affected alone. Link: http://lkml.kernel.org/r/20170809183825.GA26387@cmpxchg.org Fixes: 62cccb8c8e7a ("mm: simplify lock_page_memcg()") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Jaegeuk Kim <jaegeuk@kernel.org> Tested-by: Jaegeuk Kim <jaegeuk@kernel.org> Reported-by: Bradley Bolen <bradleybolen@gmail.com> Tested-by: Brad Bolen <bradleybolen@gmail.com> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> [4.6+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-08-18 22:15:48 +00:00
return NULL;
memcg: use new logic for page stat accounting Now, page-stat-per-memcg is recorded into per page_cgroup flag by duplicating page's status into the flag. The reason is that memcg has a feature to move a page from a group to another group and we have race between "move" and "page stat accounting", Under current logic, assume CPU-A and CPU-B. CPU-A does "move" and CPU-B does "page stat accounting". When CPU-A goes 1st, CPU-A CPU-B update "struct page" info. move_lock_mem_cgroup(memcg) see pc->flags copy page stat to new group overwrite pc->mem_cgroup. move_unlock_mem_cgroup(memcg) move_lock_mem_cgroup(mem) set pc->flags update page stat accounting move_unlock_mem_cgroup(mem) stat accounting is guarded by move_lock_mem_cgroup() and "move" logic (CPU-A) doesn't see changes in "struct page" information. But it's costly to have the same information both in 'struct page' and 'struct page_cgroup'. And, there is a potential problem. For example, assume we have PG_dirty accounting in memcg. PG_..is a flag for struct page. PCG_ is a flag for struct page_cgroup. (This is just an example. The same problem can be found in any kind of page stat accounting.) CPU-A CPU-B TestSet PG_dirty (delay) TestClear PG_dirty if (TestClear(PCG_dirty)) memcg->nr_dirty-- if (TestSet(PCG_dirty)) memcg->nr_dirty++ Here, memcg->nr_dirty = +1, this is wrong. This race was reported by Greg Thelen <gthelen@google.com>. Now, only FILE_MAPPED is supported but fortunately, it's serialized by page table lock and this is not real bug, _now_, If this potential problem is caused by having duplicated information in struct page and struct page_cgroup, we may be able to fix this by using original 'struct page' information. But we'll have a problem in "move account" Assume we use only PG_dirty. CPU-A CPU-B TestSet PG_dirty (delay) move_lock_mem_cgroup() if (PageDirty(page)) new_memcg->nr_dirty++ pc->mem_cgroup = new_memcg; move_unlock_mem_cgroup() move_lock_mem_cgroup() memcg = pc->mem_cgroup new_memcg->nr_dirty++ accounting information may be double-counted. This was original reason to have PCG_xxx flags but it seems PCG_xxx has another problem. I think we need a bigger lock as move_lock_mem_cgroup(page) TestSetPageDirty(page) update page stats (without any checks) move_unlock_mem_cgroup(page) This fixes both of problems and we don't have to duplicate page flag into page_cgroup. Please note: move_lock_mem_cgroup() is held only when there are possibility of "account move" under the system. So, in most path, status update will go without atomic locks. This patch introduces mem_cgroup_begin_update_page_stat() and mem_cgroup_end_update_page_stat() both should be called at modifying 'struct page' information if memcg takes care of it. as mem_cgroup_begin_update_page_stat() modify page information mem_cgroup_update_page_stat() => never check any 'struct page' info, just update counters. mem_cgroup_end_update_page_stat(). This patch is slow because we need to call begin_update_page_stat()/ end_update_page_stat() regardless of accounted will be changed or not. A following patch adds an easy optimization and reduces the cost. [akpm@linux-foundation.org: s/lock/locked/] [hughd@google.com: fix deadlock by avoiding stat lock when anon] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Greg Thelen <gthelen@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Ying Han <yinghan@google.com> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-21 23:34:25 +00:00
again:
memcg = head->mem_cgroup;
if (unlikely(!memcg))
mm: memcontrol: fix NULL pointer crash in test_clear_page_writeback() Jaegeuk and Brad report a NULL pointer crash when writeback ending tries to update the memcg stats: BUG: unable to handle kernel NULL pointer dereference at 00000000000003b0 IP: test_clear_page_writeback+0x12e/0x2c0 [...] RIP: 0010:test_clear_page_writeback+0x12e/0x2c0 Call Trace: <IRQ> end_page_writeback+0x47/0x70 f2fs_write_end_io+0x76/0x180 [f2fs] bio_endio+0x9f/0x120 blk_update_request+0xa8/0x2f0 scsi_end_request+0x39/0x1d0 scsi_io_completion+0x211/0x690 scsi_finish_command+0xd9/0x120 scsi_softirq_done+0x127/0x150 __blk_mq_complete_request_remote+0x13/0x20 flush_smp_call_function_queue+0x56/0x110 generic_smp_call_function_single_interrupt+0x13/0x30 smp_call_function_single_interrupt+0x27/0x40 call_function_single_interrupt+0x89/0x90 RIP: 0010:native_safe_halt+0x6/0x10 (gdb) l *(test_clear_page_writeback+0x12e) 0xffffffff811bae3e is in test_clear_page_writeback (./include/linux/memcontrol.h:619). 614 mod_node_page_state(page_pgdat(page), idx, val); 615 if (mem_cgroup_disabled() || !page->mem_cgroup) 616 return; 617 mod_memcg_state(page->mem_cgroup, idx, val); 618 pn = page->mem_cgroup->nodeinfo[page_to_nid(page)]; 619 this_cpu_add(pn->lruvec_stat->count[idx], val); 620 } 621 622 unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order, 623 gfp_t gfp_mask, The issue is that writeback doesn't hold a page reference and the page might get freed after PG_writeback is cleared (and the mapping is unlocked) in test_clear_page_writeback(). The stat functions looking up the page's node or zone are safe, as those attributes are static across allocation and free cycles. But page->mem_cgroup is not, and it will get cleared if we race with truncation or migration. It appears this race window has been around for a while, but less likely to trigger when the memcg stats were updated first thing after PG_writeback is cleared. Recent changes reshuffled this code to update the global node stats before the memcg ones, though, stretching the race window out to an extent where people can reproduce the problem. Update test_clear_page_writeback() to look up and pin page->mem_cgroup before clearing PG_writeback, then not use that pointer afterward. It is a partial revert of 62cccb8c8e7a ("mm: simplify lock_page_memcg()") but leaves the pageref-holding callsites that aren't affected alone. Link: http://lkml.kernel.org/r/20170809183825.GA26387@cmpxchg.org Fixes: 62cccb8c8e7a ("mm: simplify lock_page_memcg()") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Jaegeuk Kim <jaegeuk@kernel.org> Tested-by: Jaegeuk Kim <jaegeuk@kernel.org> Reported-by: Bradley Bolen <bradleybolen@gmail.com> Tested-by: Brad Bolen <bradleybolen@gmail.com> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> [4.6+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-08-18 22:15:48 +00:00
return NULL;
mm: memcontrol: fix missed end-writeback page accounting Commit 0a31bc97c80c ("mm: memcontrol: rewrite uncharge API") changed page migration to uncharge the old page right away. The page is locked, unmapped, truncated, and off the LRU, but it could race with writeback ending, which then doesn't unaccount the page properly: test_clear_page_writeback() migration wait_on_page_writeback() TestClearPageWriteback() mem_cgroup_migrate() clear PCG_USED mem_cgroup_update_page_stat() if (PageCgroupUsed(pc)) decrease memcg pages under writeback release pc->mem_cgroup->move_lock The per-page statistics interface is heavily optimized to avoid a function call and a lookup_page_cgroup() in the file unmap fast path, which means it doesn't verify whether a page is still charged before clearing PageWriteback() and it has to do it in the stat update later. Rework it so that it looks up the page's memcg once at the beginning of the transaction and then uses it throughout. The charge will be verified before clearing PageWriteback() and migration can't uncharge the page as long as that is still set. The RCU lock will protect the memcg past uncharge. As far as losing the optimization goes, the following test results are from a microbenchmark that maps, faults, and unmaps a 4GB sparse file three times in a nested fashion, so that there are two negative passes that don't account but still go through the new transaction overhead. There is no actual difference: old: 33.195102545 seconds time elapsed ( +- 0.01% ) new: 33.199231369 seconds time elapsed ( +- 0.03% ) The time spent in page_remove_rmap()'s callees still adds up to the same, but the time spent in the function itself seems reduced: # Children Self Command Shared Object Symbol old: 0.12% 0.11% filemapstress [kernel.kallsyms] [k] page_remove_rmap new: 0.12% 0.08% filemapstress [kernel.kallsyms] [k] page_remove_rmap Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: <stable@vger.kernel.org> [3.17.x] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-29 21:50:48 +00:00
if (atomic_read(&memcg->moving_account) <= 0)
mm: memcontrol: fix NULL pointer crash in test_clear_page_writeback() Jaegeuk and Brad report a NULL pointer crash when writeback ending tries to update the memcg stats: BUG: unable to handle kernel NULL pointer dereference at 00000000000003b0 IP: test_clear_page_writeback+0x12e/0x2c0 [...] RIP: 0010:test_clear_page_writeback+0x12e/0x2c0 Call Trace: <IRQ> end_page_writeback+0x47/0x70 f2fs_write_end_io+0x76/0x180 [f2fs] bio_endio+0x9f/0x120 blk_update_request+0xa8/0x2f0 scsi_end_request+0x39/0x1d0 scsi_io_completion+0x211/0x690 scsi_finish_command+0xd9/0x120 scsi_softirq_done+0x127/0x150 __blk_mq_complete_request_remote+0x13/0x20 flush_smp_call_function_queue+0x56/0x110 generic_smp_call_function_single_interrupt+0x13/0x30 smp_call_function_single_interrupt+0x27/0x40 call_function_single_interrupt+0x89/0x90 RIP: 0010:native_safe_halt+0x6/0x10 (gdb) l *(test_clear_page_writeback+0x12e) 0xffffffff811bae3e is in test_clear_page_writeback (./include/linux/memcontrol.h:619). 614 mod_node_page_state(page_pgdat(page), idx, val); 615 if (mem_cgroup_disabled() || !page->mem_cgroup) 616 return; 617 mod_memcg_state(page->mem_cgroup, idx, val); 618 pn = page->mem_cgroup->nodeinfo[page_to_nid(page)]; 619 this_cpu_add(pn->lruvec_stat->count[idx], val); 620 } 621 622 unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order, 623 gfp_t gfp_mask, The issue is that writeback doesn't hold a page reference and the page might get freed after PG_writeback is cleared (and the mapping is unlocked) in test_clear_page_writeback(). The stat functions looking up the page's node or zone are safe, as those attributes are static across allocation and free cycles. But page->mem_cgroup is not, and it will get cleared if we race with truncation or migration. It appears this race window has been around for a while, but less likely to trigger when the memcg stats were updated first thing after PG_writeback is cleared. Recent changes reshuffled this code to update the global node stats before the memcg ones, though, stretching the race window out to an extent where people can reproduce the problem. Update test_clear_page_writeback() to look up and pin page->mem_cgroup before clearing PG_writeback, then not use that pointer afterward. It is a partial revert of 62cccb8c8e7a ("mm: simplify lock_page_memcg()") but leaves the pageref-holding callsites that aren't affected alone. Link: http://lkml.kernel.org/r/20170809183825.GA26387@cmpxchg.org Fixes: 62cccb8c8e7a ("mm: simplify lock_page_memcg()") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Jaegeuk Kim <jaegeuk@kernel.org> Tested-by: Jaegeuk Kim <jaegeuk@kernel.org> Reported-by: Bradley Bolen <bradleybolen@gmail.com> Tested-by: Brad Bolen <bradleybolen@gmail.com> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> [4.6+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-08-18 22:15:48 +00:00
return memcg;
memcg: use new logic for page stat accounting Now, page-stat-per-memcg is recorded into per page_cgroup flag by duplicating page's status into the flag. The reason is that memcg has a feature to move a page from a group to another group and we have race between "move" and "page stat accounting", Under current logic, assume CPU-A and CPU-B. CPU-A does "move" and CPU-B does "page stat accounting". When CPU-A goes 1st, CPU-A CPU-B update "struct page" info. move_lock_mem_cgroup(memcg) see pc->flags copy page stat to new group overwrite pc->mem_cgroup. move_unlock_mem_cgroup(memcg) move_lock_mem_cgroup(mem) set pc->flags update page stat accounting move_unlock_mem_cgroup(mem) stat accounting is guarded by move_lock_mem_cgroup() and "move" logic (CPU-A) doesn't see changes in "struct page" information. But it's costly to have the same information both in 'struct page' and 'struct page_cgroup'. And, there is a potential problem. For example, assume we have PG_dirty accounting in memcg. PG_..is a flag for struct page. PCG_ is a flag for struct page_cgroup. (This is just an example. The same problem can be found in any kind of page stat accounting.) CPU-A CPU-B TestSet PG_dirty (delay) TestClear PG_dirty if (TestClear(PCG_dirty)) memcg->nr_dirty-- if (TestSet(PCG_dirty)) memcg->nr_dirty++ Here, memcg->nr_dirty = +1, this is wrong. This race was reported by Greg Thelen <gthelen@google.com>. Now, only FILE_MAPPED is supported but fortunately, it's serialized by page table lock and this is not real bug, _now_, If this potential problem is caused by having duplicated information in struct page and struct page_cgroup, we may be able to fix this by using original 'struct page' information. But we'll have a problem in "move account" Assume we use only PG_dirty. CPU-A CPU-B TestSet PG_dirty (delay) move_lock_mem_cgroup() if (PageDirty(page)) new_memcg->nr_dirty++ pc->mem_cgroup = new_memcg; move_unlock_mem_cgroup() move_lock_mem_cgroup() memcg = pc->mem_cgroup new_memcg->nr_dirty++ accounting information may be double-counted. This was original reason to have PCG_xxx flags but it seems PCG_xxx has another problem. I think we need a bigger lock as move_lock_mem_cgroup(page) TestSetPageDirty(page) update page stats (without any checks) move_unlock_mem_cgroup(page) This fixes both of problems and we don't have to duplicate page flag into page_cgroup. Please note: move_lock_mem_cgroup() is held only when there are possibility of "account move" under the system. So, in most path, status update will go without atomic locks. This patch introduces mem_cgroup_begin_update_page_stat() and mem_cgroup_end_update_page_stat() both should be called at modifying 'struct page' information if memcg takes care of it. as mem_cgroup_begin_update_page_stat() modify page information mem_cgroup_update_page_stat() => never check any 'struct page' info, just update counters. mem_cgroup_end_update_page_stat(). This patch is slow because we need to call begin_update_page_stat()/ end_update_page_stat() regardless of accounted will be changed or not. A following patch adds an easy optimization and reduces the cost. [akpm@linux-foundation.org: s/lock/locked/] [hughd@google.com: fix deadlock by avoiding stat lock when anon] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Greg Thelen <gthelen@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Ying Han <yinghan@google.com> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-21 23:34:25 +00:00
spin_lock_irqsave(&memcg->move_lock, flags);
if (memcg != head->mem_cgroup) {
spin_unlock_irqrestore(&memcg->move_lock, flags);
memcg: use new logic for page stat accounting Now, page-stat-per-memcg is recorded into per page_cgroup flag by duplicating page's status into the flag. The reason is that memcg has a feature to move a page from a group to another group and we have race between "move" and "page stat accounting", Under current logic, assume CPU-A and CPU-B. CPU-A does "move" and CPU-B does "page stat accounting". When CPU-A goes 1st, CPU-A CPU-B update "struct page" info. move_lock_mem_cgroup(memcg) see pc->flags copy page stat to new group overwrite pc->mem_cgroup. move_unlock_mem_cgroup(memcg) move_lock_mem_cgroup(mem) set pc->flags update page stat accounting move_unlock_mem_cgroup(mem) stat accounting is guarded by move_lock_mem_cgroup() and "move" logic (CPU-A) doesn't see changes in "struct page" information. But it's costly to have the same information both in 'struct page' and 'struct page_cgroup'. And, there is a potential problem. For example, assume we have PG_dirty accounting in memcg. PG_..is a flag for struct page. PCG_ is a flag for struct page_cgroup. (This is just an example. The same problem can be found in any kind of page stat accounting.) CPU-A CPU-B TestSet PG_dirty (delay) TestClear PG_dirty if (TestClear(PCG_dirty)) memcg->nr_dirty-- if (TestSet(PCG_dirty)) memcg->nr_dirty++ Here, memcg->nr_dirty = +1, this is wrong. This race was reported by Greg Thelen <gthelen@google.com>. Now, only FILE_MAPPED is supported but fortunately, it's serialized by page table lock and this is not real bug, _now_, If this potential problem is caused by having duplicated information in struct page and struct page_cgroup, we may be able to fix this by using original 'struct page' information. But we'll have a problem in "move account" Assume we use only PG_dirty. CPU-A CPU-B TestSet PG_dirty (delay) move_lock_mem_cgroup() if (PageDirty(page)) new_memcg->nr_dirty++ pc->mem_cgroup = new_memcg; move_unlock_mem_cgroup() move_lock_mem_cgroup() memcg = pc->mem_cgroup new_memcg->nr_dirty++ accounting information may be double-counted. This was original reason to have PCG_xxx flags but it seems PCG_xxx has another problem. I think we need a bigger lock as move_lock_mem_cgroup(page) TestSetPageDirty(page) update page stats (without any checks) move_unlock_mem_cgroup(page) This fixes both of problems and we don't have to duplicate page flag into page_cgroup. Please note: move_lock_mem_cgroup() is held only when there are possibility of "account move" under the system. So, in most path, status update will go without atomic locks. This patch introduces mem_cgroup_begin_update_page_stat() and mem_cgroup_end_update_page_stat() both should be called at modifying 'struct page' information if memcg takes care of it. as mem_cgroup_begin_update_page_stat() modify page information mem_cgroup_update_page_stat() => never check any 'struct page' info, just update counters. mem_cgroup_end_update_page_stat(). This patch is slow because we need to call begin_update_page_stat()/ end_update_page_stat() regardless of accounted will be changed or not. A following patch adds an easy optimization and reduces the cost. [akpm@linux-foundation.org: s/lock/locked/] [hughd@google.com: fix deadlock by avoiding stat lock when anon] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Greg Thelen <gthelen@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Ying Han <yinghan@google.com> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-21 23:34:25 +00:00
goto again;
}
/*
* When charge migration first begins, we can have locked and
* unlocked page stat updates happening concurrently. Track
* the task who has the lock for unlock_page_memcg().
*/
memcg->move_lock_task = current;
memcg->move_lock_flags = flags;
mm: memcontrol: fix missed end-writeback page accounting Commit 0a31bc97c80c ("mm: memcontrol: rewrite uncharge API") changed page migration to uncharge the old page right away. The page is locked, unmapped, truncated, and off the LRU, but it could race with writeback ending, which then doesn't unaccount the page properly: test_clear_page_writeback() migration wait_on_page_writeback() TestClearPageWriteback() mem_cgroup_migrate() clear PCG_USED mem_cgroup_update_page_stat() if (PageCgroupUsed(pc)) decrease memcg pages under writeback release pc->mem_cgroup->move_lock The per-page statistics interface is heavily optimized to avoid a function call and a lookup_page_cgroup() in the file unmap fast path, which means it doesn't verify whether a page is still charged before clearing PageWriteback() and it has to do it in the stat update later. Rework it so that it looks up the page's memcg once at the beginning of the transaction and then uses it throughout. The charge will be verified before clearing PageWriteback() and migration can't uncharge the page as long as that is still set. The RCU lock will protect the memcg past uncharge. As far as losing the optimization goes, the following test results are from a microbenchmark that maps, faults, and unmaps a 4GB sparse file three times in a nested fashion, so that there are two negative passes that don't account but still go through the new transaction overhead. There is no actual difference: old: 33.195102545 seconds time elapsed ( +- 0.01% ) new: 33.199231369 seconds time elapsed ( +- 0.03% ) The time spent in page_remove_rmap()'s callees still adds up to the same, but the time spent in the function itself seems reduced: # Children Self Command Shared Object Symbol old: 0.12% 0.11% filemapstress [kernel.kallsyms] [k] page_remove_rmap new: 0.12% 0.08% filemapstress [kernel.kallsyms] [k] page_remove_rmap Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: <stable@vger.kernel.org> [3.17.x] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-29 21:50:48 +00:00
mm: memcontrol: fix NULL pointer crash in test_clear_page_writeback() Jaegeuk and Brad report a NULL pointer crash when writeback ending tries to update the memcg stats: BUG: unable to handle kernel NULL pointer dereference at 00000000000003b0 IP: test_clear_page_writeback+0x12e/0x2c0 [...] RIP: 0010:test_clear_page_writeback+0x12e/0x2c0 Call Trace: <IRQ> end_page_writeback+0x47/0x70 f2fs_write_end_io+0x76/0x180 [f2fs] bio_endio+0x9f/0x120 blk_update_request+0xa8/0x2f0 scsi_end_request+0x39/0x1d0 scsi_io_completion+0x211/0x690 scsi_finish_command+0xd9/0x120 scsi_softirq_done+0x127/0x150 __blk_mq_complete_request_remote+0x13/0x20 flush_smp_call_function_queue+0x56/0x110 generic_smp_call_function_single_interrupt+0x13/0x30 smp_call_function_single_interrupt+0x27/0x40 call_function_single_interrupt+0x89/0x90 RIP: 0010:native_safe_halt+0x6/0x10 (gdb) l *(test_clear_page_writeback+0x12e) 0xffffffff811bae3e is in test_clear_page_writeback (./include/linux/memcontrol.h:619). 614 mod_node_page_state(page_pgdat(page), idx, val); 615 if (mem_cgroup_disabled() || !page->mem_cgroup) 616 return; 617 mod_memcg_state(page->mem_cgroup, idx, val); 618 pn = page->mem_cgroup->nodeinfo[page_to_nid(page)]; 619 this_cpu_add(pn->lruvec_stat->count[idx], val); 620 } 621 622 unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order, 623 gfp_t gfp_mask, The issue is that writeback doesn't hold a page reference and the page might get freed after PG_writeback is cleared (and the mapping is unlocked) in test_clear_page_writeback(). The stat functions looking up the page's node or zone are safe, as those attributes are static across allocation and free cycles. But page->mem_cgroup is not, and it will get cleared if we race with truncation or migration. It appears this race window has been around for a while, but less likely to trigger when the memcg stats were updated first thing after PG_writeback is cleared. Recent changes reshuffled this code to update the global node stats before the memcg ones, though, stretching the race window out to an extent where people can reproduce the problem. Update test_clear_page_writeback() to look up and pin page->mem_cgroup before clearing PG_writeback, then not use that pointer afterward. It is a partial revert of 62cccb8c8e7a ("mm: simplify lock_page_memcg()") but leaves the pageref-holding callsites that aren't affected alone. Link: http://lkml.kernel.org/r/20170809183825.GA26387@cmpxchg.org Fixes: 62cccb8c8e7a ("mm: simplify lock_page_memcg()") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Jaegeuk Kim <jaegeuk@kernel.org> Tested-by: Jaegeuk Kim <jaegeuk@kernel.org> Reported-by: Bradley Bolen <bradleybolen@gmail.com> Tested-by: Brad Bolen <bradleybolen@gmail.com> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> [4.6+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-08-18 22:15:48 +00:00
return memcg;
memcg: use new logic for page stat accounting Now, page-stat-per-memcg is recorded into per page_cgroup flag by duplicating page's status into the flag. The reason is that memcg has a feature to move a page from a group to another group and we have race between "move" and "page stat accounting", Under current logic, assume CPU-A and CPU-B. CPU-A does "move" and CPU-B does "page stat accounting". When CPU-A goes 1st, CPU-A CPU-B update "struct page" info. move_lock_mem_cgroup(memcg) see pc->flags copy page stat to new group overwrite pc->mem_cgroup. move_unlock_mem_cgroup(memcg) move_lock_mem_cgroup(mem) set pc->flags update page stat accounting move_unlock_mem_cgroup(mem) stat accounting is guarded by move_lock_mem_cgroup() and "move" logic (CPU-A) doesn't see changes in "struct page" information. But it's costly to have the same information both in 'struct page' and 'struct page_cgroup'. And, there is a potential problem. For example, assume we have PG_dirty accounting in memcg. PG_..is a flag for struct page. PCG_ is a flag for struct page_cgroup. (This is just an example. The same problem can be found in any kind of page stat accounting.) CPU-A CPU-B TestSet PG_dirty (delay) TestClear PG_dirty if (TestClear(PCG_dirty)) memcg->nr_dirty-- if (TestSet(PCG_dirty)) memcg->nr_dirty++ Here, memcg->nr_dirty = +1, this is wrong. This race was reported by Greg Thelen <gthelen@google.com>. Now, only FILE_MAPPED is supported but fortunately, it's serialized by page table lock and this is not real bug, _now_, If this potential problem is caused by having duplicated information in struct page and struct page_cgroup, we may be able to fix this by using original 'struct page' information. But we'll have a problem in "move account" Assume we use only PG_dirty. CPU-A CPU-B TestSet PG_dirty (delay) move_lock_mem_cgroup() if (PageDirty(page)) new_memcg->nr_dirty++ pc->mem_cgroup = new_memcg; move_unlock_mem_cgroup() move_lock_mem_cgroup() memcg = pc->mem_cgroup new_memcg->nr_dirty++ accounting information may be double-counted. This was original reason to have PCG_xxx flags but it seems PCG_xxx has another problem. I think we need a bigger lock as move_lock_mem_cgroup(page) TestSetPageDirty(page) update page stats (without any checks) move_unlock_mem_cgroup(page) This fixes both of problems and we don't have to duplicate page flag into page_cgroup. Please note: move_lock_mem_cgroup() is held only when there are possibility of "account move" under the system. So, in most path, status update will go without atomic locks. This patch introduces mem_cgroup_begin_update_page_stat() and mem_cgroup_end_update_page_stat() both should be called at modifying 'struct page' information if memcg takes care of it. as mem_cgroup_begin_update_page_stat() modify page information mem_cgroup_update_page_stat() => never check any 'struct page' info, just update counters. mem_cgroup_end_update_page_stat(). This patch is slow because we need to call begin_update_page_stat()/ end_update_page_stat() regardless of accounted will be changed or not. A following patch adds an easy optimization and reduces the cost. [akpm@linux-foundation.org: s/lock/locked/] [hughd@google.com: fix deadlock by avoiding stat lock when anon] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Greg Thelen <gthelen@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Ying Han <yinghan@google.com> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-21 23:34:25 +00:00
}
EXPORT_SYMBOL(lock_page_memcg);
memcg: use new logic for page stat accounting Now, page-stat-per-memcg is recorded into per page_cgroup flag by duplicating page's status into the flag. The reason is that memcg has a feature to move a page from a group to another group and we have race between "move" and "page stat accounting", Under current logic, assume CPU-A and CPU-B. CPU-A does "move" and CPU-B does "page stat accounting". When CPU-A goes 1st, CPU-A CPU-B update "struct page" info. move_lock_mem_cgroup(memcg) see pc->flags copy page stat to new group overwrite pc->mem_cgroup. move_unlock_mem_cgroup(memcg) move_lock_mem_cgroup(mem) set pc->flags update page stat accounting move_unlock_mem_cgroup(mem) stat accounting is guarded by move_lock_mem_cgroup() and "move" logic (CPU-A) doesn't see changes in "struct page" information. But it's costly to have the same information both in 'struct page' and 'struct page_cgroup'. And, there is a potential problem. For example, assume we have PG_dirty accounting in memcg. PG_..is a flag for struct page. PCG_ is a flag for struct page_cgroup. (This is just an example. The same problem can be found in any kind of page stat accounting.) CPU-A CPU-B TestSet PG_dirty (delay) TestClear PG_dirty if (TestClear(PCG_dirty)) memcg->nr_dirty-- if (TestSet(PCG_dirty)) memcg->nr_dirty++ Here, memcg->nr_dirty = +1, this is wrong. This race was reported by Greg Thelen <gthelen@google.com>. Now, only FILE_MAPPED is supported but fortunately, it's serialized by page table lock and this is not real bug, _now_, If this potential problem is caused by having duplicated information in struct page and struct page_cgroup, we may be able to fix this by using original 'struct page' information. But we'll have a problem in "move account" Assume we use only PG_dirty. CPU-A CPU-B TestSet PG_dirty (delay) move_lock_mem_cgroup() if (PageDirty(page)) new_memcg->nr_dirty++ pc->mem_cgroup = new_memcg; move_unlock_mem_cgroup() move_lock_mem_cgroup() memcg = pc->mem_cgroup new_memcg->nr_dirty++ accounting information may be double-counted. This was original reason to have PCG_xxx flags but it seems PCG_xxx has another problem. I think we need a bigger lock as move_lock_mem_cgroup(page) TestSetPageDirty(page) update page stats (without any checks) move_unlock_mem_cgroup(page) This fixes both of problems and we don't have to duplicate page flag into page_cgroup. Please note: move_lock_mem_cgroup() is held only when there are possibility of "account move" under the system. So, in most path, status update will go without atomic locks. This patch introduces mem_cgroup_begin_update_page_stat() and mem_cgroup_end_update_page_stat() both should be called at modifying 'struct page' information if memcg takes care of it. as mem_cgroup_begin_update_page_stat() modify page information mem_cgroup_update_page_stat() => never check any 'struct page' info, just update counters. mem_cgroup_end_update_page_stat(). This patch is slow because we need to call begin_update_page_stat()/ end_update_page_stat() regardless of accounted will be changed or not. A following patch adds an easy optimization and reduces the cost. [akpm@linux-foundation.org: s/lock/locked/] [hughd@google.com: fix deadlock by avoiding stat lock when anon] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Greg Thelen <gthelen@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Ying Han <yinghan@google.com> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-21 23:34:25 +00:00
mm: memcontrol: fix missed end-writeback page accounting Commit 0a31bc97c80c ("mm: memcontrol: rewrite uncharge API") changed page migration to uncharge the old page right away. The page is locked, unmapped, truncated, and off the LRU, but it could race with writeback ending, which then doesn't unaccount the page properly: test_clear_page_writeback() migration wait_on_page_writeback() TestClearPageWriteback() mem_cgroup_migrate() clear PCG_USED mem_cgroup_update_page_stat() if (PageCgroupUsed(pc)) decrease memcg pages under writeback release pc->mem_cgroup->move_lock The per-page statistics interface is heavily optimized to avoid a function call and a lookup_page_cgroup() in the file unmap fast path, which means it doesn't verify whether a page is still charged before clearing PageWriteback() and it has to do it in the stat update later. Rework it so that it looks up the page's memcg once at the beginning of the transaction and then uses it throughout. The charge will be verified before clearing PageWriteback() and migration can't uncharge the page as long as that is still set. The RCU lock will protect the memcg past uncharge. As far as losing the optimization goes, the following test results are from a microbenchmark that maps, faults, and unmaps a 4GB sparse file three times in a nested fashion, so that there are two negative passes that don't account but still go through the new transaction overhead. There is no actual difference: old: 33.195102545 seconds time elapsed ( +- 0.01% ) new: 33.199231369 seconds time elapsed ( +- 0.03% ) The time spent in page_remove_rmap()'s callees still adds up to the same, but the time spent in the function itself seems reduced: # Children Self Command Shared Object Symbol old: 0.12% 0.11% filemapstress [kernel.kallsyms] [k] page_remove_rmap new: 0.12% 0.08% filemapstress [kernel.kallsyms] [k] page_remove_rmap Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: <stable@vger.kernel.org> [3.17.x] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-29 21:50:48 +00:00
/**
mm: memcontrol: fix NULL pointer crash in test_clear_page_writeback() Jaegeuk and Brad report a NULL pointer crash when writeback ending tries to update the memcg stats: BUG: unable to handle kernel NULL pointer dereference at 00000000000003b0 IP: test_clear_page_writeback+0x12e/0x2c0 [...] RIP: 0010:test_clear_page_writeback+0x12e/0x2c0 Call Trace: <IRQ> end_page_writeback+0x47/0x70 f2fs_write_end_io+0x76/0x180 [f2fs] bio_endio+0x9f/0x120 blk_update_request+0xa8/0x2f0 scsi_end_request+0x39/0x1d0 scsi_io_completion+0x211/0x690 scsi_finish_command+0xd9/0x120 scsi_softirq_done+0x127/0x150 __blk_mq_complete_request_remote+0x13/0x20 flush_smp_call_function_queue+0x56/0x110 generic_smp_call_function_single_interrupt+0x13/0x30 smp_call_function_single_interrupt+0x27/0x40 call_function_single_interrupt+0x89/0x90 RIP: 0010:native_safe_halt+0x6/0x10 (gdb) l *(test_clear_page_writeback+0x12e) 0xffffffff811bae3e is in test_clear_page_writeback (./include/linux/memcontrol.h:619). 614 mod_node_page_state(page_pgdat(page), idx, val); 615 if (mem_cgroup_disabled() || !page->mem_cgroup) 616 return; 617 mod_memcg_state(page->mem_cgroup, idx, val); 618 pn = page->mem_cgroup->nodeinfo[page_to_nid(page)]; 619 this_cpu_add(pn->lruvec_stat->count[idx], val); 620 } 621 622 unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order, 623 gfp_t gfp_mask, The issue is that writeback doesn't hold a page reference and the page might get freed after PG_writeback is cleared (and the mapping is unlocked) in test_clear_page_writeback(). The stat functions looking up the page's node or zone are safe, as those attributes are static across allocation and free cycles. But page->mem_cgroup is not, and it will get cleared if we race with truncation or migration. It appears this race window has been around for a while, but less likely to trigger when the memcg stats were updated first thing after PG_writeback is cleared. Recent changes reshuffled this code to update the global node stats before the memcg ones, though, stretching the race window out to an extent where people can reproduce the problem. Update test_clear_page_writeback() to look up and pin page->mem_cgroup before clearing PG_writeback, then not use that pointer afterward. It is a partial revert of 62cccb8c8e7a ("mm: simplify lock_page_memcg()") but leaves the pageref-holding callsites that aren't affected alone. Link: http://lkml.kernel.org/r/20170809183825.GA26387@cmpxchg.org Fixes: 62cccb8c8e7a ("mm: simplify lock_page_memcg()") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Jaegeuk Kim <jaegeuk@kernel.org> Tested-by: Jaegeuk Kim <jaegeuk@kernel.org> Reported-by: Bradley Bolen <bradleybolen@gmail.com> Tested-by: Brad Bolen <bradleybolen@gmail.com> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> [4.6+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-08-18 22:15:48 +00:00
* __unlock_page_memcg - unlock and unpin a memcg
* @memcg: the memcg
*
* Unlock and unpin a memcg returned by lock_page_memcg().
mm: memcontrol: fix missed end-writeback page accounting Commit 0a31bc97c80c ("mm: memcontrol: rewrite uncharge API") changed page migration to uncharge the old page right away. The page is locked, unmapped, truncated, and off the LRU, but it could race with writeback ending, which then doesn't unaccount the page properly: test_clear_page_writeback() migration wait_on_page_writeback() TestClearPageWriteback() mem_cgroup_migrate() clear PCG_USED mem_cgroup_update_page_stat() if (PageCgroupUsed(pc)) decrease memcg pages under writeback release pc->mem_cgroup->move_lock The per-page statistics interface is heavily optimized to avoid a function call and a lookup_page_cgroup() in the file unmap fast path, which means it doesn't verify whether a page is still charged before clearing PageWriteback() and it has to do it in the stat update later. Rework it so that it looks up the page's memcg once at the beginning of the transaction and then uses it throughout. The charge will be verified before clearing PageWriteback() and migration can't uncharge the page as long as that is still set. The RCU lock will protect the memcg past uncharge. As far as losing the optimization goes, the following test results are from a microbenchmark that maps, faults, and unmaps a 4GB sparse file three times in a nested fashion, so that there are two negative passes that don't account but still go through the new transaction overhead. There is no actual difference: old: 33.195102545 seconds time elapsed ( +- 0.01% ) new: 33.199231369 seconds time elapsed ( +- 0.03% ) The time spent in page_remove_rmap()'s callees still adds up to the same, but the time spent in the function itself seems reduced: # Children Self Command Shared Object Symbol old: 0.12% 0.11% filemapstress [kernel.kallsyms] [k] page_remove_rmap new: 0.12% 0.08% filemapstress [kernel.kallsyms] [k] page_remove_rmap Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: <stable@vger.kernel.org> [3.17.x] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-29 21:50:48 +00:00
*/
mm: memcontrol: fix NULL pointer crash in test_clear_page_writeback() Jaegeuk and Brad report a NULL pointer crash when writeback ending tries to update the memcg stats: BUG: unable to handle kernel NULL pointer dereference at 00000000000003b0 IP: test_clear_page_writeback+0x12e/0x2c0 [...] RIP: 0010:test_clear_page_writeback+0x12e/0x2c0 Call Trace: <IRQ> end_page_writeback+0x47/0x70 f2fs_write_end_io+0x76/0x180 [f2fs] bio_endio+0x9f/0x120 blk_update_request+0xa8/0x2f0 scsi_end_request+0x39/0x1d0 scsi_io_completion+0x211/0x690 scsi_finish_command+0xd9/0x120 scsi_softirq_done+0x127/0x150 __blk_mq_complete_request_remote+0x13/0x20 flush_smp_call_function_queue+0x56/0x110 generic_smp_call_function_single_interrupt+0x13/0x30 smp_call_function_single_interrupt+0x27/0x40 call_function_single_interrupt+0x89/0x90 RIP: 0010:native_safe_halt+0x6/0x10 (gdb) l *(test_clear_page_writeback+0x12e) 0xffffffff811bae3e is in test_clear_page_writeback (./include/linux/memcontrol.h:619). 614 mod_node_page_state(page_pgdat(page), idx, val); 615 if (mem_cgroup_disabled() || !page->mem_cgroup) 616 return; 617 mod_memcg_state(page->mem_cgroup, idx, val); 618 pn = page->mem_cgroup->nodeinfo[page_to_nid(page)]; 619 this_cpu_add(pn->lruvec_stat->count[idx], val); 620 } 621 622 unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order, 623 gfp_t gfp_mask, The issue is that writeback doesn't hold a page reference and the page might get freed after PG_writeback is cleared (and the mapping is unlocked) in test_clear_page_writeback(). The stat functions looking up the page's node or zone are safe, as those attributes are static across allocation and free cycles. But page->mem_cgroup is not, and it will get cleared if we race with truncation or migration. It appears this race window has been around for a while, but less likely to trigger when the memcg stats were updated first thing after PG_writeback is cleared. Recent changes reshuffled this code to update the global node stats before the memcg ones, though, stretching the race window out to an extent where people can reproduce the problem. Update test_clear_page_writeback() to look up and pin page->mem_cgroup before clearing PG_writeback, then not use that pointer afterward. It is a partial revert of 62cccb8c8e7a ("mm: simplify lock_page_memcg()") but leaves the pageref-holding callsites that aren't affected alone. Link: http://lkml.kernel.org/r/20170809183825.GA26387@cmpxchg.org Fixes: 62cccb8c8e7a ("mm: simplify lock_page_memcg()") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Jaegeuk Kim <jaegeuk@kernel.org> Tested-by: Jaegeuk Kim <jaegeuk@kernel.org> Reported-by: Bradley Bolen <bradleybolen@gmail.com> Tested-by: Brad Bolen <bradleybolen@gmail.com> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> [4.6+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-08-18 22:15:48 +00:00
void __unlock_page_memcg(struct mem_cgroup *memcg)
memcg: use new logic for page stat accounting Now, page-stat-per-memcg is recorded into per page_cgroup flag by duplicating page's status into the flag. The reason is that memcg has a feature to move a page from a group to another group and we have race between "move" and "page stat accounting", Under current logic, assume CPU-A and CPU-B. CPU-A does "move" and CPU-B does "page stat accounting". When CPU-A goes 1st, CPU-A CPU-B update "struct page" info. move_lock_mem_cgroup(memcg) see pc->flags copy page stat to new group overwrite pc->mem_cgroup. move_unlock_mem_cgroup(memcg) move_lock_mem_cgroup(mem) set pc->flags update page stat accounting move_unlock_mem_cgroup(mem) stat accounting is guarded by move_lock_mem_cgroup() and "move" logic (CPU-A) doesn't see changes in "struct page" information. But it's costly to have the same information both in 'struct page' and 'struct page_cgroup'. And, there is a potential problem. For example, assume we have PG_dirty accounting in memcg. PG_..is a flag for struct page. PCG_ is a flag for struct page_cgroup. (This is just an example. The same problem can be found in any kind of page stat accounting.) CPU-A CPU-B TestSet PG_dirty (delay) TestClear PG_dirty if (TestClear(PCG_dirty)) memcg->nr_dirty-- if (TestSet(PCG_dirty)) memcg->nr_dirty++ Here, memcg->nr_dirty = +1, this is wrong. This race was reported by Greg Thelen <gthelen@google.com>. Now, only FILE_MAPPED is supported but fortunately, it's serialized by page table lock and this is not real bug, _now_, If this potential problem is caused by having duplicated information in struct page and struct page_cgroup, we may be able to fix this by using original 'struct page' information. But we'll have a problem in "move account" Assume we use only PG_dirty. CPU-A CPU-B TestSet PG_dirty (delay) move_lock_mem_cgroup() if (PageDirty(page)) new_memcg->nr_dirty++ pc->mem_cgroup = new_memcg; move_unlock_mem_cgroup() move_lock_mem_cgroup() memcg = pc->mem_cgroup new_memcg->nr_dirty++ accounting information may be double-counted. This was original reason to have PCG_xxx flags but it seems PCG_xxx has another problem. I think we need a bigger lock as move_lock_mem_cgroup(page) TestSetPageDirty(page) update page stats (without any checks) move_unlock_mem_cgroup(page) This fixes both of problems and we don't have to duplicate page flag into page_cgroup. Please note: move_lock_mem_cgroup() is held only when there are possibility of "account move" under the system. So, in most path, status update will go without atomic locks. This patch introduces mem_cgroup_begin_update_page_stat() and mem_cgroup_end_update_page_stat() both should be called at modifying 'struct page' information if memcg takes care of it. as mem_cgroup_begin_update_page_stat() modify page information mem_cgroup_update_page_stat() => never check any 'struct page' info, just update counters. mem_cgroup_end_update_page_stat(). This patch is slow because we need to call begin_update_page_stat()/ end_update_page_stat() regardless of accounted will be changed or not. A following patch adds an easy optimization and reduces the cost. [akpm@linux-foundation.org: s/lock/locked/] [hughd@google.com: fix deadlock by avoiding stat lock when anon] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Greg Thelen <gthelen@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Ying Han <yinghan@google.com> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-21 23:34:25 +00:00
{
if (memcg && memcg->move_lock_task == current) {
unsigned long flags = memcg->move_lock_flags;
memcg->move_lock_task = NULL;
memcg->move_lock_flags = 0;
spin_unlock_irqrestore(&memcg->move_lock, flags);
}
memcg: use new logic for page stat accounting Now, page-stat-per-memcg is recorded into per page_cgroup flag by duplicating page's status into the flag. The reason is that memcg has a feature to move a page from a group to another group and we have race between "move" and "page stat accounting", Under current logic, assume CPU-A and CPU-B. CPU-A does "move" and CPU-B does "page stat accounting". When CPU-A goes 1st, CPU-A CPU-B update "struct page" info. move_lock_mem_cgroup(memcg) see pc->flags copy page stat to new group overwrite pc->mem_cgroup. move_unlock_mem_cgroup(memcg) move_lock_mem_cgroup(mem) set pc->flags update page stat accounting move_unlock_mem_cgroup(mem) stat accounting is guarded by move_lock_mem_cgroup() and "move" logic (CPU-A) doesn't see changes in "struct page" information. But it's costly to have the same information both in 'struct page' and 'struct page_cgroup'. And, there is a potential problem. For example, assume we have PG_dirty accounting in memcg. PG_..is a flag for struct page. PCG_ is a flag for struct page_cgroup. (This is just an example. The same problem can be found in any kind of page stat accounting.) CPU-A CPU-B TestSet PG_dirty (delay) TestClear PG_dirty if (TestClear(PCG_dirty)) memcg->nr_dirty-- if (TestSet(PCG_dirty)) memcg->nr_dirty++ Here, memcg->nr_dirty = +1, this is wrong. This race was reported by Greg Thelen <gthelen@google.com>. Now, only FILE_MAPPED is supported but fortunately, it's serialized by page table lock and this is not real bug, _now_, If this potential problem is caused by having duplicated information in struct page and struct page_cgroup, we may be able to fix this by using original 'struct page' information. But we'll have a problem in "move account" Assume we use only PG_dirty. CPU-A CPU-B TestSet PG_dirty (delay) move_lock_mem_cgroup() if (PageDirty(page)) new_memcg->nr_dirty++ pc->mem_cgroup = new_memcg; move_unlock_mem_cgroup() move_lock_mem_cgroup() memcg = pc->mem_cgroup new_memcg->nr_dirty++ accounting information may be double-counted. This was original reason to have PCG_xxx flags but it seems PCG_xxx has another problem. I think we need a bigger lock as move_lock_mem_cgroup(page) TestSetPageDirty(page) update page stats (without any checks) move_unlock_mem_cgroup(page) This fixes both of problems and we don't have to duplicate page flag into page_cgroup. Please note: move_lock_mem_cgroup() is held only when there are possibility of "account move" under the system. So, in most path, status update will go without atomic locks. This patch introduces mem_cgroup_begin_update_page_stat() and mem_cgroup_end_update_page_stat() both should be called at modifying 'struct page' information if memcg takes care of it. as mem_cgroup_begin_update_page_stat() modify page information mem_cgroup_update_page_stat() => never check any 'struct page' info, just update counters. mem_cgroup_end_update_page_stat(). This patch is slow because we need to call begin_update_page_stat()/ end_update_page_stat() regardless of accounted will be changed or not. A following patch adds an easy optimization and reduces the cost. [akpm@linux-foundation.org: s/lock/locked/] [hughd@google.com: fix deadlock by avoiding stat lock when anon] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Greg Thelen <gthelen@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Ying Han <yinghan@google.com> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-21 23:34:25 +00:00
mm: memcontrol: fix missed end-writeback page accounting Commit 0a31bc97c80c ("mm: memcontrol: rewrite uncharge API") changed page migration to uncharge the old page right away. The page is locked, unmapped, truncated, and off the LRU, but it could race with writeback ending, which then doesn't unaccount the page properly: test_clear_page_writeback() migration wait_on_page_writeback() TestClearPageWriteback() mem_cgroup_migrate() clear PCG_USED mem_cgroup_update_page_stat() if (PageCgroupUsed(pc)) decrease memcg pages under writeback release pc->mem_cgroup->move_lock The per-page statistics interface is heavily optimized to avoid a function call and a lookup_page_cgroup() in the file unmap fast path, which means it doesn't verify whether a page is still charged before clearing PageWriteback() and it has to do it in the stat update later. Rework it so that it looks up the page's memcg once at the beginning of the transaction and then uses it throughout. The charge will be verified before clearing PageWriteback() and migration can't uncharge the page as long as that is still set. The RCU lock will protect the memcg past uncharge. As far as losing the optimization goes, the following test results are from a microbenchmark that maps, faults, and unmaps a 4GB sparse file three times in a nested fashion, so that there are two negative passes that don't account but still go through the new transaction overhead. There is no actual difference: old: 33.195102545 seconds time elapsed ( +- 0.01% ) new: 33.199231369 seconds time elapsed ( +- 0.03% ) The time spent in page_remove_rmap()'s callees still adds up to the same, but the time spent in the function itself seems reduced: # Children Self Command Shared Object Symbol old: 0.12% 0.11% filemapstress [kernel.kallsyms] [k] page_remove_rmap new: 0.12% 0.08% filemapstress [kernel.kallsyms] [k] page_remove_rmap Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: <stable@vger.kernel.org> [3.17.x] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-29 21:50:48 +00:00
rcu_read_unlock();
memcg: use new logic for page stat accounting Now, page-stat-per-memcg is recorded into per page_cgroup flag by duplicating page's status into the flag. The reason is that memcg has a feature to move a page from a group to another group and we have race between "move" and "page stat accounting", Under current logic, assume CPU-A and CPU-B. CPU-A does "move" and CPU-B does "page stat accounting". When CPU-A goes 1st, CPU-A CPU-B update "struct page" info. move_lock_mem_cgroup(memcg) see pc->flags copy page stat to new group overwrite pc->mem_cgroup. move_unlock_mem_cgroup(memcg) move_lock_mem_cgroup(mem) set pc->flags update page stat accounting move_unlock_mem_cgroup(mem) stat accounting is guarded by move_lock_mem_cgroup() and "move" logic (CPU-A) doesn't see changes in "struct page" information. But it's costly to have the same information both in 'struct page' and 'struct page_cgroup'. And, there is a potential problem. For example, assume we have PG_dirty accounting in memcg. PG_..is a flag for struct page. PCG_ is a flag for struct page_cgroup. (This is just an example. The same problem can be found in any kind of page stat accounting.) CPU-A CPU-B TestSet PG_dirty (delay) TestClear PG_dirty if (TestClear(PCG_dirty)) memcg->nr_dirty-- if (TestSet(PCG_dirty)) memcg->nr_dirty++ Here, memcg->nr_dirty = +1, this is wrong. This race was reported by Greg Thelen <gthelen@google.com>. Now, only FILE_MAPPED is supported but fortunately, it's serialized by page table lock and this is not real bug, _now_, If this potential problem is caused by having duplicated information in struct page and struct page_cgroup, we may be able to fix this by using original 'struct page' information. But we'll have a problem in "move account" Assume we use only PG_dirty. CPU-A CPU-B TestSet PG_dirty (delay) move_lock_mem_cgroup() if (PageDirty(page)) new_memcg->nr_dirty++ pc->mem_cgroup = new_memcg; move_unlock_mem_cgroup() move_lock_mem_cgroup() memcg = pc->mem_cgroup new_memcg->nr_dirty++ accounting information may be double-counted. This was original reason to have PCG_xxx flags but it seems PCG_xxx has another problem. I think we need a bigger lock as move_lock_mem_cgroup(page) TestSetPageDirty(page) update page stats (without any checks) move_unlock_mem_cgroup(page) This fixes both of problems and we don't have to duplicate page flag into page_cgroup. Please note: move_lock_mem_cgroup() is held only when there are possibility of "account move" under the system. So, in most path, status update will go without atomic locks. This patch introduces mem_cgroup_begin_update_page_stat() and mem_cgroup_end_update_page_stat() both should be called at modifying 'struct page' information if memcg takes care of it. as mem_cgroup_begin_update_page_stat() modify page information mem_cgroup_update_page_stat() => never check any 'struct page' info, just update counters. mem_cgroup_end_update_page_stat(). This patch is slow because we need to call begin_update_page_stat()/ end_update_page_stat() regardless of accounted will be changed or not. A following patch adds an easy optimization and reduces the cost. [akpm@linux-foundation.org: s/lock/locked/] [hughd@google.com: fix deadlock by avoiding stat lock when anon] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Greg Thelen <gthelen@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Ying Han <yinghan@google.com> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-21 23:34:25 +00:00
}
mm: memcontrol: fix NULL pointer crash in test_clear_page_writeback() Jaegeuk and Brad report a NULL pointer crash when writeback ending tries to update the memcg stats: BUG: unable to handle kernel NULL pointer dereference at 00000000000003b0 IP: test_clear_page_writeback+0x12e/0x2c0 [...] RIP: 0010:test_clear_page_writeback+0x12e/0x2c0 Call Trace: <IRQ> end_page_writeback+0x47/0x70 f2fs_write_end_io+0x76/0x180 [f2fs] bio_endio+0x9f/0x120 blk_update_request+0xa8/0x2f0 scsi_end_request+0x39/0x1d0 scsi_io_completion+0x211/0x690 scsi_finish_command+0xd9/0x120 scsi_softirq_done+0x127/0x150 __blk_mq_complete_request_remote+0x13/0x20 flush_smp_call_function_queue+0x56/0x110 generic_smp_call_function_single_interrupt+0x13/0x30 smp_call_function_single_interrupt+0x27/0x40 call_function_single_interrupt+0x89/0x90 RIP: 0010:native_safe_halt+0x6/0x10 (gdb) l *(test_clear_page_writeback+0x12e) 0xffffffff811bae3e is in test_clear_page_writeback (./include/linux/memcontrol.h:619). 614 mod_node_page_state(page_pgdat(page), idx, val); 615 if (mem_cgroup_disabled() || !page->mem_cgroup) 616 return; 617 mod_memcg_state(page->mem_cgroup, idx, val); 618 pn = page->mem_cgroup->nodeinfo[page_to_nid(page)]; 619 this_cpu_add(pn->lruvec_stat->count[idx], val); 620 } 621 622 unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order, 623 gfp_t gfp_mask, The issue is that writeback doesn't hold a page reference and the page might get freed after PG_writeback is cleared (and the mapping is unlocked) in test_clear_page_writeback(). The stat functions looking up the page's node or zone are safe, as those attributes are static across allocation and free cycles. But page->mem_cgroup is not, and it will get cleared if we race with truncation or migration. It appears this race window has been around for a while, but less likely to trigger when the memcg stats were updated first thing after PG_writeback is cleared. Recent changes reshuffled this code to update the global node stats before the memcg ones, though, stretching the race window out to an extent where people can reproduce the problem. Update test_clear_page_writeback() to look up and pin page->mem_cgroup before clearing PG_writeback, then not use that pointer afterward. It is a partial revert of 62cccb8c8e7a ("mm: simplify lock_page_memcg()") but leaves the pageref-holding callsites that aren't affected alone. Link: http://lkml.kernel.org/r/20170809183825.GA26387@cmpxchg.org Fixes: 62cccb8c8e7a ("mm: simplify lock_page_memcg()") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Jaegeuk Kim <jaegeuk@kernel.org> Tested-by: Jaegeuk Kim <jaegeuk@kernel.org> Reported-by: Bradley Bolen <bradleybolen@gmail.com> Tested-by: Brad Bolen <bradleybolen@gmail.com> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> [4.6+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-08-18 22:15:48 +00:00
/**
* unlock_page_memcg - unlock a page->mem_cgroup binding
* @page: the page
*/
void unlock_page_memcg(struct page *page)
{
struct page *head = compound_head(page);
__unlock_page_memcg(head->mem_cgroup);
mm: memcontrol: fix NULL pointer crash in test_clear_page_writeback() Jaegeuk and Brad report a NULL pointer crash when writeback ending tries to update the memcg stats: BUG: unable to handle kernel NULL pointer dereference at 00000000000003b0 IP: test_clear_page_writeback+0x12e/0x2c0 [...] RIP: 0010:test_clear_page_writeback+0x12e/0x2c0 Call Trace: <IRQ> end_page_writeback+0x47/0x70 f2fs_write_end_io+0x76/0x180 [f2fs] bio_endio+0x9f/0x120 blk_update_request+0xa8/0x2f0 scsi_end_request+0x39/0x1d0 scsi_io_completion+0x211/0x690 scsi_finish_command+0xd9/0x120 scsi_softirq_done+0x127/0x150 __blk_mq_complete_request_remote+0x13/0x20 flush_smp_call_function_queue+0x56/0x110 generic_smp_call_function_single_interrupt+0x13/0x30 smp_call_function_single_interrupt+0x27/0x40 call_function_single_interrupt+0x89/0x90 RIP: 0010:native_safe_halt+0x6/0x10 (gdb) l *(test_clear_page_writeback+0x12e) 0xffffffff811bae3e is in test_clear_page_writeback (./include/linux/memcontrol.h:619). 614 mod_node_page_state(page_pgdat(page), idx, val); 615 if (mem_cgroup_disabled() || !page->mem_cgroup) 616 return; 617 mod_memcg_state(page->mem_cgroup, idx, val); 618 pn = page->mem_cgroup->nodeinfo[page_to_nid(page)]; 619 this_cpu_add(pn->lruvec_stat->count[idx], val); 620 } 621 622 unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order, 623 gfp_t gfp_mask, The issue is that writeback doesn't hold a page reference and the page might get freed after PG_writeback is cleared (and the mapping is unlocked) in test_clear_page_writeback(). The stat functions looking up the page's node or zone are safe, as those attributes are static across allocation and free cycles. But page->mem_cgroup is not, and it will get cleared if we race with truncation or migration. It appears this race window has been around for a while, but less likely to trigger when the memcg stats were updated first thing after PG_writeback is cleared. Recent changes reshuffled this code to update the global node stats before the memcg ones, though, stretching the race window out to an extent where people can reproduce the problem. Update test_clear_page_writeback() to look up and pin page->mem_cgroup before clearing PG_writeback, then not use that pointer afterward. It is a partial revert of 62cccb8c8e7a ("mm: simplify lock_page_memcg()") but leaves the pageref-holding callsites that aren't affected alone. Link: http://lkml.kernel.org/r/20170809183825.GA26387@cmpxchg.org Fixes: 62cccb8c8e7a ("mm: simplify lock_page_memcg()") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Jaegeuk Kim <jaegeuk@kernel.org> Tested-by: Jaegeuk Kim <jaegeuk@kernel.org> Reported-by: Bradley Bolen <bradleybolen@gmail.com> Tested-by: Brad Bolen <bradleybolen@gmail.com> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: <stable@vger.kernel.org> [4.6+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-08-18 22:15:48 +00:00
}
EXPORT_SYMBOL(unlock_page_memcg);
memcg: use new logic for page stat accounting Now, page-stat-per-memcg is recorded into per page_cgroup flag by duplicating page's status into the flag. The reason is that memcg has a feature to move a page from a group to another group and we have race between "move" and "page stat accounting", Under current logic, assume CPU-A and CPU-B. CPU-A does "move" and CPU-B does "page stat accounting". When CPU-A goes 1st, CPU-A CPU-B update "struct page" info. move_lock_mem_cgroup(memcg) see pc->flags copy page stat to new group overwrite pc->mem_cgroup. move_unlock_mem_cgroup(memcg) move_lock_mem_cgroup(mem) set pc->flags update page stat accounting move_unlock_mem_cgroup(mem) stat accounting is guarded by move_lock_mem_cgroup() and "move" logic (CPU-A) doesn't see changes in "struct page" information. But it's costly to have the same information both in 'struct page' and 'struct page_cgroup'. And, there is a potential problem. For example, assume we have PG_dirty accounting in memcg. PG_..is a flag for struct page. PCG_ is a flag for struct page_cgroup. (This is just an example. The same problem can be found in any kind of page stat accounting.) CPU-A CPU-B TestSet PG_dirty (delay) TestClear PG_dirty if (TestClear(PCG_dirty)) memcg->nr_dirty-- if (TestSet(PCG_dirty)) memcg->nr_dirty++ Here, memcg->nr_dirty = +1, this is wrong. This race was reported by Greg Thelen <gthelen@google.com>. Now, only FILE_MAPPED is supported but fortunately, it's serialized by page table lock and this is not real bug, _now_, If this potential problem is caused by having duplicated information in struct page and struct page_cgroup, we may be able to fix this by using original 'struct page' information. But we'll have a problem in "move account" Assume we use only PG_dirty. CPU-A CPU-B TestSet PG_dirty (delay) move_lock_mem_cgroup() if (PageDirty(page)) new_memcg->nr_dirty++ pc->mem_cgroup = new_memcg; move_unlock_mem_cgroup() move_lock_mem_cgroup() memcg = pc->mem_cgroup new_memcg->nr_dirty++ accounting information may be double-counted. This was original reason to have PCG_xxx flags but it seems PCG_xxx has another problem. I think we need a bigger lock as move_lock_mem_cgroup(page) TestSetPageDirty(page) update page stats (without any checks) move_unlock_mem_cgroup(page) This fixes both of problems and we don't have to duplicate page flag into page_cgroup. Please note: move_lock_mem_cgroup() is held only when there are possibility of "account move" under the system. So, in most path, status update will go without atomic locks. This patch introduces mem_cgroup_begin_update_page_stat() and mem_cgroup_end_update_page_stat() both should be called at modifying 'struct page' information if memcg takes care of it. as mem_cgroup_begin_update_page_stat() modify page information mem_cgroup_update_page_stat() => never check any 'struct page' info, just update counters. mem_cgroup_end_update_page_stat(). This patch is slow because we need to call begin_update_page_stat()/ end_update_page_stat() regardless of accounted will be changed or not. A following patch adds an easy optimization and reduces the cost. [akpm@linux-foundation.org: s/lock/locked/] [hughd@google.com: fix deadlock by avoiding stat lock when anon] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Greg Thelen <gthelen@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: KOSAKI Motohiro <kosaki.motohiro@jp.fujitsu.com> Cc: Ying Han <yinghan@google.com> Signed-off-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-21 23:34:25 +00:00
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
struct memcg_stock_pcp {
struct mem_cgroup *cached; /* this never be root cgroup */
unsigned int nr_pages;
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
struct work_struct work;
memcg: fix percpu cached charge draining frequency For performance, memory cgroup caches some "charge" from res_counter into per cpu cache. This works well but because it's cache, it needs to be flushed in some cases. Typical cases are 1. when someone hit limit. 2. when rmdir() is called and need to charges to be 0. But "1" has problem. Recently, with large SMP machines, we see many kworker runs because of flushing memcg's cache. Bad things in implementation are that even if a cpu contains a cache for memcg not related to a memcg which hits limit, drain code is called. This patch does A) check percpu cache contains a useful data or not. B) check other asynchronous percpu draining doesn't run. C) don't call local cpu callback. (*)This patch avoid changing the calling condition with hard-limit. When I run "cat 1Gfile > /dev/null" under 300M limit memcg, [Before] 13767 kamezawa 20 0 98.6m 424 416 D 10.0 0.0 0:00.61 cat 58 root 20 0 0 0 0 S 0.6 0.0 0:00.09 kworker/2:1 60 root 20 0 0 0 0 S 0.6 0.0 0:00.08 kworker/4:1 4 root 20 0 0 0 0 S 0.3 0.0 0:00.02 kworker/0:0 57 root 20 0 0 0 0 S 0.3 0.0 0:00.05 kworker/1:1 61 root 20 0 0 0 0 S 0.3 0.0 0:00.05 kworker/5:1 62 root 20 0 0 0 0 S 0.3 0.0 0:00.05 kworker/6:1 63 root 20 0 0 0 0 S 0.3 0.0 0:00.05 kworker/7:1 [After] 2676 root 20 0 98.6m 416 416 D 9.3 0.0 0:00.87 cat 2626 kamezawa 20 0 15192 1312 920 R 0.3 0.0 0:00.28 top 1 root 20 0 19384 1496 1204 S 0.0 0.0 0:00.66 init 2 root 20 0 0 0 0 S 0.0 0.0 0:00.00 kthreadd 3 root 20 0 0 0 0 S 0.0 0.0 0:00.00 ksoftirqd/0 4 root 20 0 0 0 0 S 0.0 0.0 0:00.00 kworker/0:0 [akpm@linux-foundation.org: make percpu_charge_mutex static, tweak comments] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Reviewed-by: Michal Hocko <mhocko@suse.cz> Tested-by: Ying Han <yinghan@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-06-15 22:08:45 +00:00
unsigned long flags;
#define FLUSHING_CACHED_CHARGE 0
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
};
static DEFINE_PER_CPU(struct memcg_stock_pcp, memcg_stock);
Revert "memcg: get rid of percpu_charge_mutex lock" This reverts commit 8521fc50d433507a7cdc96bec280f9e5888a54cc. The patch incorrectly assumes that using atomic FLUSHING_CACHED_CHARGE bit operations is sufficient but that is not true. Johannes Weiner has reported a crash during parallel memory cgroup removal: BUG: unable to handle kernel NULL pointer dereference at 0000000000000018 IP: [<ffffffff81083b70>] css_is_ancestor+0x20/0x70 Oops: 0000 [#1] PREEMPT SMP Pid: 19677, comm: rmdir Tainted: G W 3.0.0-mm1-00188-gf38d32b #35 ECS MCP61M-M3/MCP61M-M3 RIP: 0010:[<ffffffff81083b70>] css_is_ancestor+0x20/0x70 RSP: 0018:ffff880077b09c88 EFLAGS: 00010202 Process rmdir (pid: 19677, threadinfo ffff880077b08000, task ffff8800781bb310) Call Trace: [<ffffffff810feba3>] mem_cgroup_same_or_subtree+0x33/0x40 [<ffffffff810feccf>] drain_all_stock+0x11f/0x170 [<ffffffff81103211>] mem_cgroup_force_empty+0x231/0x6d0 [<ffffffff811036c4>] mem_cgroup_pre_destroy+0x14/0x20 [<ffffffff81080559>] cgroup_rmdir+0xb9/0x500 [<ffffffff81114d26>] vfs_rmdir+0x86/0xe0 [<ffffffff81114e7b>] do_rmdir+0xfb/0x110 [<ffffffff81114ea6>] sys_rmdir+0x16/0x20 [<ffffffff8154d76b>] system_call_fastpath+0x16/0x1b We are crashing because we try to dereference cached memcg when we are checking whether we should wait for draining on the cache. The cache is already cleaned up, though. There is also a theoretical chance that the cached memcg gets freed between we test for the FLUSHING_CACHED_CHARGE and dereference it in mem_cgroup_same_or_subtree: CPU0 CPU1 CPU2 mem=stock->cached stock->cached=NULL clear_bit test_and_set_bit test_bit() ... <preempted> mem_cgroup_destroy use after free The percpu_charge_mutex protected from this race because sync draining is exclusive. It is safer to revert now and come up with a more parallel implementation later. Signed-off-by: Michal Hocko <mhocko@suse.cz> Reported-by: Johannes Weiner <jweiner@redhat.com> Acked-by: Johannes Weiner <jweiner@redhat.com> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: stable@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-08-09 09:56:26 +00:00
static DEFINE_MUTEX(percpu_charge_mutex);
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
/**
* consume_stock: Try to consume stocked charge on this cpu.
* @memcg: memcg to consume from.
* @nr_pages: how many pages to charge.
*
* The charges will only happen if @memcg matches the current cpu's memcg
* stock, and at least @nr_pages are available in that stock. Failure to
* service an allocation will refill the stock.
*
* returns true if successful, false otherwise.
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
*/
static bool consume_stock(struct mem_cgroup *memcg, unsigned int nr_pages)
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
{
struct memcg_stock_pcp *stock;
2016-09-19 21:44:36 +00:00
unsigned long flags;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
bool ret = false;
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
mm: memcontrol: fix excessive complexity in memory.stat reporting We've seen memory.stat reads in top-level cgroups take up to fourteen seconds during a userspace bug that created tens of thousands of ghost cgroups pinned by lingering page cache. Even with a more reasonable number of cgroups, aggregating memory.stat is unnecessarily heavy. The complexity is this: nr_cgroups * nr_stat_items * nr_possible_cpus where the stat items are ~70 at this point. With 128 cgroups and 128 CPUs - decent, not enormous setups - reading the top-level memory.stat has to aggregate over a million per-cpu counters. This doesn't scale. Instead of spreading the source of truth across all CPUs, use the per-cpu counters merely to batch updates to shared atomic counters. This is the same as the per-cpu stocks we use for charging memory to the shared atomic page_counters, and also the way the global vmstat counters are implemented. Vmstat has elaborate spilling thresholds that depend on the number of CPUs, amount of memory, and memory pressure - carefully balancing the cost of counter updates with the amount of per-cpu error. That's because the vmstat counters are system-wide, but also used for decisions inside the kernel (e.g. NR_FREE_PAGES in the allocator). Neither is true for the memory controller. Use the same static batch size we already use for page_counter updates during charging. The per-cpu error in the stats will be 128k, which is an acceptable ratio of cores to memory accounting granularity. [hannes@cmpxchg.org: fix warning in __this_cpu_xchg() calls] Link: http://lkml.kernel.org/r/20171201135750.GB8097@cmpxchg.org Link: http://lkml.kernel.org/r/20171103153336.24044-3-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Michal Hocko <mhocko@suse.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-02-01 00:16:45 +00:00
if (nr_pages > MEMCG_CHARGE_BATCH)
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
return ret;
2016-09-19 21:44:36 +00:00
local_irq_save(flags);
stock = this_cpu_ptr(&memcg_stock);
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
if (memcg == stock->cached && stock->nr_pages >= nr_pages) {
stock->nr_pages -= nr_pages;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
ret = true;
}
2016-09-19 21:44:36 +00:00
local_irq_restore(flags);
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
return ret;
}
/*
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
* Returns stocks cached in percpu and reset cached information.
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
*/
static void drain_stock(struct memcg_stock_pcp *stock)
{
struct mem_cgroup *old = stock->cached;
if (stock->nr_pages) {
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
page_counter_uncharge(&old->memory, stock->nr_pages);
if (do_memsw_account())
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
page_counter_uncharge(&old->memsw, stock->nr_pages);
css_put_many(&old->css, stock->nr_pages);
stock->nr_pages = 0;
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
}
stock->cached = NULL;
}
static void drain_local_stock(struct work_struct *dummy)
{
2016-09-19 21:44:36 +00:00
struct memcg_stock_pcp *stock;
unsigned long flags;
mm, memcg: remove hotplug locking from try_charge The following lockdep splat has been noticed during LTP testing ====================================================== WARNING: possible circular locking dependency detected 4.13.0-rc3-next-20170807 #12 Not tainted ------------------------------------------------------ a.out/4771 is trying to acquire lock: (cpu_hotplug_lock.rw_sem){++++++}, at: [<ffffffff812b4668>] drain_all_stock.part.35+0x18/0x140 but task is already holding lock: (&mm->mmap_sem){++++++}, at: [<ffffffff8106eb35>] __do_page_fault+0x175/0x530 which lock already depends on the new lock. the existing dependency chain (in reverse order) is: -> #3 (&mm->mmap_sem){++++++}: lock_acquire+0xc9/0x230 __might_fault+0x70/0xa0 _copy_to_user+0x23/0x70 filldir+0xa7/0x110 xfs_dir2_sf_getdents.isra.10+0x20c/0x2c0 [xfs] xfs_readdir+0x1fa/0x2c0 [xfs] xfs_file_readdir+0x30/0x40 [xfs] iterate_dir+0x17a/0x1a0 SyS_getdents+0xb0/0x160 entry_SYSCALL_64_fastpath+0x1f/0xbe -> #2 (&type->i_mutex_dir_key#3){++++++}: lock_acquire+0xc9/0x230 down_read+0x51/0xb0 lookup_slow+0xde/0x210 walk_component+0x160/0x250 link_path_walk+0x1a6/0x610 path_openat+0xe4/0xd50 do_filp_open+0x91/0x100 file_open_name+0xf5/0x130 filp_open+0x33/0x50 kernel_read_file_from_path+0x39/0x80 _request_firmware+0x39f/0x880 request_firmware_direct+0x37/0x50 request_microcode_fw+0x64/0xe0 reload_store+0xf7/0x180 dev_attr_store+0x18/0x30 sysfs_kf_write+0x44/0x60 kernfs_fop_write+0x113/0x1a0 __vfs_write+0x37/0x170 vfs_write+0xc7/0x1c0 SyS_write+0x58/0xc0 do_syscall_64+0x6c/0x1f0 return_from_SYSCALL_64+0x0/0x7a -> #1 (microcode_mutex){+.+.+.}: lock_acquire+0xc9/0x230 __mutex_lock+0x88/0x960 mutex_lock_nested+0x1b/0x20 microcode_init+0xbb/0x208 do_one_initcall+0x51/0x1a9 kernel_init_freeable+0x208/0x2a7 kernel_init+0xe/0x104 ret_from_fork+0x2a/0x40 -> #0 (cpu_hotplug_lock.rw_sem){++++++}: __lock_acquire+0x153c/0x1550 lock_acquire+0xc9/0x230 cpus_read_lock+0x4b/0x90 drain_all_stock.part.35+0x18/0x140 try_charge+0x3ab/0x6e0 mem_cgroup_try_charge+0x7f/0x2c0 shmem_getpage_gfp+0x25f/0x1050 shmem_fault+0x96/0x200 __do_fault+0x1e/0xa0 __handle_mm_fault+0x9c3/0xe00 handle_mm_fault+0x16e/0x380 __do_page_fault+0x24a/0x530 do_page_fault+0x30/0x80 page_fault+0x28/0x30 other info that might help us debug this: Chain exists of: cpu_hotplug_lock.rw_sem --> &type->i_mutex_dir_key#3 --> &mm->mmap_sem Possible unsafe locking scenario: CPU0 CPU1 ---- ---- lock(&mm->mmap_sem); lock(&type->i_mutex_dir_key#3); lock(&mm->mmap_sem); lock(cpu_hotplug_lock.rw_sem); *** DEADLOCK *** 2 locks held by a.out/4771: #0: (&mm->mmap_sem){++++++}, at: [<ffffffff8106eb35>] __do_page_fault+0x175/0x530 #1: (percpu_charge_mutex){+.+...}, at: [<ffffffff812b4c97>] try_charge+0x397/0x6e0 The problem is very similar to the one fixed by commit a459eeb7b852 ("mm, page_alloc: do not depend on cpu hotplug locks inside the allocator"). We are taking hotplug locks while we can be sitting on top of basically arbitrary locks. This just calls for problems. We can get rid of {get,put}_online_cpus, fortunately. We do not have to be worried about races with memory hotplug because drain_local_stock, which is called from both the WQ draining and the memory hotplug contexts, is always operating on the local cpu stock with IRQs disabled. The only thing to be careful about is that the target memcg doesn't vanish while we are still in drain_all_stock so take a reference on it. Link: http://lkml.kernel.org/r/20170913090023.28322-1-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Reported-by: Artem Savkov <asavkov@redhat.com> Tested-by: Artem Savkov <asavkov@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-10-03 23:14:53 +00:00
/*
* The only protection from memory hotplug vs. drain_stock races is
* that we always operate on local CPU stock here with IRQ disabled
*/
2016-09-19 21:44:36 +00:00
local_irq_save(flags);
stock = this_cpu_ptr(&memcg_stock);
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
drain_stock(stock);
memcg: fix percpu cached charge draining frequency For performance, memory cgroup caches some "charge" from res_counter into per cpu cache. This works well but because it's cache, it needs to be flushed in some cases. Typical cases are 1. when someone hit limit. 2. when rmdir() is called and need to charges to be 0. But "1" has problem. Recently, with large SMP machines, we see many kworker runs because of flushing memcg's cache. Bad things in implementation are that even if a cpu contains a cache for memcg not related to a memcg which hits limit, drain code is called. This patch does A) check percpu cache contains a useful data or not. B) check other asynchronous percpu draining doesn't run. C) don't call local cpu callback. (*)This patch avoid changing the calling condition with hard-limit. When I run "cat 1Gfile > /dev/null" under 300M limit memcg, [Before] 13767 kamezawa 20 0 98.6m 424 416 D 10.0 0.0 0:00.61 cat 58 root 20 0 0 0 0 S 0.6 0.0 0:00.09 kworker/2:1 60 root 20 0 0 0 0 S 0.6 0.0 0:00.08 kworker/4:1 4 root 20 0 0 0 0 S 0.3 0.0 0:00.02 kworker/0:0 57 root 20 0 0 0 0 S 0.3 0.0 0:00.05 kworker/1:1 61 root 20 0 0 0 0 S 0.3 0.0 0:00.05 kworker/5:1 62 root 20 0 0 0 0 S 0.3 0.0 0:00.05 kworker/6:1 63 root 20 0 0 0 0 S 0.3 0.0 0:00.05 kworker/7:1 [After] 2676 root 20 0 98.6m 416 416 D 9.3 0.0 0:00.87 cat 2626 kamezawa 20 0 15192 1312 920 R 0.3 0.0 0:00.28 top 1 root 20 0 19384 1496 1204 S 0.0 0.0 0:00.66 init 2 root 20 0 0 0 0 S 0.0 0.0 0:00.00 kthreadd 3 root 20 0 0 0 0 S 0.0 0.0 0:00.00 ksoftirqd/0 4 root 20 0 0 0 0 S 0.0 0.0 0:00.00 kworker/0:0 [akpm@linux-foundation.org: make percpu_charge_mutex static, tweak comments] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Reviewed-by: Michal Hocko <mhocko@suse.cz> Tested-by: Ying Han <yinghan@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-06-15 22:08:45 +00:00
clear_bit(FLUSHING_CACHED_CHARGE, &stock->flags);
2016-09-19 21:44:36 +00:00
local_irq_restore(flags);
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
}
/*
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
* Cache charges(val) to local per_cpu area.
* This will be consumed by consume_stock() function, later.
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
*/
static void refill_stock(struct mem_cgroup *memcg, unsigned int nr_pages)
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
{
2016-09-19 21:44:36 +00:00
struct memcg_stock_pcp *stock;
unsigned long flags;
local_irq_save(flags);
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
2016-09-19 21:44:36 +00:00
stock = this_cpu_ptr(&memcg_stock);
if (stock->cached != memcg) { /* reset if necessary */
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
drain_stock(stock);
stock->cached = memcg;
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
}
stock->nr_pages += nr_pages;
2016-09-19 21:44:36 +00:00
mm: memcontrol: fix excessive complexity in memory.stat reporting We've seen memory.stat reads in top-level cgroups take up to fourteen seconds during a userspace bug that created tens of thousands of ghost cgroups pinned by lingering page cache. Even with a more reasonable number of cgroups, aggregating memory.stat is unnecessarily heavy. The complexity is this: nr_cgroups * nr_stat_items * nr_possible_cpus where the stat items are ~70 at this point. With 128 cgroups and 128 CPUs - decent, not enormous setups - reading the top-level memory.stat has to aggregate over a million per-cpu counters. This doesn't scale. Instead of spreading the source of truth across all CPUs, use the per-cpu counters merely to batch updates to shared atomic counters. This is the same as the per-cpu stocks we use for charging memory to the shared atomic page_counters, and also the way the global vmstat counters are implemented. Vmstat has elaborate spilling thresholds that depend on the number of CPUs, amount of memory, and memory pressure - carefully balancing the cost of counter updates with the amount of per-cpu error. That's because the vmstat counters are system-wide, but also used for decisions inside the kernel (e.g. NR_FREE_PAGES in the allocator). Neither is true for the memory controller. Use the same static batch size we already use for page_counter updates during charging. The per-cpu error in the stats will be 128k, which is an acceptable ratio of cores to memory accounting granularity. [hannes@cmpxchg.org: fix warning in __this_cpu_xchg() calls] Link: http://lkml.kernel.org/r/20171201135750.GB8097@cmpxchg.org Link: http://lkml.kernel.org/r/20171103153336.24044-3-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Michal Hocko <mhocko@suse.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-02-01 00:16:45 +00:00
if (stock->nr_pages > MEMCG_CHARGE_BATCH)
mm: memcontrol: use per-cpu stocks for socket memory uncharging We've noticed a quite noticeable performance overhead on some hosts with significant network traffic when socket memory accounting is enabled. Perf top shows that socket memory uncharging path is hot: 2.13% [kernel] [k] page_counter_cancel 1.14% [kernel] [k] __sk_mem_reduce_allocated 1.14% [kernel] [k] _raw_spin_lock 0.87% [kernel] [k] _raw_spin_lock_irqsave 0.84% [kernel] [k] tcp_ack 0.84% [kernel] [k] ixgbe_poll 0.83% < workload > 0.82% [kernel] [k] enqueue_entity 0.68% [kernel] [k] __fget 0.68% [kernel] [k] tcp_delack_timer_handler 0.67% [kernel] [k] __schedule 0.60% < workload > 0.59% [kernel] [k] __inet6_lookup_established 0.55% [kernel] [k] __switch_to 0.55% [kernel] [k] menu_select 0.54% libc-2.20.so [.] __memcpy_avx_unaligned To address this issue, the existing per-cpu stock infrastructure can be used. refill_stock() can be called from mem_cgroup_uncharge_skmem() to move charge to a per-cpu stock instead of calling atomic page_counter_uncharge(). To prevent the uncontrolled growth of per-cpu stocks, refill_stock() will explicitly drain the cached charge, if the cached value exceeds CHARGE_BATCH. This allows significantly optimize the load: 1.21% [kernel] [k] _raw_spin_lock 1.01% [kernel] [k] ixgbe_poll 0.92% [kernel] [k] _raw_spin_lock_irqsave 0.90% [kernel] [k] enqueue_entity 0.86% [kernel] [k] tcp_ack 0.85% < workload > 0.74% perf-11120.map [.] 0x000000000061bf24 0.73% [kernel] [k] __schedule 0.67% [kernel] [k] __fget 0.63% [kernel] [k] __inet6_lookup_established 0.62% [kernel] [k] menu_select 0.59% < workload > 0.59% [kernel] [k] __switch_to 0.57% libc-2.20.so [.] __memcpy_avx_unaligned Link: http://lkml.kernel.org/r/20170829100150.4580-1-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-09-08 23:13:09 +00:00
drain_stock(stock);
2016-09-19 21:44:36 +00:00
local_irq_restore(flags);
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
}
/*
* Drains all per-CPU charge caches for given root_memcg resp. subtree
* of the hierarchy under it.
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
*/
static void drain_all_stock(struct mem_cgroup *root_memcg)
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
{
memcg: fix percpu cached charge draining frequency For performance, memory cgroup caches some "charge" from res_counter into per cpu cache. This works well but because it's cache, it needs to be flushed in some cases. Typical cases are 1. when someone hit limit. 2. when rmdir() is called and need to charges to be 0. But "1" has problem. Recently, with large SMP machines, we see many kworker runs because of flushing memcg's cache. Bad things in implementation are that even if a cpu contains a cache for memcg not related to a memcg which hits limit, drain code is called. This patch does A) check percpu cache contains a useful data or not. B) check other asynchronous percpu draining doesn't run. C) don't call local cpu callback. (*)This patch avoid changing the calling condition with hard-limit. When I run "cat 1Gfile > /dev/null" under 300M limit memcg, [Before] 13767 kamezawa 20 0 98.6m 424 416 D 10.0 0.0 0:00.61 cat 58 root 20 0 0 0 0 S 0.6 0.0 0:00.09 kworker/2:1 60 root 20 0 0 0 0 S 0.6 0.0 0:00.08 kworker/4:1 4 root 20 0 0 0 0 S 0.3 0.0 0:00.02 kworker/0:0 57 root 20 0 0 0 0 S 0.3 0.0 0:00.05 kworker/1:1 61 root 20 0 0 0 0 S 0.3 0.0 0:00.05 kworker/5:1 62 root 20 0 0 0 0 S 0.3 0.0 0:00.05 kworker/6:1 63 root 20 0 0 0 0 S 0.3 0.0 0:00.05 kworker/7:1 [After] 2676 root 20 0 98.6m 416 416 D 9.3 0.0 0:00.87 cat 2626 kamezawa 20 0 15192 1312 920 R 0.3 0.0 0:00.28 top 1 root 20 0 19384 1496 1204 S 0.0 0.0 0:00.66 init 2 root 20 0 0 0 0 S 0.0 0.0 0:00.00 kthreadd 3 root 20 0 0 0 0 S 0.0 0.0 0:00.00 ksoftirqd/0 4 root 20 0 0 0 0 S 0.0 0.0 0:00.00 kworker/0:0 [akpm@linux-foundation.org: make percpu_charge_mutex static, tweak comments] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Reviewed-by: Michal Hocko <mhocko@suse.cz> Tested-by: Ying Han <yinghan@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-06-15 22:08:45 +00:00
int cpu, curcpu;
memcg: unify sync and async per-cpu charge cache draining Currently we have two ways how to drain per-CPU caches for charges. drain_all_stock_sync will synchronously drain all caches while drain_all_stock_async will asynchronously drain only those that refer to a given memory cgroup or its subtree in hierarchy. Targeted async draining has been introduced by 26fe6168 (memcg: fix percpu cached charge draining frequency) to reduce the cpu workers number. sync draining is currently triggered only from mem_cgroup_force_empty which is triggered only by userspace (mem_cgroup_force_empty_write) or when a cgroup is removed (mem_cgroup_pre_destroy). Although these are not usually frequent operations it still makes some sense to do targeted draining as well, especially if the box has many CPUs. This patch unifies both methods to use the single code (drain_all_stock) which relies on the original async implementation and just adds flush_work to wait on all caches that are still under work for the sync mode. We are using FLUSHING_CACHED_CHARGE bit check to prevent from waiting on a work that we haven't triggered. Please note that both sync and async functions are currently protected by percpu_charge_mutex so we cannot race with other drainers. Signed-off-by: Michal Hocko <mhocko@suse.cz> Reviewed-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <bsingharora@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-07-26 23:08:28 +00:00
/* If someone's already draining, avoid adding running more workers. */
if (!mutex_trylock(&percpu_charge_mutex))
return;
mm, memcg: remove hotplug locking from try_charge The following lockdep splat has been noticed during LTP testing ====================================================== WARNING: possible circular locking dependency detected 4.13.0-rc3-next-20170807 #12 Not tainted ------------------------------------------------------ a.out/4771 is trying to acquire lock: (cpu_hotplug_lock.rw_sem){++++++}, at: [<ffffffff812b4668>] drain_all_stock.part.35+0x18/0x140 but task is already holding lock: (&mm->mmap_sem){++++++}, at: [<ffffffff8106eb35>] __do_page_fault+0x175/0x530 which lock already depends on the new lock. the existing dependency chain (in reverse order) is: -> #3 (&mm->mmap_sem){++++++}: lock_acquire+0xc9/0x230 __might_fault+0x70/0xa0 _copy_to_user+0x23/0x70 filldir+0xa7/0x110 xfs_dir2_sf_getdents.isra.10+0x20c/0x2c0 [xfs] xfs_readdir+0x1fa/0x2c0 [xfs] xfs_file_readdir+0x30/0x40 [xfs] iterate_dir+0x17a/0x1a0 SyS_getdents+0xb0/0x160 entry_SYSCALL_64_fastpath+0x1f/0xbe -> #2 (&type->i_mutex_dir_key#3){++++++}: lock_acquire+0xc9/0x230 down_read+0x51/0xb0 lookup_slow+0xde/0x210 walk_component+0x160/0x250 link_path_walk+0x1a6/0x610 path_openat+0xe4/0xd50 do_filp_open+0x91/0x100 file_open_name+0xf5/0x130 filp_open+0x33/0x50 kernel_read_file_from_path+0x39/0x80 _request_firmware+0x39f/0x880 request_firmware_direct+0x37/0x50 request_microcode_fw+0x64/0xe0 reload_store+0xf7/0x180 dev_attr_store+0x18/0x30 sysfs_kf_write+0x44/0x60 kernfs_fop_write+0x113/0x1a0 __vfs_write+0x37/0x170 vfs_write+0xc7/0x1c0 SyS_write+0x58/0xc0 do_syscall_64+0x6c/0x1f0 return_from_SYSCALL_64+0x0/0x7a -> #1 (microcode_mutex){+.+.+.}: lock_acquire+0xc9/0x230 __mutex_lock+0x88/0x960 mutex_lock_nested+0x1b/0x20 microcode_init+0xbb/0x208 do_one_initcall+0x51/0x1a9 kernel_init_freeable+0x208/0x2a7 kernel_init+0xe/0x104 ret_from_fork+0x2a/0x40 -> #0 (cpu_hotplug_lock.rw_sem){++++++}: __lock_acquire+0x153c/0x1550 lock_acquire+0xc9/0x230 cpus_read_lock+0x4b/0x90 drain_all_stock.part.35+0x18/0x140 try_charge+0x3ab/0x6e0 mem_cgroup_try_charge+0x7f/0x2c0 shmem_getpage_gfp+0x25f/0x1050 shmem_fault+0x96/0x200 __do_fault+0x1e/0xa0 __handle_mm_fault+0x9c3/0xe00 handle_mm_fault+0x16e/0x380 __do_page_fault+0x24a/0x530 do_page_fault+0x30/0x80 page_fault+0x28/0x30 other info that might help us debug this: Chain exists of: cpu_hotplug_lock.rw_sem --> &type->i_mutex_dir_key#3 --> &mm->mmap_sem Possible unsafe locking scenario: CPU0 CPU1 ---- ---- lock(&mm->mmap_sem); lock(&type->i_mutex_dir_key#3); lock(&mm->mmap_sem); lock(cpu_hotplug_lock.rw_sem); *** DEADLOCK *** 2 locks held by a.out/4771: #0: (&mm->mmap_sem){++++++}, at: [<ffffffff8106eb35>] __do_page_fault+0x175/0x530 #1: (percpu_charge_mutex){+.+...}, at: [<ffffffff812b4c97>] try_charge+0x397/0x6e0 The problem is very similar to the one fixed by commit a459eeb7b852 ("mm, page_alloc: do not depend on cpu hotplug locks inside the allocator"). We are taking hotplug locks while we can be sitting on top of basically arbitrary locks. This just calls for problems. We can get rid of {get,put}_online_cpus, fortunately. We do not have to be worried about races with memory hotplug because drain_local_stock, which is called from both the WQ draining and the memory hotplug contexts, is always operating on the local cpu stock with IRQs disabled. The only thing to be careful about is that the target memcg doesn't vanish while we are still in drain_all_stock so take a reference on it. Link: http://lkml.kernel.org/r/20170913090023.28322-1-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Reported-by: Artem Savkov <asavkov@redhat.com> Tested-by: Artem Savkov <asavkov@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Thomas Gleixner <tglx@linutronix.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-10-03 23:14:53 +00:00
/*
* Notify other cpus that system-wide "drain" is running
* We do not care about races with the cpu hotplug because cpu down
* as well as workers from this path always operate on the local
* per-cpu data. CPU up doesn't touch memcg_stock at all.
*/
curcpu = get_cpu();
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
for_each_online_cpu(cpu) {
struct memcg_stock_pcp *stock = &per_cpu(memcg_stock, cpu);
struct mem_cgroup *memcg;
mm: memcontrol: switch to rcu protection in drain_all_stock() Commit 72f0184c8a00 ("mm, memcg: remove hotplug locking from try_charge") introduced css_tryget()/css_put() calls in drain_all_stock(), which are supposed to protect the target memory cgroup from being released during the mem_cgroup_is_descendant() call. However, it's not completely safe. In theory, memcg can go away between reading stock->cached pointer and calling css_tryget(). This can happen if drain_all_stock() races with drain_local_stock() performed on the remote cpu as a result of a work, scheduled by the previous invocation of drain_all_stock(). The race is a bit theoretical and there are few chances to trigger it, but the current code looks a bit confusing, so it makes sense to fix it anyway. The code looks like as if css_tryget() and css_put() are used to protect stocks drainage. It's not necessary because stocked pages are holding references to the cached cgroup. And it obviously won't work for works, scheduled on other cpus. So, let's read the stock->cached pointer and evaluate the memory cgroup inside a rcu read section, and get rid of css_tryget()/css_put() calls. Link: http://lkml.kernel.org/r/20190802192241.3253165-1-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Hillf Danton <hdanton@sina.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-09-23 22:34:58 +00:00
bool flush = false;
memcg: fix percpu cached charge draining frequency For performance, memory cgroup caches some "charge" from res_counter into per cpu cache. This works well but because it's cache, it needs to be flushed in some cases. Typical cases are 1. when someone hit limit. 2. when rmdir() is called and need to charges to be 0. But "1" has problem. Recently, with large SMP machines, we see many kworker runs because of flushing memcg's cache. Bad things in implementation are that even if a cpu contains a cache for memcg not related to a memcg which hits limit, drain code is called. This patch does A) check percpu cache contains a useful data or not. B) check other asynchronous percpu draining doesn't run. C) don't call local cpu callback. (*)This patch avoid changing the calling condition with hard-limit. When I run "cat 1Gfile > /dev/null" under 300M limit memcg, [Before] 13767 kamezawa 20 0 98.6m 424 416 D 10.0 0.0 0:00.61 cat 58 root 20 0 0 0 0 S 0.6 0.0 0:00.09 kworker/2:1 60 root 20 0 0 0 0 S 0.6 0.0 0:00.08 kworker/4:1 4 root 20 0 0 0 0 S 0.3 0.0 0:00.02 kworker/0:0 57 root 20 0 0 0 0 S 0.3 0.0 0:00.05 kworker/1:1 61 root 20 0 0 0 0 S 0.3 0.0 0:00.05 kworker/5:1 62 root 20 0 0 0 0 S 0.3 0.0 0:00.05 kworker/6:1 63 root 20 0 0 0 0 S 0.3 0.0 0:00.05 kworker/7:1 [After] 2676 root 20 0 98.6m 416 416 D 9.3 0.0 0:00.87 cat 2626 kamezawa 20 0 15192 1312 920 R 0.3 0.0 0:00.28 top 1 root 20 0 19384 1496 1204 S 0.0 0.0 0:00.66 init 2 root 20 0 0 0 0 S 0.0 0.0 0:00.00 kthreadd 3 root 20 0 0 0 0 S 0.0 0.0 0:00.00 ksoftirqd/0 4 root 20 0 0 0 0 S 0.0 0.0 0:00.00 kworker/0:0 [akpm@linux-foundation.org: make percpu_charge_mutex static, tweak comments] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Reviewed-by: Michal Hocko <mhocko@suse.cz> Tested-by: Ying Han <yinghan@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-06-15 22:08:45 +00:00
mm: memcontrol: switch to rcu protection in drain_all_stock() Commit 72f0184c8a00 ("mm, memcg: remove hotplug locking from try_charge") introduced css_tryget()/css_put() calls in drain_all_stock(), which are supposed to protect the target memory cgroup from being released during the mem_cgroup_is_descendant() call. However, it's not completely safe. In theory, memcg can go away between reading stock->cached pointer and calling css_tryget(). This can happen if drain_all_stock() races with drain_local_stock() performed on the remote cpu as a result of a work, scheduled by the previous invocation of drain_all_stock(). The race is a bit theoretical and there are few chances to trigger it, but the current code looks a bit confusing, so it makes sense to fix it anyway. The code looks like as if css_tryget() and css_put() are used to protect stocks drainage. It's not necessary because stocked pages are holding references to the cached cgroup. And it obviously won't work for works, scheduled on other cpus. So, let's read the stock->cached pointer and evaluate the memory cgroup inside a rcu read section, and get rid of css_tryget()/css_put() calls. Link: http://lkml.kernel.org/r/20190802192241.3253165-1-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Hillf Danton <hdanton@sina.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-09-23 22:34:58 +00:00
rcu_read_lock();
memcg = stock->cached;
mm: memcontrol: switch to rcu protection in drain_all_stock() Commit 72f0184c8a00 ("mm, memcg: remove hotplug locking from try_charge") introduced css_tryget()/css_put() calls in drain_all_stock(), which are supposed to protect the target memory cgroup from being released during the mem_cgroup_is_descendant() call. However, it's not completely safe. In theory, memcg can go away between reading stock->cached pointer and calling css_tryget(). This can happen if drain_all_stock() races with drain_local_stock() performed on the remote cpu as a result of a work, scheduled by the previous invocation of drain_all_stock(). The race is a bit theoretical and there are few chances to trigger it, but the current code looks a bit confusing, so it makes sense to fix it anyway. The code looks like as if css_tryget() and css_put() are used to protect stocks drainage. It's not necessary because stocked pages are holding references to the cached cgroup. And it obviously won't work for works, scheduled on other cpus. So, let's read the stock->cached pointer and evaluate the memory cgroup inside a rcu read section, and get rid of css_tryget()/css_put() calls. Link: http://lkml.kernel.org/r/20190802192241.3253165-1-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Hillf Danton <hdanton@sina.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-09-23 22:34:58 +00:00
if (memcg && stock->nr_pages &&
mem_cgroup_is_descendant(memcg, root_memcg))
flush = true;
rcu_read_unlock();
if (flush &&
!test_and_set_bit(FLUSHING_CACHED_CHARGE, &stock->flags)) {
if (cpu == curcpu)
drain_local_stock(&stock->work);
else
schedule_work_on(cpu, &stock->work);
}
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
}
put_cpu();
Revert "memcg: get rid of percpu_charge_mutex lock" This reverts commit 8521fc50d433507a7cdc96bec280f9e5888a54cc. The patch incorrectly assumes that using atomic FLUSHING_CACHED_CHARGE bit operations is sufficient but that is not true. Johannes Weiner has reported a crash during parallel memory cgroup removal: BUG: unable to handle kernel NULL pointer dereference at 0000000000000018 IP: [<ffffffff81083b70>] css_is_ancestor+0x20/0x70 Oops: 0000 [#1] PREEMPT SMP Pid: 19677, comm: rmdir Tainted: G W 3.0.0-mm1-00188-gf38d32b #35 ECS MCP61M-M3/MCP61M-M3 RIP: 0010:[<ffffffff81083b70>] css_is_ancestor+0x20/0x70 RSP: 0018:ffff880077b09c88 EFLAGS: 00010202 Process rmdir (pid: 19677, threadinfo ffff880077b08000, task ffff8800781bb310) Call Trace: [<ffffffff810feba3>] mem_cgroup_same_or_subtree+0x33/0x40 [<ffffffff810feccf>] drain_all_stock+0x11f/0x170 [<ffffffff81103211>] mem_cgroup_force_empty+0x231/0x6d0 [<ffffffff811036c4>] mem_cgroup_pre_destroy+0x14/0x20 [<ffffffff81080559>] cgroup_rmdir+0xb9/0x500 [<ffffffff81114d26>] vfs_rmdir+0x86/0xe0 [<ffffffff81114e7b>] do_rmdir+0xfb/0x110 [<ffffffff81114ea6>] sys_rmdir+0x16/0x20 [<ffffffff8154d76b>] system_call_fastpath+0x16/0x1b We are crashing because we try to dereference cached memcg when we are checking whether we should wait for draining on the cache. The cache is already cleaned up, though. There is also a theoretical chance that the cached memcg gets freed between we test for the FLUSHING_CACHED_CHARGE and dereference it in mem_cgroup_same_or_subtree: CPU0 CPU1 CPU2 mem=stock->cached stock->cached=NULL clear_bit test_and_set_bit test_bit() ... <preempted> mem_cgroup_destroy use after free The percpu_charge_mutex protected from this race because sync draining is exclusive. It is safer to revert now and come up with a more parallel implementation later. Signed-off-by: Michal Hocko <mhocko@suse.cz> Reported-by: Johannes Weiner <jweiner@redhat.com> Acked-by: Johannes Weiner <jweiner@redhat.com> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: stable@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-08-09 09:56:26 +00:00
mutex_unlock(&percpu_charge_mutex);
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
}
static int memcg_hotplug_cpu_dead(unsigned int cpu)
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
{
struct memcg_stock_pcp *stock;
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
struct mem_cgroup *memcg, *mi;
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
stock = &per_cpu(memcg_stock, cpu);
drain_stock(stock);
mm: memcontrol: fix excessive complexity in memory.stat reporting We've seen memory.stat reads in top-level cgroups take up to fourteen seconds during a userspace bug that created tens of thousands of ghost cgroups pinned by lingering page cache. Even with a more reasonable number of cgroups, aggregating memory.stat is unnecessarily heavy. The complexity is this: nr_cgroups * nr_stat_items * nr_possible_cpus where the stat items are ~70 at this point. With 128 cgroups and 128 CPUs - decent, not enormous setups - reading the top-level memory.stat has to aggregate over a million per-cpu counters. This doesn't scale. Instead of spreading the source of truth across all CPUs, use the per-cpu counters merely to batch updates to shared atomic counters. This is the same as the per-cpu stocks we use for charging memory to the shared atomic page_counters, and also the way the global vmstat counters are implemented. Vmstat has elaborate spilling thresholds that depend on the number of CPUs, amount of memory, and memory pressure - carefully balancing the cost of counter updates with the amount of per-cpu error. That's because the vmstat counters are system-wide, but also used for decisions inside the kernel (e.g. NR_FREE_PAGES in the allocator). Neither is true for the memory controller. Use the same static batch size we already use for page_counter updates during charging. The per-cpu error in the stats will be 128k, which is an acceptable ratio of cores to memory accounting granularity. [hannes@cmpxchg.org: fix warning in __this_cpu_xchg() calls] Link: http://lkml.kernel.org/r/20171201135750.GB8097@cmpxchg.org Link: http://lkml.kernel.org/r/20171103153336.24044-3-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Michal Hocko <mhocko@suse.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-02-01 00:16:45 +00:00
for_each_mem_cgroup(memcg) {
int i;
for (i = 0; i < MEMCG_NR_STAT; i++) {
int nid;
long x;
x = this_cpu_xchg(memcg->vmstats_percpu->stat[i], 0);
mm: memcontrol: don't batch updates of local VM stats and events The kernel test robot noticed a 26% will-it-scale pagefault regression from commit 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty"). This appears to be caused by bouncing the additional cachelines from the new hierarchical statistics counters. We can fix this by getting rid of the batched local counters instead. Originally, there were *only* group-local counters, and they were fully maintained per cpu. A reader of a stats file high up in the cgroup tree would have to walk the entire subtree and collect each level's per-cpu counters to get the recursive view. This was prohibitively expensive, and so we switched to per-cpu batched updates of the local counters during a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting"), reducing the complexity from nr_subgroups * nr_cpus to nr_subgroups. With growing machines and cgroup trees, the tree walk itself became too expensive for monitoring top-level groups, and this is when the culprit patch added hierarchy counters on each cgroup level. When the per-cpu batch size would be reached, both the local and the hierarchy counters would get batch-updated from the per-cpu delta simultaneously. This makes local and hierarchical counter reads blazingly fast, but it unfortunately makes the write-side too cache line intense. Since local counter reads were never a problem - we only centralized them to accelerate the hierarchy walk - and use of the local counters are becoming rarer due to replacement with hierarchical views (ongoing rework in the page reclaim and workingset code), we can make those local counters unbatched per-cpu counters again. The scheme will then be as such: when a memcg statistic changes, the writer will: - update the local counter (per-cpu) - update the batch counter (per-cpu). If the batch is full: - spill the batch into the group's atomic_t - spill the batch into all ancestors' atomic_ts - empty out the batch counter (per-cpu) when a local memcg counter is read, the reader will: - collect the local counter from all cpus when a hiearchy memcg counter is read, the reader will: - read the atomic_t We might be able to simplify this further and make the recursive counters unbatched per-cpu counters as well (batch upward propagation, but leave per-cpu collection to the readers), but that will require a more in-depth analysis and testing of all the callsites. Deal with the immediate regression for now. Link: http://lkml.kernel.org/r/20190521151647.GB2870@cmpxchg.org Fixes: 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: kernel test robot <rong.a.chen@intel.com> Tested-by: kernel test robot <rong.a.chen@intel.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Shakeel Butt <shakeelb@google.com> Cc: Roman Gushchin <guro@fb.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-06-13 22:55:46 +00:00
if (x)
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
for (mi = memcg; mi; mi = parent_mem_cgroup(mi))
atomic_long_add(x, &memcg->vmstats[i]);
mm: memcontrol: fix excessive complexity in memory.stat reporting We've seen memory.stat reads in top-level cgroups take up to fourteen seconds during a userspace bug that created tens of thousands of ghost cgroups pinned by lingering page cache. Even with a more reasonable number of cgroups, aggregating memory.stat is unnecessarily heavy. The complexity is this: nr_cgroups * nr_stat_items * nr_possible_cpus where the stat items are ~70 at this point. With 128 cgroups and 128 CPUs - decent, not enormous setups - reading the top-level memory.stat has to aggregate over a million per-cpu counters. This doesn't scale. Instead of spreading the source of truth across all CPUs, use the per-cpu counters merely to batch updates to shared atomic counters. This is the same as the per-cpu stocks we use for charging memory to the shared atomic page_counters, and also the way the global vmstat counters are implemented. Vmstat has elaborate spilling thresholds that depend on the number of CPUs, amount of memory, and memory pressure - carefully balancing the cost of counter updates with the amount of per-cpu error. That's because the vmstat counters are system-wide, but also used for decisions inside the kernel (e.g. NR_FREE_PAGES in the allocator). Neither is true for the memory controller. Use the same static batch size we already use for page_counter updates during charging. The per-cpu error in the stats will be 128k, which is an acceptable ratio of cores to memory accounting granularity. [hannes@cmpxchg.org: fix warning in __this_cpu_xchg() calls] Link: http://lkml.kernel.org/r/20171201135750.GB8097@cmpxchg.org Link: http://lkml.kernel.org/r/20171103153336.24044-3-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Michal Hocko <mhocko@suse.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-02-01 00:16:45 +00:00
if (i >= NR_VM_NODE_STAT_ITEMS)
continue;
for_each_node(nid) {
struct mem_cgroup_per_node *pn;
pn = mem_cgroup_nodeinfo(memcg, nid);
x = this_cpu_xchg(pn->lruvec_stat_cpu->count[i], 0);
mm: memcontrol: don't batch updates of local VM stats and events The kernel test robot noticed a 26% will-it-scale pagefault regression from commit 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty"). This appears to be caused by bouncing the additional cachelines from the new hierarchical statistics counters. We can fix this by getting rid of the batched local counters instead. Originally, there were *only* group-local counters, and they were fully maintained per cpu. A reader of a stats file high up in the cgroup tree would have to walk the entire subtree and collect each level's per-cpu counters to get the recursive view. This was prohibitively expensive, and so we switched to per-cpu batched updates of the local counters during a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting"), reducing the complexity from nr_subgroups * nr_cpus to nr_subgroups. With growing machines and cgroup trees, the tree walk itself became too expensive for monitoring top-level groups, and this is when the culprit patch added hierarchy counters on each cgroup level. When the per-cpu batch size would be reached, both the local and the hierarchy counters would get batch-updated from the per-cpu delta simultaneously. This makes local and hierarchical counter reads blazingly fast, but it unfortunately makes the write-side too cache line intense. Since local counter reads were never a problem - we only centralized them to accelerate the hierarchy walk - and use of the local counters are becoming rarer due to replacement with hierarchical views (ongoing rework in the page reclaim and workingset code), we can make those local counters unbatched per-cpu counters again. The scheme will then be as such: when a memcg statistic changes, the writer will: - update the local counter (per-cpu) - update the batch counter (per-cpu). If the batch is full: - spill the batch into the group's atomic_t - spill the batch into all ancestors' atomic_ts - empty out the batch counter (per-cpu) when a local memcg counter is read, the reader will: - collect the local counter from all cpus when a hiearchy memcg counter is read, the reader will: - read the atomic_t We might be able to simplify this further and make the recursive counters unbatched per-cpu counters as well (batch upward propagation, but leave per-cpu collection to the readers), but that will require a more in-depth analysis and testing of all the callsites. Deal with the immediate regression for now. Link: http://lkml.kernel.org/r/20190521151647.GB2870@cmpxchg.org Fixes: 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: kernel test robot <rong.a.chen@intel.com> Tested-by: kernel test robot <rong.a.chen@intel.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Shakeel Butt <shakeelb@google.com> Cc: Roman Gushchin <guro@fb.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-06-13 22:55:46 +00:00
if (x)
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
do {
atomic_long_add(x, &pn->lruvec_stat[i]);
} while ((pn = parent_nodeinfo(pn, nid)));
mm: memcontrol: fix excessive complexity in memory.stat reporting We've seen memory.stat reads in top-level cgroups take up to fourteen seconds during a userspace bug that created tens of thousands of ghost cgroups pinned by lingering page cache. Even with a more reasonable number of cgroups, aggregating memory.stat is unnecessarily heavy. The complexity is this: nr_cgroups * nr_stat_items * nr_possible_cpus where the stat items are ~70 at this point. With 128 cgroups and 128 CPUs - decent, not enormous setups - reading the top-level memory.stat has to aggregate over a million per-cpu counters. This doesn't scale. Instead of spreading the source of truth across all CPUs, use the per-cpu counters merely to batch updates to shared atomic counters. This is the same as the per-cpu stocks we use for charging memory to the shared atomic page_counters, and also the way the global vmstat counters are implemented. Vmstat has elaborate spilling thresholds that depend on the number of CPUs, amount of memory, and memory pressure - carefully balancing the cost of counter updates with the amount of per-cpu error. That's because the vmstat counters are system-wide, but also used for decisions inside the kernel (e.g. NR_FREE_PAGES in the allocator). Neither is true for the memory controller. Use the same static batch size we already use for page_counter updates during charging. The per-cpu error in the stats will be 128k, which is an acceptable ratio of cores to memory accounting granularity. [hannes@cmpxchg.org: fix warning in __this_cpu_xchg() calls] Link: http://lkml.kernel.org/r/20171201135750.GB8097@cmpxchg.org Link: http://lkml.kernel.org/r/20171103153336.24044-3-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Michal Hocko <mhocko@suse.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-02-01 00:16:45 +00:00
}
}
mm: memcg: make sure memory.events is uptodate when waking pollers Commit a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting") added per-cpu drift to all memory cgroup stats and events shown in memory.stat and memory.events. For memory.stat this is acceptable. But memory.events issues file notifications, and somebody polling the file for changes will be confused when the counters in it are unchanged after a wakeup. Luckily, the events in memory.events - MEMCG_LOW, MEMCG_HIGH, MEMCG_MAX, MEMCG_OOM - are sufficiently rare and high-level that we don't need per-cpu buffering for them: MEMCG_HIGH and MEMCG_MAX would be the most frequent, but they're counting invocations of reclaim, which is a complex operation that touches many shared cachelines. This splits memory.events from the generic VM events and tracks them in their own, unbuffered atomic counters. That's also cleaner, as it eliminates the ugly enum nesting of VM and cgroup events. [hannes@cmpxchg.org: "array subscript is above array bounds"] Link: http://lkml.kernel.org/r/20180406155441.GA20806@cmpxchg.org Link: http://lkml.kernel.org/r/20180405175507.GA24817@cmpxchg.org Fixes: a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Tejun Heo <tj@kernel.org> Acked-by: Tejun Heo <tj@kernel.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Cc: Rik van Riel <riel@surriel.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-10 23:29:45 +00:00
for (i = 0; i < NR_VM_EVENT_ITEMS; i++) {
mm: memcontrol: fix excessive complexity in memory.stat reporting We've seen memory.stat reads in top-level cgroups take up to fourteen seconds during a userspace bug that created tens of thousands of ghost cgroups pinned by lingering page cache. Even with a more reasonable number of cgroups, aggregating memory.stat is unnecessarily heavy. The complexity is this: nr_cgroups * nr_stat_items * nr_possible_cpus where the stat items are ~70 at this point. With 128 cgroups and 128 CPUs - decent, not enormous setups - reading the top-level memory.stat has to aggregate over a million per-cpu counters. This doesn't scale. Instead of spreading the source of truth across all CPUs, use the per-cpu counters merely to batch updates to shared atomic counters. This is the same as the per-cpu stocks we use for charging memory to the shared atomic page_counters, and also the way the global vmstat counters are implemented. Vmstat has elaborate spilling thresholds that depend on the number of CPUs, amount of memory, and memory pressure - carefully balancing the cost of counter updates with the amount of per-cpu error. That's because the vmstat counters are system-wide, but also used for decisions inside the kernel (e.g. NR_FREE_PAGES in the allocator). Neither is true for the memory controller. Use the same static batch size we already use for page_counter updates during charging. The per-cpu error in the stats will be 128k, which is an acceptable ratio of cores to memory accounting granularity. [hannes@cmpxchg.org: fix warning in __this_cpu_xchg() calls] Link: http://lkml.kernel.org/r/20171201135750.GB8097@cmpxchg.org Link: http://lkml.kernel.org/r/20171103153336.24044-3-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Michal Hocko <mhocko@suse.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-02-01 00:16:45 +00:00
long x;
x = this_cpu_xchg(memcg->vmstats_percpu->events[i], 0);
mm: memcontrol: don't batch updates of local VM stats and events The kernel test robot noticed a 26% will-it-scale pagefault regression from commit 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty"). This appears to be caused by bouncing the additional cachelines from the new hierarchical statistics counters. We can fix this by getting rid of the batched local counters instead. Originally, there were *only* group-local counters, and they were fully maintained per cpu. A reader of a stats file high up in the cgroup tree would have to walk the entire subtree and collect each level's per-cpu counters to get the recursive view. This was prohibitively expensive, and so we switched to per-cpu batched updates of the local counters during a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting"), reducing the complexity from nr_subgroups * nr_cpus to nr_subgroups. With growing machines and cgroup trees, the tree walk itself became too expensive for monitoring top-level groups, and this is when the culprit patch added hierarchy counters on each cgroup level. When the per-cpu batch size would be reached, both the local and the hierarchy counters would get batch-updated from the per-cpu delta simultaneously. This makes local and hierarchical counter reads blazingly fast, but it unfortunately makes the write-side too cache line intense. Since local counter reads were never a problem - we only centralized them to accelerate the hierarchy walk - and use of the local counters are becoming rarer due to replacement with hierarchical views (ongoing rework in the page reclaim and workingset code), we can make those local counters unbatched per-cpu counters again. The scheme will then be as such: when a memcg statistic changes, the writer will: - update the local counter (per-cpu) - update the batch counter (per-cpu). If the batch is full: - spill the batch into the group's atomic_t - spill the batch into all ancestors' atomic_ts - empty out the batch counter (per-cpu) when a local memcg counter is read, the reader will: - collect the local counter from all cpus when a hiearchy memcg counter is read, the reader will: - read the atomic_t We might be able to simplify this further and make the recursive counters unbatched per-cpu counters as well (batch upward propagation, but leave per-cpu collection to the readers), but that will require a more in-depth analysis and testing of all the callsites. Deal with the immediate regression for now. Link: http://lkml.kernel.org/r/20190521151647.GB2870@cmpxchg.org Fixes: 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: kernel test robot <rong.a.chen@intel.com> Tested-by: kernel test robot <rong.a.chen@intel.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Shakeel Butt <shakeelb@google.com> Cc: Roman Gushchin <guro@fb.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-06-13 22:55:46 +00:00
if (x)
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
for (mi = memcg; mi; mi = parent_mem_cgroup(mi))
atomic_long_add(x, &memcg->vmevents[i]);
mm: memcontrol: fix excessive complexity in memory.stat reporting We've seen memory.stat reads in top-level cgroups take up to fourteen seconds during a userspace bug that created tens of thousands of ghost cgroups pinned by lingering page cache. Even with a more reasonable number of cgroups, aggregating memory.stat is unnecessarily heavy. The complexity is this: nr_cgroups * nr_stat_items * nr_possible_cpus where the stat items are ~70 at this point. With 128 cgroups and 128 CPUs - decent, not enormous setups - reading the top-level memory.stat has to aggregate over a million per-cpu counters. This doesn't scale. Instead of spreading the source of truth across all CPUs, use the per-cpu counters merely to batch updates to shared atomic counters. This is the same as the per-cpu stocks we use for charging memory to the shared atomic page_counters, and also the way the global vmstat counters are implemented. Vmstat has elaborate spilling thresholds that depend on the number of CPUs, amount of memory, and memory pressure - carefully balancing the cost of counter updates with the amount of per-cpu error. That's because the vmstat counters are system-wide, but also used for decisions inside the kernel (e.g. NR_FREE_PAGES in the allocator). Neither is true for the memory controller. Use the same static batch size we already use for page_counter updates during charging. The per-cpu error in the stats will be 128k, which is an acceptable ratio of cores to memory accounting granularity. [hannes@cmpxchg.org: fix warning in __this_cpu_xchg() calls] Link: http://lkml.kernel.org/r/20171201135750.GB8097@cmpxchg.org Link: http://lkml.kernel.org/r/20171103153336.24044-3-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Michal Hocko <mhocko@suse.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-02-01 00:16:45 +00:00
}
}
return 0;
memcg: coalesce charging via percpu storage This is a patch for coalescing access to res_counter at charging by percpu caching. At charge, memcg charges 64pages and remember it in percpu cache. Because it's cache, drain/flush if necessary. This version uses public percpu area. 2 benefits for using public percpu area. 1. Sum of stocked charge in the system is limited to # of cpus not to the number of memcg. This shows better synchonization. 2. drain code for flush/cpuhotplug is very easy (and quick) The most important point of this patch is that we never touch res_counter in fast path. The res_counter is system-wide shared counter which is modified very frequently. We shouldn't touch it as far as we can for avoiding false sharing. On x86-64 8cpu server, I tested overheads of memcg at page fault by running a program which does map/fault/unmap in a loop. Running a task per a cpu by taskset and see sum of the number of page faults in 60secs. [without memcg config] 40156968 page-faults # 0.085 M/sec ( +- 0.046% ) 27.67 cache-miss/faults [root cgroup] 36659599 page-faults # 0.077 M/sec ( +- 0.247% ) 31.58 cache miss/faults [in a child cgroup] 18444157 page-faults # 0.039 M/sec ( +- 0.133% ) 69.96 cache miss/faults [ + coalescing uncharge patch] 27133719 page-faults # 0.057 M/sec ( +- 0.155% ) 47.16 cache miss/faults [ + coalescing uncharge patch + this patch ] 34224709 page-faults # 0.072 M/sec ( +- 0.173% ) 34.69 cache miss/faults Changelog (since Oct/2): - updated comments - replaced get_cpu_var() with __get_cpu_var() if possible. - removed mutex for system-wide drain. adds a counter instead of it. - removed CONFIG_HOTPLUG_CPU Changelog (old): - rebased onto the latest mmotm - moved charge size check before __GFP_WAIT check for avoiding unnecesary - added asynchronous flush routine. - fixed bugs pointed out by Nishimura-san. [akpm@linux-foundation.org: tweak comments] [nishimura@mxp.nes.nec.co.jp: don't do INIT_WORK() repeatedly against the same work_struct] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-12-16 00:47:08 +00:00
}
static void reclaim_high(struct mem_cgroup *memcg,
unsigned int nr_pages,
gfp_t gfp_mask)
{
do {
if (page_counter_read(&memcg->memory) <=
READ_ONCE(memcg->memory.high))
continue;
mm: memcg: make sure memory.events is uptodate when waking pollers Commit a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting") added per-cpu drift to all memory cgroup stats and events shown in memory.stat and memory.events. For memory.stat this is acceptable. But memory.events issues file notifications, and somebody polling the file for changes will be confused when the counters in it are unchanged after a wakeup. Luckily, the events in memory.events - MEMCG_LOW, MEMCG_HIGH, MEMCG_MAX, MEMCG_OOM - are sufficiently rare and high-level that we don't need per-cpu buffering for them: MEMCG_HIGH and MEMCG_MAX would be the most frequent, but they're counting invocations of reclaim, which is a complex operation that touches many shared cachelines. This splits memory.events from the generic VM events and tracks them in their own, unbuffered atomic counters. That's also cleaner, as it eliminates the ugly enum nesting of VM and cgroup events. [hannes@cmpxchg.org: "array subscript is above array bounds"] Link: http://lkml.kernel.org/r/20180406155441.GA20806@cmpxchg.org Link: http://lkml.kernel.org/r/20180405175507.GA24817@cmpxchg.org Fixes: a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Tejun Heo <tj@kernel.org> Acked-by: Tejun Heo <tj@kernel.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Cc: Rik van Riel <riel@surriel.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-10 23:29:45 +00:00
memcg_memory_event(memcg, MEMCG_HIGH);
try_to_free_mem_cgroup_pages(memcg, nr_pages, gfp_mask, true);
} while ((memcg = parent_mem_cgroup(memcg)) &&
!mem_cgroup_is_root(memcg));
}
static void high_work_func(struct work_struct *work)
{
struct mem_cgroup *memcg;
memcg = container_of(work, struct mem_cgroup, high_work);
mm: memcontrol: fix excessive complexity in memory.stat reporting We've seen memory.stat reads in top-level cgroups take up to fourteen seconds during a userspace bug that created tens of thousands of ghost cgroups pinned by lingering page cache. Even with a more reasonable number of cgroups, aggregating memory.stat is unnecessarily heavy. The complexity is this: nr_cgroups * nr_stat_items * nr_possible_cpus where the stat items are ~70 at this point. With 128 cgroups and 128 CPUs - decent, not enormous setups - reading the top-level memory.stat has to aggregate over a million per-cpu counters. This doesn't scale. Instead of spreading the source of truth across all CPUs, use the per-cpu counters merely to batch updates to shared atomic counters. This is the same as the per-cpu stocks we use for charging memory to the shared atomic page_counters, and also the way the global vmstat counters are implemented. Vmstat has elaborate spilling thresholds that depend on the number of CPUs, amount of memory, and memory pressure - carefully balancing the cost of counter updates with the amount of per-cpu error. That's because the vmstat counters are system-wide, but also used for decisions inside the kernel (e.g. NR_FREE_PAGES in the allocator). Neither is true for the memory controller. Use the same static batch size we already use for page_counter updates during charging. The per-cpu error in the stats will be 128k, which is an acceptable ratio of cores to memory accounting granularity. [hannes@cmpxchg.org: fix warning in __this_cpu_xchg() calls] Link: http://lkml.kernel.org/r/20171201135750.GB8097@cmpxchg.org Link: http://lkml.kernel.org/r/20171103153336.24044-3-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Michal Hocko <mhocko@suse.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-02-01 00:16:45 +00:00
reclaim_high(memcg, MEMCG_CHARGE_BATCH, GFP_KERNEL);
}
mm, memcg: throttle allocators when failing reclaim over memory.high We're trying to use memory.high to limit workloads, but have found that containment can frequently fail completely and cause OOM situations outside of the cgroup. This happens especially with swap space -- either when none is configured, or swap is full. These failures often also don't have enough warning to allow one to react, whether for a human or for a daemon monitoring PSI. Here is output from a simple program showing how long it takes in usec (column 2) to allocate a megabyte of anonymous memory (column 1) when a cgroup is already beyond its memory high setting, and no swap is available: [root@ktst ~]# systemd-run -p MemoryHigh=100M -p MemorySwapMax=1 \ > --wait -t timeout 300 /root/mdf [...] 95 1035 96 1038 97 1000 98 1036 99 1048 100 1590 101 1968 102 1776 103 1863 104 1757 105 1921 106 1893 107 1760 108 1748 109 1843 110 1716 111 1924 112 1776 113 1831 114 1766 115 1836 116 1588 117 1912 118 1802 119 1857 120 1731 [...] [System OOM in 2-3 seconds] The delay does go up extremely marginally past the 100MB memory.high threshold, as now we spend time scanning before returning to usermode, but it's nowhere near enough to contain growth. It also doesn't get worse the more pages you have, since it only considers nr_pages. The current situation goes against both the expectations of users of memory.high, and our intentions as cgroup v2 developers. In cgroup-v2.txt, we claim that we will throttle and only under "extreme conditions" will memory.high protection be breached. Likewise, cgroup v2 users generally also expect that memory.high should throttle workloads as they exceed their high threshold. However, as seen above, this isn't always how it works in practice -- even on banal setups like those with no swap, or where swap has become exhausted, we can end up with memory.high being breached and us having no weapons left in our arsenal to combat runaway growth with, since reclaim is futile. It's also hard for system monitoring software or users to tell how bad the situation is, as "high" events for the memcg may in some cases be benign, and in others be catastrophic. The current status quo is that we fail containment in a way that doesn't provide any advance warning that things are about to go horribly wrong (for example, we are about to invoke the kernel OOM killer). This patch introduces explicit throttling when reclaim is failing to keep memcg size contained at the memory.high setting. It does so by applying an exponential delay curve derived from the memcg's overage compared to memory.high. In the normal case where the memcg is either below or only marginally over its memory.high setting, no throttling will be performed. This composes well with system health monitoring and remediation, as these allocator delays are factored into PSI's memory pressure calculations. This both creates a mechanism system administrators or applications consuming the PSI interface to trivially see that the memcg in question is struggling and use that to make more reasonable decisions, and permits them enough time to act. Either of these can act with significantly more nuance than that we can provide using the system OOM killer. This is a similar idea to memory.oom_control in cgroup v1 which would put the cgroup to sleep if the threshold was violated, but it's also significantly improved as it results in visible memory pressure, and also doesn't schedule indefinitely, which previously made tracing and other introspection difficult (ie. it's clamped at 2*HZ per allocation through MEMCG_MAX_HIGH_DELAY_JIFFIES). Contrast the previous results with a kernel with this patch: [root@ktst ~]# systemd-run -p MemoryHigh=100M -p MemorySwapMax=1 \ > --wait -t timeout 300 /root/mdf [...] 95 1002 96 1000 97 1002 98 1003 99 1000 100 1043 101 84724 102 330628 103 610511 104 1016265 105 1503969 106 2391692 107 2872061 108 3248003 109 4791904 110 5759832 111 6912509 112 8127818 113 9472203 114 12287622 115 12480079 116 14144008 117 15808029 118 16384500 119 16383242 120 16384979 [...] As you can see, in the normal case, memory allocation takes around 1000 usec. However, as we exceed our memory.high, things start to increase exponentially, but fairly leniently at first. Our first megabyte over memory.high takes us 0.16 seconds, then the next is 0.46 seconds, then the next is almost an entire second. This gets worse until we reach our eventual 2*HZ clamp per batch, resulting in 16 seconds per megabyte. However, this is still making forward progress, so permits tracing or further analysis with programs like GDB. We use an exponential curve for our delay penalty for a few reasons: 1. We run mem_cgroup_handle_over_high to potentially do reclaim after we've already performed allocations, which means that temporarily going over memory.high by a small amount may be perfectly legitimate, even for compliant workloads. We don't want to unduly penalise such cases. 2. An exponential curve (as opposed to a static or linear delay) allows ramping up memory pressure stats more gradually, which can be useful to work out that you have set memory.high too low, without destroying application performance entirely. This patch expands on earlier work by Johannes Weiner. Thanks! [akpm@linux-foundation.org: fix max() warning] [akpm@linux-foundation.org: fix __udivdi3 ref on 32-bit] [akpm@linux-foundation.org: fix it even more] [chris@chrisdown.name: fix 64-bit divide even more] Link: http://lkml.kernel.org/r/20190723180700.GA29459@chrisdown.name Signed-off-by: Chris Down <chris@chrisdown.name> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Nathan Chancellor <natechancellor@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-09-23 22:34:55 +00:00
/*
* Clamp the maximum sleep time per allocation batch to 2 seconds. This is
* enough to still cause a significant slowdown in most cases, while still
* allowing diagnostics and tracing to proceed without becoming stuck.
*/
#define MEMCG_MAX_HIGH_DELAY_JIFFIES (2UL*HZ)
/*
* When calculating the delay, we use these either side of the exponentiation to
* maintain precision and scale to a reasonable number of jiffies (see the table
* below.
*
* - MEMCG_DELAY_PRECISION_SHIFT: Extra precision bits while translating the
* overage ratio to a delay.
* - MEMCG_DELAY_SCALING_SHIFT: The number of bits to scale down down the
* proposed penalty in order to reduce to a reasonable number of jiffies, and
* to produce a reasonable delay curve.
*
* MEMCG_DELAY_SCALING_SHIFT just happens to be a number that produces a
* reasonable delay curve compared to precision-adjusted overage, not
* penalising heavily at first, but still making sure that growth beyond the
* limit penalises misbehaviour cgroups by slowing them down exponentially. For
* example, with a high of 100 megabytes:
*
* +-------+------------------------+
* | usage | time to allocate in ms |
* +-------+------------------------+
* | 100M | 0 |
* | 101M | 6 |
* | 102M | 25 |
* | 103M | 57 |
* | 104M | 102 |
* | 105M | 159 |
* | 106M | 230 |
* | 107M | 313 |
* | 108M | 409 |
* | 109M | 518 |
* | 110M | 639 |
* | 111M | 774 |
* | 112M | 921 |
* | 113M | 1081 |
* | 114M | 1254 |
* | 115M | 1439 |
* | 116M | 1638 |
* | 117M | 1849 |
* | 118M | 2000 |
* | 119M | 2000 |
* | 120M | 2000 |
* +-------+------------------------+
*/
#define MEMCG_DELAY_PRECISION_SHIFT 20
#define MEMCG_DELAY_SCALING_SHIFT 14
mm/memcg: prepare for swap over-high accounting and penalty calculation Patch series "memcg: Slow down swap allocation as the available space gets depleted", v6. Tejun describes the problem as follows: When swap runs out, there's an abrupt change in system behavior - the anonymous memory suddenly becomes unmanageable which readily breaks any sort of memory isolation and can bring down the whole system. To avoid that, oomd [1] monitors free swap space and triggers kills when it drops below the specific threshold (e.g. 15%). While this works, it's far from ideal: - Depending on IO performance and total swap size, a given headroom might not be enough or too much. - oomd has to monitor swap depletion in addition to the usual pressure metrics and it currently doesn't consider memory.swap.max. Solve this by adapting parts of the approach that memory.high uses - slow down allocation as the resource gets depleted turning the depletion behavior from abrupt cliff one to gradual degradation observable through memory pressure metric. [1] https://github.com/facebookincubator/oomd This patch (of 4): Slice the memory overage calculation logic a little bit so we can reuse it to apply a similar penalty to the swap. The logic which accesses the memory-specific fields (use and high values) has to be taken out of calculate_high_delay(). Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Chris Down <chris@chrisdown.name> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Link: http://lkml.kernel.org/r/20200527195846.102707-1-kuba@kernel.org Link: http://lkml.kernel.org/r/20200527195846.102707-2-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:42 +00:00
static u64 calculate_overage(unsigned long usage, unsigned long high)
memcg: punt high overage reclaim to return-to-userland path Currently, try_charge() tries to reclaim memory synchronously when the high limit is breached; however, if the allocation doesn't have __GFP_WAIT, synchronous reclaim is skipped. If a process performs only speculative allocations, it can blow way past the high limit. This is actually easily reproducible by simply doing "find /". slab/slub allocator tries speculative allocations first, so as long as there's memory which can be consumed without blocking, it can keep allocating memory regardless of the high limit. This patch makes try_charge() always punt the over-high reclaim to the return-to-userland path. If try_charge() detects that high limit is breached, it adds the overage to current->memcg_nr_pages_over_high and schedules execution of mem_cgroup_handle_over_high() which performs synchronous reclaim from the return-to-userland path. As long as kernel doesn't have a run-away allocation spree, this should provide enough protection while making kmemcg behave more consistently. It also has the following benefits. - All over-high reclaims can use GFP_KERNEL regardless of the specific gfp mask in use, e.g. GFP_NOFS, when the limit was breached. - It copes with prio inversion. Previously, a low-prio task with small memory.high might perform over-high reclaim with a bunch of locks held. If a higher prio task needed any of these locks, it would have to wait until the low prio task finished reclaim and released the locks. By handing over-high reclaim to the task exit path this issue can be avoided. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Michal Hocko <mhocko@kernel.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:46:11 +00:00
{
mm/memcg: prepare for swap over-high accounting and penalty calculation Patch series "memcg: Slow down swap allocation as the available space gets depleted", v6. Tejun describes the problem as follows: When swap runs out, there's an abrupt change in system behavior - the anonymous memory suddenly becomes unmanageable which readily breaks any sort of memory isolation and can bring down the whole system. To avoid that, oomd [1] monitors free swap space and triggers kills when it drops below the specific threshold (e.g. 15%). While this works, it's far from ideal: - Depending on IO performance and total swap size, a given headroom might not be enough or too much. - oomd has to monitor swap depletion in addition to the usual pressure metrics and it currently doesn't consider memory.swap.max. Solve this by adapting parts of the approach that memory.high uses - slow down allocation as the resource gets depleted turning the depletion behavior from abrupt cliff one to gradual degradation observable through memory pressure metric. [1] https://github.com/facebookincubator/oomd This patch (of 4): Slice the memory overage calculation logic a little bit so we can reuse it to apply a similar penalty to the swap. The logic which accesses the memory-specific fields (use and high values) has to be taken out of calculate_high_delay(). Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Chris Down <chris@chrisdown.name> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Link: http://lkml.kernel.org/r/20200527195846.102707-1-kuba@kernel.org Link: http://lkml.kernel.org/r/20200527195846.102707-2-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:42 +00:00
u64 overage;
memcg: punt high overage reclaim to return-to-userland path Currently, try_charge() tries to reclaim memory synchronously when the high limit is breached; however, if the allocation doesn't have __GFP_WAIT, synchronous reclaim is skipped. If a process performs only speculative allocations, it can blow way past the high limit. This is actually easily reproducible by simply doing "find /". slab/slub allocator tries speculative allocations first, so as long as there's memory which can be consumed without blocking, it can keep allocating memory regardless of the high limit. This patch makes try_charge() always punt the over-high reclaim to the return-to-userland path. If try_charge() detects that high limit is breached, it adds the overage to current->memcg_nr_pages_over_high and schedules execution of mem_cgroup_handle_over_high() which performs synchronous reclaim from the return-to-userland path. As long as kernel doesn't have a run-away allocation spree, this should provide enough protection while making kmemcg behave more consistently. It also has the following benefits. - All over-high reclaims can use GFP_KERNEL regardless of the specific gfp mask in use, e.g. GFP_NOFS, when the limit was breached. - It copes with prio inversion. Previously, a low-prio task with small memory.high might perform over-high reclaim with a bunch of locks held. If a higher prio task needed any of these locks, it would have to wait until the low prio task finished reclaim and released the locks. By handing over-high reclaim to the task exit path this issue can be avoided. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Michal Hocko <mhocko@kernel.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:46:11 +00:00
mm/memcg: prepare for swap over-high accounting and penalty calculation Patch series "memcg: Slow down swap allocation as the available space gets depleted", v6. Tejun describes the problem as follows: When swap runs out, there's an abrupt change in system behavior - the anonymous memory suddenly becomes unmanageable which readily breaks any sort of memory isolation and can bring down the whole system. To avoid that, oomd [1] monitors free swap space and triggers kills when it drops below the specific threshold (e.g. 15%). While this works, it's far from ideal: - Depending on IO performance and total swap size, a given headroom might not be enough or too much. - oomd has to monitor swap depletion in addition to the usual pressure metrics and it currently doesn't consider memory.swap.max. Solve this by adapting parts of the approach that memory.high uses - slow down allocation as the resource gets depleted turning the depletion behavior from abrupt cliff one to gradual degradation observable through memory pressure metric. [1] https://github.com/facebookincubator/oomd This patch (of 4): Slice the memory overage calculation logic a little bit so we can reuse it to apply a similar penalty to the swap. The logic which accesses the memory-specific fields (use and high values) has to be taken out of calculate_high_delay(). Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Chris Down <chris@chrisdown.name> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Link: http://lkml.kernel.org/r/20200527195846.102707-1-kuba@kernel.org Link: http://lkml.kernel.org/r/20200527195846.102707-2-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:42 +00:00
if (usage <= high)
return 0;
mm/memcg: prepare for swap over-high accounting and penalty calculation Patch series "memcg: Slow down swap allocation as the available space gets depleted", v6. Tejun describes the problem as follows: When swap runs out, there's an abrupt change in system behavior - the anonymous memory suddenly becomes unmanageable which readily breaks any sort of memory isolation and can bring down the whole system. To avoid that, oomd [1] monitors free swap space and triggers kills when it drops below the specific threshold (e.g. 15%). While this works, it's far from ideal: - Depending on IO performance and total swap size, a given headroom might not be enough or too much. - oomd has to monitor swap depletion in addition to the usual pressure metrics and it currently doesn't consider memory.swap.max. Solve this by adapting parts of the approach that memory.high uses - slow down allocation as the resource gets depleted turning the depletion behavior from abrupt cliff one to gradual degradation observable through memory pressure metric. [1] https://github.com/facebookincubator/oomd This patch (of 4): Slice the memory overage calculation logic a little bit so we can reuse it to apply a similar penalty to the swap. The logic which accesses the memory-specific fields (use and high values) has to be taken out of calculate_high_delay(). Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Chris Down <chris@chrisdown.name> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Link: http://lkml.kernel.org/r/20200527195846.102707-1-kuba@kernel.org Link: http://lkml.kernel.org/r/20200527195846.102707-2-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:42 +00:00
/*
* Prevent division by 0 in overage calculation by acting as if
* it was a threshold of 1 page
*/
high = max(high, 1UL);
mm/memcg: prepare for swap over-high accounting and penalty calculation Patch series "memcg: Slow down swap allocation as the available space gets depleted", v6. Tejun describes the problem as follows: When swap runs out, there's an abrupt change in system behavior - the anonymous memory suddenly becomes unmanageable which readily breaks any sort of memory isolation and can bring down the whole system. To avoid that, oomd [1] monitors free swap space and triggers kills when it drops below the specific threshold (e.g. 15%). While this works, it's far from ideal: - Depending on IO performance and total swap size, a given headroom might not be enough or too much. - oomd has to monitor swap depletion in addition to the usual pressure metrics and it currently doesn't consider memory.swap.max. Solve this by adapting parts of the approach that memory.high uses - slow down allocation as the resource gets depleted turning the depletion behavior from abrupt cliff one to gradual degradation observable through memory pressure metric. [1] https://github.com/facebookincubator/oomd This patch (of 4): Slice the memory overage calculation logic a little bit so we can reuse it to apply a similar penalty to the swap. The logic which accesses the memory-specific fields (use and high values) has to be taken out of calculate_high_delay(). Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Chris Down <chris@chrisdown.name> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Link: http://lkml.kernel.org/r/20200527195846.102707-1-kuba@kernel.org Link: http://lkml.kernel.org/r/20200527195846.102707-2-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:42 +00:00
overage = usage - high;
overage <<= MEMCG_DELAY_PRECISION_SHIFT;
return div64_u64(overage, high);
}
mm/memcg: prepare for swap over-high accounting and penalty calculation Patch series "memcg: Slow down swap allocation as the available space gets depleted", v6. Tejun describes the problem as follows: When swap runs out, there's an abrupt change in system behavior - the anonymous memory suddenly becomes unmanageable which readily breaks any sort of memory isolation and can bring down the whole system. To avoid that, oomd [1] monitors free swap space and triggers kills when it drops below the specific threshold (e.g. 15%). While this works, it's far from ideal: - Depending on IO performance and total swap size, a given headroom might not be enough or too much. - oomd has to monitor swap depletion in addition to the usual pressure metrics and it currently doesn't consider memory.swap.max. Solve this by adapting parts of the approach that memory.high uses - slow down allocation as the resource gets depleted turning the depletion behavior from abrupt cliff one to gradual degradation observable through memory pressure metric. [1] https://github.com/facebookincubator/oomd This patch (of 4): Slice the memory overage calculation logic a little bit so we can reuse it to apply a similar penalty to the swap. The logic which accesses the memory-specific fields (use and high values) has to be taken out of calculate_high_delay(). Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Chris Down <chris@chrisdown.name> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Link: http://lkml.kernel.org/r/20200527195846.102707-1-kuba@kernel.org Link: http://lkml.kernel.org/r/20200527195846.102707-2-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:42 +00:00
static u64 mem_find_max_overage(struct mem_cgroup *memcg)
{
u64 overage, max_overage = 0;
mm/memcg: prepare for swap over-high accounting and penalty calculation Patch series "memcg: Slow down swap allocation as the available space gets depleted", v6. Tejun describes the problem as follows: When swap runs out, there's an abrupt change in system behavior - the anonymous memory suddenly becomes unmanageable which readily breaks any sort of memory isolation and can bring down the whole system. To avoid that, oomd [1] monitors free swap space and triggers kills when it drops below the specific threshold (e.g. 15%). While this works, it's far from ideal: - Depending on IO performance and total swap size, a given headroom might not be enough or too much. - oomd has to monitor swap depletion in addition to the usual pressure metrics and it currently doesn't consider memory.swap.max. Solve this by adapting parts of the approach that memory.high uses - slow down allocation as the resource gets depleted turning the depletion behavior from abrupt cliff one to gradual degradation observable through memory pressure metric. [1] https://github.com/facebookincubator/oomd This patch (of 4): Slice the memory overage calculation logic a little bit so we can reuse it to apply a similar penalty to the swap. The logic which accesses the memory-specific fields (use and high values) has to be taken out of calculate_high_delay(). Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Chris Down <chris@chrisdown.name> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Link: http://lkml.kernel.org/r/20200527195846.102707-1-kuba@kernel.org Link: http://lkml.kernel.org/r/20200527195846.102707-2-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:42 +00:00
do {
overage = calculate_overage(page_counter_read(&memcg->memory),
READ_ONCE(memcg->memory.high));
mm/memcg: prepare for swap over-high accounting and penalty calculation Patch series "memcg: Slow down swap allocation as the available space gets depleted", v6. Tejun describes the problem as follows: When swap runs out, there's an abrupt change in system behavior - the anonymous memory suddenly becomes unmanageable which readily breaks any sort of memory isolation and can bring down the whole system. To avoid that, oomd [1] monitors free swap space and triggers kills when it drops below the specific threshold (e.g. 15%). While this works, it's far from ideal: - Depending on IO performance and total swap size, a given headroom might not be enough or too much. - oomd has to monitor swap depletion in addition to the usual pressure metrics and it currently doesn't consider memory.swap.max. Solve this by adapting parts of the approach that memory.high uses - slow down allocation as the resource gets depleted turning the depletion behavior from abrupt cliff one to gradual degradation observable through memory pressure metric. [1] https://github.com/facebookincubator/oomd This patch (of 4): Slice the memory overage calculation logic a little bit so we can reuse it to apply a similar penalty to the swap. The logic which accesses the memory-specific fields (use and high values) has to be taken out of calculate_high_delay(). Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Chris Down <chris@chrisdown.name> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Link: http://lkml.kernel.org/r/20200527195846.102707-1-kuba@kernel.org Link: http://lkml.kernel.org/r/20200527195846.102707-2-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:42 +00:00
max_overage = max(overage, max_overage);
} while ((memcg = parent_mem_cgroup(memcg)) &&
!mem_cgroup_is_root(memcg));
mm/memcg: prepare for swap over-high accounting and penalty calculation Patch series "memcg: Slow down swap allocation as the available space gets depleted", v6. Tejun describes the problem as follows: When swap runs out, there's an abrupt change in system behavior - the anonymous memory suddenly becomes unmanageable which readily breaks any sort of memory isolation and can bring down the whole system. To avoid that, oomd [1] monitors free swap space and triggers kills when it drops below the specific threshold (e.g. 15%). While this works, it's far from ideal: - Depending on IO performance and total swap size, a given headroom might not be enough or too much. - oomd has to monitor swap depletion in addition to the usual pressure metrics and it currently doesn't consider memory.swap.max. Solve this by adapting parts of the approach that memory.high uses - slow down allocation as the resource gets depleted turning the depletion behavior from abrupt cliff one to gradual degradation observable through memory pressure metric. [1] https://github.com/facebookincubator/oomd This patch (of 4): Slice the memory overage calculation logic a little bit so we can reuse it to apply a similar penalty to the swap. The logic which accesses the memory-specific fields (use and high values) has to be taken out of calculate_high_delay(). Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Chris Down <chris@chrisdown.name> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Link: http://lkml.kernel.org/r/20200527195846.102707-1-kuba@kernel.org Link: http://lkml.kernel.org/r/20200527195846.102707-2-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:42 +00:00
return max_overage;
}
mm/memcg: automatically penalize tasks with high swap use Add a memory.swap.high knob, which can be used to protect the system from SWAP exhaustion. The mechanism used for penalizing is similar to memory.high penalty (sleep on return to user space). That is not to say that the knob itself is equivalent to memory.high. The objective is more to protect the system from potentially buggy tasks consuming a lot of swap and impacting other tasks, or even bringing the whole system to stand still with complete SWAP exhaustion. Hopefully without the need to find per-task hard limits. Slowing misbehaving tasks down gradually allows user space oom killers or other protection mechanisms to react. oomd and earlyoom already do killing based on swap exhaustion, and memory.swap.high protection will help implement such userspace oom policies more reliably. We can use one counter for number of pages allocated under pressure to save struct task space and avoid two separate hierarchy walks on the hot path. The exact overage is calculated on return to user space, anyway. Take the new high limit into account when determining if swap is "full". Borrowing the explanation from Johannes: The idea behind "swap full" is that as long as the workload has plenty of swap space available and it's not changing its memory contents, it makes sense to generously hold on to copies of data in the swap device, even after the swapin. A later reclaim cycle can drop the page without any IO. Trading disk space for IO. But the only two ways to reclaim a swap slot is when they're faulted in and the references go away, or by scanning the virtual address space like swapoff does - which is very expensive (one could argue it's too expensive even for swapoff, it's often more practical to just reboot). So at some point in the fill level, we have to start freeing up swap slots on fault/swapin. Otherwise we could eventually run out of swap slots while they're filled with copies of data that is also in RAM. We don't want to OOM a workload because its available swap space is filled with redundant cache. Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Chris Down <chris@chrisdown.name> Cc: Shakeel Butt <shakeelb@google.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Hugh Dickins <hughd@google.com> Link: http://lkml.kernel.org/r/20200527195846.102707-5-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:52 +00:00
static u64 swap_find_max_overage(struct mem_cgroup *memcg)
{
u64 overage, max_overage = 0;
do {
overage = calculate_overage(page_counter_read(&memcg->swap),
READ_ONCE(memcg->swap.high));
if (overage)
memcg_memory_event(memcg, MEMCG_SWAP_HIGH);
max_overage = max(overage, max_overage);
} while ((memcg = parent_mem_cgroup(memcg)) &&
!mem_cgroup_is_root(memcg));
return max_overage;
}
mm/memcg: prepare for swap over-high accounting and penalty calculation Patch series "memcg: Slow down swap allocation as the available space gets depleted", v6. Tejun describes the problem as follows: When swap runs out, there's an abrupt change in system behavior - the anonymous memory suddenly becomes unmanageable which readily breaks any sort of memory isolation and can bring down the whole system. To avoid that, oomd [1] monitors free swap space and triggers kills when it drops below the specific threshold (e.g. 15%). While this works, it's far from ideal: - Depending on IO performance and total swap size, a given headroom might not be enough or too much. - oomd has to monitor swap depletion in addition to the usual pressure metrics and it currently doesn't consider memory.swap.max. Solve this by adapting parts of the approach that memory.high uses - slow down allocation as the resource gets depleted turning the depletion behavior from abrupt cliff one to gradual degradation observable through memory pressure metric. [1] https://github.com/facebookincubator/oomd This patch (of 4): Slice the memory overage calculation logic a little bit so we can reuse it to apply a similar penalty to the swap. The logic which accesses the memory-specific fields (use and high values) has to be taken out of calculate_high_delay(). Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Chris Down <chris@chrisdown.name> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Link: http://lkml.kernel.org/r/20200527195846.102707-1-kuba@kernel.org Link: http://lkml.kernel.org/r/20200527195846.102707-2-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:42 +00:00
/*
* Get the number of jiffies that we should penalise a mischievous cgroup which
* is exceeding its memory.high by checking both it and its ancestors.
*/
static unsigned long calculate_high_delay(struct mem_cgroup *memcg,
unsigned int nr_pages,
u64 max_overage)
{
unsigned long penalty_jiffies;
if (!max_overage)
return 0;
mm, memcg: throttle allocators when failing reclaim over memory.high We're trying to use memory.high to limit workloads, but have found that containment can frequently fail completely and cause OOM situations outside of the cgroup. This happens especially with swap space -- either when none is configured, or swap is full. These failures often also don't have enough warning to allow one to react, whether for a human or for a daemon monitoring PSI. Here is output from a simple program showing how long it takes in usec (column 2) to allocate a megabyte of anonymous memory (column 1) when a cgroup is already beyond its memory high setting, and no swap is available: [root@ktst ~]# systemd-run -p MemoryHigh=100M -p MemorySwapMax=1 \ > --wait -t timeout 300 /root/mdf [...] 95 1035 96 1038 97 1000 98 1036 99 1048 100 1590 101 1968 102 1776 103 1863 104 1757 105 1921 106 1893 107 1760 108 1748 109 1843 110 1716 111 1924 112 1776 113 1831 114 1766 115 1836 116 1588 117 1912 118 1802 119 1857 120 1731 [...] [System OOM in 2-3 seconds] The delay does go up extremely marginally past the 100MB memory.high threshold, as now we spend time scanning before returning to usermode, but it's nowhere near enough to contain growth. It also doesn't get worse the more pages you have, since it only considers nr_pages. The current situation goes against both the expectations of users of memory.high, and our intentions as cgroup v2 developers. In cgroup-v2.txt, we claim that we will throttle and only under "extreme conditions" will memory.high protection be breached. Likewise, cgroup v2 users generally also expect that memory.high should throttle workloads as they exceed their high threshold. However, as seen above, this isn't always how it works in practice -- even on banal setups like those with no swap, or where swap has become exhausted, we can end up with memory.high being breached and us having no weapons left in our arsenal to combat runaway growth with, since reclaim is futile. It's also hard for system monitoring software or users to tell how bad the situation is, as "high" events for the memcg may in some cases be benign, and in others be catastrophic. The current status quo is that we fail containment in a way that doesn't provide any advance warning that things are about to go horribly wrong (for example, we are about to invoke the kernel OOM killer). This patch introduces explicit throttling when reclaim is failing to keep memcg size contained at the memory.high setting. It does so by applying an exponential delay curve derived from the memcg's overage compared to memory.high. In the normal case where the memcg is either below or only marginally over its memory.high setting, no throttling will be performed. This composes well with system health monitoring and remediation, as these allocator delays are factored into PSI's memory pressure calculations. This both creates a mechanism system administrators or applications consuming the PSI interface to trivially see that the memcg in question is struggling and use that to make more reasonable decisions, and permits them enough time to act. Either of these can act with significantly more nuance than that we can provide using the system OOM killer. This is a similar idea to memory.oom_control in cgroup v1 which would put the cgroup to sleep if the threshold was violated, but it's also significantly improved as it results in visible memory pressure, and also doesn't schedule indefinitely, which previously made tracing and other introspection difficult (ie. it's clamped at 2*HZ per allocation through MEMCG_MAX_HIGH_DELAY_JIFFIES). Contrast the previous results with a kernel with this patch: [root@ktst ~]# systemd-run -p MemoryHigh=100M -p MemorySwapMax=1 \ > --wait -t timeout 300 /root/mdf [...] 95 1002 96 1000 97 1002 98 1003 99 1000 100 1043 101 84724 102 330628 103 610511 104 1016265 105 1503969 106 2391692 107 2872061 108 3248003 109 4791904 110 5759832 111 6912509 112 8127818 113 9472203 114 12287622 115 12480079 116 14144008 117 15808029 118 16384500 119 16383242 120 16384979 [...] As you can see, in the normal case, memory allocation takes around 1000 usec. However, as we exceed our memory.high, things start to increase exponentially, but fairly leniently at first. Our first megabyte over memory.high takes us 0.16 seconds, then the next is 0.46 seconds, then the next is almost an entire second. This gets worse until we reach our eventual 2*HZ clamp per batch, resulting in 16 seconds per megabyte. However, this is still making forward progress, so permits tracing or further analysis with programs like GDB. We use an exponential curve for our delay penalty for a few reasons: 1. We run mem_cgroup_handle_over_high to potentially do reclaim after we've already performed allocations, which means that temporarily going over memory.high by a small amount may be perfectly legitimate, even for compliant workloads. We don't want to unduly penalise such cases. 2. An exponential curve (as opposed to a static or linear delay) allows ramping up memory pressure stats more gradually, which can be useful to work out that you have set memory.high too low, without destroying application performance entirely. This patch expands on earlier work by Johannes Weiner. Thanks! [akpm@linux-foundation.org: fix max() warning] [akpm@linux-foundation.org: fix __udivdi3 ref on 32-bit] [akpm@linux-foundation.org: fix it even more] [chris@chrisdown.name: fix 64-bit divide even more] Link: http://lkml.kernel.org/r/20190723180700.GA29459@chrisdown.name Signed-off-by: Chris Down <chris@chrisdown.name> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Nathan Chancellor <natechancellor@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-09-23 22:34:55 +00:00
/*
* We use overage compared to memory.high to calculate the number of
* jiffies to sleep (penalty_jiffies). Ideally this value should be
* fairly lenient on small overages, and increasingly harsh when the
* memcg in question makes it clear that it has no intention of stopping
* its crazy behaviour, so we exponentially increase the delay based on
* overage amount.
*/
penalty_jiffies = max_overage * max_overage * HZ;
penalty_jiffies >>= MEMCG_DELAY_PRECISION_SHIFT;
penalty_jiffies >>= MEMCG_DELAY_SCALING_SHIFT;
mm, memcg: throttle allocators when failing reclaim over memory.high We're trying to use memory.high to limit workloads, but have found that containment can frequently fail completely and cause OOM situations outside of the cgroup. This happens especially with swap space -- either when none is configured, or swap is full. These failures often also don't have enough warning to allow one to react, whether for a human or for a daemon monitoring PSI. Here is output from a simple program showing how long it takes in usec (column 2) to allocate a megabyte of anonymous memory (column 1) when a cgroup is already beyond its memory high setting, and no swap is available: [root@ktst ~]# systemd-run -p MemoryHigh=100M -p MemorySwapMax=1 \ > --wait -t timeout 300 /root/mdf [...] 95 1035 96 1038 97 1000 98 1036 99 1048 100 1590 101 1968 102 1776 103 1863 104 1757 105 1921 106 1893 107 1760 108 1748 109 1843 110 1716 111 1924 112 1776 113 1831 114 1766 115 1836 116 1588 117 1912 118 1802 119 1857 120 1731 [...] [System OOM in 2-3 seconds] The delay does go up extremely marginally past the 100MB memory.high threshold, as now we spend time scanning before returning to usermode, but it's nowhere near enough to contain growth. It also doesn't get worse the more pages you have, since it only considers nr_pages. The current situation goes against both the expectations of users of memory.high, and our intentions as cgroup v2 developers. In cgroup-v2.txt, we claim that we will throttle and only under "extreme conditions" will memory.high protection be breached. Likewise, cgroup v2 users generally also expect that memory.high should throttle workloads as they exceed their high threshold. However, as seen above, this isn't always how it works in practice -- even on banal setups like those with no swap, or where swap has become exhausted, we can end up with memory.high being breached and us having no weapons left in our arsenal to combat runaway growth with, since reclaim is futile. It's also hard for system monitoring software or users to tell how bad the situation is, as "high" events for the memcg may in some cases be benign, and in others be catastrophic. The current status quo is that we fail containment in a way that doesn't provide any advance warning that things are about to go horribly wrong (for example, we are about to invoke the kernel OOM killer). This patch introduces explicit throttling when reclaim is failing to keep memcg size contained at the memory.high setting. It does so by applying an exponential delay curve derived from the memcg's overage compared to memory.high. In the normal case where the memcg is either below or only marginally over its memory.high setting, no throttling will be performed. This composes well with system health monitoring and remediation, as these allocator delays are factored into PSI's memory pressure calculations. This both creates a mechanism system administrators or applications consuming the PSI interface to trivially see that the memcg in question is struggling and use that to make more reasonable decisions, and permits them enough time to act. Either of these can act with significantly more nuance than that we can provide using the system OOM killer. This is a similar idea to memory.oom_control in cgroup v1 which would put the cgroup to sleep if the threshold was violated, but it's also significantly improved as it results in visible memory pressure, and also doesn't schedule indefinitely, which previously made tracing and other introspection difficult (ie. it's clamped at 2*HZ per allocation through MEMCG_MAX_HIGH_DELAY_JIFFIES). Contrast the previous results with a kernel with this patch: [root@ktst ~]# systemd-run -p MemoryHigh=100M -p MemorySwapMax=1 \ > --wait -t timeout 300 /root/mdf [...] 95 1002 96 1000 97 1002 98 1003 99 1000 100 1043 101 84724 102 330628 103 610511 104 1016265 105 1503969 106 2391692 107 2872061 108 3248003 109 4791904 110 5759832 111 6912509 112 8127818 113 9472203 114 12287622 115 12480079 116 14144008 117 15808029 118 16384500 119 16383242 120 16384979 [...] As you can see, in the normal case, memory allocation takes around 1000 usec. However, as we exceed our memory.high, things start to increase exponentially, but fairly leniently at first. Our first megabyte over memory.high takes us 0.16 seconds, then the next is 0.46 seconds, then the next is almost an entire second. This gets worse until we reach our eventual 2*HZ clamp per batch, resulting in 16 seconds per megabyte. However, this is still making forward progress, so permits tracing or further analysis with programs like GDB. We use an exponential curve for our delay penalty for a few reasons: 1. We run mem_cgroup_handle_over_high to potentially do reclaim after we've already performed allocations, which means that temporarily going over memory.high by a small amount may be perfectly legitimate, even for compliant workloads. We don't want to unduly penalise such cases. 2. An exponential curve (as opposed to a static or linear delay) allows ramping up memory pressure stats more gradually, which can be useful to work out that you have set memory.high too low, without destroying application performance entirely. This patch expands on earlier work by Johannes Weiner. Thanks! [akpm@linux-foundation.org: fix max() warning] [akpm@linux-foundation.org: fix __udivdi3 ref on 32-bit] [akpm@linux-foundation.org: fix it even more] [chris@chrisdown.name: fix 64-bit divide even more] Link: http://lkml.kernel.org/r/20190723180700.GA29459@chrisdown.name Signed-off-by: Chris Down <chris@chrisdown.name> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Nathan Chancellor <natechancellor@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-09-23 22:34:55 +00:00
/*
* Factor in the task's own contribution to the overage, such that four
* N-sized allocations are throttled approximately the same as one
* 4N-sized allocation.
*
* MEMCG_CHARGE_BATCH pages is nominal, so work out how much smaller or
* larger the current charge patch is than that.
*/
return penalty_jiffies * nr_pages / MEMCG_CHARGE_BATCH;
}
/*
* Scheduled by try_charge() to be executed from the userland return path
* and reclaims memory over the high limit.
*/
void mem_cgroup_handle_over_high(void)
{
unsigned long penalty_jiffies;
unsigned long pflags;
unsigned int nr_pages = current->memcg_nr_pages_over_high;
struct mem_cgroup *memcg;
if (likely(!nr_pages))
return;
memcg = get_mem_cgroup_from_mm(current->mm);
reclaim_high(memcg, nr_pages, GFP_KERNEL);
current->memcg_nr_pages_over_high = 0;
/*
* memory.high is breached and reclaim is unable to keep up. Throttle
* allocators proactively to slow down excessive growth.
*/
mm/memcg: prepare for swap over-high accounting and penalty calculation Patch series "memcg: Slow down swap allocation as the available space gets depleted", v6. Tejun describes the problem as follows: When swap runs out, there's an abrupt change in system behavior - the anonymous memory suddenly becomes unmanageable which readily breaks any sort of memory isolation and can bring down the whole system. To avoid that, oomd [1] monitors free swap space and triggers kills when it drops below the specific threshold (e.g. 15%). While this works, it's far from ideal: - Depending on IO performance and total swap size, a given headroom might not be enough or too much. - oomd has to monitor swap depletion in addition to the usual pressure metrics and it currently doesn't consider memory.swap.max. Solve this by adapting parts of the approach that memory.high uses - slow down allocation as the resource gets depleted turning the depletion behavior from abrupt cliff one to gradual degradation observable through memory pressure metric. [1] https://github.com/facebookincubator/oomd This patch (of 4): Slice the memory overage calculation logic a little bit so we can reuse it to apply a similar penalty to the swap. The logic which accesses the memory-specific fields (use and high values) has to be taken out of calculate_high_delay(). Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Chris Down <chris@chrisdown.name> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Link: http://lkml.kernel.org/r/20200527195846.102707-1-kuba@kernel.org Link: http://lkml.kernel.org/r/20200527195846.102707-2-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:42 +00:00
penalty_jiffies = calculate_high_delay(memcg, nr_pages,
mem_find_max_overage(memcg));
mm, memcg: throttle allocators when failing reclaim over memory.high We're trying to use memory.high to limit workloads, but have found that containment can frequently fail completely and cause OOM situations outside of the cgroup. This happens especially with swap space -- either when none is configured, or swap is full. These failures often also don't have enough warning to allow one to react, whether for a human or for a daemon monitoring PSI. Here is output from a simple program showing how long it takes in usec (column 2) to allocate a megabyte of anonymous memory (column 1) when a cgroup is already beyond its memory high setting, and no swap is available: [root@ktst ~]# systemd-run -p MemoryHigh=100M -p MemorySwapMax=1 \ > --wait -t timeout 300 /root/mdf [...] 95 1035 96 1038 97 1000 98 1036 99 1048 100 1590 101 1968 102 1776 103 1863 104 1757 105 1921 106 1893 107 1760 108 1748 109 1843 110 1716 111 1924 112 1776 113 1831 114 1766 115 1836 116 1588 117 1912 118 1802 119 1857 120 1731 [...] [System OOM in 2-3 seconds] The delay does go up extremely marginally past the 100MB memory.high threshold, as now we spend time scanning before returning to usermode, but it's nowhere near enough to contain growth. It also doesn't get worse the more pages you have, since it only considers nr_pages. The current situation goes against both the expectations of users of memory.high, and our intentions as cgroup v2 developers. In cgroup-v2.txt, we claim that we will throttle and only under "extreme conditions" will memory.high protection be breached. Likewise, cgroup v2 users generally also expect that memory.high should throttle workloads as they exceed their high threshold. However, as seen above, this isn't always how it works in practice -- even on banal setups like those with no swap, or where swap has become exhausted, we can end up with memory.high being breached and us having no weapons left in our arsenal to combat runaway growth with, since reclaim is futile. It's also hard for system monitoring software or users to tell how bad the situation is, as "high" events for the memcg may in some cases be benign, and in others be catastrophic. The current status quo is that we fail containment in a way that doesn't provide any advance warning that things are about to go horribly wrong (for example, we are about to invoke the kernel OOM killer). This patch introduces explicit throttling when reclaim is failing to keep memcg size contained at the memory.high setting. It does so by applying an exponential delay curve derived from the memcg's overage compared to memory.high. In the normal case where the memcg is either below or only marginally over its memory.high setting, no throttling will be performed. This composes well with system health monitoring and remediation, as these allocator delays are factored into PSI's memory pressure calculations. This both creates a mechanism system administrators or applications consuming the PSI interface to trivially see that the memcg in question is struggling and use that to make more reasonable decisions, and permits them enough time to act. Either of these can act with significantly more nuance than that we can provide using the system OOM killer. This is a similar idea to memory.oom_control in cgroup v1 which would put the cgroup to sleep if the threshold was violated, but it's also significantly improved as it results in visible memory pressure, and also doesn't schedule indefinitely, which previously made tracing and other introspection difficult (ie. it's clamped at 2*HZ per allocation through MEMCG_MAX_HIGH_DELAY_JIFFIES). Contrast the previous results with a kernel with this patch: [root@ktst ~]# systemd-run -p MemoryHigh=100M -p MemorySwapMax=1 \ > --wait -t timeout 300 /root/mdf [...] 95 1002 96 1000 97 1002 98 1003 99 1000 100 1043 101 84724 102 330628 103 610511 104 1016265 105 1503969 106 2391692 107 2872061 108 3248003 109 4791904 110 5759832 111 6912509 112 8127818 113 9472203 114 12287622 115 12480079 116 14144008 117 15808029 118 16384500 119 16383242 120 16384979 [...] As you can see, in the normal case, memory allocation takes around 1000 usec. However, as we exceed our memory.high, things start to increase exponentially, but fairly leniently at first. Our first megabyte over memory.high takes us 0.16 seconds, then the next is 0.46 seconds, then the next is almost an entire second. This gets worse until we reach our eventual 2*HZ clamp per batch, resulting in 16 seconds per megabyte. However, this is still making forward progress, so permits tracing or further analysis with programs like GDB. We use an exponential curve for our delay penalty for a few reasons: 1. We run mem_cgroup_handle_over_high to potentially do reclaim after we've already performed allocations, which means that temporarily going over memory.high by a small amount may be perfectly legitimate, even for compliant workloads. We don't want to unduly penalise such cases. 2. An exponential curve (as opposed to a static or linear delay) allows ramping up memory pressure stats more gradually, which can be useful to work out that you have set memory.high too low, without destroying application performance entirely. This patch expands on earlier work by Johannes Weiner. Thanks! [akpm@linux-foundation.org: fix max() warning] [akpm@linux-foundation.org: fix __udivdi3 ref on 32-bit] [akpm@linux-foundation.org: fix it even more] [chris@chrisdown.name: fix 64-bit divide even more] Link: http://lkml.kernel.org/r/20190723180700.GA29459@chrisdown.name Signed-off-by: Chris Down <chris@chrisdown.name> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Nathan Chancellor <natechancellor@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-09-23 22:34:55 +00:00
mm/memcg: automatically penalize tasks with high swap use Add a memory.swap.high knob, which can be used to protect the system from SWAP exhaustion. The mechanism used for penalizing is similar to memory.high penalty (sleep on return to user space). That is not to say that the knob itself is equivalent to memory.high. The objective is more to protect the system from potentially buggy tasks consuming a lot of swap and impacting other tasks, or even bringing the whole system to stand still with complete SWAP exhaustion. Hopefully without the need to find per-task hard limits. Slowing misbehaving tasks down gradually allows user space oom killers or other protection mechanisms to react. oomd and earlyoom already do killing based on swap exhaustion, and memory.swap.high protection will help implement such userspace oom policies more reliably. We can use one counter for number of pages allocated under pressure to save struct task space and avoid two separate hierarchy walks on the hot path. The exact overage is calculated on return to user space, anyway. Take the new high limit into account when determining if swap is "full". Borrowing the explanation from Johannes: The idea behind "swap full" is that as long as the workload has plenty of swap space available and it's not changing its memory contents, it makes sense to generously hold on to copies of data in the swap device, even after the swapin. A later reclaim cycle can drop the page without any IO. Trading disk space for IO. But the only two ways to reclaim a swap slot is when they're faulted in and the references go away, or by scanning the virtual address space like swapoff does - which is very expensive (one could argue it's too expensive even for swapoff, it's often more practical to just reboot). So at some point in the fill level, we have to start freeing up swap slots on fault/swapin. Otherwise we could eventually run out of swap slots while they're filled with copies of data that is also in RAM. We don't want to OOM a workload because its available swap space is filled with redundant cache. Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Chris Down <chris@chrisdown.name> Cc: Shakeel Butt <shakeelb@google.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Hugh Dickins <hughd@google.com> Link: http://lkml.kernel.org/r/20200527195846.102707-5-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:52 +00:00
penalty_jiffies += calculate_high_delay(memcg, nr_pages,
swap_find_max_overage(memcg));
/*
* Clamp the max delay per usermode return so as to still keep the
* application moving forwards and also permit diagnostics, albeit
* extremely slowly.
*/
penalty_jiffies = min(penalty_jiffies, MEMCG_MAX_HIGH_DELAY_JIFFIES);
mm, memcg: throttle allocators when failing reclaim over memory.high We're trying to use memory.high to limit workloads, but have found that containment can frequently fail completely and cause OOM situations outside of the cgroup. This happens especially with swap space -- either when none is configured, or swap is full. These failures often also don't have enough warning to allow one to react, whether for a human or for a daemon monitoring PSI. Here is output from a simple program showing how long it takes in usec (column 2) to allocate a megabyte of anonymous memory (column 1) when a cgroup is already beyond its memory high setting, and no swap is available: [root@ktst ~]# systemd-run -p MemoryHigh=100M -p MemorySwapMax=1 \ > --wait -t timeout 300 /root/mdf [...] 95 1035 96 1038 97 1000 98 1036 99 1048 100 1590 101 1968 102 1776 103 1863 104 1757 105 1921 106 1893 107 1760 108 1748 109 1843 110 1716 111 1924 112 1776 113 1831 114 1766 115 1836 116 1588 117 1912 118 1802 119 1857 120 1731 [...] [System OOM in 2-3 seconds] The delay does go up extremely marginally past the 100MB memory.high threshold, as now we spend time scanning before returning to usermode, but it's nowhere near enough to contain growth. It also doesn't get worse the more pages you have, since it only considers nr_pages. The current situation goes against both the expectations of users of memory.high, and our intentions as cgroup v2 developers. In cgroup-v2.txt, we claim that we will throttle and only under "extreme conditions" will memory.high protection be breached. Likewise, cgroup v2 users generally also expect that memory.high should throttle workloads as they exceed their high threshold. However, as seen above, this isn't always how it works in practice -- even on banal setups like those with no swap, or where swap has become exhausted, we can end up with memory.high being breached and us having no weapons left in our arsenal to combat runaway growth with, since reclaim is futile. It's also hard for system monitoring software or users to tell how bad the situation is, as "high" events for the memcg may in some cases be benign, and in others be catastrophic. The current status quo is that we fail containment in a way that doesn't provide any advance warning that things are about to go horribly wrong (for example, we are about to invoke the kernel OOM killer). This patch introduces explicit throttling when reclaim is failing to keep memcg size contained at the memory.high setting. It does so by applying an exponential delay curve derived from the memcg's overage compared to memory.high. In the normal case where the memcg is either below or only marginally over its memory.high setting, no throttling will be performed. This composes well with system health monitoring and remediation, as these allocator delays are factored into PSI's memory pressure calculations. This both creates a mechanism system administrators or applications consuming the PSI interface to trivially see that the memcg in question is struggling and use that to make more reasonable decisions, and permits them enough time to act. Either of these can act with significantly more nuance than that we can provide using the system OOM killer. This is a similar idea to memory.oom_control in cgroup v1 which would put the cgroup to sleep if the threshold was violated, but it's also significantly improved as it results in visible memory pressure, and also doesn't schedule indefinitely, which previously made tracing and other introspection difficult (ie. it's clamped at 2*HZ per allocation through MEMCG_MAX_HIGH_DELAY_JIFFIES). Contrast the previous results with a kernel with this patch: [root@ktst ~]# systemd-run -p MemoryHigh=100M -p MemorySwapMax=1 \ > --wait -t timeout 300 /root/mdf [...] 95 1002 96 1000 97 1002 98 1003 99 1000 100 1043 101 84724 102 330628 103 610511 104 1016265 105 1503969 106 2391692 107 2872061 108 3248003 109 4791904 110 5759832 111 6912509 112 8127818 113 9472203 114 12287622 115 12480079 116 14144008 117 15808029 118 16384500 119 16383242 120 16384979 [...] As you can see, in the normal case, memory allocation takes around 1000 usec. However, as we exceed our memory.high, things start to increase exponentially, but fairly leniently at first. Our first megabyte over memory.high takes us 0.16 seconds, then the next is 0.46 seconds, then the next is almost an entire second. This gets worse until we reach our eventual 2*HZ clamp per batch, resulting in 16 seconds per megabyte. However, this is still making forward progress, so permits tracing or further analysis with programs like GDB. We use an exponential curve for our delay penalty for a few reasons: 1. We run mem_cgroup_handle_over_high to potentially do reclaim after we've already performed allocations, which means that temporarily going over memory.high by a small amount may be perfectly legitimate, even for compliant workloads. We don't want to unduly penalise such cases. 2. An exponential curve (as opposed to a static or linear delay) allows ramping up memory pressure stats more gradually, which can be useful to work out that you have set memory.high too low, without destroying application performance entirely. This patch expands on earlier work by Johannes Weiner. Thanks! [akpm@linux-foundation.org: fix max() warning] [akpm@linux-foundation.org: fix __udivdi3 ref on 32-bit] [akpm@linux-foundation.org: fix it even more] [chris@chrisdown.name: fix 64-bit divide even more] Link: http://lkml.kernel.org/r/20190723180700.GA29459@chrisdown.name Signed-off-by: Chris Down <chris@chrisdown.name> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Nathan Chancellor <natechancellor@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-09-23 22:34:55 +00:00
/*
* Don't sleep if the amount of jiffies this memcg owes us is so low
* that it's not even worth doing, in an attempt to be nice to those who
* go only a small amount over their memory.high value and maybe haven't
* been aggressively reclaimed enough yet.
*/
if (penalty_jiffies <= HZ / 100)
goto out;
/*
* If we exit early, we're guaranteed to die (since
* schedule_timeout_killable sets TASK_KILLABLE). This means we don't
* need to account for any ill-begotten jiffies to pay them off later.
*/
psi_memstall_enter(&pflags);
schedule_timeout_killable(penalty_jiffies);
psi_memstall_leave(&pflags);
out:
css_put(&memcg->css);
memcg: punt high overage reclaim to return-to-userland path Currently, try_charge() tries to reclaim memory synchronously when the high limit is breached; however, if the allocation doesn't have __GFP_WAIT, synchronous reclaim is skipped. If a process performs only speculative allocations, it can blow way past the high limit. This is actually easily reproducible by simply doing "find /". slab/slub allocator tries speculative allocations first, so as long as there's memory which can be consumed without blocking, it can keep allocating memory regardless of the high limit. This patch makes try_charge() always punt the over-high reclaim to the return-to-userland path. If try_charge() detects that high limit is breached, it adds the overage to current->memcg_nr_pages_over_high and schedules execution of mem_cgroup_handle_over_high() which performs synchronous reclaim from the return-to-userland path. As long as kernel doesn't have a run-away allocation spree, this should provide enough protection while making kmemcg behave more consistently. It also has the following benefits. - All over-high reclaims can use GFP_KERNEL regardless of the specific gfp mask in use, e.g. GFP_NOFS, when the limit was breached. - It copes with prio inversion. Previously, a low-prio task with small memory.high might perform over-high reclaim with a bunch of locks held. If a higher prio task needed any of these locks, it would have to wait until the low prio task finished reclaim and released the locks. By handing over-high reclaim to the task exit path this issue can be avoided. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Michal Hocko <mhocko@kernel.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:46:11 +00:00
}
mm: memcontrol: rewrite charge API These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 4): The memcg charge API charges pages before they are rmapped - i.e. have an actual "type" - and so every callsite needs its own set of charge and uncharge functions to know what type is being operated on. Worse, uncharge has to happen from a context that is still type-specific, rather than at the end of the page's lifetime with exclusive access, and so requires a lot of synchronization. Rewrite the charge API to provide a generic set of try_charge(), commit_charge() and cancel_charge() transaction operations, much like what's currently done for swap-in: mem_cgroup_try_charge() attempts to reserve a charge, reclaiming pages from the memcg if necessary. mem_cgroup_commit_charge() commits the page to the charge once it has a valid page->mapping and PageAnon() reliably tells the type. mem_cgroup_cancel_charge() aborts the transaction. This reduces the charge API and enables subsequent patches to drastically simplify uncharging. As pages need to be committed after rmap is established but before they are added to the LRU, page_add_new_anon_rmap() must stop doing LRU additions again. Revive lru_cache_add_active_or_unevictable(). [hughd@google.com: fix shmem_unuse] [hughd@google.com: Add comments on the private use of -EAGAIN] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Hugh Dickins <hughd@google.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:20 +00:00
static int try_charge(struct mem_cgroup *memcg, gfp_t gfp_mask,
unsigned int nr_pages)
Memory controller: memory accounting Add the accounting hooks. The accounting is carried out for RSS and Page Cache (unmapped) pages. There is now a common limit and accounting for both. The RSS accounting is accounted at page_add_*_rmap() and page_remove_rmap() time. Page cache is accounted at add_to_page_cache(), __delete_from_page_cache(). Swap cache is also accounted for. Each page's page_cgroup is protected with the last bit of the page_cgroup pointer, this makes handling of race conditions involving simultaneous mappings of a page easier. A reference count is kept in the page_cgroup to deal with cases where a page might be unmapped from the RSS of all tasks, but still lives in the page cache. Credits go to Vaidyanathan Srinivasan for helping with reference counting work of the page cgroup. Almost all of the page cache accounting code has help from Vaidyanathan Srinivasan. [hugh@veritas.com: fix swapoff breakage] [akpm@linux-foundation.org: fix locking] Signed-off-by: Vaidyanathan Srinivasan <svaidy@linux.vnet.ibm.com> Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com> Cc: Pavel Emelianov <xemul@openvz.org> Cc: Paul Menage <menage@google.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Cc: Kirill Korotaev <dev@sw.ru> Cc: Herbert Poetzl <herbert@13thfloor.at> Cc: David Rientjes <rientjes@google.com> Cc: <Valdis.Kletnieks@vt.edu> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-07 08:13:53 +00:00
{
mm: memcontrol: fix excessive complexity in memory.stat reporting We've seen memory.stat reads in top-level cgroups take up to fourteen seconds during a userspace bug that created tens of thousands of ghost cgroups pinned by lingering page cache. Even with a more reasonable number of cgroups, aggregating memory.stat is unnecessarily heavy. The complexity is this: nr_cgroups * nr_stat_items * nr_possible_cpus where the stat items are ~70 at this point. With 128 cgroups and 128 CPUs - decent, not enormous setups - reading the top-level memory.stat has to aggregate over a million per-cpu counters. This doesn't scale. Instead of spreading the source of truth across all CPUs, use the per-cpu counters merely to batch updates to shared atomic counters. This is the same as the per-cpu stocks we use for charging memory to the shared atomic page_counters, and also the way the global vmstat counters are implemented. Vmstat has elaborate spilling thresholds that depend on the number of CPUs, amount of memory, and memory pressure - carefully balancing the cost of counter updates with the amount of per-cpu error. That's because the vmstat counters are system-wide, but also used for decisions inside the kernel (e.g. NR_FREE_PAGES in the allocator). Neither is true for the memory controller. Use the same static batch size we already use for page_counter updates during charging. The per-cpu error in the stats will be 128k, which is an acceptable ratio of cores to memory accounting granularity. [hannes@cmpxchg.org: fix warning in __this_cpu_xchg() calls] Link: http://lkml.kernel.org/r/20171201135750.GB8097@cmpxchg.org Link: http://lkml.kernel.org/r/20171103153336.24044-3-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Michal Hocko <mhocko@suse.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-02-01 00:16:45 +00:00
unsigned int batch = max(MEMCG_CHARGE_BATCH, nr_pages);
int nr_retries = MEM_CGROUP_RECLAIM_RETRIES;
mm: memcontrol: fold mem_cgroup_do_charge() These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 13): This function was split out because mem_cgroup_try_charge() got too big. But having essentially one sequence of operations arbitrarily split in half is not good for reworking the code. Fold it back in. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-06 23:05:42 +00:00
struct mem_cgroup *mem_over_limit;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
struct page_counter *counter;
mm: memcontrol: fold mem_cgroup_do_charge() These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 13): This function was split out because mem_cgroup_try_charge() got too big. But having essentially one sequence of operations arbitrarily split in half is not good for reworking the code. Fold it back in. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-06 23:05:42 +00:00
unsigned long nr_reclaimed;
mm: memcontrol: fix transparent huge page allocations under pressure In a memcg with even just moderate cache pressure, success rates for transparent huge page allocations drop to zero, wasting a lot of effort that the allocator puts into assembling these pages. The reason for this is that the memcg reclaim code was never designed for higher-order charges. It reclaims in small batches until there is room for at least one page. Huge page charges only succeed when these batches add up over a series of huge faults, which is unlikely under any significant load involving order-0 allocations in the group. Remove that loop on the memcg side in favor of passing the actual reclaim goal to direct reclaim, which is already set up and optimized to meet higher-order goals efficiently. This brings memcg's THP policy in line with the system policy: if the allocator painstakingly assembles a hugepage, memcg will at least make an honest effort to charge it. As a result, transparent hugepage allocation rates amid cache activity are drastically improved: vanilla patched pgalloc 4717530.80 ( +0.00%) 4451376.40 ( -5.64%) pgfault 491370.60 ( +0.00%) 225477.40 ( -54.11%) pgmajfault 2.00 ( +0.00%) 1.80 ( -6.67%) thp_fault_alloc 0.00 ( +0.00%) 531.60 (+100.00%) thp_fault_fallback 749.00 ( +0.00%) 217.40 ( -70.88%) [ Note: this may in turn increase memory consumption from internal fragmentation, which is an inherent risk of transparent hugepages. Some setups may have to adjust the memcg limits accordingly to accomodate this - or, if the machine is already packed to capacity, disable the transparent huge page feature. ] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Dave Hansen <dave@sr71.net> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:28:56 +00:00
bool may_swap = true;
bool drained = false;
memcg, oom: move out_of_memory back to the charge path Commit 3812c8c8f395 ("mm: memcg: do not trap chargers with full callstack on OOM") has changed the ENOMEM semantic of memcg charges. Rather than invoking the oom killer from the charging context it delays the oom killer to the page fault path (pagefault_out_of_memory). This in turn means that many users (e.g. slab or g-u-p) will get ENOMEM when the corresponding memcg hits the hard limit and the memcg is is OOM. This is behavior is inconsistent with !memcg case where the oom killer is invoked from the allocation context and the allocator keeps retrying until it succeeds. The difference in the behavior is user visible. mmap(MAP_POPULATE) might result in not fully populated ranges while the mmap return code doesn't tell that to the userspace. Random syscalls might fail with ENOMEM etc. The primary motivation of the different memcg oom semantic was the deadlock avoidance. Things have changed since then, though. We have an async oom teardown by the oom reaper now and so we do not have to rely on the victim to tear down its memory anymore. Therefore we can return to the original semantic as long as the memcg oom killer is not handed over to the users space. There is still one thing to be careful about here though. If the oom killer is not able to make any forward progress - e.g. because there is no eligible task to kill - then we have to bail out of the charge path to prevent from same class of deadlocks. We have basically two options here. Either we fail the charge with ENOMEM or force the charge and allow overcharge. The first option has been considered more harmful than useful because rare inconsistencies in the ENOMEM behavior is hard to test for and error prone. Basically the same reason why the page allocator doesn't fail allocations under such conditions. The later might allow runaways but those should be really unlikely unless somebody misconfigures the system. E.g. allowing to migrate tasks away from the memcg to a different unlimited memcg with move_charge_at_immigrate disabled. Link: http://lkml.kernel.org/r/20180628151101.25307-1-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Acked-by: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Shakeel Butt <shakeelb@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:11 +00:00
enum oom_status oom_status;
if (mem_cgroup_is_root(memcg))
memcg: ratify and consolidate over-charge handling try_charge() is the main charging logic of memcg. When it hits the limit but either can't fail the allocation due to __GFP_NOFAIL or the task is likely to free memory very soon, being OOM killed, has SIGKILL pending or exiting, it "bypasses" the charge to the root memcg and returns -EINTR. While this is one approach which can be taken for these situations, it has several issues. * It unnecessarily lies about the reality. The number itself doesn't go over the limit but the actual usage does. memcg is either forced to or actively chooses to go over the limit because that is the right behavior under the circumstances, which is completely fine, but, if at all avoidable, it shouldn't be misrepresenting what's happening by sneaking the charges into the root memcg. * Despite trying, we already do over-charge. kmemcg can't deal with switching over to the root memcg by the point try_charge() returns -EINTR, so it open-codes over-charing. * It complicates the callers. Each try_charge() user has to handle the weird -EINTR exception. memcg_charge_kmem() does the manual over-charging. mem_cgroup_do_precharge() performs unnecessary uncharging of root memcg, which BTW is inconsistent with what memcg_charge_kmem() does but not broken as [un]charging are noops on root memcg. mem_cgroup_try_charge() needs to switch the returned cgroup to the root one. The reality is that in memcg there are cases where we are forced and/or willing to go over the limit. Each such case needs to be scrutinized and justified but there definitely are situations where that is the right thing to do. We alredy do this but with a superficial and inconsistent disguise which leads to unnecessary complications. This patch updates try_charge() so that it over-charges and returns 0 when deemed necessary. -EINTR return is removed along with all special case handling in the callers. While at it, remove the local variable @ret, which was initialized to zero and never changed, along with done: label which just returned the always zero @ret. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:46:17 +00:00
return 0;
mm: memcontrol: fold mem_cgroup_do_charge() These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 13): This function was split out because mem_cgroup_try_charge() got too big. But having essentially one sequence of operations arbitrarily split in half is not good for reworking the code. Fold it back in. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-06 23:05:42 +00:00
retry:
if (consume_stock(memcg, nr_pages))
memcg: ratify and consolidate over-charge handling try_charge() is the main charging logic of memcg. When it hits the limit but either can't fail the allocation due to __GFP_NOFAIL or the task is likely to free memory very soon, being OOM killed, has SIGKILL pending or exiting, it "bypasses" the charge to the root memcg and returns -EINTR. While this is one approach which can be taken for these situations, it has several issues. * It unnecessarily lies about the reality. The number itself doesn't go over the limit but the actual usage does. memcg is either forced to or actively chooses to go over the limit because that is the right behavior under the circumstances, which is completely fine, but, if at all avoidable, it shouldn't be misrepresenting what's happening by sneaking the charges into the root memcg. * Despite trying, we already do over-charge. kmemcg can't deal with switching over to the root memcg by the point try_charge() returns -EINTR, so it open-codes over-charing. * It complicates the callers. Each try_charge() user has to handle the weird -EINTR exception. memcg_charge_kmem() does the manual over-charging. mem_cgroup_do_precharge() performs unnecessary uncharging of root memcg, which BTW is inconsistent with what memcg_charge_kmem() does but not broken as [un]charging are noops on root memcg. mem_cgroup_try_charge() needs to switch the returned cgroup to the root one. The reality is that in memcg there are cases where we are forced and/or willing to go over the limit. Each such case needs to be scrutinized and justified but there definitely are situations where that is the right thing to do. We alredy do this but with a superficial and inconsistent disguise which leads to unnecessary complications. This patch updates try_charge() so that it over-charges and returns 0 when deemed necessary. -EINTR return is removed along with all special case handling in the callers. While at it, remove the local variable @ret, which was initialized to zero and never changed, along with done: label which just returned the always zero @ret. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:46:17 +00:00
return 0;
Memory controller: memory accounting Add the accounting hooks. The accounting is carried out for RSS and Page Cache (unmapped) pages. There is now a common limit and accounting for both. The RSS accounting is accounted at page_add_*_rmap() and page_remove_rmap() time. Page cache is accounted at add_to_page_cache(), __delete_from_page_cache(). Swap cache is also accounted for. Each page's page_cgroup is protected with the last bit of the page_cgroup pointer, this makes handling of race conditions involving simultaneous mappings of a page easier. A reference count is kept in the page_cgroup to deal with cases where a page might be unmapped from the RSS of all tasks, but still lives in the page cache. Credits go to Vaidyanathan Srinivasan for helping with reference counting work of the page cgroup. Almost all of the page cache accounting code has help from Vaidyanathan Srinivasan. [hugh@veritas.com: fix swapoff breakage] [akpm@linux-foundation.org: fix locking] Signed-off-by: Vaidyanathan Srinivasan <svaidy@linux.vnet.ibm.com> Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com> Cc: Pavel Emelianov <xemul@openvz.org> Cc: Paul Menage <menage@google.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Cc: Kirill Korotaev <dev@sw.ru> Cc: Herbert Poetzl <herbert@13thfloor.at> Cc: David Rientjes <rientjes@google.com> Cc: <Valdis.Kletnieks@vt.edu> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-07 08:13:53 +00:00
if (!do_memsw_account() ||
page_counter_try_charge(&memcg->memsw, batch, &counter)) {
if (page_counter_try_charge(&memcg->memory, batch, &counter))
mm: memcontrol: fold mem_cgroup_do_charge() These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 13): This function was split out because mem_cgroup_try_charge() got too big. But having essentially one sequence of operations arbitrarily split in half is not good for reworking the code. Fold it back in. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-06 23:05:42 +00:00
goto done_restock;
if (do_memsw_account())
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
page_counter_uncharge(&memcg->memsw, batch);
mem_over_limit = mem_cgroup_from_counter(counter, memory);
} else {
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
mem_over_limit = mem_cgroup_from_counter(counter, memsw);
mm: memcontrol: fix transparent huge page allocations under pressure In a memcg with even just moderate cache pressure, success rates for transparent huge page allocations drop to zero, wasting a lot of effort that the allocator puts into assembling these pages. The reason for this is that the memcg reclaim code was never designed for higher-order charges. It reclaims in small batches until there is room for at least one page. Huge page charges only succeed when these batches add up over a series of huge faults, which is unlikely under any significant load involving order-0 allocations in the group. Remove that loop on the memcg side in favor of passing the actual reclaim goal to direct reclaim, which is already set up and optimized to meet higher-order goals efficiently. This brings memcg's THP policy in line with the system policy: if the allocator painstakingly assembles a hugepage, memcg will at least make an honest effort to charge it. As a result, transparent hugepage allocation rates amid cache activity are drastically improved: vanilla patched pgalloc 4717530.80 ( +0.00%) 4451376.40 ( -5.64%) pgfault 491370.60 ( +0.00%) 225477.40 ( -54.11%) pgmajfault 2.00 ( +0.00%) 1.80 ( -6.67%) thp_fault_alloc 0.00 ( +0.00%) 531.60 (+100.00%) thp_fault_fallback 749.00 ( +0.00%) 217.40 ( -70.88%) [ Note: this may in turn increase memory consumption from internal fragmentation, which is an inherent risk of transparent hugepages. Some setups may have to adjust the memcg limits accordingly to accomodate this - or, if the machine is already packed to capacity, disable the transparent huge page feature. ] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Dave Hansen <dave@sr71.net> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:28:56 +00:00
may_swap = false;
}
memcg: introduce charge-commit-cancel style of functions There is a small race in do_swap_page(). When the page swapped-in is charged, the mapcount can be greater than 0. But, at the same time some process (shares it ) call unmap and make mapcount 1->0 and the page is uncharged. CPUA CPUB mapcount == 1. (1) charge if mapcount==0 zap_pte_range() (2) mapcount 1 => 0. (3) uncharge(). (success) (4) set page's rmap() mapcount 0=>1 Then, this swap page's account is leaked. For fixing this, I added a new interface. - charge account to res_counter by PAGE_SIZE and try to free pages if necessary. - commit register page_cgroup and add to LRU if necessary. - cancel uncharge PAGE_SIZE because of do_swap_page failure. CPUA (1) charge (always) (2) set page's rmap (mapcount > 0) (3) commit charge was necessary or not after set_pte(). This protocol uses PCG_USED bit on page_cgroup for avoiding over accounting. Usual mem_cgroup_charge_common() does charge -> commit at a time. And this patch also adds following function to clarify all charges. - mem_cgroup_newpage_charge() ....replacement for mem_cgroup_charge() called against newly allocated anon pages. - mem_cgroup_charge_migrate_fixup() called only from remove_migration_ptes(). we'll have to rewrite this later.(this patch just keeps old behavior) This function will be removed by additional patch to make migration clearer. Good for clarifying "what we do" Then, we have 4 following charge points. - newpage - swap-in - add-to-cache. - migration. [akpm@linux-foundation.org: add missing inline directives to stubs] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:07:48 +00:00
mm: memcontrol: fold mem_cgroup_do_charge() These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 13): This function was split out because mem_cgroup_try_charge() got too big. But having essentially one sequence of operations arbitrarily split in half is not good for reworking the code. Fold it back in. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-06 23:05:42 +00:00
if (batch > nr_pages) {
batch = nr_pages;
goto retry;
}
mm: memcontrol: fix network errors from failing __GFP_ATOMIC charges While upgrading from 4.16 to 5.2, we noticed these allocation errors in the log of the new kernel: SLUB: Unable to allocate memory on node -1, gfp=0xa20(GFP_ATOMIC) cache: tw_sock_TCPv6(960:helper-logs), object size: 232, buffer size: 240, default order: 1, min order: 0 node 0: slabs: 5, objs: 170, free: 0 slab_out_of_memory+1 ___slab_alloc+969 __slab_alloc+14 kmem_cache_alloc+346 inet_twsk_alloc+60 tcp_time_wait+46 tcp_fin+206 tcp_data_queue+2034 tcp_rcv_state_process+784 tcp_v6_do_rcv+405 __release_sock+118 tcp_close+385 inet_release+46 __sock_release+55 sock_close+17 __fput+170 task_work_run+127 exit_to_usermode_loop+191 do_syscall_64+212 entry_SYSCALL_64_after_hwframe+68 accompanied by an increase in machines going completely radio silent under memory pressure. One thing that changed since 4.16 is e699e2c6a654 ("net, mm: account sock objects to kmemcg"), which made these slab caches subject to cgroup memory accounting and control. The problem with that is that cgroups, unlike the page allocator, do not maintain dedicated atomic reserves. As a cgroup's usage hovers at its limit, atomic allocations - such as done during network rx - can fail consistently for extended periods of time. The kernel is not able to operate under these conditions. We don't want to revert the culprit patch, because it indeed tracks a potentially substantial amount of memory used by a cgroup. We also don't want to implement dedicated atomic reserves for cgroups. There is no point in keeping a fixed margin of unused bytes in the cgroup's memory budget to accomodate a consumer that is impossible to predict - we'd be wasting memory and get into configuration headaches, not unlike what we have going with min_free_kbytes. We do this for physical mem because we have to, but cgroups are an accounting game. Instead, account these privileged allocations to the cgroup, but let them bypass the configured limit if they have to. This way, we get the benefits of accounting the consumed memory and have it exert pressure on the rest of the cgroup, but like with the page allocator, we shift the burden of reclaimining on behalf of atomic allocations onto the regular allocations that can block. Link: http://lkml.kernel.org/r/20191022233708.365764-1-hannes@cmpxchg.org Fixes: e699e2c6a654 ("net, mm: account sock objects to kmemcg") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: <stable@vger.kernel.org> [4.18+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-11-06 05:17:13 +00:00
/*
* Memcg doesn't have a dedicated reserve for atomic
* allocations. But like the global atomic pool, we need to
* put the burden of reclaim on regular allocation requests
* and let these go through as privileged allocations.
*/
if (gfp_mask & __GFP_ATOMIC)
goto force;
/*
* Unlike in global OOM situations, memcg is not in a physical
* memory shortage. Allow dying and OOM-killed tasks to
* bypass the last charges so that they can exit quickly and
* free their memory.
*/
memcg: killed threads should not invoke memcg OOM killer If a memory cgroup contains a single process with many threads (including different process group sharing the mm) then it is possible to trigger a race when the oom killer complains that there are no oom elible tasks and complain into the log which is both annoying and confusing because there is no actual problem. The race looks as follows: P1 oom_reaper P2 try_charge try_charge mem_cgroup_out_of_memory mutex_lock(oom_lock) out_of_memory oom_kill_process(P1,P2) wake_oom_reaper mutex_unlock(oom_lock) oom_reap_task mutex_lock(oom_lock) select_bad_process # no victim The problem is more visible with many threads. Fix this by checking for fatal_signal_pending from mem_cgroup_out_of_memory when the oom_lock is already held. The oom bypass is safe because we do the same early in the try_charge path already. The situation migh have changed in the mean time. It should be safe to check for fatal_signal_pending and tsk_is_oom_victim but for a better code readability abstract the current charge bypass condition into should_force_charge and reuse it from that path. " Link: http://lkml.kernel.org/r/01370f70-e1f6-ebe4-b95e-0df21a0bc15e@i-love.sakura.ne.jp Signed-off-by: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: David Rientjes <rientjes@google.com> Cc: Kirill Tkhai <ktkhai@virtuozzo.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-03-05 23:46:47 +00:00
if (unlikely(should_force_charge()))
memcg: ratify and consolidate over-charge handling try_charge() is the main charging logic of memcg. When it hits the limit but either can't fail the allocation due to __GFP_NOFAIL or the task is likely to free memory very soon, being OOM killed, has SIGKILL pending or exiting, it "bypasses" the charge to the root memcg and returns -EINTR. While this is one approach which can be taken for these situations, it has several issues. * It unnecessarily lies about the reality. The number itself doesn't go over the limit but the actual usage does. memcg is either forced to or actively chooses to go over the limit because that is the right behavior under the circumstances, which is completely fine, but, if at all avoidable, it shouldn't be misrepresenting what's happening by sneaking the charges into the root memcg. * Despite trying, we already do over-charge. kmemcg can't deal with switching over to the root memcg by the point try_charge() returns -EINTR, so it open-codes over-charing. * It complicates the callers. Each try_charge() user has to handle the weird -EINTR exception. memcg_charge_kmem() does the manual over-charging. mem_cgroup_do_precharge() performs unnecessary uncharging of root memcg, which BTW is inconsistent with what memcg_charge_kmem() does but not broken as [un]charging are noops on root memcg. mem_cgroup_try_charge() needs to switch the returned cgroup to the root one. The reality is that in memcg there are cases where we are forced and/or willing to go over the limit. Each such case needs to be scrutinized and justified but there definitely are situations where that is the right thing to do. We alredy do this but with a superficial and inconsistent disguise which leads to unnecessary complications. This patch updates try_charge() so that it over-charges and returns 0 when deemed necessary. -EINTR return is removed along with all special case handling in the callers. While at it, remove the local variable @ret, which was initialized to zero and never changed, along with done: label which just returned the always zero @ret. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:46:17 +00:00
goto force;
mm: memcontrol: do not recurse in direct reclaim On 4.0, we saw a stack corruption from a page fault entering direct memory cgroup reclaim, calling into btrfs_releasepage(), which then tried to allocate an extent and recursed back into a kmem charge ad nauseam: [...] btrfs_releasepage+0x2c/0x30 try_to_release_page+0x32/0x50 shrink_page_list+0x6da/0x7a0 shrink_inactive_list+0x1e5/0x510 shrink_lruvec+0x605/0x7f0 shrink_zone+0xee/0x320 do_try_to_free_pages+0x174/0x440 try_to_free_mem_cgroup_pages+0xa7/0x130 try_charge+0x17b/0x830 memcg_charge_kmem+0x40/0x80 new_slab+0x2d9/0x5a0 __slab_alloc+0x2fd/0x44f kmem_cache_alloc+0x193/0x1e0 alloc_extent_state+0x21/0xc0 __clear_extent_bit+0x2b5/0x400 try_release_extent_mapping+0x1a3/0x220 __btrfs_releasepage+0x31/0x70 btrfs_releasepage+0x2c/0x30 try_to_release_page+0x32/0x50 shrink_page_list+0x6da/0x7a0 shrink_inactive_list+0x1e5/0x510 shrink_lruvec+0x605/0x7f0 shrink_zone+0xee/0x320 do_try_to_free_pages+0x174/0x440 try_to_free_mem_cgroup_pages+0xa7/0x130 try_charge+0x17b/0x830 mem_cgroup_try_charge+0x65/0x1c0 handle_mm_fault+0x117f/0x1510 __do_page_fault+0x177/0x420 do_page_fault+0xc/0x10 page_fault+0x22/0x30 On later kernels, kmem charging is opt-in rather than opt-out, and that particular kmem allocation in btrfs_releasepage() is no longer being charged and won't recurse and overrun the stack anymore. But it's not impossible for an accounted allocation to happen from the memcg direct reclaim context, and we needed to reproduce this crash many times before we even got a useful stack trace out of it. Like other direct reclaimers, mark tasks in memcg reclaim PF_MEMALLOC to avoid recursing into any other form of direct reclaim. Then let recursive charges from PF_MEMALLOC contexts bypass the cgroup limit. Link: http://lkml.kernel.org/r/20161025141050.GA13019@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-10-28 00:46:56 +00:00
/*
* Prevent unbounded recursion when reclaim operations need to
* allocate memory. This might exceed the limits temporarily,
* but we prefer facilitating memory reclaim and getting back
* under the limit over triggering OOM kills in these cases.
*/
if (unlikely(current->flags & PF_MEMALLOC))
goto force;
if (unlikely(task_in_memcg_oom(current)))
goto nomem;
mm, page_alloc: distinguish between being unable to sleep, unwilling to sleep and avoiding waking kswapd __GFP_WAIT has been used to identify atomic context in callers that hold spinlocks or are in interrupts. They are expected to be high priority and have access one of two watermarks lower than "min" which can be referred to as the "atomic reserve". __GFP_HIGH users get access to the first lower watermark and can be called the "high priority reserve". Over time, callers had a requirement to not block when fallback options were available. Some have abused __GFP_WAIT leading to a situation where an optimisitic allocation with a fallback option can access atomic reserves. This patch uses __GFP_ATOMIC to identify callers that are truely atomic, cannot sleep and have no alternative. High priority users continue to use __GFP_HIGH. __GFP_DIRECT_RECLAIM identifies callers that can sleep and are willing to enter direct reclaim. __GFP_KSWAPD_RECLAIM to identify callers that want to wake kswapd for background reclaim. __GFP_WAIT is redefined as a caller that is willing to enter direct reclaim and wake kswapd for background reclaim. This patch then converts a number of sites o __GFP_ATOMIC is used by callers that are high priority and have memory pools for those requests. GFP_ATOMIC uses this flag. o Callers that have a limited mempool to guarantee forward progress clear __GFP_DIRECT_RECLAIM but keep __GFP_KSWAPD_RECLAIM. bio allocations fall into this category where kswapd will still be woken but atomic reserves are not used as there is a one-entry mempool to guarantee progress. o Callers that are checking if they are non-blocking should use the helper gfpflags_allow_blocking() where possible. This is because checking for __GFP_WAIT as was done historically now can trigger false positives. Some exceptions like dm-crypt.c exist where the code intent is clearer if __GFP_DIRECT_RECLAIM is used instead of the helper due to flag manipulations. o Callers that built their own GFP flags instead of starting with GFP_KERNEL and friends now also need to specify __GFP_KSWAPD_RECLAIM. The first key hazard to watch out for is callers that removed __GFP_WAIT and was depending on access to atomic reserves for inconspicuous reasons. In some cases it may be appropriate for them to use __GFP_HIGH. The second key hazard is callers that assembled their own combination of GFP flags instead of starting with something like GFP_KERNEL. They may now wish to specify __GFP_KSWAPD_RECLAIM. It's almost certainly harmless if it's missed in most cases as other activity will wake kswapd. Signed-off-by: Mel Gorman <mgorman@techsingularity.net> Acked-by: Vlastimil Babka <vbabka@suse.cz> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Vitaly Wool <vitalywool@gmail.com> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-07 00:28:21 +00:00
if (!gfpflags_allow_blocking(gfp_mask))
mm: memcontrol: fold mem_cgroup_do_charge() These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 13): This function was split out because mem_cgroup_try_charge() got too big. But having essentially one sequence of operations arbitrarily split in half is not good for reworking the code. Fold it back in. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-06 23:05:42 +00:00
goto nomem;
mm: memcg: make sure memory.events is uptodate when waking pollers Commit a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting") added per-cpu drift to all memory cgroup stats and events shown in memory.stat and memory.events. For memory.stat this is acceptable. But memory.events issues file notifications, and somebody polling the file for changes will be confused when the counters in it are unchanged after a wakeup. Luckily, the events in memory.events - MEMCG_LOW, MEMCG_HIGH, MEMCG_MAX, MEMCG_OOM - are sufficiently rare and high-level that we don't need per-cpu buffering for them: MEMCG_HIGH and MEMCG_MAX would be the most frequent, but they're counting invocations of reclaim, which is a complex operation that touches many shared cachelines. This splits memory.events from the generic VM events and tracks them in their own, unbuffered atomic counters. That's also cleaner, as it eliminates the ugly enum nesting of VM and cgroup events. [hannes@cmpxchg.org: "array subscript is above array bounds"] Link: http://lkml.kernel.org/r/20180406155441.GA20806@cmpxchg.org Link: http://lkml.kernel.org/r/20180405175507.GA24817@cmpxchg.org Fixes: a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Tejun Heo <tj@kernel.org> Acked-by: Tejun Heo <tj@kernel.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Cc: Rik van Riel <riel@surriel.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-10 23:29:45 +00:00
memcg_memory_event(mem_over_limit, MEMCG_MAX);
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
mm: memcontrol: fix transparent huge page allocations under pressure In a memcg with even just moderate cache pressure, success rates for transparent huge page allocations drop to zero, wasting a lot of effort that the allocator puts into assembling these pages. The reason for this is that the memcg reclaim code was never designed for higher-order charges. It reclaims in small batches until there is room for at least one page. Huge page charges only succeed when these batches add up over a series of huge faults, which is unlikely under any significant load involving order-0 allocations in the group. Remove that loop on the memcg side in favor of passing the actual reclaim goal to direct reclaim, which is already set up and optimized to meet higher-order goals efficiently. This brings memcg's THP policy in line with the system policy: if the allocator painstakingly assembles a hugepage, memcg will at least make an honest effort to charge it. As a result, transparent hugepage allocation rates amid cache activity are drastically improved: vanilla patched pgalloc 4717530.80 ( +0.00%) 4451376.40 ( -5.64%) pgfault 491370.60 ( +0.00%) 225477.40 ( -54.11%) pgmajfault 2.00 ( +0.00%) 1.80 ( -6.67%) thp_fault_alloc 0.00 ( +0.00%) 531.60 (+100.00%) thp_fault_fallback 749.00 ( +0.00%) 217.40 ( -70.88%) [ Note: this may in turn increase memory consumption from internal fragmentation, which is an inherent risk of transparent hugepages. Some setups may have to adjust the memcg limits accordingly to accomodate this - or, if the machine is already packed to capacity, disable the transparent huge page feature. ] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Dave Hansen <dave@sr71.net> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:28:56 +00:00
nr_reclaimed = try_to_free_mem_cgroup_pages(mem_over_limit, nr_pages,
gfp_mask, may_swap);
mm: memcontrol: fold mem_cgroup_do_charge() These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 13): This function was split out because mem_cgroup_try_charge() got too big. But having essentially one sequence of operations arbitrarily split in half is not good for reworking the code. Fold it back in. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-06 23:05:42 +00:00
if (mem_cgroup_margin(mem_over_limit) >= nr_pages)
mm: memcontrol: fold mem_cgroup_do_charge() These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 13): This function was split out because mem_cgroup_try_charge() got too big. But having essentially one sequence of operations arbitrarily split in half is not good for reworking the code. Fold it back in. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-06 23:05:42 +00:00
goto retry;
mm: memcontrol: fix transparent huge page allocations under pressure In a memcg with even just moderate cache pressure, success rates for transparent huge page allocations drop to zero, wasting a lot of effort that the allocator puts into assembling these pages. The reason for this is that the memcg reclaim code was never designed for higher-order charges. It reclaims in small batches until there is room for at least one page. Huge page charges only succeed when these batches add up over a series of huge faults, which is unlikely under any significant load involving order-0 allocations in the group. Remove that loop on the memcg side in favor of passing the actual reclaim goal to direct reclaim, which is already set up and optimized to meet higher-order goals efficiently. This brings memcg's THP policy in line with the system policy: if the allocator painstakingly assembles a hugepage, memcg will at least make an honest effort to charge it. As a result, transparent hugepage allocation rates amid cache activity are drastically improved: vanilla patched pgalloc 4717530.80 ( +0.00%) 4451376.40 ( -5.64%) pgfault 491370.60 ( +0.00%) 225477.40 ( -54.11%) pgmajfault 2.00 ( +0.00%) 1.80 ( -6.67%) thp_fault_alloc 0.00 ( +0.00%) 531.60 (+100.00%) thp_fault_fallback 749.00 ( +0.00%) 217.40 ( -70.88%) [ Note: this may in turn increase memory consumption from internal fragmentation, which is an inherent risk of transparent hugepages. Some setups may have to adjust the memcg limits accordingly to accomodate this - or, if the machine is already packed to capacity, disable the transparent huge page feature. ] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Dave Hansen <dave@sr71.net> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:28:56 +00:00
if (!drained) {
drain_all_stock(mem_over_limit);
mm: memcontrol: fix transparent huge page allocations under pressure In a memcg with even just moderate cache pressure, success rates for transparent huge page allocations drop to zero, wasting a lot of effort that the allocator puts into assembling these pages. The reason for this is that the memcg reclaim code was never designed for higher-order charges. It reclaims in small batches until there is room for at least one page. Huge page charges only succeed when these batches add up over a series of huge faults, which is unlikely under any significant load involving order-0 allocations in the group. Remove that loop on the memcg side in favor of passing the actual reclaim goal to direct reclaim, which is already set up and optimized to meet higher-order goals efficiently. This brings memcg's THP policy in line with the system policy: if the allocator painstakingly assembles a hugepage, memcg will at least make an honest effort to charge it. As a result, transparent hugepage allocation rates amid cache activity are drastically improved: vanilla patched pgalloc 4717530.80 ( +0.00%) 4451376.40 ( -5.64%) pgfault 491370.60 ( +0.00%) 225477.40 ( -54.11%) pgmajfault 2.00 ( +0.00%) 1.80 ( -6.67%) thp_fault_alloc 0.00 ( +0.00%) 531.60 (+100.00%) thp_fault_fallback 749.00 ( +0.00%) 217.40 ( -70.88%) [ Note: this may in turn increase memory consumption from internal fragmentation, which is an inherent risk of transparent hugepages. Some setups may have to adjust the memcg limits accordingly to accomodate this - or, if the machine is already packed to capacity, disable the transparent huge page feature. ] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Dave Hansen <dave@sr71.net> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:28:56 +00:00
drained = true;
goto retry;
}
if (gfp_mask & __GFP_NORETRY)
goto nomem;
mm: memcontrol: fold mem_cgroup_do_charge() These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 13): This function was split out because mem_cgroup_try_charge() got too big. But having essentially one sequence of operations arbitrarily split in half is not good for reworking the code. Fold it back in. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-06 23:05:42 +00:00
/*
* Even though the limit is exceeded at this point, reclaim
* may have been able to free some pages. Retry the charge
* before killing the task.
*
* Only for regular pages, though: huge pages are rather
* unlikely to succeed so close to the limit, and we fall back
* to regular pages anyway in case of failure.
*/
if (nr_reclaimed && nr_pages <= (1 << PAGE_ALLOC_COSTLY_ORDER))
mm: memcontrol: fold mem_cgroup_do_charge() These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 13): This function was split out because mem_cgroup_try_charge() got too big. But having essentially one sequence of operations arbitrarily split in half is not good for reworking the code. Fold it back in. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-06 23:05:42 +00:00
goto retry;
/*
* At task move, charge accounts can be doubly counted. So, it's
* better to wait until the end of task_move if something is going on.
*/
if (mem_cgroup_wait_acct_move(mem_over_limit))
goto retry;
if (nr_retries--)
goto retry;
if (gfp_mask & __GFP_RETRY_MAYFAIL)
memcg, oom: move out_of_memory back to the charge path Commit 3812c8c8f395 ("mm: memcg: do not trap chargers with full callstack on OOM") has changed the ENOMEM semantic of memcg charges. Rather than invoking the oom killer from the charging context it delays the oom killer to the page fault path (pagefault_out_of_memory). This in turn means that many users (e.g. slab or g-u-p) will get ENOMEM when the corresponding memcg hits the hard limit and the memcg is is OOM. This is behavior is inconsistent with !memcg case where the oom killer is invoked from the allocation context and the allocator keeps retrying until it succeeds. The difference in the behavior is user visible. mmap(MAP_POPULATE) might result in not fully populated ranges while the mmap return code doesn't tell that to the userspace. Random syscalls might fail with ENOMEM etc. The primary motivation of the different memcg oom semantic was the deadlock avoidance. Things have changed since then, though. We have an async oom teardown by the oom reaper now and so we do not have to rely on the victim to tear down its memory anymore. Therefore we can return to the original semantic as long as the memcg oom killer is not handed over to the users space. There is still one thing to be careful about here though. If the oom killer is not able to make any forward progress - e.g. because there is no eligible task to kill - then we have to bail out of the charge path to prevent from same class of deadlocks. We have basically two options here. Either we fail the charge with ENOMEM or force the charge and allow overcharge. The first option has been considered more harmful than useful because rare inconsistencies in the ENOMEM behavior is hard to test for and error prone. Basically the same reason why the page allocator doesn't fail allocations under such conditions. The later might allow runaways but those should be really unlikely unless somebody misconfigures the system. E.g. allowing to migrate tasks away from the memcg to a different unlimited memcg with move_charge_at_immigrate disabled. Link: http://lkml.kernel.org/r/20180628151101.25307-1-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Acked-by: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Shakeel Butt <shakeelb@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:11 +00:00
goto nomem;
if (gfp_mask & __GFP_NOFAIL)
memcg: ratify and consolidate over-charge handling try_charge() is the main charging logic of memcg. When it hits the limit but either can't fail the allocation due to __GFP_NOFAIL or the task is likely to free memory very soon, being OOM killed, has SIGKILL pending or exiting, it "bypasses" the charge to the root memcg and returns -EINTR. While this is one approach which can be taken for these situations, it has several issues. * It unnecessarily lies about the reality. The number itself doesn't go over the limit but the actual usage does. memcg is either forced to or actively chooses to go over the limit because that is the right behavior under the circumstances, which is completely fine, but, if at all avoidable, it shouldn't be misrepresenting what's happening by sneaking the charges into the root memcg. * Despite trying, we already do over-charge. kmemcg can't deal with switching over to the root memcg by the point try_charge() returns -EINTR, so it open-codes over-charing. * It complicates the callers. Each try_charge() user has to handle the weird -EINTR exception. memcg_charge_kmem() does the manual over-charging. mem_cgroup_do_precharge() performs unnecessary uncharging of root memcg, which BTW is inconsistent with what memcg_charge_kmem() does but not broken as [un]charging are noops on root memcg. mem_cgroup_try_charge() needs to switch the returned cgroup to the root one. The reality is that in memcg there are cases where we are forced and/or willing to go over the limit. Each such case needs to be scrutinized and justified but there definitely are situations where that is the right thing to do. We alredy do this but with a superficial and inconsistent disguise which leads to unnecessary complications. This patch updates try_charge() so that it over-charges and returns 0 when deemed necessary. -EINTR return is removed along with all special case handling in the callers. While at it, remove the local variable @ret, which was initialized to zero and never changed, along with done: label which just returned the always zero @ret. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:46:17 +00:00
goto force;
mm: memcontrol: fold mem_cgroup_do_charge() These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 13): This function was split out because mem_cgroup_try_charge() got too big. But having essentially one sequence of operations arbitrarily split in half is not good for reworking the code. Fold it back in. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-06 23:05:42 +00:00
if (fatal_signal_pending(current))
memcg: ratify and consolidate over-charge handling try_charge() is the main charging logic of memcg. When it hits the limit but either can't fail the allocation due to __GFP_NOFAIL or the task is likely to free memory very soon, being OOM killed, has SIGKILL pending or exiting, it "bypasses" the charge to the root memcg and returns -EINTR. While this is one approach which can be taken for these situations, it has several issues. * It unnecessarily lies about the reality. The number itself doesn't go over the limit but the actual usage does. memcg is either forced to or actively chooses to go over the limit because that is the right behavior under the circumstances, which is completely fine, but, if at all avoidable, it shouldn't be misrepresenting what's happening by sneaking the charges into the root memcg. * Despite trying, we already do over-charge. kmemcg can't deal with switching over to the root memcg by the point try_charge() returns -EINTR, so it open-codes over-charing. * It complicates the callers. Each try_charge() user has to handle the weird -EINTR exception. memcg_charge_kmem() does the manual over-charging. mem_cgroup_do_precharge() performs unnecessary uncharging of root memcg, which BTW is inconsistent with what memcg_charge_kmem() does but not broken as [un]charging are noops on root memcg. mem_cgroup_try_charge() needs to switch the returned cgroup to the root one. The reality is that in memcg there are cases where we are forced and/or willing to go over the limit. Each such case needs to be scrutinized and justified but there definitely are situations where that is the right thing to do. We alredy do this but with a superficial and inconsistent disguise which leads to unnecessary complications. This patch updates try_charge() so that it over-charges and returns 0 when deemed necessary. -EINTR return is removed along with all special case handling in the callers. While at it, remove the local variable @ret, which was initialized to zero and never changed, along with done: label which just returned the always zero @ret. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:46:17 +00:00
goto force;
mm: memcontrol: fold mem_cgroup_do_charge() These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 13): This function was split out because mem_cgroup_try_charge() got too big. But having essentially one sequence of operations arbitrarily split in half is not good for reworking the code. Fold it back in. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-06 23:05:42 +00:00
memcg, oom: move out_of_memory back to the charge path Commit 3812c8c8f395 ("mm: memcg: do not trap chargers with full callstack on OOM") has changed the ENOMEM semantic of memcg charges. Rather than invoking the oom killer from the charging context it delays the oom killer to the page fault path (pagefault_out_of_memory). This in turn means that many users (e.g. slab or g-u-p) will get ENOMEM when the corresponding memcg hits the hard limit and the memcg is is OOM. This is behavior is inconsistent with !memcg case where the oom killer is invoked from the allocation context and the allocator keeps retrying until it succeeds. The difference in the behavior is user visible. mmap(MAP_POPULATE) might result in not fully populated ranges while the mmap return code doesn't tell that to the userspace. Random syscalls might fail with ENOMEM etc. The primary motivation of the different memcg oom semantic was the deadlock avoidance. Things have changed since then, though. We have an async oom teardown by the oom reaper now and so we do not have to rely on the victim to tear down its memory anymore. Therefore we can return to the original semantic as long as the memcg oom killer is not handed over to the users space. There is still one thing to be careful about here though. If the oom killer is not able to make any forward progress - e.g. because there is no eligible task to kill - then we have to bail out of the charge path to prevent from same class of deadlocks. We have basically two options here. Either we fail the charge with ENOMEM or force the charge and allow overcharge. The first option has been considered more harmful than useful because rare inconsistencies in the ENOMEM behavior is hard to test for and error prone. Basically the same reason why the page allocator doesn't fail allocations under such conditions. The later might allow runaways but those should be really unlikely unless somebody misconfigures the system. E.g. allowing to migrate tasks away from the memcg to a different unlimited memcg with move_charge_at_immigrate disabled. Link: http://lkml.kernel.org/r/20180628151101.25307-1-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Acked-by: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Shakeel Butt <shakeelb@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:11 +00:00
/*
* keep retrying as long as the memcg oom killer is able to make
* a forward progress or bypass the charge if the oom killer
* couldn't make any progress.
*/
oom_status = mem_cgroup_oom(mem_over_limit, gfp_mask,
get_order(nr_pages * PAGE_SIZE));
memcg, oom: move out_of_memory back to the charge path Commit 3812c8c8f395 ("mm: memcg: do not trap chargers with full callstack on OOM") has changed the ENOMEM semantic of memcg charges. Rather than invoking the oom killer from the charging context it delays the oom killer to the page fault path (pagefault_out_of_memory). This in turn means that many users (e.g. slab or g-u-p) will get ENOMEM when the corresponding memcg hits the hard limit and the memcg is is OOM. This is behavior is inconsistent with !memcg case where the oom killer is invoked from the allocation context and the allocator keeps retrying until it succeeds. The difference in the behavior is user visible. mmap(MAP_POPULATE) might result in not fully populated ranges while the mmap return code doesn't tell that to the userspace. Random syscalls might fail with ENOMEM etc. The primary motivation of the different memcg oom semantic was the deadlock avoidance. Things have changed since then, though. We have an async oom teardown by the oom reaper now and so we do not have to rely on the victim to tear down its memory anymore. Therefore we can return to the original semantic as long as the memcg oom killer is not handed over to the users space. There is still one thing to be careful about here though. If the oom killer is not able to make any forward progress - e.g. because there is no eligible task to kill - then we have to bail out of the charge path to prevent from same class of deadlocks. We have basically two options here. Either we fail the charge with ENOMEM or force the charge and allow overcharge. The first option has been considered more harmful than useful because rare inconsistencies in the ENOMEM behavior is hard to test for and error prone. Basically the same reason why the page allocator doesn't fail allocations under such conditions. The later might allow runaways but those should be really unlikely unless somebody misconfigures the system. E.g. allowing to migrate tasks away from the memcg to a different unlimited memcg with move_charge_at_immigrate disabled. Link: http://lkml.kernel.org/r/20180628151101.25307-1-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Acked-by: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Shakeel Butt <shakeelb@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:11 +00:00
switch (oom_status) {
case OOM_SUCCESS:
nr_retries = MEM_CGROUP_RECLAIM_RETRIES;
goto retry;
case OOM_FAILED:
goto force;
default:
goto nomem;
}
memcg: introduce charge-commit-cancel style of functions There is a small race in do_swap_page(). When the page swapped-in is charged, the mapcount can be greater than 0. But, at the same time some process (shares it ) call unmap and make mapcount 1->0 and the page is uncharged. CPUA CPUB mapcount == 1. (1) charge if mapcount==0 zap_pte_range() (2) mapcount 1 => 0. (3) uncharge(). (success) (4) set page's rmap() mapcount 0=>1 Then, this swap page's account is leaked. For fixing this, I added a new interface. - charge account to res_counter by PAGE_SIZE and try to free pages if necessary. - commit register page_cgroup and add to LRU if necessary. - cancel uncharge PAGE_SIZE because of do_swap_page failure. CPUA (1) charge (always) (2) set page's rmap (mapcount > 0) (3) commit charge was necessary or not after set_pte(). This protocol uses PCG_USED bit on page_cgroup for avoiding over accounting. Usual mem_cgroup_charge_common() does charge -> commit at a time. And this patch also adds following function to clarify all charges. - mem_cgroup_newpage_charge() ....replacement for mem_cgroup_charge() called against newly allocated anon pages. - mem_cgroup_charge_migrate_fixup() called only from remove_migration_ptes(). we'll have to rewrite this later.(this patch just keeps old behavior) This function will be removed by additional patch to make migration clearer. Good for clarifying "what we do" Then, we have 4 following charge points. - newpage - swap-in - add-to-cache. - migration. [akpm@linux-foundation.org: add missing inline directives to stubs] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:07:48 +00:00
nomem:
if (!(gfp_mask & __GFP_NOFAIL))
return -ENOMEM;
memcg: ratify and consolidate over-charge handling try_charge() is the main charging logic of memcg. When it hits the limit but either can't fail the allocation due to __GFP_NOFAIL or the task is likely to free memory very soon, being OOM killed, has SIGKILL pending or exiting, it "bypasses" the charge to the root memcg and returns -EINTR. While this is one approach which can be taken for these situations, it has several issues. * It unnecessarily lies about the reality. The number itself doesn't go over the limit but the actual usage does. memcg is either forced to or actively chooses to go over the limit because that is the right behavior under the circumstances, which is completely fine, but, if at all avoidable, it shouldn't be misrepresenting what's happening by sneaking the charges into the root memcg. * Despite trying, we already do over-charge. kmemcg can't deal with switching over to the root memcg by the point try_charge() returns -EINTR, so it open-codes over-charing. * It complicates the callers. Each try_charge() user has to handle the weird -EINTR exception. memcg_charge_kmem() does the manual over-charging. mem_cgroup_do_precharge() performs unnecessary uncharging of root memcg, which BTW is inconsistent with what memcg_charge_kmem() does but not broken as [un]charging are noops on root memcg. mem_cgroup_try_charge() needs to switch the returned cgroup to the root one. The reality is that in memcg there are cases where we are forced and/or willing to go over the limit. Each such case needs to be scrutinized and justified but there definitely are situations where that is the right thing to do. We alredy do this but with a superficial and inconsistent disguise which leads to unnecessary complications. This patch updates try_charge() so that it over-charges and returns 0 when deemed necessary. -EINTR return is removed along with all special case handling in the callers. While at it, remove the local variable @ret, which was initialized to zero and never changed, along with done: label which just returned the always zero @ret. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:46:17 +00:00
force:
/*
* The allocation either can't fail or will lead to more memory
* being freed very soon. Allow memory usage go over the limit
* temporarily by force charging it.
*/
page_counter_charge(&memcg->memory, nr_pages);
if (do_memsw_account())
memcg: ratify and consolidate over-charge handling try_charge() is the main charging logic of memcg. When it hits the limit but either can't fail the allocation due to __GFP_NOFAIL or the task is likely to free memory very soon, being OOM killed, has SIGKILL pending or exiting, it "bypasses" the charge to the root memcg and returns -EINTR. While this is one approach which can be taken for these situations, it has several issues. * It unnecessarily lies about the reality. The number itself doesn't go over the limit but the actual usage does. memcg is either forced to or actively chooses to go over the limit because that is the right behavior under the circumstances, which is completely fine, but, if at all avoidable, it shouldn't be misrepresenting what's happening by sneaking the charges into the root memcg. * Despite trying, we already do over-charge. kmemcg can't deal with switching over to the root memcg by the point try_charge() returns -EINTR, so it open-codes over-charing. * It complicates the callers. Each try_charge() user has to handle the weird -EINTR exception. memcg_charge_kmem() does the manual over-charging. mem_cgroup_do_precharge() performs unnecessary uncharging of root memcg, which BTW is inconsistent with what memcg_charge_kmem() does but not broken as [un]charging are noops on root memcg. mem_cgroup_try_charge() needs to switch the returned cgroup to the root one. The reality is that in memcg there are cases where we are forced and/or willing to go over the limit. Each such case needs to be scrutinized and justified but there definitely are situations where that is the right thing to do. We alredy do this but with a superficial and inconsistent disguise which leads to unnecessary complications. This patch updates try_charge() so that it over-charges and returns 0 when deemed necessary. -EINTR return is removed along with all special case handling in the callers. While at it, remove the local variable @ret, which was initialized to zero and never changed, along with done: label which just returned the always zero @ret. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:46:17 +00:00
page_counter_charge(&memcg->memsw, nr_pages);
css_get_many(&memcg->css, nr_pages);
return 0;
mm: memcontrol: fold mem_cgroup_do_charge() These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 13): This function was split out because mem_cgroup_try_charge() got too big. But having essentially one sequence of operations arbitrarily split in half is not good for reworking the code. Fold it back in. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-06 23:05:42 +00:00
done_restock:
css_get_many(&memcg->css, batch);
mm: memcontrol: fold mem_cgroup_do_charge() These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 13): This function was split out because mem_cgroup_try_charge() got too big. But having essentially one sequence of operations arbitrarily split in half is not good for reworking the code. Fold it back in. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-06 23:05:42 +00:00
if (batch > nr_pages)
refill_stock(memcg, batch - nr_pages);
memcg: punt high overage reclaim to return-to-userland path Currently, try_charge() tries to reclaim memory synchronously when the high limit is breached; however, if the allocation doesn't have __GFP_WAIT, synchronous reclaim is skipped. If a process performs only speculative allocations, it can blow way past the high limit. This is actually easily reproducible by simply doing "find /". slab/slub allocator tries speculative allocations first, so as long as there's memory which can be consumed without blocking, it can keep allocating memory regardless of the high limit. This patch makes try_charge() always punt the over-high reclaim to the return-to-userland path. If try_charge() detects that high limit is breached, it adds the overage to current->memcg_nr_pages_over_high and schedules execution of mem_cgroup_handle_over_high() which performs synchronous reclaim from the return-to-userland path. As long as kernel doesn't have a run-away allocation spree, this should provide enough protection while making kmemcg behave more consistently. It also has the following benefits. - All over-high reclaims can use GFP_KERNEL regardless of the specific gfp mask in use, e.g. GFP_NOFS, when the limit was breached. - It copes with prio inversion. Previously, a low-prio task with small memory.high might perform over-high reclaim with a bunch of locks held. If a higher prio task needed any of these locks, it would have to wait until the low prio task finished reclaim and released the locks. By handing over-high reclaim to the task exit path this issue can be avoided. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Michal Hocko <mhocko@kernel.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:46:11 +00:00
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
/*
memcg: punt high overage reclaim to return-to-userland path Currently, try_charge() tries to reclaim memory synchronously when the high limit is breached; however, if the allocation doesn't have __GFP_WAIT, synchronous reclaim is skipped. If a process performs only speculative allocations, it can blow way past the high limit. This is actually easily reproducible by simply doing "find /". slab/slub allocator tries speculative allocations first, so as long as there's memory which can be consumed without blocking, it can keep allocating memory regardless of the high limit. This patch makes try_charge() always punt the over-high reclaim to the return-to-userland path. If try_charge() detects that high limit is breached, it adds the overage to current->memcg_nr_pages_over_high and schedules execution of mem_cgroup_handle_over_high() which performs synchronous reclaim from the return-to-userland path. As long as kernel doesn't have a run-away allocation spree, this should provide enough protection while making kmemcg behave more consistently. It also has the following benefits. - All over-high reclaims can use GFP_KERNEL regardless of the specific gfp mask in use, e.g. GFP_NOFS, when the limit was breached. - It copes with prio inversion. Previously, a low-prio task with small memory.high might perform over-high reclaim with a bunch of locks held. If a higher prio task needed any of these locks, it would have to wait until the low prio task finished reclaim and released the locks. By handing over-high reclaim to the task exit path this issue can be avoided. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Michal Hocko <mhocko@kernel.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:46:11 +00:00
* If the hierarchy is above the normal consumption range, schedule
* reclaim on returning to userland. We can perform reclaim here
* if __GFP_RECLAIM but let's always punt for simplicity and so that
memcg: punt high overage reclaim to return-to-userland path Currently, try_charge() tries to reclaim memory synchronously when the high limit is breached; however, if the allocation doesn't have __GFP_WAIT, synchronous reclaim is skipped. If a process performs only speculative allocations, it can blow way past the high limit. This is actually easily reproducible by simply doing "find /". slab/slub allocator tries speculative allocations first, so as long as there's memory which can be consumed without blocking, it can keep allocating memory regardless of the high limit. This patch makes try_charge() always punt the over-high reclaim to the return-to-userland path. If try_charge() detects that high limit is breached, it adds the overage to current->memcg_nr_pages_over_high and schedules execution of mem_cgroup_handle_over_high() which performs synchronous reclaim from the return-to-userland path. As long as kernel doesn't have a run-away allocation spree, this should provide enough protection while making kmemcg behave more consistently. It also has the following benefits. - All over-high reclaims can use GFP_KERNEL regardless of the specific gfp mask in use, e.g. GFP_NOFS, when the limit was breached. - It copes with prio inversion. Previously, a low-prio task with small memory.high might perform over-high reclaim with a bunch of locks held. If a higher prio task needed any of these locks, it would have to wait until the low prio task finished reclaim and released the locks. By handing over-high reclaim to the task exit path this issue can be avoided. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Michal Hocko <mhocko@kernel.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:46:11 +00:00
* GFP_KERNEL can consistently be used during reclaim. @memcg is
* not recorded as it most likely matches current's and won't
* change in the meantime. As high limit is checked again before
* reclaim, the cost of mismatch is negligible.
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
*/
do {
mm/memcg: automatically penalize tasks with high swap use Add a memory.swap.high knob, which can be used to protect the system from SWAP exhaustion. The mechanism used for penalizing is similar to memory.high penalty (sleep on return to user space). That is not to say that the knob itself is equivalent to memory.high. The objective is more to protect the system from potentially buggy tasks consuming a lot of swap and impacting other tasks, or even bringing the whole system to stand still with complete SWAP exhaustion. Hopefully without the need to find per-task hard limits. Slowing misbehaving tasks down gradually allows user space oom killers or other protection mechanisms to react. oomd and earlyoom already do killing based on swap exhaustion, and memory.swap.high protection will help implement such userspace oom policies more reliably. We can use one counter for number of pages allocated under pressure to save struct task space and avoid two separate hierarchy walks on the hot path. The exact overage is calculated on return to user space, anyway. Take the new high limit into account when determining if swap is "full". Borrowing the explanation from Johannes: The idea behind "swap full" is that as long as the workload has plenty of swap space available and it's not changing its memory contents, it makes sense to generously hold on to copies of data in the swap device, even after the swapin. A later reclaim cycle can drop the page without any IO. Trading disk space for IO. But the only two ways to reclaim a swap slot is when they're faulted in and the references go away, or by scanning the virtual address space like swapoff does - which is very expensive (one could argue it's too expensive even for swapoff, it's often more practical to just reboot). So at some point in the fill level, we have to start freeing up swap slots on fault/swapin. Otherwise we could eventually run out of swap slots while they're filled with copies of data that is also in RAM. We don't want to OOM a workload because its available swap space is filled with redundant cache. Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Chris Down <chris@chrisdown.name> Cc: Shakeel Butt <shakeelb@google.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Hugh Dickins <hughd@google.com> Link: http://lkml.kernel.org/r/20200527195846.102707-5-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:52 +00:00
bool mem_high, swap_high;
mem_high = page_counter_read(&memcg->memory) >
READ_ONCE(memcg->memory.high);
swap_high = page_counter_read(&memcg->swap) >
READ_ONCE(memcg->swap.high);
/* Don't bother a random interrupted task */
if (in_interrupt()) {
if (mem_high) {
schedule_work(&memcg->high_work);
break;
}
mm/memcg: automatically penalize tasks with high swap use Add a memory.swap.high knob, which can be used to protect the system from SWAP exhaustion. The mechanism used for penalizing is similar to memory.high penalty (sleep on return to user space). That is not to say that the knob itself is equivalent to memory.high. The objective is more to protect the system from potentially buggy tasks consuming a lot of swap and impacting other tasks, or even bringing the whole system to stand still with complete SWAP exhaustion. Hopefully without the need to find per-task hard limits. Slowing misbehaving tasks down gradually allows user space oom killers or other protection mechanisms to react. oomd and earlyoom already do killing based on swap exhaustion, and memory.swap.high protection will help implement such userspace oom policies more reliably. We can use one counter for number of pages allocated under pressure to save struct task space and avoid two separate hierarchy walks on the hot path. The exact overage is calculated on return to user space, anyway. Take the new high limit into account when determining if swap is "full". Borrowing the explanation from Johannes: The idea behind "swap full" is that as long as the workload has plenty of swap space available and it's not changing its memory contents, it makes sense to generously hold on to copies of data in the swap device, even after the swapin. A later reclaim cycle can drop the page without any IO. Trading disk space for IO. But the only two ways to reclaim a swap slot is when they're faulted in and the references go away, or by scanning the virtual address space like swapoff does - which is very expensive (one could argue it's too expensive even for swapoff, it's often more practical to just reboot). So at some point in the fill level, we have to start freeing up swap slots on fault/swapin. Otherwise we could eventually run out of swap slots while they're filled with copies of data that is also in RAM. We don't want to OOM a workload because its available swap space is filled with redundant cache. Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Chris Down <chris@chrisdown.name> Cc: Shakeel Butt <shakeelb@google.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Hugh Dickins <hughd@google.com> Link: http://lkml.kernel.org/r/20200527195846.102707-5-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:52 +00:00
continue;
}
if (mem_high || swap_high) {
/*
* The allocating tasks in this cgroup will need to do
* reclaim or be throttled to prevent further growth
* of the memory or swap footprints.
*
* Target some best-effort fairness between the tasks,
* and distribute reclaim work and delay penalties
* based on how much each task is actually allocating.
*/
current->memcg_nr_pages_over_high += batch;
memcg: punt high overage reclaim to return-to-userland path Currently, try_charge() tries to reclaim memory synchronously when the high limit is breached; however, if the allocation doesn't have __GFP_WAIT, synchronous reclaim is skipped. If a process performs only speculative allocations, it can blow way past the high limit. This is actually easily reproducible by simply doing "find /". slab/slub allocator tries speculative allocations first, so as long as there's memory which can be consumed without blocking, it can keep allocating memory regardless of the high limit. This patch makes try_charge() always punt the over-high reclaim to the return-to-userland path. If try_charge() detects that high limit is breached, it adds the overage to current->memcg_nr_pages_over_high and schedules execution of mem_cgroup_handle_over_high() which performs synchronous reclaim from the return-to-userland path. As long as kernel doesn't have a run-away allocation spree, this should provide enough protection while making kmemcg behave more consistently. It also has the following benefits. - All over-high reclaims can use GFP_KERNEL regardless of the specific gfp mask in use, e.g. GFP_NOFS, when the limit was breached. - It copes with prio inversion. Previously, a low-prio task with small memory.high might perform over-high reclaim with a bunch of locks held. If a higher prio task needed any of these locks, it would have to wait until the low prio task finished reclaim and released the locks. By handing over-high reclaim to the task exit path this issue can be avoided. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Michal Hocko <mhocko@kernel.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:46:11 +00:00
set_notify_resume(current);
break;
}
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
} while ((memcg = parent_mem_cgroup(memcg)));
memcg: ratify and consolidate over-charge handling try_charge() is the main charging logic of memcg. When it hits the limit but either can't fail the allocation due to __GFP_NOFAIL or the task is likely to free memory very soon, being OOM killed, has SIGKILL pending or exiting, it "bypasses" the charge to the root memcg and returns -EINTR. While this is one approach which can be taken for these situations, it has several issues. * It unnecessarily lies about the reality. The number itself doesn't go over the limit but the actual usage does. memcg is either forced to or actively chooses to go over the limit because that is the right behavior under the circumstances, which is completely fine, but, if at all avoidable, it shouldn't be misrepresenting what's happening by sneaking the charges into the root memcg. * Despite trying, we already do over-charge. kmemcg can't deal with switching over to the root memcg by the point try_charge() returns -EINTR, so it open-codes over-charing. * It complicates the callers. Each try_charge() user has to handle the weird -EINTR exception. memcg_charge_kmem() does the manual over-charging. mem_cgroup_do_precharge() performs unnecessary uncharging of root memcg, which BTW is inconsistent with what memcg_charge_kmem() does but not broken as [un]charging are noops on root memcg. mem_cgroup_try_charge() needs to switch the returned cgroup to the root one. The reality is that in memcg there are cases where we are forced and/or willing to go over the limit. Each such case needs to be scrutinized and justified but there definitely are situations where that is the right thing to do. We alredy do this but with a superficial and inconsistent disguise which leads to unnecessary complications. This patch updates try_charge() so that it over-charges and returns 0 when deemed necessary. -EINTR return is removed along with all special case handling in the callers. While at it, remove the local variable @ret, which was initialized to zero and never changed, along with done: label which just returned the always zero @ret. Signed-off-by: Tejun Heo <tj@kernel.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:46:17 +00:00
return 0;
memcg: introduce charge-commit-cancel style of functions There is a small race in do_swap_page(). When the page swapped-in is charged, the mapcount can be greater than 0. But, at the same time some process (shares it ) call unmap and make mapcount 1->0 and the page is uncharged. CPUA CPUB mapcount == 1. (1) charge if mapcount==0 zap_pte_range() (2) mapcount 1 => 0. (3) uncharge(). (success) (4) set page's rmap() mapcount 0=>1 Then, this swap page's account is leaked. For fixing this, I added a new interface. - charge account to res_counter by PAGE_SIZE and try to free pages if necessary. - commit register page_cgroup and add to LRU if necessary. - cancel uncharge PAGE_SIZE because of do_swap_page failure. CPUA (1) charge (always) (2) set page's rmap (mapcount > 0) (3) commit charge was necessary or not after set_pte(). This protocol uses PCG_USED bit on page_cgroup for avoiding over accounting. Usual mem_cgroup_charge_common() does charge -> commit at a time. And this patch also adds following function to clarify all charges. - mem_cgroup_newpage_charge() ....replacement for mem_cgroup_charge() called against newly allocated anon pages. - mem_cgroup_charge_migrate_fixup() called only from remove_migration_ptes(). we'll have to rewrite this later.(this patch just keeps old behavior) This function will be removed by additional patch to make migration clearer. Good for clarifying "what we do" Then, we have 4 following charge points. - newpage - swap-in - add-to-cache. - migration. [akpm@linux-foundation.org: add missing inline directives to stubs] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:07:48 +00:00
}
Memory controller: memory accounting Add the accounting hooks. The accounting is carried out for RSS and Page Cache (unmapped) pages. There is now a common limit and accounting for both. The RSS accounting is accounted at page_add_*_rmap() and page_remove_rmap() time. Page cache is accounted at add_to_page_cache(), __delete_from_page_cache(). Swap cache is also accounted for. Each page's page_cgroup is protected with the last bit of the page_cgroup pointer, this makes handling of race conditions involving simultaneous mappings of a page easier. A reference count is kept in the page_cgroup to deal with cases where a page might be unmapped from the RSS of all tasks, but still lives in the page cache. Credits go to Vaidyanathan Srinivasan for helping with reference counting work of the page cgroup. Almost all of the page cache accounting code has help from Vaidyanathan Srinivasan. [hugh@veritas.com: fix swapoff breakage] [akpm@linux-foundation.org: fix locking] Signed-off-by: Vaidyanathan Srinivasan <svaidy@linux.vnet.ibm.com> Signed-off-by: Balbir Singh <balbir@linux.vnet.ibm.com> Cc: Pavel Emelianov <xemul@openvz.org> Cc: Paul Menage <menage@google.com> Cc: Peter Zijlstra <a.p.zijlstra@chello.nl> Cc: "Eric W. Biederman" <ebiederm@xmission.com> Cc: Nick Piggin <nickpiggin@yahoo.com.au> Cc: Kirill Korotaev <dev@sw.ru> Cc: Herbert Poetzl <herbert@13thfloor.at> Cc: David Rientjes <rientjes@google.com> Cc: <Valdis.Kletnieks@vt.edu> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2008-02-07 08:13:53 +00:00
#if defined(CONFIG_MEMCG_KMEM) || defined(CONFIG_MMU)
mm: memcontrol: rewrite charge API These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 4): The memcg charge API charges pages before they are rmapped - i.e. have an actual "type" - and so every callsite needs its own set of charge and uncharge functions to know what type is being operated on. Worse, uncharge has to happen from a context that is still type-specific, rather than at the end of the page's lifetime with exclusive access, and so requires a lot of synchronization. Rewrite the charge API to provide a generic set of try_charge(), commit_charge() and cancel_charge() transaction operations, much like what's currently done for swap-in: mem_cgroup_try_charge() attempts to reserve a charge, reclaiming pages from the memcg if necessary. mem_cgroup_commit_charge() commits the page to the charge once it has a valid page->mapping and PageAnon() reliably tells the type. mem_cgroup_cancel_charge() aborts the transaction. This reduces the charge API and enables subsequent patches to drastically simplify uncharging. As pages need to be committed after rmap is established but before they are added to the LRU, page_add_new_anon_rmap() must stop doing LRU additions again. Revive lru_cache_add_active_or_unevictable(). [hughd@google.com: fix shmem_unuse] [hughd@google.com: Add comments on the private use of -EAGAIN] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Hugh Dickins <hughd@google.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:20 +00:00
static void cancel_charge(struct mem_cgroup *memcg, unsigned int nr_pages)
{
if (mem_cgroup_is_root(memcg))
return;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
page_counter_uncharge(&memcg->memory, nr_pages);
if (do_memsw_account())
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
page_counter_uncharge(&memcg->memsw, nr_pages);
css_put_many(&memcg->css, nr_pages);
}
#endif
static void commit_charge(struct page *page, struct mem_cgroup *memcg)
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
{
mm: embed the memcg pointer directly into struct page Memory cgroups used to have 5 per-page pointers. To allow users to disable that amount of overhead during runtime, those pointers were allocated in a separate array, with a translation layer between them and struct page. There is now only one page pointer remaining: the memcg pointer, that indicates which cgroup the page is associated with when charged. The complexity of runtime allocation and the runtime translation overhead is no longer justified to save that *potential* 0.19% of memory. With CONFIG_SLUB, page->mem_cgroup actually sits in the doubleword padding after the page->private member and doesn't even increase struct page, and then this patch actually saves space. Remaining users that care can still compile their kernels without CONFIG_MEMCG. text data bss dec hex filename 8828345 1725264 983040 11536649 b00909 vmlinux.old 8827425 1725264 966656 11519345 afc571 vmlinux.new [mhocko@suse.cz: update Documentation/cgroups/memory.txt] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: David S. Miller <davem@davemloft.net> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Acked-by: Konstantin Khlebnikov <koct9i@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:44:52 +00:00
VM_BUG_ON_PAGE(page->mem_cgroup, page);
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
/*
* Any of the following ensures page->mem_cgroup stability:
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
*
* - the page lock
* - LRU isolation
* - lock_page_memcg()
* - exclusive reference
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
*/
mm: embed the memcg pointer directly into struct page Memory cgroups used to have 5 per-page pointers. To allow users to disable that amount of overhead during runtime, those pointers were allocated in a separate array, with a translation layer between them and struct page. There is now only one page pointer remaining: the memcg pointer, that indicates which cgroup the page is associated with when charged. The complexity of runtime allocation and the runtime translation overhead is no longer justified to save that *potential* 0.19% of memory. With CONFIG_SLUB, page->mem_cgroup actually sits in the doubleword padding after the page->private member and doesn't even increase struct page, and then this patch actually saves space. Remaining users that care can still compile their kernels without CONFIG_MEMCG. text data bss dec hex filename 8828345 1725264 983040 11536649 b00909 vmlinux.old 8827425 1725264 966656 11519345 afc571 vmlinux.new [mhocko@suse.cz: update Documentation/cgroups/memory.txt] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: David S. Miller <davem@davemloft.net> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Acked-by: Konstantin Khlebnikov <koct9i@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:44:52 +00:00
page->mem_cgroup = memcg;
memcg: introduce charge-commit-cancel style of functions There is a small race in do_swap_page(). When the page swapped-in is charged, the mapcount can be greater than 0. But, at the same time some process (shares it ) call unmap and make mapcount 1->0 and the page is uncharged. CPUA CPUB mapcount == 1. (1) charge if mapcount==0 zap_pte_range() (2) mapcount 1 => 0. (3) uncharge(). (success) (4) set page's rmap() mapcount 0=>1 Then, this swap page's account is leaked. For fixing this, I added a new interface. - charge account to res_counter by PAGE_SIZE and try to free pages if necessary. - commit register page_cgroup and add to LRU if necessary. - cancel uncharge PAGE_SIZE because of do_swap_page failure. CPUA (1) charge (always) (2) set page's rmap (mapcount > 0) (3) commit charge was necessary or not after set_pte(). This protocol uses PCG_USED bit on page_cgroup for avoiding over accounting. Usual mem_cgroup_charge_common() does charge -> commit at a time. And this patch also adds following function to clarify all charges. - mem_cgroup_newpage_charge() ....replacement for mem_cgroup_charge() called against newly allocated anon pages. - mem_cgroup_charge_migrate_fixup() called only from remove_migration_ptes(). we'll have to rewrite this later.(this patch just keeps old behavior) This function will be removed by additional patch to make migration clearer. Good for clarifying "what we do" Then, we have 4 following charge points. - newpage - swap-in - add-to-cache. - migration. [akpm@linux-foundation.org: add missing inline directives to stubs] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:07:48 +00:00
}
mm: introduce CONFIG_MEMCG_KMEM as combination of CONFIG_MEMCG && !CONFIG_SLOB Introduce new config option, which is used to replace repeating CONFIG_MEMCG && !CONFIG_SLOB pattern. Next patches add a little more memcg+kmem related code, so let's keep the defines more clearly. Link: http://lkml.kernel.org/r/153063053670.1818.15013136946600481138.stgit@localhost.localdomain Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Tested-by: Shakeel Butt <shakeelb@google.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Guenter Roeck <linux@roeck-us.net> Cc: "Huang, Ying" <ying.huang@intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Josef Bacik <jbacik@fb.com> Cc: Li RongQing <lirongqing@baidu.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Matthias Kaehlcke <mka@chromium.org> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Michal Hocko <mhocko@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Philippe Ombredanne <pombredanne@nexb.com> Cc: Roman Gushchin <guro@fb.com> Cc: Sahitya Tummala <stummala@codeaurora.org> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Waiman Long <longman@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:25 +00:00
#ifdef CONFIG_MEMCG_KMEM
mm: fork: fix kernel_stack memcg stats for various stack implementations Depending on CONFIG_VMAP_STACK and the THREAD_SIZE / PAGE_SIZE ratio the space for task stacks can be allocated using __vmalloc_node_range(), alloc_pages_node() and kmem_cache_alloc_node(). In the first and the second cases page->mem_cgroup pointer is set, but in the third it's not: memcg membership of a slab page should be determined using the memcg_from_slab_page() function, which looks at page->slab_cache->memcg_params.memcg . In this case, using mod_memcg_page_state() (as in account_kernel_stack()) is incorrect: page->mem_cgroup pointer is NULL even for pages charged to a non-root memory cgroup. It can lead to kernel_stack per-memcg counters permanently showing 0 on some architectures (depending on the configuration). In order to fix it, let's introduce a mod_memcg_obj_state() helper, which takes a pointer to a kernel object as a first argument, uses mem_cgroup_from_obj() to get a RCU-protected memcg pointer and calls mod_memcg_state(). It allows to handle all possible configurations (CONFIG_VMAP_STACK and various THREAD_SIZE/PAGE_SIZE values) without spilling any memcg/kmem specifics into fork.c . Note: This is a special version of the patch created for stable backports. It contains code from the following two patches: - mm: memcg/slab: introduce mem_cgroup_from_obj() - mm: fork: fix kernel_stack memcg stats for various stack implementations [guro@fb.com: introduce mem_cgroup_from_obj()] Link: http://lkml.kernel.org/r/20200324004221.GA36662@carbon.dhcp.thefacebook.com Fixes: 4d96ba353075 ("mm: memcg/slab: stop setting page->mem_cgroup pointer for slab pages") Signed-off-by: Roman Gushchin <guro@fb.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Bharata B Rao <bharata@linux.ibm.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: <stable@vger.kernel.org> Link: http://lkml.kernel.org/r/20200303233550.251375-1-guro@fb.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-03-29 02:17:25 +00:00
/*
* Returns a pointer to the memory cgroup to which the kernel object is charged.
*
* The caller must ensure the memcg lifetime, e.g. by taking rcu_read_lock(),
* cgroup_mutex, etc.
*/
struct mem_cgroup *mem_cgroup_from_obj(void *p)
{
struct page *page;
if (mem_cgroup_disabled())
return NULL;
page = virt_to_head_page(p);
/*
* Slab pages don't have page->mem_cgroup set because corresponding
* kmem caches can be reparented during the lifetime. That's why
* memcg_from_slab_page() should be used instead.
*/
if (PageSlab(page))
return memcg_from_slab_page(page);
/* All other pages use page->mem_cgroup */
return page->mem_cgroup;
}
static int memcg_alloc_cache_id(void)
memcg: allocate memory for memcg caches whenever a new memcg appears Every cache that is considered a root cache (basically the "original" caches, tied to the root memcg/no-memcg) will have an array that should be large enough to store a cache pointer per each memcg in the system. Theoreticaly, this is as high as 1 << sizeof(css_id), which is currently in the 64k pointers range. Most of the time, we won't be using that much. What goes in this patch, is a simple scheme to dynamically allocate such an array, in order to minimize memory usage for memcg caches. Because we would also like to avoid allocations all the time, at least for now, the array will only grow. It will tend to be big enough to hold the maximum number of kmem-limited memcgs ever achieved. We'll allocate it to be a minimum of 64 kmem-limited memcgs. When we have more than that, we'll start doubling the size of this array every time the limit is reached. Because we are only considering kmem limited memcgs, a natural point for this to happen is when we write to the limit. At that point, we already have set_limit_mutex held, so that will become our natural synchronization mechanism. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:38 +00:00
{
int id, size;
int err;
id = ida_simple_get(&memcg_cache_ida,
0, MEMCG_CACHES_MAX_SIZE, GFP_KERNEL);
if (id < 0)
return id;
memcg: allocate memory for memcg caches whenever a new memcg appears Every cache that is considered a root cache (basically the "original" caches, tied to the root memcg/no-memcg) will have an array that should be large enough to store a cache pointer per each memcg in the system. Theoreticaly, this is as high as 1 << sizeof(css_id), which is currently in the 64k pointers range. Most of the time, we won't be using that much. What goes in this patch, is a simple scheme to dynamically allocate such an array, in order to minimize memory usage for memcg caches. Because we would also like to avoid allocations all the time, at least for now, the array will only grow. It will tend to be big enough to hold the maximum number of kmem-limited memcgs ever achieved. We'll allocate it to be a minimum of 64 kmem-limited memcgs. When we have more than that, we'll start doubling the size of this array every time the limit is reached. Because we are only considering kmem limited memcgs, a natural point for this to happen is when we write to the limit. At that point, we already have set_limit_mutex held, so that will become our natural synchronization mechanism. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:38 +00:00
if (id < memcg_nr_cache_ids)
return id;
/*
* There's no space for the new id in memcg_caches arrays,
* so we have to grow them.
*/
memcg: add rwsem to synchronize against memcg_caches arrays relocation We need a stable value of memcg_nr_cache_ids in kmem_cache_create() (memcg_alloc_cache_params() wants it for root caches), where we only hold the slab_mutex and no memcg-related locks. As a result, we have to update memcg_nr_cache_ids under the slab_mutex, which we can only take on the slab's side (see memcg_update_array_size). This looks awkward and will become even worse when per-memcg list_lru is introduced, which also wants stable access to memcg_nr_cache_ids. To get rid of this dependency between the memcg_nr_cache_ids and the slab_mutex, this patch introduces a special rwsem. The rwsem is held for writing during memcg_caches arrays relocation and memcg_nr_cache_ids updates. Therefore one can take it for reading to get a stable access to memcg_caches arrays and/or memcg_nr_cache_ids. Currently the semaphore is taken for reading only from kmem_cache_create, right before taking the slab_mutex, so right now there's no much point in using rwsem instead of mutex. However, once list_lru is made per-memcg it will allow list_lru initializations to proceed concurrently. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Dave Chinner <david@fromorbit.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Greg Thelen <gthelen@google.com> Cc: Glauber Costa <glommer@gmail.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-12 22:59:01 +00:00
down_write(&memcg_cache_ids_sem);
size = 2 * (id + 1);
memcg: allocate memory for memcg caches whenever a new memcg appears Every cache that is considered a root cache (basically the "original" caches, tied to the root memcg/no-memcg) will have an array that should be large enough to store a cache pointer per each memcg in the system. Theoreticaly, this is as high as 1 << sizeof(css_id), which is currently in the 64k pointers range. Most of the time, we won't be using that much. What goes in this patch, is a simple scheme to dynamically allocate such an array, in order to minimize memory usage for memcg caches. Because we would also like to avoid allocations all the time, at least for now, the array will only grow. It will tend to be big enough to hold the maximum number of kmem-limited memcgs ever achieved. We'll allocate it to be a minimum of 64 kmem-limited memcgs. When we have more than that, we'll start doubling the size of this array every time the limit is reached. Because we are only considering kmem limited memcgs, a natural point for this to happen is when we write to the limit. At that point, we already have set_limit_mutex held, so that will become our natural synchronization mechanism. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:38 +00:00
if (size < MEMCG_CACHES_MIN_SIZE)
size = MEMCG_CACHES_MIN_SIZE;
else if (size > MEMCG_CACHES_MAX_SIZE)
size = MEMCG_CACHES_MAX_SIZE;
err = memcg_update_all_caches(size);
list_lru: introduce per-memcg lists There are several FS shrinkers, including super_block::s_shrink, that keep reclaimable objects in the list_lru structure. Hence to turn them to memcg-aware shrinkers, it is enough to make list_lru per-memcg. This patch does the trick. It adds an array of lru lists to the list_lru_node structure (per-node part of the list_lru), one for each kmem-active memcg, and dispatches every item addition or removal to the list corresponding to the memcg which the item is accounted to. So now the list_lru structure is not just per node, but per node and per memcg. Not all list_lrus need this feature, so this patch also adds a new method, list_lru_init_memcg, which initializes a list_lru as memcg aware. Otherwise (i.e. if initialized with old list_lru_init), the list_lru won't have per memcg lists. Just like per memcg caches arrays, the arrays of per-memcg lists are indexed by memcg_cache_id, so we must grow them whenever memcg_nr_cache_ids is increased. So we introduce a callback, memcg_update_all_list_lrus, invoked by memcg_alloc_cache_id if the id space is full. The locking is implemented in a manner similar to lruvecs, i.e. we have one lock per node that protects all lists (both global and per cgroup) on the node. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Dave Chinner <david@fromorbit.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Greg Thelen <gthelen@google.com> Cc: Glauber Costa <glommer@gmail.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-12 22:59:10 +00:00
if (!err)
err = memcg_update_all_list_lrus(size);
memcg: add rwsem to synchronize against memcg_caches arrays relocation We need a stable value of memcg_nr_cache_ids in kmem_cache_create() (memcg_alloc_cache_params() wants it for root caches), where we only hold the slab_mutex and no memcg-related locks. As a result, we have to update memcg_nr_cache_ids under the slab_mutex, which we can only take on the slab's side (see memcg_update_array_size). This looks awkward and will become even worse when per-memcg list_lru is introduced, which also wants stable access to memcg_nr_cache_ids. To get rid of this dependency between the memcg_nr_cache_ids and the slab_mutex, this patch introduces a special rwsem. The rwsem is held for writing during memcg_caches arrays relocation and memcg_nr_cache_ids updates. Therefore one can take it for reading to get a stable access to memcg_caches arrays and/or memcg_nr_cache_ids. Currently the semaphore is taken for reading only from kmem_cache_create, right before taking the slab_mutex, so right now there's no much point in using rwsem instead of mutex. However, once list_lru is made per-memcg it will allow list_lru initializations to proceed concurrently. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Dave Chinner <david@fromorbit.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Greg Thelen <gthelen@google.com> Cc: Glauber Costa <glommer@gmail.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-12 22:59:01 +00:00
if (!err)
memcg_nr_cache_ids = size;
up_write(&memcg_cache_ids_sem);
if (err) {
ida_simple_remove(&memcg_cache_ida, id);
return err;
}
return id;
}
static void memcg_free_cache_id(int id)
{
ida_simple_remove(&memcg_cache_ida, id);
memcg: allocate memory for memcg caches whenever a new memcg appears Every cache that is considered a root cache (basically the "original" caches, tied to the root memcg/no-memcg) will have an array that should be large enough to store a cache pointer per each memcg in the system. Theoreticaly, this is as high as 1 << sizeof(css_id), which is currently in the 64k pointers range. Most of the time, we won't be using that much. What goes in this patch, is a simple scheme to dynamically allocate such an array, in order to minimize memory usage for memcg caches. Because we would also like to avoid allocations all the time, at least for now, the array will only grow. It will tend to be big enough to hold the maximum number of kmem-limited memcgs ever achieved. We'll allocate it to be a minimum of 64 kmem-limited memcgs. When we have more than that, we'll start doubling the size of this array every time the limit is reached. Because we are only considering kmem limited memcgs, a natural point for this to happen is when we write to the limit. At that point, we already have set_limit_mutex held, so that will become our natural synchronization mechanism. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:38 +00:00
}
memcg: zap memcg_slab_caches and memcg_slab_mutex mem_cgroup->memcg_slab_caches is a list of kmem caches corresponding to the given cgroup. Currently, it is only used on css free in order to destroy all caches corresponding to the memory cgroup being freed. The list is protected by memcg_slab_mutex. The mutex is also used to protect kmem_cache->memcg_params->memcg_caches arrays and synchronizes kmem_cache_destroy vs memcg_unregister_all_caches. However, we can perfectly get on without these two. To destroy all caches corresponding to a memory cgroup, we can walk over the global list of kmem caches, slab_caches, and we can do all the synchronization stuff using the slab_mutex instead of the memcg_slab_mutex. This patch therefore gets rid of the memcg_slab_caches and memcg_slab_mutex. Apart from this nice cleanup, it also: - assures that rcu_barrier() is called once at max when a root cache is destroyed or a memory cgroup is freed, no matter how many caches have SLAB_DESTROY_BY_RCU flag set; - fixes the race between kmem_cache_destroy and kmem_cache_create that exists, because memcg_cleanup_cache_params, which is called from kmem_cache_destroy after checking that kmem_cache->refcount=0, releases the slab_mutex, which gives kmem_cache_create a chance to make an alias to a cache doomed to be destroyed. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Acked-by: Christoph Lameter <cl@linux.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-10 22:11:47 +00:00
struct memcg_kmem_cache_create_work {
struct mem_cgroup *memcg;
struct kmem_cache *cachep;
struct work_struct work;
};
memcg: zap memcg_slab_caches and memcg_slab_mutex mem_cgroup->memcg_slab_caches is a list of kmem caches corresponding to the given cgroup. Currently, it is only used on css free in order to destroy all caches corresponding to the memory cgroup being freed. The list is protected by memcg_slab_mutex. The mutex is also used to protect kmem_cache->memcg_params->memcg_caches arrays and synchronizes kmem_cache_destroy vs memcg_unregister_all_caches. However, we can perfectly get on without these two. To destroy all caches corresponding to a memory cgroup, we can walk over the global list of kmem caches, slab_caches, and we can do all the synchronization stuff using the slab_mutex instead of the memcg_slab_mutex. This patch therefore gets rid of the memcg_slab_caches and memcg_slab_mutex. Apart from this nice cleanup, it also: - assures that rcu_barrier() is called once at max when a root cache is destroyed or a memory cgroup is freed, no matter how many caches have SLAB_DESTROY_BY_RCU flag set; - fixes the race between kmem_cache_destroy and kmem_cache_create that exists, because memcg_cleanup_cache_params, which is called from kmem_cache_destroy after checking that kmem_cache->refcount=0, releases the slab_mutex, which gives kmem_cache_create a chance to make an alias to a cache doomed to be destroyed. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Acked-by: Christoph Lameter <cl@linux.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-10 22:11:47 +00:00
static void memcg_kmem_cache_create_func(struct work_struct *w)
memcg: infrastructure to match an allocation to the right cache The page allocator is able to bind a page to a memcg when it is allocated. But for the caches, we'd like to have as many objects as possible in a page belonging to the same cache. This is done in this patch by calling memcg_kmem_get_cache in the beginning of every allocation function. This function is patched out by static branches when kernel memory controller is not being used. It assumes that the task allocating, which determines the memcg in the page allocator, belongs to the same cgroup throughout the whole process. Misaccounting can happen if the task calls memcg_kmem_get_cache() while belonging to a cgroup, and later on changes. This is considered acceptable, and should only happen upon task migration. Before the cache is created by the memcg core, there is also a possible imbalance: the task belongs to a memcg, but the cache being allocated from is the global cache, since the child cache is not yet guaranteed to be ready. This case is also fine, since in this case the GFP_KMEMCG will not be passed and the page allocator will not attempt any cgroup accounting. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:40 +00:00
{
memcg: zap memcg_slab_caches and memcg_slab_mutex mem_cgroup->memcg_slab_caches is a list of kmem caches corresponding to the given cgroup. Currently, it is only used on css free in order to destroy all caches corresponding to the memory cgroup being freed. The list is protected by memcg_slab_mutex. The mutex is also used to protect kmem_cache->memcg_params->memcg_caches arrays and synchronizes kmem_cache_destroy vs memcg_unregister_all_caches. However, we can perfectly get on without these two. To destroy all caches corresponding to a memory cgroup, we can walk over the global list of kmem caches, slab_caches, and we can do all the synchronization stuff using the slab_mutex instead of the memcg_slab_mutex. This patch therefore gets rid of the memcg_slab_caches and memcg_slab_mutex. Apart from this nice cleanup, it also: - assures that rcu_barrier() is called once at max when a root cache is destroyed or a memory cgroup is freed, no matter how many caches have SLAB_DESTROY_BY_RCU flag set; - fixes the race between kmem_cache_destroy and kmem_cache_create that exists, because memcg_cleanup_cache_params, which is called from kmem_cache_destroy after checking that kmem_cache->refcount=0, releases the slab_mutex, which gives kmem_cache_create a chance to make an alias to a cache doomed to be destroyed. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Acked-by: Christoph Lameter <cl@linux.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-10 22:11:47 +00:00
struct memcg_kmem_cache_create_work *cw =
container_of(w, struct memcg_kmem_cache_create_work, work);
struct mem_cgroup *memcg = cw->memcg;
struct kmem_cache *cachep = cw->cachep;
memcg: infrastructure to match an allocation to the right cache The page allocator is able to bind a page to a memcg when it is allocated. But for the caches, we'd like to have as many objects as possible in a page belonging to the same cache. This is done in this patch by calling memcg_kmem_get_cache in the beginning of every allocation function. This function is patched out by static branches when kernel memory controller is not being used. It assumes that the task allocating, which determines the memcg in the page allocator, belongs to the same cgroup throughout the whole process. Misaccounting can happen if the task calls memcg_kmem_get_cache() while belonging to a cgroup, and later on changes. This is considered acceptable, and should only happen upon task migration. Before the cache is created by the memcg core, there is also a possible imbalance: the task belongs to a memcg, but the cache being allocated from is the global cache, since the child cache is not yet guaranteed to be ready. This case is also fine, since in this case the GFP_KMEMCG will not be passed and the page allocator will not attempt any cgroup accounting. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:40 +00:00
memcg: zap memcg_slab_caches and memcg_slab_mutex mem_cgroup->memcg_slab_caches is a list of kmem caches corresponding to the given cgroup. Currently, it is only used on css free in order to destroy all caches corresponding to the memory cgroup being freed. The list is protected by memcg_slab_mutex. The mutex is also used to protect kmem_cache->memcg_params->memcg_caches arrays and synchronizes kmem_cache_destroy vs memcg_unregister_all_caches. However, we can perfectly get on without these two. To destroy all caches corresponding to a memory cgroup, we can walk over the global list of kmem caches, slab_caches, and we can do all the synchronization stuff using the slab_mutex instead of the memcg_slab_mutex. This patch therefore gets rid of the memcg_slab_caches and memcg_slab_mutex. Apart from this nice cleanup, it also: - assures that rcu_barrier() is called once at max when a root cache is destroyed or a memory cgroup is freed, no matter how many caches have SLAB_DESTROY_BY_RCU flag set; - fixes the race between kmem_cache_destroy and kmem_cache_create that exists, because memcg_cleanup_cache_params, which is called from kmem_cache_destroy after checking that kmem_cache->refcount=0, releases the slab_mutex, which gives kmem_cache_create a chance to make an alias to a cache doomed to be destroyed. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Acked-by: Christoph Lameter <cl@linux.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-10 22:11:47 +00:00
memcg_create_kmem_cache(memcg, cachep);
memcg, slab: simplify synchronization scheme At present, we have the following mutexes protecting data related to per memcg kmem caches: - slab_mutex. This one is held during the whole kmem cache creation and destruction paths. We also take it when updating per root cache memcg_caches arrays (see memcg_update_all_caches). As a result, taking it guarantees there will be no changes to any kmem cache (including per memcg). Why do we need something else then? The point is it is private to slab implementation and has some internal dependencies with other mutexes (get_online_cpus). So we just don't want to rely upon it and prefer to introduce additional mutexes instead. - activate_kmem_mutex. Initially it was added to synchronize initializing kmem limit (memcg_activate_kmem). However, since we can grow per root cache memcg_caches arrays only on kmem limit initialization (see memcg_update_all_caches), we also employ it to protect against memcg_caches arrays relocation (e.g. see __kmem_cache_destroy_memcg_children). - We have a convention not to take slab_mutex in memcontrol.c, but we want to walk over per memcg memcg_slab_caches lists there (e.g. for destroying all memcg caches on offline). So we have per memcg slab_caches_mutex's protecting those lists. The mutexes are taken in the following order: activate_kmem_mutex -> slab_mutex -> memcg::slab_caches_mutex Such a syncrhonization scheme has a number of flaws, for instance: - We can't call kmem_cache_{destroy,shrink} while walking over a memcg::memcg_slab_caches list due to locking order. As a result, in mem_cgroup_destroy_all_caches we schedule the memcg_cache_params::destroy work shrinking and destroying the cache. - We don't have a mutex to synchronize per memcg caches destruction between memcg offline (mem_cgroup_destroy_all_caches) and root cache destruction (__kmem_cache_destroy_memcg_children). Currently we just don't bother about it. This patch simplifies it by substituting per memcg slab_caches_mutex's with the global memcg_slab_mutex. It will be held whenever a new per memcg cache is created or destroyed, so it protects per root cache memcg_caches arrays and per memcg memcg_slab_caches lists. The locking order is following: activate_kmem_mutex -> memcg_slab_mutex -> slab_mutex This allows us to call kmem_cache_{create,shrink,destroy} under the memcg_slab_mutex. As a result, we don't need memcg_cache_params::destroy work any more - we can simply destroy caches while iterating over a per memcg slab caches list. Also using the global mutex simplifies synchronization between concurrent per memcg caches creation/destruction, e.g. mem_cgroup_destroy_all_caches vs __kmem_cache_destroy_memcg_children. The downside of this is that we substitute per-memcg slab_caches_mutex's with a hummer-like global mutex, but since we already take either the slab_mutex or the cgroup_mutex along with a memcg::slab_caches_mutex, it shouldn't hurt concurrency a lot. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Glauber Costa <glommer@gmail.com> Cc: Pekka Enberg <penberg@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-06-04 23:07:40 +00:00
css_put(&memcg->css);
memcg: infrastructure to match an allocation to the right cache The page allocator is able to bind a page to a memcg when it is allocated. But for the caches, we'd like to have as many objects as possible in a page belonging to the same cache. This is done in this patch by calling memcg_kmem_get_cache in the beginning of every allocation function. This function is patched out by static branches when kernel memory controller is not being used. It assumes that the task allocating, which determines the memcg in the page allocator, belongs to the same cgroup throughout the whole process. Misaccounting can happen if the task calls memcg_kmem_get_cache() while belonging to a cgroup, and later on changes. This is considered acceptable, and should only happen upon task migration. Before the cache is created by the memcg core, there is also a possible imbalance: the task belongs to a memcg, but the cache being allocated from is the global cache, since the child cache is not yet guaranteed to be ready. This case is also fine, since in this case the GFP_KMEMCG will not be passed and the page allocator will not attempt any cgroup accounting. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:40 +00:00
kfree(cw);
}
/*
* Enqueue the creation of a per-memcg kmem_cache.
*/
static void memcg_schedule_kmem_cache_create(struct mem_cgroup *memcg,
memcg: zap memcg_slab_caches and memcg_slab_mutex mem_cgroup->memcg_slab_caches is a list of kmem caches corresponding to the given cgroup. Currently, it is only used on css free in order to destroy all caches corresponding to the memory cgroup being freed. The list is protected by memcg_slab_mutex. The mutex is also used to protect kmem_cache->memcg_params->memcg_caches arrays and synchronizes kmem_cache_destroy vs memcg_unregister_all_caches. However, we can perfectly get on without these two. To destroy all caches corresponding to a memory cgroup, we can walk over the global list of kmem caches, slab_caches, and we can do all the synchronization stuff using the slab_mutex instead of the memcg_slab_mutex. This patch therefore gets rid of the memcg_slab_caches and memcg_slab_mutex. Apart from this nice cleanup, it also: - assures that rcu_barrier() is called once at max when a root cache is destroyed or a memory cgroup is freed, no matter how many caches have SLAB_DESTROY_BY_RCU flag set; - fixes the race between kmem_cache_destroy and kmem_cache_create that exists, because memcg_cleanup_cache_params, which is called from kmem_cache_destroy after checking that kmem_cache->refcount=0, releases the slab_mutex, which gives kmem_cache_create a chance to make an alias to a cache doomed to be destroyed. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Acked-by: Christoph Lameter <cl@linux.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-10 22:11:47 +00:00
struct kmem_cache *cachep)
memcg: infrastructure to match an allocation to the right cache The page allocator is able to bind a page to a memcg when it is allocated. But for the caches, we'd like to have as many objects as possible in a page belonging to the same cache. This is done in this patch by calling memcg_kmem_get_cache in the beginning of every allocation function. This function is patched out by static branches when kernel memory controller is not being used. It assumes that the task allocating, which determines the memcg in the page allocator, belongs to the same cgroup throughout the whole process. Misaccounting can happen if the task calls memcg_kmem_get_cache() while belonging to a cgroup, and later on changes. This is considered acceptable, and should only happen upon task migration. Before the cache is created by the memcg core, there is also a possible imbalance: the task belongs to a memcg, but the cache being allocated from is the global cache, since the child cache is not yet guaranteed to be ready. This case is also fine, since in this case the GFP_KMEMCG will not be passed and the page allocator will not attempt any cgroup accounting. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:40 +00:00
{
memcg: zap memcg_slab_caches and memcg_slab_mutex mem_cgroup->memcg_slab_caches is a list of kmem caches corresponding to the given cgroup. Currently, it is only used on css free in order to destroy all caches corresponding to the memory cgroup being freed. The list is protected by memcg_slab_mutex. The mutex is also used to protect kmem_cache->memcg_params->memcg_caches arrays and synchronizes kmem_cache_destroy vs memcg_unregister_all_caches. However, we can perfectly get on without these two. To destroy all caches corresponding to a memory cgroup, we can walk over the global list of kmem caches, slab_caches, and we can do all the synchronization stuff using the slab_mutex instead of the memcg_slab_mutex. This patch therefore gets rid of the memcg_slab_caches and memcg_slab_mutex. Apart from this nice cleanup, it also: - assures that rcu_barrier() is called once at max when a root cache is destroyed or a memory cgroup is freed, no matter how many caches have SLAB_DESTROY_BY_RCU flag set; - fixes the race between kmem_cache_destroy and kmem_cache_create that exists, because memcg_cleanup_cache_params, which is called from kmem_cache_destroy after checking that kmem_cache->refcount=0, releases the slab_mutex, which gives kmem_cache_create a chance to make an alias to a cache doomed to be destroyed. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Acked-by: Christoph Lameter <cl@linux.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-10 22:11:47 +00:00
struct memcg_kmem_cache_create_work *cw;
memcg: infrastructure to match an allocation to the right cache The page allocator is able to bind a page to a memcg when it is allocated. But for the caches, we'd like to have as many objects as possible in a page belonging to the same cache. This is done in this patch by calling memcg_kmem_get_cache in the beginning of every allocation function. This function is patched out by static branches when kernel memory controller is not being used. It assumes that the task allocating, which determines the memcg in the page allocator, belongs to the same cgroup throughout the whole process. Misaccounting can happen if the task calls memcg_kmem_get_cache() while belonging to a cgroup, and later on changes. This is considered acceptable, and should only happen upon task migration. Before the cache is created by the memcg core, there is also a possible imbalance: the task belongs to a memcg, but the cache being allocated from is the global cache, since the child cache is not yet guaranteed to be ready. This case is also fine, since in this case the GFP_KMEMCG will not be passed and the page allocator will not attempt any cgroup accounting. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:40 +00:00
mm: memcg/slab: rework non-root kmem_cache lifecycle management Currently each charged slab page holds a reference to the cgroup to which it's charged. Kmem_caches are held by the memcg and are released all together with the memory cgroup. It means that none of kmem_caches are released unless at least one reference to the memcg exists, which is very far from optimal. Let's rework it in a way that allows releasing individual kmem_caches as soon as the cgroup is offline, the kmem_cache is empty and there are no pending allocations. To make it possible, let's introduce a new percpu refcounter for non-root kmem caches. The counter is initialized to the percpu mode, and is switched to the atomic mode during kmem_cache deactivation. The counter is bumped for every charged page and also for every running allocation. So the kmem_cache can't be released unless all allocations complete. To shutdown non-active empty kmem_caches, let's reuse the work queue, previously used for the kmem_cache deactivation. Once the reference counter reaches 0, let's schedule an asynchronous kmem_cache release. * I used the following simple approach to test the performance (stolen from another patchset by T. Harding): time find / -name fname-no-exist echo 2 > /proc/sys/vm/drop_caches repeat 10 times Results: orig patched real 0m1.455s real 0m1.355s user 0m0.206s user 0m0.219s sys 0m0.855s sys 0m0.807s real 0m1.487s real 0m1.699s user 0m0.221s user 0m0.256s sys 0m0.806s sys 0m0.948s real 0m1.515s real 0m1.505s user 0m0.183s user 0m0.215s sys 0m0.876s sys 0m0.858s real 0m1.291s real 0m1.380s user 0m0.193s user 0m0.198s sys 0m0.843s sys 0m0.786s real 0m1.364s real 0m1.374s user 0m0.180s user 0m0.182s sys 0m0.868s sys 0m0.806s real 0m1.352s real 0m1.312s user 0m0.201s user 0m0.212s sys 0m0.820s sys 0m0.761s real 0m1.302s real 0m1.349s user 0m0.205s user 0m0.203s sys 0m0.803s sys 0m0.792s real 0m1.334s real 0m1.301s user 0m0.194s user 0m0.201s sys 0m0.806s sys 0m0.779s real 0m1.426s real 0m1.434s user 0m0.216s user 0m0.181s sys 0m0.824s sys 0m0.864s real 0m1.350s real 0m1.295s user 0m0.200s user 0m0.190s sys 0m0.842s sys 0m0.811s So it looks like the difference is not noticeable in this test. [cai@lca.pw: fix an use-after-free in kmemcg_workfn()] Link: http://lkml.kernel.org/r/1560977573-10715-1-git-send-email-cai@lca.pw Link: http://lkml.kernel.org/r/20190611231813.3148843-9-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Signed-off-by: Qian Cai <cai@lca.pw> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Christoph Lameter <cl@linux.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Waiman Long <longman@redhat.com> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: Andrei Vagin <avagin@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:56:27 +00:00
if (!css_tryget_online(&memcg->css))
return;
mm: memcg: add __GFP_NOWARN in __memcg_schedule_kmem_cache_create() If there is heavy memory pressure, page allocation with __GFP_NOWAIT fails easily although it's order-0 request. I got below warning 9 times for normal boot. <snip >: page allocation failure: order:0, mode:0x2200000(GFP_NOWAIT|__GFP_NOTRACK) .. snip .. Call trace: dump_backtrace+0x0/0x4 dump_stack+0xa4/0xc0 warn_alloc+0xd4/0x15c __alloc_pages_nodemask+0xf88/0x10fc alloc_slab_page+0x40/0x18c new_slab+0x2b8/0x2e0 ___slab_alloc+0x25c/0x464 __kmalloc+0x394/0x498 memcg_kmem_get_cache+0x114/0x2b8 kmem_cache_alloc+0x98/0x3e8 mmap_region+0x3bc/0x8c0 do_mmap+0x40c/0x43c vm_mmap_pgoff+0x15c/0x1e4 sys_mmap+0xb0/0xc8 el0_svc_naked+0x24/0x28 Mem-Info: active_anon:17124 inactive_anon:193 isolated_anon:0 active_file:7898 inactive_file:712955 isolated_file:55 unevictable:0 dirty:27 writeback:18 unstable:0 slab_reclaimable:12250 slab_unreclaimable:23334 mapped:19310 shmem:212 pagetables:816 bounce:0 free:36561 free_pcp:1205 free_cma:35615 Node 0 active_anon:68496kB inactive_anon:772kB active_file:31592kB inactive_file:2851820kB unevictable:0kB isolated(anon):0kB isolated(file):220kB mapped:77240kB dirty:108kB writeback:72kB shmem:848kB writeback_tmp:0kB unstable:0kB all_unreclaimable? no DMA free:142188kB min:3056kB low:3820kB high:4584kB active_anon:10052kB inactive_anon:12kB active_file:312kB inactive_file:1412620kB unevictable:0kB writepending:0kB present:1781412kB managed:1604728kB mlocked:0kB slab_reclaimable:3592kB slab_unreclaimable:876kB kernel_stack:400kB pagetables:52kB bounce:0kB free_pcp:1436kB local_pcp:124kB free_cma:142492kB lowmem_reserve[]: 0 1842 1842 Normal free:4056kB min:4172kB low:5212kB high:6252kB active_anon:58376kB inactive_anon:760kB active_file:31348kB inactive_file:1439040kB unevictable:0kB writepending:180kB present:2000636kB managed:1923688kB mlocked:0kB slab_reclaimable:45408kB slab_unreclaimable:92460kB kernel_stack:9680kB pagetables:3212kB bounce:0kB free_pcp:3392kB local_pcp:688kB free_cma:0kB lowmem_reserve[]: 0 0 0 DMA: 0*4kB 0*8kB 1*16kB (C) 0*32kB 0*64kB 0*128kB 1*256kB (C) 1*512kB (C) 0*1024kB 1*2048kB (C) 34*4096kB (C) = 142096kB Normal: 228*4kB (UMEH) 172*8kB (UMH) 23*16kB (UH) 24*32kB (H) 5*64kB (H) 1*128kB (H) 0*256kB 0*512kB 0*1024kB 0*2048kB 0*4096kB = 3872kB 721350 total pagecache pages 0 pages in swap cache Swap cache stats: add 0, delete 0, find 0/0 Free swap = 0kB Total swap = 0kB 945512 pages RAM 0 pages HighMem/MovableOnly 63408 pages reserved 51200 pages cma reserved __memcg_schedule_kmem_cache_create() tries to create a shadow slab cache and the worker allocation failure is not really critical because we will retry on the next kmem charge. We might miss some charges but that shouldn't be critical. The excessive allocation failure report is not very helpful. [mhocko@kernel.org: changelog update] Link: http://lkml.kernel.org/r/20180418022912.248417-1-minchan@kernel.org Signed-off-by: Minchan Kim <minchan@kernel.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Andrew Morton <akpm@linux-foundation.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Matthew Wilcox <willy@infradead.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-20 21:56:17 +00:00
cw = kmalloc(sizeof(*cw), GFP_NOWAIT | __GFP_NOWARN);
memcg: fix possible use-after-free in memcg_kmem_get_cache() Suppose task @t that belongs to a memory cgroup @memcg is going to allocate an object from a kmem cache @c. The copy of @c corresponding to @memcg, @mc, is empty. Then if kmem_cache_alloc races with the memory cgroup destruction we can access the memory cgroup's copy of the cache after it was destroyed: CPU0 CPU1 ---- ---- [ current=@t @mc->memcg_params->nr_pages=0 ] kmem_cache_alloc(@c): call memcg_kmem_get_cache(@c); proceed to allocation from @mc: alloc a page for @mc: ... move @t from @memcg destroy @memcg: mem_cgroup_css_offline(@memcg): memcg_unregister_all_caches(@memcg): kmem_cache_destroy(@mc) add page to @mc We could fix this issue by taking a reference to a per-memcg cache, but that would require adding a per-cpu reference counter to per-memcg caches, which would look cumbersome. Instead, let's take a reference to a memory cgroup, which already has a per-cpu reference counter, in the beginning of kmem_cache_alloc to be dropped in the end, and move per memcg caches destruction from css offline to css free. As a side effect, per-memcg caches will be destroyed not one by one, but all at once when the last page accounted to the memory cgroup is freed. This doesn't sound as a high price for code readability though. Note, this patch does add some overhead to the kmem_cache_alloc hot path, but it is pretty negligible - it's just a function call plus a per cpu counter decrement, which is comparable to what we already have in memcg_kmem_get_cache. Besides, it's only relevant if there are memory cgroups with kmem accounting enabled. I don't think we can find a way to handle this race w/o it, because alloc_page called from kmem_cache_alloc may sleep so we can't flush all pending kmallocs w/o reference counting. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Christoph Lameter <cl@linux.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-13 00:56:38 +00:00
if (!cw)
memcg: infrastructure to match an allocation to the right cache The page allocator is able to bind a page to a memcg when it is allocated. But for the caches, we'd like to have as many objects as possible in a page belonging to the same cache. This is done in this patch by calling memcg_kmem_get_cache in the beginning of every allocation function. This function is patched out by static branches when kernel memory controller is not being used. It assumes that the task allocating, which determines the memcg in the page allocator, belongs to the same cgroup throughout the whole process. Misaccounting can happen if the task calls memcg_kmem_get_cache() while belonging to a cgroup, and later on changes. This is considered acceptable, and should only happen upon task migration. Before the cache is created by the memcg core, there is also a possible imbalance: the task belongs to a memcg, but the cache being allocated from is the global cache, since the child cache is not yet guaranteed to be ready. This case is also fine, since in this case the GFP_KMEMCG will not be passed and the page allocator will not attempt any cgroup accounting. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:40 +00:00
return;
memcg: fix possible use-after-free in memcg_kmem_get_cache() Suppose task @t that belongs to a memory cgroup @memcg is going to allocate an object from a kmem cache @c. The copy of @c corresponding to @memcg, @mc, is empty. Then if kmem_cache_alloc races with the memory cgroup destruction we can access the memory cgroup's copy of the cache after it was destroyed: CPU0 CPU1 ---- ---- [ current=@t @mc->memcg_params->nr_pages=0 ] kmem_cache_alloc(@c): call memcg_kmem_get_cache(@c); proceed to allocation from @mc: alloc a page for @mc: ... move @t from @memcg destroy @memcg: mem_cgroup_css_offline(@memcg): memcg_unregister_all_caches(@memcg): kmem_cache_destroy(@mc) add page to @mc We could fix this issue by taking a reference to a per-memcg cache, but that would require adding a per-cpu reference counter to per-memcg caches, which would look cumbersome. Instead, let's take a reference to a memory cgroup, which already has a per-cpu reference counter, in the beginning of kmem_cache_alloc to be dropped in the end, and move per memcg caches destruction from css offline to css free. As a side effect, per-memcg caches will be destroyed not one by one, but all at once when the last page accounted to the memory cgroup is freed. This doesn't sound as a high price for code readability though. Note, this patch does add some overhead to the kmem_cache_alloc hot path, but it is pretty negligible - it's just a function call plus a per cpu counter decrement, which is comparable to what we already have in memcg_kmem_get_cache. Besides, it's only relevant if there are memory cgroups with kmem accounting enabled. I don't think we can find a way to handle this race w/o it, because alloc_page called from kmem_cache_alloc may sleep so we can't flush all pending kmallocs w/o reference counting. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Christoph Lameter <cl@linux.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-13 00:56:38 +00:00
memcg: infrastructure to match an allocation to the right cache The page allocator is able to bind a page to a memcg when it is allocated. But for the caches, we'd like to have as many objects as possible in a page belonging to the same cache. This is done in this patch by calling memcg_kmem_get_cache in the beginning of every allocation function. This function is patched out by static branches when kernel memory controller is not being used. It assumes that the task allocating, which determines the memcg in the page allocator, belongs to the same cgroup throughout the whole process. Misaccounting can happen if the task calls memcg_kmem_get_cache() while belonging to a cgroup, and later on changes. This is considered acceptable, and should only happen upon task migration. Before the cache is created by the memcg core, there is also a possible imbalance: the task belongs to a memcg, but the cache being allocated from is the global cache, since the child cache is not yet guaranteed to be ready. This case is also fine, since in this case the GFP_KMEMCG will not be passed and the page allocator will not attempt any cgroup accounting. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:40 +00:00
cw->memcg = memcg;
cw->cachep = cachep;
memcg: zap memcg_slab_caches and memcg_slab_mutex mem_cgroup->memcg_slab_caches is a list of kmem caches corresponding to the given cgroup. Currently, it is only used on css free in order to destroy all caches corresponding to the memory cgroup being freed. The list is protected by memcg_slab_mutex. The mutex is also used to protect kmem_cache->memcg_params->memcg_caches arrays and synchronizes kmem_cache_destroy vs memcg_unregister_all_caches. However, we can perfectly get on without these two. To destroy all caches corresponding to a memory cgroup, we can walk over the global list of kmem caches, slab_caches, and we can do all the synchronization stuff using the slab_mutex instead of the memcg_slab_mutex. This patch therefore gets rid of the memcg_slab_caches and memcg_slab_mutex. Apart from this nice cleanup, it also: - assures that rcu_barrier() is called once at max when a root cache is destroyed or a memory cgroup is freed, no matter how many caches have SLAB_DESTROY_BY_RCU flag set; - fixes the race between kmem_cache_destroy and kmem_cache_create that exists, because memcg_cleanup_cache_params, which is called from kmem_cache_destroy after checking that kmem_cache->refcount=0, releases the slab_mutex, which gives kmem_cache_create a chance to make an alias to a cache doomed to be destroyed. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Acked-by: Christoph Lameter <cl@linux.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-10 22:11:47 +00:00
INIT_WORK(&cw->work, memcg_kmem_cache_create_func);
memcg: infrastructure to match an allocation to the right cache The page allocator is able to bind a page to a memcg when it is allocated. But for the caches, we'd like to have as many objects as possible in a page belonging to the same cache. This is done in this patch by calling memcg_kmem_get_cache in the beginning of every allocation function. This function is patched out by static branches when kernel memory controller is not being used. It assumes that the task allocating, which determines the memcg in the page allocator, belongs to the same cgroup throughout the whole process. Misaccounting can happen if the task calls memcg_kmem_get_cache() while belonging to a cgroup, and later on changes. This is considered acceptable, and should only happen upon task migration. Before the cache is created by the memcg core, there is also a possible imbalance: the task belongs to a memcg, but the cache being allocated from is the global cache, since the child cache is not yet guaranteed to be ready. This case is also fine, since in this case the GFP_KMEMCG will not be passed and the page allocator will not attempt any cgroup accounting. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:40 +00:00
queue_work(memcg_kmem_cache_wq, &cw->work);
memcg: infrastructure to match an allocation to the right cache The page allocator is able to bind a page to a memcg when it is allocated. But for the caches, we'd like to have as many objects as possible in a page belonging to the same cache. This is done in this patch by calling memcg_kmem_get_cache in the beginning of every allocation function. This function is patched out by static branches when kernel memory controller is not being used. It assumes that the task allocating, which determines the memcg in the page allocator, belongs to the same cgroup throughout the whole process. Misaccounting can happen if the task calls memcg_kmem_get_cache() while belonging to a cgroup, and later on changes. This is considered acceptable, and should only happen upon task migration. Before the cache is created by the memcg core, there is also a possible imbalance: the task belongs to a memcg, but the cache being allocated from is the global cache, since the child cache is not yet guaranteed to be ready. This case is also fine, since in this case the GFP_KMEMCG will not be passed and the page allocator will not attempt any cgroup accounting. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:40 +00:00
}
static inline bool memcg_kmem_bypass(void)
{
if (in_interrupt())
return true;
/* Allow remote memcg charging in kthread contexts. */
if ((!current->mm || (current->flags & PF_KTHREAD)) &&
!current->active_memcg)
return true;
return false;
}
/**
* memcg_kmem_get_cache: select the correct per-memcg cache for allocation
* @cachep: the original global kmem cache
*
memcg: infrastructure to match an allocation to the right cache The page allocator is able to bind a page to a memcg when it is allocated. But for the caches, we'd like to have as many objects as possible in a page belonging to the same cache. This is done in this patch by calling memcg_kmem_get_cache in the beginning of every allocation function. This function is patched out by static branches when kernel memory controller is not being used. It assumes that the task allocating, which determines the memcg in the page allocator, belongs to the same cgroup throughout the whole process. Misaccounting can happen if the task calls memcg_kmem_get_cache() while belonging to a cgroup, and later on changes. This is considered acceptable, and should only happen upon task migration. Before the cache is created by the memcg core, there is also a possible imbalance: the task belongs to a memcg, but the cache being allocated from is the global cache, since the child cache is not yet guaranteed to be ready. This case is also fine, since in this case the GFP_KMEMCG will not be passed and the page allocator will not attempt any cgroup accounting. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:40 +00:00
* Return the kmem_cache we're supposed to use for a slab allocation.
* We try to use the current memcg's version of the cache.
*
* If the cache does not exist yet, if we are the first user of it, we
* create it asynchronously in a workqueue and let the current allocation
* go through with the original cache.
memcg: infrastructure to match an allocation to the right cache The page allocator is able to bind a page to a memcg when it is allocated. But for the caches, we'd like to have as many objects as possible in a page belonging to the same cache. This is done in this patch by calling memcg_kmem_get_cache in the beginning of every allocation function. This function is patched out by static branches when kernel memory controller is not being used. It assumes that the task allocating, which determines the memcg in the page allocator, belongs to the same cgroup throughout the whole process. Misaccounting can happen if the task calls memcg_kmem_get_cache() while belonging to a cgroup, and later on changes. This is considered acceptable, and should only happen upon task migration. Before the cache is created by the memcg core, there is also a possible imbalance: the task belongs to a memcg, but the cache being allocated from is the global cache, since the child cache is not yet guaranteed to be ready. This case is also fine, since in this case the GFP_KMEMCG will not be passed and the page allocator will not attempt any cgroup accounting. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:40 +00:00
*
* This function takes a reference to the cache it returns to assure it
* won't get destroyed while we are working with it. Once the caller is
* done with it, memcg_kmem_put_cache() must be called to release the
* reference.
memcg: infrastructure to match an allocation to the right cache The page allocator is able to bind a page to a memcg when it is allocated. But for the caches, we'd like to have as many objects as possible in a page belonging to the same cache. This is done in this patch by calling memcg_kmem_get_cache in the beginning of every allocation function. This function is patched out by static branches when kernel memory controller is not being used. It assumes that the task allocating, which determines the memcg in the page allocator, belongs to the same cgroup throughout the whole process. Misaccounting can happen if the task calls memcg_kmem_get_cache() while belonging to a cgroup, and later on changes. This is considered acceptable, and should only happen upon task migration. Before the cache is created by the memcg core, there is also a possible imbalance: the task belongs to a memcg, but the cache being allocated from is the global cache, since the child cache is not yet guaranteed to be ready. This case is also fine, since in this case the GFP_KMEMCG will not be passed and the page allocator will not attempt any cgroup accounting. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:40 +00:00
*/
struct kmem_cache *memcg_kmem_get_cache(struct kmem_cache *cachep)
memcg: infrastructure to match an allocation to the right cache The page allocator is able to bind a page to a memcg when it is allocated. But for the caches, we'd like to have as many objects as possible in a page belonging to the same cache. This is done in this patch by calling memcg_kmem_get_cache in the beginning of every allocation function. This function is patched out by static branches when kernel memory controller is not being used. It assumes that the task allocating, which determines the memcg in the page allocator, belongs to the same cgroup throughout the whole process. Misaccounting can happen if the task calls memcg_kmem_get_cache() while belonging to a cgroup, and later on changes. This is considered acceptable, and should only happen upon task migration. Before the cache is created by the memcg core, there is also a possible imbalance: the task belongs to a memcg, but the cache being allocated from is the global cache, since the child cache is not yet guaranteed to be ready. This case is also fine, since in this case the GFP_KMEMCG will not be passed and the page allocator will not attempt any cgroup accounting. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:40 +00:00
{
struct mem_cgroup *memcg;
struct kmem_cache *memcg_cachep;
mm: memcg/slab: rework non-root kmem_cache lifecycle management Currently each charged slab page holds a reference to the cgroup to which it's charged. Kmem_caches are held by the memcg and are released all together with the memory cgroup. It means that none of kmem_caches are released unless at least one reference to the memcg exists, which is very far from optimal. Let's rework it in a way that allows releasing individual kmem_caches as soon as the cgroup is offline, the kmem_cache is empty and there are no pending allocations. To make it possible, let's introduce a new percpu refcounter for non-root kmem caches. The counter is initialized to the percpu mode, and is switched to the atomic mode during kmem_cache deactivation. The counter is bumped for every charged page and also for every running allocation. So the kmem_cache can't be released unless all allocations complete. To shutdown non-active empty kmem_caches, let's reuse the work queue, previously used for the kmem_cache deactivation. Once the reference counter reaches 0, let's schedule an asynchronous kmem_cache release. * I used the following simple approach to test the performance (stolen from another patchset by T. Harding): time find / -name fname-no-exist echo 2 > /proc/sys/vm/drop_caches repeat 10 times Results: orig patched real 0m1.455s real 0m1.355s user 0m0.206s user 0m0.219s sys 0m0.855s sys 0m0.807s real 0m1.487s real 0m1.699s user 0m0.221s user 0m0.256s sys 0m0.806s sys 0m0.948s real 0m1.515s real 0m1.505s user 0m0.183s user 0m0.215s sys 0m0.876s sys 0m0.858s real 0m1.291s real 0m1.380s user 0m0.193s user 0m0.198s sys 0m0.843s sys 0m0.786s real 0m1.364s real 0m1.374s user 0m0.180s user 0m0.182s sys 0m0.868s sys 0m0.806s real 0m1.352s real 0m1.312s user 0m0.201s user 0m0.212s sys 0m0.820s sys 0m0.761s real 0m1.302s real 0m1.349s user 0m0.205s user 0m0.203s sys 0m0.803s sys 0m0.792s real 0m1.334s real 0m1.301s user 0m0.194s user 0m0.201s sys 0m0.806s sys 0m0.779s real 0m1.426s real 0m1.434s user 0m0.216s user 0m0.181s sys 0m0.824s sys 0m0.864s real 0m1.350s real 0m1.295s user 0m0.200s user 0m0.190s sys 0m0.842s sys 0m0.811s So it looks like the difference is not noticeable in this test. [cai@lca.pw: fix an use-after-free in kmemcg_workfn()] Link: http://lkml.kernel.org/r/1560977573-10715-1-git-send-email-cai@lca.pw Link: http://lkml.kernel.org/r/20190611231813.3148843-9-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Signed-off-by: Qian Cai <cai@lca.pw> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Christoph Lameter <cl@linux.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Waiman Long <longman@redhat.com> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: Andrei Vagin <avagin@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:56:27 +00:00
struct memcg_cache_array *arr;
int kmemcg_id;
memcg: infrastructure to match an allocation to the right cache The page allocator is able to bind a page to a memcg when it is allocated. But for the caches, we'd like to have as many objects as possible in a page belonging to the same cache. This is done in this patch by calling memcg_kmem_get_cache in the beginning of every allocation function. This function is patched out by static branches when kernel memory controller is not being used. It assumes that the task allocating, which determines the memcg in the page allocator, belongs to the same cgroup throughout the whole process. Misaccounting can happen if the task calls memcg_kmem_get_cache() while belonging to a cgroup, and later on changes. This is considered acceptable, and should only happen upon task migration. Before the cache is created by the memcg core, there is also a possible imbalance: the task belongs to a memcg, but the cache being allocated from is the global cache, since the child cache is not yet guaranteed to be ready. This case is also fine, since in this case the GFP_KMEMCG will not be passed and the page allocator will not attempt any cgroup accounting. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:40 +00:00
slab: embed memcg_cache_params to kmem_cache Currently, kmem_cache stores a pointer to struct memcg_cache_params instead of embedding it. The rationale is to save memory when kmem accounting is disabled. However, the memcg_cache_params has shrivelled drastically since it was first introduced: * Initially: struct memcg_cache_params { bool is_root_cache; union { struct kmem_cache *memcg_caches[0]; struct { struct mem_cgroup *memcg; struct list_head list; struct kmem_cache *root_cache; bool dead; atomic_t nr_pages; struct work_struct destroy; }; }; }; * Now: struct memcg_cache_params { bool is_root_cache; union { struct { struct rcu_head rcu_head; struct kmem_cache *memcg_caches[0]; }; struct { struct mem_cgroup *memcg; struct kmem_cache *root_cache; }; }; }; So the memory saving does not seem to be a clear win anymore. OTOH, keeping a pointer to memcg_cache_params struct instead of embedding it results in touching one more cache line on kmem alloc/free hot paths. Besides, it makes linking kmem caches in a list chained by a field of struct memcg_cache_params really painful due to a level of indirection, while I want to make them linked in the following patch. That said, let us embed it. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Dave Chinner <david@fromorbit.com> Cc: Dan Carpenter <dan.carpenter@oracle.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-12 22:59:20 +00:00
VM_BUG_ON(!is_root_cache(cachep));
memcg: infrastructure to match an allocation to the right cache The page allocator is able to bind a page to a memcg when it is allocated. But for the caches, we'd like to have as many objects as possible in a page belonging to the same cache. This is done in this patch by calling memcg_kmem_get_cache in the beginning of every allocation function. This function is patched out by static branches when kernel memory controller is not being used. It assumes that the task allocating, which determines the memcg in the page allocator, belongs to the same cgroup throughout the whole process. Misaccounting can happen if the task calls memcg_kmem_get_cache() while belonging to a cgroup, and later on changes. This is considered acceptable, and should only happen upon task migration. Before the cache is created by the memcg core, there is also a possible imbalance: the task belongs to a memcg, but the cache being allocated from is the global cache, since the child cache is not yet guaranteed to be ready. This case is also fine, since in this case the GFP_KMEMCG will not be passed and the page allocator will not attempt any cgroup accounting. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:40 +00:00
if (memcg_kmem_bypass())
return cachep;
mm: memcg/slab: rework non-root kmem_cache lifecycle management Currently each charged slab page holds a reference to the cgroup to which it's charged. Kmem_caches are held by the memcg and are released all together with the memory cgroup. It means that none of kmem_caches are released unless at least one reference to the memcg exists, which is very far from optimal. Let's rework it in a way that allows releasing individual kmem_caches as soon as the cgroup is offline, the kmem_cache is empty and there are no pending allocations. To make it possible, let's introduce a new percpu refcounter for non-root kmem caches. The counter is initialized to the percpu mode, and is switched to the atomic mode during kmem_cache deactivation. The counter is bumped for every charged page and also for every running allocation. So the kmem_cache can't be released unless all allocations complete. To shutdown non-active empty kmem_caches, let's reuse the work queue, previously used for the kmem_cache deactivation. Once the reference counter reaches 0, let's schedule an asynchronous kmem_cache release. * I used the following simple approach to test the performance (stolen from another patchset by T. Harding): time find / -name fname-no-exist echo 2 > /proc/sys/vm/drop_caches repeat 10 times Results: orig patched real 0m1.455s real 0m1.355s user 0m0.206s user 0m0.219s sys 0m0.855s sys 0m0.807s real 0m1.487s real 0m1.699s user 0m0.221s user 0m0.256s sys 0m0.806s sys 0m0.948s real 0m1.515s real 0m1.505s user 0m0.183s user 0m0.215s sys 0m0.876s sys 0m0.858s real 0m1.291s real 0m1.380s user 0m0.193s user 0m0.198s sys 0m0.843s sys 0m0.786s real 0m1.364s real 0m1.374s user 0m0.180s user 0m0.182s sys 0m0.868s sys 0m0.806s real 0m1.352s real 0m1.312s user 0m0.201s user 0m0.212s sys 0m0.820s sys 0m0.761s real 0m1.302s real 0m1.349s user 0m0.205s user 0m0.203s sys 0m0.803s sys 0m0.792s real 0m1.334s real 0m1.301s user 0m0.194s user 0m0.201s sys 0m0.806s sys 0m0.779s real 0m1.426s real 0m1.434s user 0m0.216s user 0m0.181s sys 0m0.824s sys 0m0.864s real 0m1.350s real 0m1.295s user 0m0.200s user 0m0.190s sys 0m0.842s sys 0m0.811s So it looks like the difference is not noticeable in this test. [cai@lca.pw: fix an use-after-free in kmemcg_workfn()] Link: http://lkml.kernel.org/r/1560977573-10715-1-git-send-email-cai@lca.pw Link: http://lkml.kernel.org/r/20190611231813.3148843-9-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Signed-off-by: Qian Cai <cai@lca.pw> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Christoph Lameter <cl@linux.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Waiman Long <longman@redhat.com> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: Andrei Vagin <avagin@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:56:27 +00:00
rcu_read_lock();
if (unlikely(current->active_memcg))
memcg = current->active_memcg;
else
memcg = mem_cgroup_from_task(current);
if (!memcg || memcg == root_mem_cgroup)
goto out_unlock;
kmemcg_id = READ_ONCE(memcg->kmemcg_id);
if (kmemcg_id < 0)
mm: memcg/slab: rework non-root kmem_cache lifecycle management Currently each charged slab page holds a reference to the cgroup to which it's charged. Kmem_caches are held by the memcg and are released all together with the memory cgroup. It means that none of kmem_caches are released unless at least one reference to the memcg exists, which is very far from optimal. Let's rework it in a way that allows releasing individual kmem_caches as soon as the cgroup is offline, the kmem_cache is empty and there are no pending allocations. To make it possible, let's introduce a new percpu refcounter for non-root kmem caches. The counter is initialized to the percpu mode, and is switched to the atomic mode during kmem_cache deactivation. The counter is bumped for every charged page and also for every running allocation. So the kmem_cache can't be released unless all allocations complete. To shutdown non-active empty kmem_caches, let's reuse the work queue, previously used for the kmem_cache deactivation. Once the reference counter reaches 0, let's schedule an asynchronous kmem_cache release. * I used the following simple approach to test the performance (stolen from another patchset by T. Harding): time find / -name fname-no-exist echo 2 > /proc/sys/vm/drop_caches repeat 10 times Results: orig patched real 0m1.455s real 0m1.355s user 0m0.206s user 0m0.219s sys 0m0.855s sys 0m0.807s real 0m1.487s real 0m1.699s user 0m0.221s user 0m0.256s sys 0m0.806s sys 0m0.948s real 0m1.515s real 0m1.505s user 0m0.183s user 0m0.215s sys 0m0.876s sys 0m0.858s real 0m1.291s real 0m1.380s user 0m0.193s user 0m0.198s sys 0m0.843s sys 0m0.786s real 0m1.364s real 0m1.374s user 0m0.180s user 0m0.182s sys 0m0.868s sys 0m0.806s real 0m1.352s real 0m1.312s user 0m0.201s user 0m0.212s sys 0m0.820s sys 0m0.761s real 0m1.302s real 0m1.349s user 0m0.205s user 0m0.203s sys 0m0.803s sys 0m0.792s real 0m1.334s real 0m1.301s user 0m0.194s user 0m0.201s sys 0m0.806s sys 0m0.779s real 0m1.426s real 0m1.434s user 0m0.216s user 0m0.181s sys 0m0.824s sys 0m0.864s real 0m1.350s real 0m1.295s user 0m0.200s user 0m0.190s sys 0m0.842s sys 0m0.811s So it looks like the difference is not noticeable in this test. [cai@lca.pw: fix an use-after-free in kmemcg_workfn()] Link: http://lkml.kernel.org/r/1560977573-10715-1-git-send-email-cai@lca.pw Link: http://lkml.kernel.org/r/20190611231813.3148843-9-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Signed-off-by: Qian Cai <cai@lca.pw> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Christoph Lameter <cl@linux.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Waiman Long <longman@redhat.com> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: Andrei Vagin <avagin@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:56:27 +00:00
goto out_unlock;
memcg: infrastructure to match an allocation to the right cache The page allocator is able to bind a page to a memcg when it is allocated. But for the caches, we'd like to have as many objects as possible in a page belonging to the same cache. This is done in this patch by calling memcg_kmem_get_cache in the beginning of every allocation function. This function is patched out by static branches when kernel memory controller is not being used. It assumes that the task allocating, which determines the memcg in the page allocator, belongs to the same cgroup throughout the whole process. Misaccounting can happen if the task calls memcg_kmem_get_cache() while belonging to a cgroup, and later on changes. This is considered acceptable, and should only happen upon task migration. Before the cache is created by the memcg core, there is also a possible imbalance: the task belongs to a memcg, but the cache being allocated from is the global cache, since the child cache is not yet guaranteed to be ready. This case is also fine, since in this case the GFP_KMEMCG will not be passed and the page allocator will not attempt any cgroup accounting. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:40 +00:00
mm: memcg/slab: rework non-root kmem_cache lifecycle management Currently each charged slab page holds a reference to the cgroup to which it's charged. Kmem_caches are held by the memcg and are released all together with the memory cgroup. It means that none of kmem_caches are released unless at least one reference to the memcg exists, which is very far from optimal. Let's rework it in a way that allows releasing individual kmem_caches as soon as the cgroup is offline, the kmem_cache is empty and there are no pending allocations. To make it possible, let's introduce a new percpu refcounter for non-root kmem caches. The counter is initialized to the percpu mode, and is switched to the atomic mode during kmem_cache deactivation. The counter is bumped for every charged page and also for every running allocation. So the kmem_cache can't be released unless all allocations complete. To shutdown non-active empty kmem_caches, let's reuse the work queue, previously used for the kmem_cache deactivation. Once the reference counter reaches 0, let's schedule an asynchronous kmem_cache release. * I used the following simple approach to test the performance (stolen from another patchset by T. Harding): time find / -name fname-no-exist echo 2 > /proc/sys/vm/drop_caches repeat 10 times Results: orig patched real 0m1.455s real 0m1.355s user 0m0.206s user 0m0.219s sys 0m0.855s sys 0m0.807s real 0m1.487s real 0m1.699s user 0m0.221s user 0m0.256s sys 0m0.806s sys 0m0.948s real 0m1.515s real 0m1.505s user 0m0.183s user 0m0.215s sys 0m0.876s sys 0m0.858s real 0m1.291s real 0m1.380s user 0m0.193s user 0m0.198s sys 0m0.843s sys 0m0.786s real 0m1.364s real 0m1.374s user 0m0.180s user 0m0.182s sys 0m0.868s sys 0m0.806s real 0m1.352s real 0m1.312s user 0m0.201s user 0m0.212s sys 0m0.820s sys 0m0.761s real 0m1.302s real 0m1.349s user 0m0.205s user 0m0.203s sys 0m0.803s sys 0m0.792s real 0m1.334s real 0m1.301s user 0m0.194s user 0m0.201s sys 0m0.806s sys 0m0.779s real 0m1.426s real 0m1.434s user 0m0.216s user 0m0.181s sys 0m0.824s sys 0m0.864s real 0m1.350s real 0m1.295s user 0m0.200s user 0m0.190s sys 0m0.842s sys 0m0.811s So it looks like the difference is not noticeable in this test. [cai@lca.pw: fix an use-after-free in kmemcg_workfn()] Link: http://lkml.kernel.org/r/1560977573-10715-1-git-send-email-cai@lca.pw Link: http://lkml.kernel.org/r/20190611231813.3148843-9-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Signed-off-by: Qian Cai <cai@lca.pw> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Christoph Lameter <cl@linux.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Waiman Long <longman@redhat.com> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: Andrei Vagin <avagin@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:56:27 +00:00
arr = rcu_dereference(cachep->memcg_params.memcg_caches);
/*
* Make sure we will access the up-to-date value. The code updating
* memcg_caches issues a write barrier to match the data dependency
* barrier inside READ_ONCE() (see memcg_create_kmem_cache()).
*/
memcg_cachep = READ_ONCE(arr->entries[kmemcg_id]);
/*
* If we are in a safe context (can wait, and not in interrupt
* context), we could be be predictable and return right away.
* This would guarantee that the allocation being performed
* already belongs in the new cache.
*
* However, there are some clashes that can arrive from locking.
* For instance, because we acquire the slab_mutex while doing
* memcg_create_kmem_cache, this means no further allocation
* could happen with the slab_mutex held. So it's better to
* defer everything.
mm: memcg/slab: rework non-root kmem_cache lifecycle management Currently each charged slab page holds a reference to the cgroup to which it's charged. Kmem_caches are held by the memcg and are released all together with the memory cgroup. It means that none of kmem_caches are released unless at least one reference to the memcg exists, which is very far from optimal. Let's rework it in a way that allows releasing individual kmem_caches as soon as the cgroup is offline, the kmem_cache is empty and there are no pending allocations. To make it possible, let's introduce a new percpu refcounter for non-root kmem caches. The counter is initialized to the percpu mode, and is switched to the atomic mode during kmem_cache deactivation. The counter is bumped for every charged page and also for every running allocation. So the kmem_cache can't be released unless all allocations complete. To shutdown non-active empty kmem_caches, let's reuse the work queue, previously used for the kmem_cache deactivation. Once the reference counter reaches 0, let's schedule an asynchronous kmem_cache release. * I used the following simple approach to test the performance (stolen from another patchset by T. Harding): time find / -name fname-no-exist echo 2 > /proc/sys/vm/drop_caches repeat 10 times Results: orig patched real 0m1.455s real 0m1.355s user 0m0.206s user 0m0.219s sys 0m0.855s sys 0m0.807s real 0m1.487s real 0m1.699s user 0m0.221s user 0m0.256s sys 0m0.806s sys 0m0.948s real 0m1.515s real 0m1.505s user 0m0.183s user 0m0.215s sys 0m0.876s sys 0m0.858s real 0m1.291s real 0m1.380s user 0m0.193s user 0m0.198s sys 0m0.843s sys 0m0.786s real 0m1.364s real 0m1.374s user 0m0.180s user 0m0.182s sys 0m0.868s sys 0m0.806s real 0m1.352s real 0m1.312s user 0m0.201s user 0m0.212s sys 0m0.820s sys 0m0.761s real 0m1.302s real 0m1.349s user 0m0.205s user 0m0.203s sys 0m0.803s sys 0m0.792s real 0m1.334s real 0m1.301s user 0m0.194s user 0m0.201s sys 0m0.806s sys 0m0.779s real 0m1.426s real 0m1.434s user 0m0.216s user 0m0.181s sys 0m0.824s sys 0m0.864s real 0m1.350s real 0m1.295s user 0m0.200s user 0m0.190s sys 0m0.842s sys 0m0.811s So it looks like the difference is not noticeable in this test. [cai@lca.pw: fix an use-after-free in kmemcg_workfn()] Link: http://lkml.kernel.org/r/1560977573-10715-1-git-send-email-cai@lca.pw Link: http://lkml.kernel.org/r/20190611231813.3148843-9-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Signed-off-by: Qian Cai <cai@lca.pw> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Christoph Lameter <cl@linux.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Waiman Long <longman@redhat.com> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: Andrei Vagin <avagin@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:56:27 +00:00
*
* If the memcg is dying or memcg_cache is about to be released,
* don't bother creating new kmem_caches. Because memcg_cachep
* is ZEROed as the fist step of kmem offlining, we don't need
* percpu_ref_tryget_live() here. css_tryget_online() check in
* memcg_schedule_kmem_cache_create() will prevent us from
* creation of a new kmem_cache.
*/
mm: memcg/slab: rework non-root kmem_cache lifecycle management Currently each charged slab page holds a reference to the cgroup to which it's charged. Kmem_caches are held by the memcg and are released all together with the memory cgroup. It means that none of kmem_caches are released unless at least one reference to the memcg exists, which is very far from optimal. Let's rework it in a way that allows releasing individual kmem_caches as soon as the cgroup is offline, the kmem_cache is empty and there are no pending allocations. To make it possible, let's introduce a new percpu refcounter for non-root kmem caches. The counter is initialized to the percpu mode, and is switched to the atomic mode during kmem_cache deactivation. The counter is bumped for every charged page and also for every running allocation. So the kmem_cache can't be released unless all allocations complete. To shutdown non-active empty kmem_caches, let's reuse the work queue, previously used for the kmem_cache deactivation. Once the reference counter reaches 0, let's schedule an asynchronous kmem_cache release. * I used the following simple approach to test the performance (stolen from another patchset by T. Harding): time find / -name fname-no-exist echo 2 > /proc/sys/vm/drop_caches repeat 10 times Results: orig patched real 0m1.455s real 0m1.355s user 0m0.206s user 0m0.219s sys 0m0.855s sys 0m0.807s real 0m1.487s real 0m1.699s user 0m0.221s user 0m0.256s sys 0m0.806s sys 0m0.948s real 0m1.515s real 0m1.505s user 0m0.183s user 0m0.215s sys 0m0.876s sys 0m0.858s real 0m1.291s real 0m1.380s user 0m0.193s user 0m0.198s sys 0m0.843s sys 0m0.786s real 0m1.364s real 0m1.374s user 0m0.180s user 0m0.182s sys 0m0.868s sys 0m0.806s real 0m1.352s real 0m1.312s user 0m0.201s user 0m0.212s sys 0m0.820s sys 0m0.761s real 0m1.302s real 0m1.349s user 0m0.205s user 0m0.203s sys 0m0.803s sys 0m0.792s real 0m1.334s real 0m1.301s user 0m0.194s user 0m0.201s sys 0m0.806s sys 0m0.779s real 0m1.426s real 0m1.434s user 0m0.216s user 0m0.181s sys 0m0.824s sys 0m0.864s real 0m1.350s real 0m1.295s user 0m0.200s user 0m0.190s sys 0m0.842s sys 0m0.811s So it looks like the difference is not noticeable in this test. [cai@lca.pw: fix an use-after-free in kmemcg_workfn()] Link: http://lkml.kernel.org/r/1560977573-10715-1-git-send-email-cai@lca.pw Link: http://lkml.kernel.org/r/20190611231813.3148843-9-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Signed-off-by: Qian Cai <cai@lca.pw> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Christoph Lameter <cl@linux.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Waiman Long <longman@redhat.com> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: Andrei Vagin <avagin@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:56:27 +00:00
if (unlikely(!memcg_cachep))
memcg_schedule_kmem_cache_create(memcg, cachep);
else if (percpu_ref_tryget(&memcg_cachep->memcg_params.refcnt))
cachep = memcg_cachep;
out_unlock:
rcu_read_unlock();
return cachep;
memcg: infrastructure to match an allocation to the right cache The page allocator is able to bind a page to a memcg when it is allocated. But for the caches, we'd like to have as many objects as possible in a page belonging to the same cache. This is done in this patch by calling memcg_kmem_get_cache in the beginning of every allocation function. This function is patched out by static branches when kernel memory controller is not being used. It assumes that the task allocating, which determines the memcg in the page allocator, belongs to the same cgroup throughout the whole process. Misaccounting can happen if the task calls memcg_kmem_get_cache() while belonging to a cgroup, and later on changes. This is considered acceptable, and should only happen upon task migration. Before the cache is created by the memcg core, there is also a possible imbalance: the task belongs to a memcg, but the cache being allocated from is the global cache, since the child cache is not yet guaranteed to be ready. This case is also fine, since in this case the GFP_KMEMCG will not be passed and the page allocator will not attempt any cgroup accounting. Signed-off-by: Glauber Costa <glommer@parallels.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:40 +00:00
}
/**
* memcg_kmem_put_cache: drop reference taken by memcg_kmem_get_cache
* @cachep: the cache returned by memcg_kmem_get_cache
*/
void memcg_kmem_put_cache(struct kmem_cache *cachep)
memcg: fix possible use-after-free in memcg_kmem_get_cache() Suppose task @t that belongs to a memory cgroup @memcg is going to allocate an object from a kmem cache @c. The copy of @c corresponding to @memcg, @mc, is empty. Then if kmem_cache_alloc races with the memory cgroup destruction we can access the memory cgroup's copy of the cache after it was destroyed: CPU0 CPU1 ---- ---- [ current=@t @mc->memcg_params->nr_pages=0 ] kmem_cache_alloc(@c): call memcg_kmem_get_cache(@c); proceed to allocation from @mc: alloc a page for @mc: ... move @t from @memcg destroy @memcg: mem_cgroup_css_offline(@memcg): memcg_unregister_all_caches(@memcg): kmem_cache_destroy(@mc) add page to @mc We could fix this issue by taking a reference to a per-memcg cache, but that would require adding a per-cpu reference counter to per-memcg caches, which would look cumbersome. Instead, let's take a reference to a memory cgroup, which already has a per-cpu reference counter, in the beginning of kmem_cache_alloc to be dropped in the end, and move per memcg caches destruction from css offline to css free. As a side effect, per-memcg caches will be destroyed not one by one, but all at once when the last page accounted to the memory cgroup is freed. This doesn't sound as a high price for code readability though. Note, this patch does add some overhead to the kmem_cache_alloc hot path, but it is pretty negligible - it's just a function call plus a per cpu counter decrement, which is comparable to what we already have in memcg_kmem_get_cache. Besides, it's only relevant if there are memory cgroups with kmem accounting enabled. I don't think we can find a way to handle this race w/o it, because alloc_page called from kmem_cache_alloc may sleep so we can't flush all pending kmallocs w/o reference counting. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Christoph Lameter <cl@linux.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-13 00:56:38 +00:00
{
if (!is_root_cache(cachep))
mm: memcg/slab: rework non-root kmem_cache lifecycle management Currently each charged slab page holds a reference to the cgroup to which it's charged. Kmem_caches are held by the memcg and are released all together with the memory cgroup. It means that none of kmem_caches are released unless at least one reference to the memcg exists, which is very far from optimal. Let's rework it in a way that allows releasing individual kmem_caches as soon as the cgroup is offline, the kmem_cache is empty and there are no pending allocations. To make it possible, let's introduce a new percpu refcounter for non-root kmem caches. The counter is initialized to the percpu mode, and is switched to the atomic mode during kmem_cache deactivation. The counter is bumped for every charged page and also for every running allocation. So the kmem_cache can't be released unless all allocations complete. To shutdown non-active empty kmem_caches, let's reuse the work queue, previously used for the kmem_cache deactivation. Once the reference counter reaches 0, let's schedule an asynchronous kmem_cache release. * I used the following simple approach to test the performance (stolen from another patchset by T. Harding): time find / -name fname-no-exist echo 2 > /proc/sys/vm/drop_caches repeat 10 times Results: orig patched real 0m1.455s real 0m1.355s user 0m0.206s user 0m0.219s sys 0m0.855s sys 0m0.807s real 0m1.487s real 0m1.699s user 0m0.221s user 0m0.256s sys 0m0.806s sys 0m0.948s real 0m1.515s real 0m1.505s user 0m0.183s user 0m0.215s sys 0m0.876s sys 0m0.858s real 0m1.291s real 0m1.380s user 0m0.193s user 0m0.198s sys 0m0.843s sys 0m0.786s real 0m1.364s real 0m1.374s user 0m0.180s user 0m0.182s sys 0m0.868s sys 0m0.806s real 0m1.352s real 0m1.312s user 0m0.201s user 0m0.212s sys 0m0.820s sys 0m0.761s real 0m1.302s real 0m1.349s user 0m0.205s user 0m0.203s sys 0m0.803s sys 0m0.792s real 0m1.334s real 0m1.301s user 0m0.194s user 0m0.201s sys 0m0.806s sys 0m0.779s real 0m1.426s real 0m1.434s user 0m0.216s user 0m0.181s sys 0m0.824s sys 0m0.864s real 0m1.350s real 0m1.295s user 0m0.200s user 0m0.190s sys 0m0.842s sys 0m0.811s So it looks like the difference is not noticeable in this test. [cai@lca.pw: fix an use-after-free in kmemcg_workfn()] Link: http://lkml.kernel.org/r/1560977573-10715-1-git-send-email-cai@lca.pw Link: http://lkml.kernel.org/r/20190611231813.3148843-9-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Signed-off-by: Qian Cai <cai@lca.pw> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Christoph Lameter <cl@linux.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Waiman Long <longman@redhat.com> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: Andrei Vagin <avagin@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:56:27 +00:00
percpu_ref_put(&cachep->memcg_params.refcnt);
memcg: fix possible use-after-free in memcg_kmem_get_cache() Suppose task @t that belongs to a memory cgroup @memcg is going to allocate an object from a kmem cache @c. The copy of @c corresponding to @memcg, @mc, is empty. Then if kmem_cache_alloc races with the memory cgroup destruction we can access the memory cgroup's copy of the cache after it was destroyed: CPU0 CPU1 ---- ---- [ current=@t @mc->memcg_params->nr_pages=0 ] kmem_cache_alloc(@c): call memcg_kmem_get_cache(@c); proceed to allocation from @mc: alloc a page for @mc: ... move @t from @memcg destroy @memcg: mem_cgroup_css_offline(@memcg): memcg_unregister_all_caches(@memcg): kmem_cache_destroy(@mc) add page to @mc We could fix this issue by taking a reference to a per-memcg cache, but that would require adding a per-cpu reference counter to per-memcg caches, which would look cumbersome. Instead, let's take a reference to a memory cgroup, which already has a per-cpu reference counter, in the beginning of kmem_cache_alloc to be dropped in the end, and move per memcg caches destruction from css offline to css free. As a side effect, per-memcg caches will be destroyed not one by one, but all at once when the last page accounted to the memory cgroup is freed. This doesn't sound as a high price for code readability though. Note, this patch does add some overhead to the kmem_cache_alloc hot path, but it is pretty negligible - it's just a function call plus a per cpu counter decrement, which is comparable to what we already have in memcg_kmem_get_cache. Besides, it's only relevant if there are memory cgroups with kmem accounting enabled. I don't think we can find a way to handle this race w/o it, because alloc_page called from kmem_cache_alloc may sleep so we can't flush all pending kmallocs w/o reference counting. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: Christoph Lameter <cl@linux.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-13 00:56:38 +00:00
}
/**
* __memcg_kmem_charge: charge a number of kernel pages to a memcg
mm: kmem: cleanup (__)memcg_kmem_charge_memcg() arguments Patch series "mm: memcg: kmem API cleanup", v2. This patchset aims to clean up the kernel memory charging API. It doesn't bring any functional changes, just removes unused arguments, renames some functions and fixes some comments. Currently it's not obvious which functions are most basic (memcg_kmem_(un)charge_memcg()) and which are based on them (memcg_kmem_(un)charge()). The patchset renames these functions and removes unused arguments: TL;DR: was: memcg_kmem_charge_memcg(page, gfp, order, memcg) memcg_kmem_uncharge_memcg(memcg, nr_pages) memcg_kmem_charge(page, gfp, order) memcg_kmem_uncharge(page, order) now: memcg_kmem_charge(memcg, gfp, nr_pages) memcg_kmem_uncharge(memcg, nr_pages) memcg_kmem_charge_page(page, gfp, order) memcg_kmem_uncharge_page(page, order) This patch (of 6): The first argument of memcg_kmem_charge_memcg() and __memcg_kmem_charge_memcg() is the page pointer and it's not used. Let's drop it. Memcg pointer is passed as the last argument. Move it to the first place for consistency with other memcg functions, e.g. __memcg_kmem_uncharge_memcg() or try_charge(). Signed-off-by: Roman Gushchin <guro@fb.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Link: http://lkml.kernel.org/r/20200109202659.752357-2-guro@fb.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-04-02 04:06:39 +00:00
* @memcg: memory cgroup to charge
* @gfp: reclaim mode
* @nr_pages: number of pages to charge
*
* Returns 0 on success, an error code on failure.
*/
int __memcg_kmem_charge(struct mem_cgroup *memcg, gfp_t gfp,
unsigned int nr_pages)
memcg: kmem controller infrastructure Introduce infrastructure for tracking kernel memory pages to a given memcg. This will happen whenever the caller includes the flag __GFP_KMEMCG flag, and the task belong to a memcg other than the root. In memcontrol.h those functions are wrapped in inline acessors. The idea is to later on, patch those with static branches, so we don't incur any overhead when no mem cgroups with limited kmem are being used. Users of this functionality shall interact with the memcg core code through the following functions: memcg_kmem_newpage_charge: will return true if the group can handle the allocation. At this point, struct page is not yet allocated. memcg_kmem_commit_charge: will either revert the charge, if struct page allocation failed, or embed memcg information into page_cgroup. memcg_kmem_uncharge_page: called at free time, will revert the charge. Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:56 +00:00
{
memcg: unify slab and other kmem pages charging We have memcg_kmem_charge and memcg_kmem_uncharge methods for charging and uncharging kmem pages to memcg, but currently they are not used for charging slab pages (i.e. they are only used for charging pages allocated with alloc_kmem_pages). The only reason why the slab subsystem uses special helpers, memcg_charge_slab and memcg_uncharge_slab, is that it needs to charge to the memcg of kmem cache while memcg_charge_kmem charges to the memcg that the current task belongs to. To remove this diversity, this patch adds an extra argument to __memcg_kmem_charge that can be a pointer to a memcg or NULL. If it is not NULL, the function tries to charge to the memcg it points to, otherwise it charge to the current context. Next, it makes the slab subsystem use this function to charge slab pages. Since memcg_charge_kmem and memcg_uncharge_kmem helpers are now used only in __memcg_kmem_charge and __memcg_kmem_uncharge, they are inlined. Since __memcg_kmem_charge stores a pointer to the memcg in the page struct, we don't need memcg_uncharge_slab anymore and can use free_kmem_pages. Besides, one can now detect which memcg a slab page belongs to by reading /proc/kpagecgroup. Note, this patch switches slab to charge-after-alloc design. Since this design is already used for all other memcg charges, it should not make any difference. [hannes@cmpxchg.org: better to have an outer function than a magic parameter for the memcg lookup] Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:49:01 +00:00
struct page_counter *counter;
memcg: kmem controller infrastructure Introduce infrastructure for tracking kernel memory pages to a given memcg. This will happen whenever the caller includes the flag __GFP_KMEMCG flag, and the task belong to a memcg other than the root. In memcontrol.h those functions are wrapped in inline acessors. The idea is to later on, patch those with static branches, so we don't incur any overhead when no mem cgroups with limited kmem are being used. Users of this functionality shall interact with the memcg core code through the following functions: memcg_kmem_newpage_charge: will return true if the group can handle the allocation. At this point, struct page is not yet allocated. memcg_kmem_commit_charge: will either revert the charge, if struct page allocation failed, or embed memcg information into page_cgroup. memcg_kmem_uncharge_page: called at free time, will revert the charge. Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:56 +00:00
int ret;
memcg: unify slab and other kmem pages charging We have memcg_kmem_charge and memcg_kmem_uncharge methods for charging and uncharging kmem pages to memcg, but currently they are not used for charging slab pages (i.e. they are only used for charging pages allocated with alloc_kmem_pages). The only reason why the slab subsystem uses special helpers, memcg_charge_slab and memcg_uncharge_slab, is that it needs to charge to the memcg of kmem cache while memcg_charge_kmem charges to the memcg that the current task belongs to. To remove this diversity, this patch adds an extra argument to __memcg_kmem_charge that can be a pointer to a memcg or NULL. If it is not NULL, the function tries to charge to the memcg it points to, otherwise it charge to the current context. Next, it makes the slab subsystem use this function to charge slab pages. Since memcg_charge_kmem and memcg_uncharge_kmem helpers are now used only in __memcg_kmem_charge and __memcg_kmem_uncharge, they are inlined. Since __memcg_kmem_charge stores a pointer to the memcg in the page struct, we don't need memcg_uncharge_slab anymore and can use free_kmem_pages. Besides, one can now detect which memcg a slab page belongs to by reading /proc/kpagecgroup. Note, this patch switches slab to charge-after-alloc design. Since this design is already used for all other memcg charges, it should not make any difference. [hannes@cmpxchg.org: better to have an outer function than a magic parameter for the memcg lookup] Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:49:01 +00:00
ret = try_charge(memcg, gfp, nr_pages);
mm: memcontrol: account "kmem" consumers in cgroup2 memory controller The original cgroup memory controller has an extension to account slab memory (and other "kernel memory" consumers) in a separate "kmem" counter, once the user set an explicit limit on that "kmem" pool. However, this includes various consumers whose sizes are directly linked to userspace activity. Accounting them as an optional "kmem" extension is problematic for several reasons: 1. It leaves the main memory interface with incomplete semantics. A user who puts their workload into a cgroup and configures a memory limit does not expect us to leave holes in the containment as big as the dentry and inode cache, or the kernel stack pages. 2. If the limit set on this random historical subgroup of consumers is reached, subsequent allocations will fail even when the main memory pool available to the cgroup is not yet exhausted and/or has reclaimable memory in it. 3. Calling it 'kernel memory' is misleading. The dentry and inode caches are no more 'kernel' (or no less 'user') memory than the page cache itself. Treating these consumers as different classes is a historical implementation detail that should not leak to users. So, in addition to page cache, anonymous memory, and network socket memory, account the following memory consumers per default in the cgroup2 memory controller: - threadinfo - task_struct - task_delay_info - pid - cred - mm_struct - vm_area_struct and vm_region (nommu) - anon_vma and anon_vma_chain - signal_struct - sighand_struct - fs_struct - files_struct - fdtable and fdtable->full_fds_bits - dentry and external_name - inode for all filesystems. This should give us reasonable memory isolation for most common workloads out of the box. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Tejun Heo <tj@kernel.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:35 +00:00
if (ret)
memcg: unify slab and other kmem pages charging We have memcg_kmem_charge and memcg_kmem_uncharge methods for charging and uncharging kmem pages to memcg, but currently they are not used for charging slab pages (i.e. they are only used for charging pages allocated with alloc_kmem_pages). The only reason why the slab subsystem uses special helpers, memcg_charge_slab and memcg_uncharge_slab, is that it needs to charge to the memcg of kmem cache while memcg_charge_kmem charges to the memcg that the current task belongs to. To remove this diversity, this patch adds an extra argument to __memcg_kmem_charge that can be a pointer to a memcg or NULL. If it is not NULL, the function tries to charge to the memcg it points to, otherwise it charge to the current context. Next, it makes the slab subsystem use this function to charge slab pages. Since memcg_charge_kmem and memcg_uncharge_kmem helpers are now used only in __memcg_kmem_charge and __memcg_kmem_uncharge, they are inlined. Since __memcg_kmem_charge stores a pointer to the memcg in the page struct, we don't need memcg_uncharge_slab anymore and can use free_kmem_pages. Besides, one can now detect which memcg a slab page belongs to by reading /proc/kpagecgroup. Note, this patch switches slab to charge-after-alloc design. Since this design is already used for all other memcg charges, it should not make any difference. [hannes@cmpxchg.org: better to have an outer function than a magic parameter for the memcg lookup] Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:49:01 +00:00
return ret;
mm: memcontrol: account "kmem" consumers in cgroup2 memory controller The original cgroup memory controller has an extension to account slab memory (and other "kernel memory" consumers) in a separate "kmem" counter, once the user set an explicit limit on that "kmem" pool. However, this includes various consumers whose sizes are directly linked to userspace activity. Accounting them as an optional "kmem" extension is problematic for several reasons: 1. It leaves the main memory interface with incomplete semantics. A user who puts their workload into a cgroup and configures a memory limit does not expect us to leave holes in the containment as big as the dentry and inode cache, or the kernel stack pages. 2. If the limit set on this random historical subgroup of consumers is reached, subsequent allocations will fail even when the main memory pool available to the cgroup is not yet exhausted and/or has reclaimable memory in it. 3. Calling it 'kernel memory' is misleading. The dentry and inode caches are no more 'kernel' (or no less 'user') memory than the page cache itself. Treating these consumers as different classes is a historical implementation detail that should not leak to users. So, in addition to page cache, anonymous memory, and network socket memory, account the following memory consumers per default in the cgroup2 memory controller: - threadinfo - task_struct - task_delay_info - pid - cred - mm_struct - vm_area_struct and vm_region (nommu) - anon_vma and anon_vma_chain - signal_struct - sighand_struct - fs_struct - files_struct - fdtable and fdtable->full_fds_bits - dentry and external_name - inode for all filesystems. This should give us reasonable memory isolation for most common workloads out of the box. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Tejun Heo <tj@kernel.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:35 +00:00
if (!cgroup_subsys_on_dfl(memory_cgrp_subsys) &&
!page_counter_try_charge(&memcg->kmem, nr_pages, &counter)) {
memcg, kmem: do not fail __GFP_NOFAIL charges Thomas has noticed the following NULL ptr dereference when using cgroup v1 kmem limit: BUG: unable to handle kernel NULL pointer dereference at 0000000000000008 PGD 0 P4D 0 Oops: 0000 [#1] PREEMPT SMP PTI CPU: 3 PID: 16923 Comm: gtk-update-icon Not tainted 4.19.51 #42 Hardware name: Gigabyte Technology Co., Ltd. Z97X-Gaming G1/Z97X-Gaming G1, BIOS F9 07/31/2015 RIP: 0010:create_empty_buffers+0x24/0x100 Code: cd 0f 1f 44 00 00 0f 1f 44 00 00 41 54 49 89 d4 ba 01 00 00 00 55 53 48 89 fb e8 97 fe ff ff 48 89 c5 48 89 c2 eb 03 48 89 ca <48> 8b 4a 08 4c 09 22 48 85 c9 75 f1 48 89 6a 08 48 8b 43 18 48 8d RSP: 0018:ffff927ac1b37bf8 EFLAGS: 00010286 RAX: 0000000000000000 RBX: fffff2d4429fd740 RCX: 0000000100097149 RDX: 0000000000000000 RSI: 0000000000000082 RDI: ffff9075a99fbe00 RBP: 0000000000000000 R08: fffff2d440949cc8 R09: 00000000000960c0 R10: 0000000000000002 R11: 0000000000000000 R12: 0000000000000000 R13: ffff907601f18360 R14: 0000000000002000 R15: 0000000000001000 FS: 00007fb55b288bc0(0000) GS:ffff90761f8c0000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000008 CR3: 000000007aebc002 CR4: 00000000001606e0 Call Trace: create_page_buffers+0x4d/0x60 __block_write_begin_int+0x8e/0x5a0 ? ext4_inode_attach_jinode.part.82+0xb0/0xb0 ? jbd2__journal_start+0xd7/0x1f0 ext4_da_write_begin+0x112/0x3d0 generic_perform_write+0xf1/0x1b0 ? file_update_time+0x70/0x140 __generic_file_write_iter+0x141/0x1a0 ext4_file_write_iter+0xef/0x3b0 __vfs_write+0x17e/0x1e0 vfs_write+0xa5/0x1a0 ksys_write+0x57/0xd0 do_syscall_64+0x55/0x160 entry_SYSCALL_64_after_hwframe+0x44/0xa9 Tetsuo then noticed that this is because the __memcg_kmem_charge_memcg fails __GFP_NOFAIL charge when the kmem limit is reached. This is a wrong behavior because nofail allocations are not allowed to fail. Normal charge path simply forces the charge even if that means to cross the limit. Kmem accounting should be doing the same. Link: http://lkml.kernel.org/r/20190906125608.32129-1-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Reported-by: Thomas Lindroth <thomas.lindroth@gmail.com> Debugged-by: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Thomas Lindroth <thomas.lindroth@gmail.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-09-25 23:45:53 +00:00
/*
* Enforce __GFP_NOFAIL allocation because callers are not
* prepared to see failures and likely do not have any failure
* handling code.
*/
if (gfp & __GFP_NOFAIL) {
page_counter_charge(&memcg->kmem, nr_pages);
return 0;
}
mm: memcontrol: account "kmem" consumers in cgroup2 memory controller The original cgroup memory controller has an extension to account slab memory (and other "kernel memory" consumers) in a separate "kmem" counter, once the user set an explicit limit on that "kmem" pool. However, this includes various consumers whose sizes are directly linked to userspace activity. Accounting them as an optional "kmem" extension is problematic for several reasons: 1. It leaves the main memory interface with incomplete semantics. A user who puts their workload into a cgroup and configures a memory limit does not expect us to leave holes in the containment as big as the dentry and inode cache, or the kernel stack pages. 2. If the limit set on this random historical subgroup of consumers is reached, subsequent allocations will fail even when the main memory pool available to the cgroup is not yet exhausted and/or has reclaimable memory in it. 3. Calling it 'kernel memory' is misleading. The dentry and inode caches are no more 'kernel' (or no less 'user') memory than the page cache itself. Treating these consumers as different classes is a historical implementation detail that should not leak to users. So, in addition to page cache, anonymous memory, and network socket memory, account the following memory consumers per default in the cgroup2 memory controller: - threadinfo - task_struct - task_delay_info - pid - cred - mm_struct - vm_area_struct and vm_region (nommu) - anon_vma and anon_vma_chain - signal_struct - sighand_struct - fs_struct - files_struct - fdtable and fdtable->full_fds_bits - dentry and external_name - inode for all filesystems. This should give us reasonable memory isolation for most common workloads out of the box. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Tejun Heo <tj@kernel.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:35 +00:00
cancel_charge(memcg, nr_pages);
return -ENOMEM;
memcg: kmem controller infrastructure Introduce infrastructure for tracking kernel memory pages to a given memcg. This will happen whenever the caller includes the flag __GFP_KMEMCG flag, and the task belong to a memcg other than the root. In memcontrol.h those functions are wrapped in inline acessors. The idea is to later on, patch those with static branches, so we don't incur any overhead when no mem cgroups with limited kmem are being used. Users of this functionality shall interact with the memcg core code through the following functions: memcg_kmem_newpage_charge: will return true if the group can handle the allocation. At this point, struct page is not yet allocated. memcg_kmem_commit_charge: will either revert the charge, if struct page allocation failed, or embed memcg information into page_cgroup. memcg_kmem_uncharge_page: called at free time, will revert the charge. Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:56 +00:00
}
memcg: unify slab and other kmem pages charging We have memcg_kmem_charge and memcg_kmem_uncharge methods for charging and uncharging kmem pages to memcg, but currently they are not used for charging slab pages (i.e. they are only used for charging pages allocated with alloc_kmem_pages). The only reason why the slab subsystem uses special helpers, memcg_charge_slab and memcg_uncharge_slab, is that it needs to charge to the memcg of kmem cache while memcg_charge_kmem charges to the memcg that the current task belongs to. To remove this diversity, this patch adds an extra argument to __memcg_kmem_charge that can be a pointer to a memcg or NULL. If it is not NULL, the function tries to charge to the memcg it points to, otherwise it charge to the current context. Next, it makes the slab subsystem use this function to charge slab pages. Since memcg_charge_kmem and memcg_uncharge_kmem helpers are now used only in __memcg_kmem_charge and __memcg_kmem_uncharge, they are inlined. Since __memcg_kmem_charge stores a pointer to the memcg in the page struct, we don't need memcg_uncharge_slab anymore and can use free_kmem_pages. Besides, one can now detect which memcg a slab page belongs to by reading /proc/kpagecgroup. Note, this patch switches slab to charge-after-alloc design. Since this design is already used for all other memcg charges, it should not make any difference. [hannes@cmpxchg.org: better to have an outer function than a magic parameter for the memcg lookup] Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:49:01 +00:00
return 0;
memcg: kmem controller infrastructure Introduce infrastructure for tracking kernel memory pages to a given memcg. This will happen whenever the caller includes the flag __GFP_KMEMCG flag, and the task belong to a memcg other than the root. In memcontrol.h those functions are wrapped in inline acessors. The idea is to later on, patch those with static branches, so we don't incur any overhead when no mem cgroups with limited kmem are being used. Users of this functionality shall interact with the memcg core code through the following functions: memcg_kmem_newpage_charge: will return true if the group can handle the allocation. At this point, struct page is not yet allocated. memcg_kmem_commit_charge: will either revert the charge, if struct page allocation failed, or embed memcg information into page_cgroup. memcg_kmem_uncharge_page: called at free time, will revert the charge. Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:56 +00:00
}
/**
* __memcg_kmem_uncharge: uncharge a number of kernel pages from a memcg
* @memcg: memcg to uncharge
* @nr_pages: number of pages to uncharge
*/
void __memcg_kmem_uncharge(struct mem_cgroup *memcg, unsigned int nr_pages)
{
if (!cgroup_subsys_on_dfl(memory_cgrp_subsys))
page_counter_uncharge(&memcg->kmem, nr_pages);
page_counter_uncharge(&memcg->memory, nr_pages);
if (do_memsw_account())
page_counter_uncharge(&memcg->memsw, nr_pages);
}
/**
* __memcg_kmem_charge_page: charge a kmem page to the current memory cgroup
* @page: page to charge
* @gfp: reclaim mode
* @order: allocation order
*
* Returns 0 on success, an error code on failure.
*/
int __memcg_kmem_charge_page(struct page *page, gfp_t gfp, int order)
memcg: kmem controller infrastructure Introduce infrastructure for tracking kernel memory pages to a given memcg. This will happen whenever the caller includes the flag __GFP_KMEMCG flag, and the task belong to a memcg other than the root. In memcontrol.h those functions are wrapped in inline acessors. The idea is to later on, patch those with static branches, so we don't incur any overhead when no mem cgroups with limited kmem are being used. Users of this functionality shall interact with the memcg core code through the following functions: memcg_kmem_newpage_charge: will return true if the group can handle the allocation. At this point, struct page is not yet allocated. memcg_kmem_commit_charge: will either revert the charge, if struct page allocation failed, or embed memcg information into page_cgroup. memcg_kmem_uncharge_page: called at free time, will revert the charge. Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:56 +00:00
{
memcg: unify slab and other kmem pages charging We have memcg_kmem_charge and memcg_kmem_uncharge methods for charging and uncharging kmem pages to memcg, but currently they are not used for charging slab pages (i.e. they are only used for charging pages allocated with alloc_kmem_pages). The only reason why the slab subsystem uses special helpers, memcg_charge_slab and memcg_uncharge_slab, is that it needs to charge to the memcg of kmem cache while memcg_charge_kmem charges to the memcg that the current task belongs to. To remove this diversity, this patch adds an extra argument to __memcg_kmem_charge that can be a pointer to a memcg or NULL. If it is not NULL, the function tries to charge to the memcg it points to, otherwise it charge to the current context. Next, it makes the slab subsystem use this function to charge slab pages. Since memcg_charge_kmem and memcg_uncharge_kmem helpers are now used only in __memcg_kmem_charge and __memcg_kmem_uncharge, they are inlined. Since __memcg_kmem_charge stores a pointer to the memcg in the page struct, we don't need memcg_uncharge_slab anymore and can use free_kmem_pages. Besides, one can now detect which memcg a slab page belongs to by reading /proc/kpagecgroup. Note, this patch switches slab to charge-after-alloc design. Since this design is already used for all other memcg charges, it should not make any difference. [hannes@cmpxchg.org: better to have an outer function than a magic parameter for the memcg lookup] Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:49:01 +00:00
struct mem_cgroup *memcg;
2016-03-17 21:17:29 +00:00
int ret = 0;
memcg: kmem controller infrastructure Introduce infrastructure for tracking kernel memory pages to a given memcg. This will happen whenever the caller includes the flag __GFP_KMEMCG flag, and the task belong to a memcg other than the root. In memcontrol.h those functions are wrapped in inline acessors. The idea is to later on, patch those with static branches, so we don't incur any overhead when no mem cgroups with limited kmem are being used. Users of this functionality shall interact with the memcg core code through the following functions: memcg_kmem_newpage_charge: will return true if the group can handle the allocation. At this point, struct page is not yet allocated. memcg_kmem_commit_charge: will either revert the charge, if struct page allocation failed, or embed memcg information into page_cgroup. memcg_kmem_uncharge_page: called at free time, will revert the charge. Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:56 +00:00
if (memcg_kmem_bypass())
return 0;
fs: fsnotify: account fsnotify metadata to kmemcg Patch series "Directed kmem charging", v8. The Linux kernel's memory cgroup allows limiting the memory usage of the jobs running on the system to provide isolation between the jobs. All the kernel memory allocated in the context of the job and marked with __GFP_ACCOUNT will also be included in the memory usage and be limited by the job's limit. The kernel memory can only be charged to the memcg of the process in whose context kernel memory was allocated. However there are cases where the allocated kernel memory should be charged to the memcg different from the current processes's memcg. This patch series contains two such concrete use-cases i.e. fsnotify and buffer_head. The fsnotify event objects can consume a lot of system memory for large or unlimited queues if there is either no or slow listener. The events are allocated in the context of the event producer. However they should be charged to the event consumer. Similarly the buffer_head objects can be allocated in a memcg different from the memcg of the page for which buffer_head objects are being allocated. To solve this issue, this patch series introduces mechanism to charge kernel memory to a given memcg. In case of fsnotify events, the memcg of the consumer can be used for charging and for buffer_head, the memcg of the page can be charged. For directed charging, the caller can use the scope API memalloc_[un]use_memcg() to specify the memcg to charge for all the __GFP_ACCOUNT allocations within the scope. This patch (of 2): A lot of memory can be consumed by the events generated for the huge or unlimited queues if there is either no or slow listener. This can cause system level memory pressure or OOMs. So, it's better to account the fsnotify kmem caches to the memcg of the listener. However the listener can be in a different memcg than the memcg of the producer and these allocations happen in the context of the event producer. This patch introduces remote memcg charging API which the producer can use to charge the allocations to the memcg of the listener. There are seven fsnotify kmem caches and among them allocations from dnotify_struct_cache, dnotify_mark_cache, fanotify_mark_cache and inotify_inode_mark_cachep happens in the context of syscall from the listener. So, SLAB_ACCOUNT is enough for these caches. The objects from fsnotify_mark_connector_cachep are not accounted as they are small compared to the notification mark or events and it is unclear whom to account connector to since it is shared by all events attached to the inode. The allocations from the event caches happen in the context of the event producer. For such caches we will need to remote charge the allocations to the listener's memcg. Thus we save the memcg reference in the fsnotify_group structure of the listener. This patch has also moved the members of fsnotify_group to keep the size same, at least for 64 bit build, even with additional member by filling the holes. [shakeelb@google.com: use GFP_KERNEL_ACCOUNT rather than open-coding it] Link: http://lkml.kernel.org/r/20180702215439.211597-1-shakeelb@google.com Link: http://lkml.kernel.org/r/20180627191250.209150-2-shakeelb@google.com Signed-off-by: Shakeel Butt <shakeelb@google.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Amir Goldstein <amir73il@gmail.com> Cc: Greg Thelen <gthelen@google.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:46:39 +00:00
memcg = get_mem_cgroup_from_current();
mm: memcontrol: only mark charged pages with PageKmemcg To distinguish non-slab pages charged to kmemcg we mark them PageKmemcg, which sets page->_mapcount to -512. Currently, we set/clear PageKmemcg in __alloc_pages_nodemask()/free_pages_prepare() for any page allocated with __GFP_ACCOUNT, including those that aren't actually charged to any cgroup, i.e. allocated from the root cgroup context. To avoid overhead in case cgroups are not used, we only do that if memcg_kmem_enabled() is true. The latter is set iff there are kmem-enabled memory cgroups (online or offline). The root cgroup is not considered kmem-enabled. As a result, if a page is allocated with __GFP_ACCOUNT for the root cgroup when there are kmem-enabled memory cgroups and is freed after all kmem-enabled memory cgroups were removed, e.g. # no memory cgroups has been created yet, create one mkdir /sys/fs/cgroup/memory/test # run something allocating pages with __GFP_ACCOUNT, e.g. # a program using pipe dmesg | tail # remove the memory cgroup rmdir /sys/fs/cgroup/memory/test we'll get bad page state bug complaining about page->_mapcount != -1: BUG: Bad page state in process swapper/0 pfn:1fd945c page:ffffea007f651700 count:0 mapcount:-511 mapping: (null) index:0x0 flags: 0x1000000000000000() To avoid that, let's mark with PageKmemcg only those pages that are actually charged to and hence pin a non-root memory cgroup. Fixes: 4949148ad433 ("mm: charge/uncharge kmemcg from generic page allocator paths") Reported-and-tested-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-08-08 20:03:12 +00:00
if (!mem_cgroup_is_root(memcg)) {
ret = __memcg_kmem_charge(memcg, gfp, 1 << order);
mm: memcg/slab: stop setting page->mem_cgroup pointer for slab pages Every slab page charged to a non-root memory cgroup has a pointer to the memory cgroup and holds a reference to it, which protects a non-empty memory cgroup from being released. At the same time the page has a pointer to the corresponding kmem_cache, and also hold a reference to the kmem_cache. And kmem_cache by itself holds a reference to the cgroup. So there is clearly some redundancy, which allows to stop setting the page->mem_cgroup pointer and rely on getting memcg pointer indirectly via kmem_cache. Further it will allow to change this pointer easier, without a need to go over all charged pages. So let's stop setting page->mem_cgroup pointer for slab pages, and stop using the css refcounter directly for protecting the memory cgroup from going away. Instead rely on kmem_cache as an intermediate object. Make sure that vmstats and shrinker lists are working as previously, as well as /proc/kpagecgroup interface. Link: http://lkml.kernel.org/r/20190611231813.3148843-10-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Reviewed-by: Shakeel Butt <shakeelb@google.com> Cc: Christoph Lameter <cl@linux.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Waiman Long <longman@redhat.com> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: Andrei Vagin <avagin@gmail.com> Cc: Qian Cai <cai@lca.pw> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:56:31 +00:00
if (!ret) {
page->mem_cgroup = memcg;
mm: memcontrol: only mark charged pages with PageKmemcg To distinguish non-slab pages charged to kmemcg we mark them PageKmemcg, which sets page->_mapcount to -512. Currently, we set/clear PageKmemcg in __alloc_pages_nodemask()/free_pages_prepare() for any page allocated with __GFP_ACCOUNT, including those that aren't actually charged to any cgroup, i.e. allocated from the root cgroup context. To avoid overhead in case cgroups are not used, we only do that if memcg_kmem_enabled() is true. The latter is set iff there are kmem-enabled memory cgroups (online or offline). The root cgroup is not considered kmem-enabled. As a result, if a page is allocated with __GFP_ACCOUNT for the root cgroup when there are kmem-enabled memory cgroups and is freed after all kmem-enabled memory cgroups were removed, e.g. # no memory cgroups has been created yet, create one mkdir /sys/fs/cgroup/memory/test # run something allocating pages with __GFP_ACCOUNT, e.g. # a program using pipe dmesg | tail # remove the memory cgroup rmdir /sys/fs/cgroup/memory/test we'll get bad page state bug complaining about page->_mapcount != -1: BUG: Bad page state in process swapper/0 pfn:1fd945c page:ffffea007f651700 count:0 mapcount:-511 mapping: (null) index:0x0 flags: 0x1000000000000000() To avoid that, let's mark with PageKmemcg only those pages that are actually charged to and hence pin a non-root memory cgroup. Fixes: 4949148ad433 ("mm: charge/uncharge kmemcg from generic page allocator paths") Reported-and-tested-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-08-08 20:03:12 +00:00
__SetPageKmemcg(page);
mm: memcg/slab: stop setting page->mem_cgroup pointer for slab pages Every slab page charged to a non-root memory cgroup has a pointer to the memory cgroup and holds a reference to it, which protects a non-empty memory cgroup from being released. At the same time the page has a pointer to the corresponding kmem_cache, and also hold a reference to the kmem_cache. And kmem_cache by itself holds a reference to the cgroup. So there is clearly some redundancy, which allows to stop setting the page->mem_cgroup pointer and rely on getting memcg pointer indirectly via kmem_cache. Further it will allow to change this pointer easier, without a need to go over all charged pages. So let's stop setting page->mem_cgroup pointer for slab pages, and stop using the css refcounter directly for protecting the memory cgroup from going away. Instead rely on kmem_cache as an intermediate object. Make sure that vmstats and shrinker lists are working as previously, as well as /proc/kpagecgroup interface. Link: http://lkml.kernel.org/r/20190611231813.3148843-10-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Reviewed-by: Shakeel Butt <shakeelb@google.com> Cc: Christoph Lameter <cl@linux.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Waiman Long <longman@redhat.com> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: Andrei Vagin <avagin@gmail.com> Cc: Qian Cai <cai@lca.pw> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:56:31 +00:00
}
mm: memcontrol: only mark charged pages with PageKmemcg To distinguish non-slab pages charged to kmemcg we mark them PageKmemcg, which sets page->_mapcount to -512. Currently, we set/clear PageKmemcg in __alloc_pages_nodemask()/free_pages_prepare() for any page allocated with __GFP_ACCOUNT, including those that aren't actually charged to any cgroup, i.e. allocated from the root cgroup context. To avoid overhead in case cgroups are not used, we only do that if memcg_kmem_enabled() is true. The latter is set iff there are kmem-enabled memory cgroups (online or offline). The root cgroup is not considered kmem-enabled. As a result, if a page is allocated with __GFP_ACCOUNT for the root cgroup when there are kmem-enabled memory cgroups and is freed after all kmem-enabled memory cgroups were removed, e.g. # no memory cgroups has been created yet, create one mkdir /sys/fs/cgroup/memory/test # run something allocating pages with __GFP_ACCOUNT, e.g. # a program using pipe dmesg | tail # remove the memory cgroup rmdir /sys/fs/cgroup/memory/test we'll get bad page state bug complaining about page->_mapcount != -1: BUG: Bad page state in process swapper/0 pfn:1fd945c page:ffffea007f651700 count:0 mapcount:-511 mapping: (null) index:0x0 flags: 0x1000000000000000() To avoid that, let's mark with PageKmemcg only those pages that are actually charged to and hence pin a non-root memory cgroup. Fixes: 4949148ad433 ("mm: charge/uncharge kmemcg from generic page allocator paths") Reported-and-tested-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-08-08 20:03:12 +00:00
}
memcg: kmem controller infrastructure Introduce infrastructure for tracking kernel memory pages to a given memcg. This will happen whenever the caller includes the flag __GFP_KMEMCG flag, and the task belong to a memcg other than the root. In memcontrol.h those functions are wrapped in inline acessors. The idea is to later on, patch those with static branches, so we don't incur any overhead when no mem cgroups with limited kmem are being used. Users of this functionality shall interact with the memcg core code through the following functions: memcg_kmem_newpage_charge: will return true if the group can handle the allocation. At this point, struct page is not yet allocated. memcg_kmem_commit_charge: will either revert the charge, if struct page allocation failed, or embed memcg information into page_cgroup. memcg_kmem_uncharge_page: called at free time, will revert the charge. Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:56 +00:00
css_put(&memcg->css);
return ret;
memcg: kmem controller infrastructure Introduce infrastructure for tracking kernel memory pages to a given memcg. This will happen whenever the caller includes the flag __GFP_KMEMCG flag, and the task belong to a memcg other than the root. In memcontrol.h those functions are wrapped in inline acessors. The idea is to later on, patch those with static branches, so we don't incur any overhead when no mem cgroups with limited kmem are being used. Users of this functionality shall interact with the memcg core code through the following functions: memcg_kmem_newpage_charge: will return true if the group can handle the allocation. At this point, struct page is not yet allocated. memcg_kmem_commit_charge: will either revert the charge, if struct page allocation failed, or embed memcg information into page_cgroup. memcg_kmem_uncharge_page: called at free time, will revert the charge. Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:56 +00:00
}
/**
* __memcg_kmem_uncharge_page: uncharge a kmem page
* @page: page to uncharge
* @order: allocation order
*/
void __memcg_kmem_uncharge_page(struct page *page, int order)
memcg: kmem controller infrastructure Introduce infrastructure for tracking kernel memory pages to a given memcg. This will happen whenever the caller includes the flag __GFP_KMEMCG flag, and the task belong to a memcg other than the root. In memcontrol.h those functions are wrapped in inline acessors. The idea is to later on, patch those with static branches, so we don't incur any overhead when no mem cgroups with limited kmem are being used. Users of this functionality shall interact with the memcg core code through the following functions: memcg_kmem_newpage_charge: will return true if the group can handle the allocation. At this point, struct page is not yet allocated. memcg_kmem_commit_charge: will either revert the charge, if struct page allocation failed, or embed memcg information into page_cgroup. memcg_kmem_uncharge_page: called at free time, will revert the charge. Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:56 +00:00
{
mm: embed the memcg pointer directly into struct page Memory cgroups used to have 5 per-page pointers. To allow users to disable that amount of overhead during runtime, those pointers were allocated in a separate array, with a translation layer between them and struct page. There is now only one page pointer remaining: the memcg pointer, that indicates which cgroup the page is associated with when charged. The complexity of runtime allocation and the runtime translation overhead is no longer justified to save that *potential* 0.19% of memory. With CONFIG_SLUB, page->mem_cgroup actually sits in the doubleword padding after the page->private member and doesn't even increase struct page, and then this patch actually saves space. Remaining users that care can still compile their kernels without CONFIG_MEMCG. text data bss dec hex filename 8828345 1725264 983040 11536649 b00909 vmlinux.old 8827425 1725264 966656 11519345 afc571 vmlinux.new [mhocko@suse.cz: update Documentation/cgroups/memory.txt] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: David S. Miller <davem@davemloft.net> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Acked-by: Konstantin Khlebnikov <koct9i@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:44:52 +00:00
struct mem_cgroup *memcg = page->mem_cgroup;
memcg: unify slab and other kmem pages charging We have memcg_kmem_charge and memcg_kmem_uncharge methods for charging and uncharging kmem pages to memcg, but currently they are not used for charging slab pages (i.e. they are only used for charging pages allocated with alloc_kmem_pages). The only reason why the slab subsystem uses special helpers, memcg_charge_slab and memcg_uncharge_slab, is that it needs to charge to the memcg of kmem cache while memcg_charge_kmem charges to the memcg that the current task belongs to. To remove this diversity, this patch adds an extra argument to __memcg_kmem_charge that can be a pointer to a memcg or NULL. If it is not NULL, the function tries to charge to the memcg it points to, otherwise it charge to the current context. Next, it makes the slab subsystem use this function to charge slab pages. Since memcg_charge_kmem and memcg_uncharge_kmem helpers are now used only in __memcg_kmem_charge and __memcg_kmem_uncharge, they are inlined. Since __memcg_kmem_charge stores a pointer to the memcg in the page struct, we don't need memcg_uncharge_slab anymore and can use free_kmem_pages. Besides, one can now detect which memcg a slab page belongs to by reading /proc/kpagecgroup. Note, this patch switches slab to charge-after-alloc design. Since this design is already used for all other memcg charges, it should not make any difference. [hannes@cmpxchg.org: better to have an outer function than a magic parameter for the memcg lookup] Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:49:01 +00:00
unsigned int nr_pages = 1 << order;
memcg: kmem controller infrastructure Introduce infrastructure for tracking kernel memory pages to a given memcg. This will happen whenever the caller includes the flag __GFP_KMEMCG flag, and the task belong to a memcg other than the root. In memcontrol.h those functions are wrapped in inline acessors. The idea is to later on, patch those with static branches, so we don't incur any overhead when no mem cgroups with limited kmem are being used. Users of this functionality shall interact with the memcg core code through the following functions: memcg_kmem_newpage_charge: will return true if the group can handle the allocation. At this point, struct page is not yet allocated. memcg_kmem_commit_charge: will either revert the charge, if struct page allocation failed, or embed memcg information into page_cgroup. memcg_kmem_uncharge_page: called at free time, will revert the charge. Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:56 +00:00
if (!memcg)
return;
VM_BUG_ON_PAGE(mem_cgroup_is_root(memcg), page);
__memcg_kmem_uncharge(memcg, nr_pages);
mm: embed the memcg pointer directly into struct page Memory cgroups used to have 5 per-page pointers. To allow users to disable that amount of overhead during runtime, those pointers were allocated in a separate array, with a translation layer between them and struct page. There is now only one page pointer remaining: the memcg pointer, that indicates which cgroup the page is associated with when charged. The complexity of runtime allocation and the runtime translation overhead is no longer justified to save that *potential* 0.19% of memory. With CONFIG_SLUB, page->mem_cgroup actually sits in the doubleword padding after the page->private member and doesn't even increase struct page, and then this patch actually saves space. Remaining users that care can still compile their kernels without CONFIG_MEMCG. text data bss dec hex filename 8828345 1725264 983040 11536649 b00909 vmlinux.old 8827425 1725264 966656 11519345 afc571 vmlinux.new [mhocko@suse.cz: update Documentation/cgroups/memory.txt] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: David S. Miller <davem@davemloft.net> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Acked-by: Konstantin Khlebnikov <koct9i@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:44:52 +00:00
page->mem_cgroup = NULL;
mm: memcontrol: only mark charged pages with PageKmemcg To distinguish non-slab pages charged to kmemcg we mark them PageKmemcg, which sets page->_mapcount to -512. Currently, we set/clear PageKmemcg in __alloc_pages_nodemask()/free_pages_prepare() for any page allocated with __GFP_ACCOUNT, including those that aren't actually charged to any cgroup, i.e. allocated from the root cgroup context. To avoid overhead in case cgroups are not used, we only do that if memcg_kmem_enabled() is true. The latter is set iff there are kmem-enabled memory cgroups (online or offline). The root cgroup is not considered kmem-enabled. As a result, if a page is allocated with __GFP_ACCOUNT for the root cgroup when there are kmem-enabled memory cgroups and is freed after all kmem-enabled memory cgroups were removed, e.g. # no memory cgroups has been created yet, create one mkdir /sys/fs/cgroup/memory/test # run something allocating pages with __GFP_ACCOUNT, e.g. # a program using pipe dmesg | tail # remove the memory cgroup rmdir /sys/fs/cgroup/memory/test we'll get bad page state bug complaining about page->_mapcount != -1: BUG: Bad page state in process swapper/0 pfn:1fd945c page:ffffea007f651700 count:0 mapcount:-511 mapping: (null) index:0x0 flags: 0x1000000000000000() To avoid that, let's mark with PageKmemcg only those pages that are actually charged to and hence pin a non-root memory cgroup. Fixes: 4949148ad433 ("mm: charge/uncharge kmemcg from generic page allocator paths") Reported-and-tested-by: Eric Dumazet <eric.dumazet@gmail.com> Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-08-08 20:03:12 +00:00
/* slab pages do not have PageKmemcg flag set */
if (PageKmemcg(page))
__ClearPageKmemcg(page);
memcg: unify slab and other kmem pages charging We have memcg_kmem_charge and memcg_kmem_uncharge methods for charging and uncharging kmem pages to memcg, but currently they are not used for charging slab pages (i.e. they are only used for charging pages allocated with alloc_kmem_pages). The only reason why the slab subsystem uses special helpers, memcg_charge_slab and memcg_uncharge_slab, is that it needs to charge to the memcg of kmem cache while memcg_charge_kmem charges to the memcg that the current task belongs to. To remove this diversity, this patch adds an extra argument to __memcg_kmem_charge that can be a pointer to a memcg or NULL. If it is not NULL, the function tries to charge to the memcg it points to, otherwise it charge to the current context. Next, it makes the slab subsystem use this function to charge slab pages. Since memcg_charge_kmem and memcg_uncharge_kmem helpers are now used only in __memcg_kmem_charge and __memcg_kmem_uncharge, they are inlined. Since __memcg_kmem_charge stores a pointer to the memcg in the page struct, we don't need memcg_uncharge_slab anymore and can use free_kmem_pages. Besides, one can now detect which memcg a slab page belongs to by reading /proc/kpagecgroup. Note, this patch switches slab to charge-after-alloc design. Since this design is already used for all other memcg charges, it should not make any difference. [hannes@cmpxchg.org: better to have an outer function than a magic parameter for the memcg lookup] Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:49:01 +00:00
css_put_many(&memcg->css, nr_pages);
list_lru: introduce per-memcg lists There are several FS shrinkers, including super_block::s_shrink, that keep reclaimable objects in the list_lru structure. Hence to turn them to memcg-aware shrinkers, it is enough to make list_lru per-memcg. This patch does the trick. It adds an array of lru lists to the list_lru_node structure (per-node part of the list_lru), one for each kmem-active memcg, and dispatches every item addition or removal to the list corresponding to the memcg which the item is accounted to. So now the list_lru structure is not just per node, but per node and per memcg. Not all list_lrus need this feature, so this patch also adds a new method, list_lru_init_memcg, which initializes a list_lru as memcg aware. Otherwise (i.e. if initialized with old list_lru_init), the list_lru won't have per memcg lists. Just like per memcg caches arrays, the arrays of per-memcg lists are indexed by memcg_cache_id, so we must grow them whenever memcg_nr_cache_ids is increased. So we introduce a callback, memcg_update_all_list_lrus, invoked by memcg_alloc_cache_id if the id space is full. The locking is implemented in a manner similar to lruvecs, i.e. we have one lock per node that protects all lists (both global and per cgroup) on the node. Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Dave Chinner <david@fromorbit.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Greg Thelen <gthelen@google.com> Cc: Glauber Costa <glommer@gmail.com> Cc: Alexander Viro <viro@zeniv.linux.org.uk> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-12 22:59:10 +00:00
}
mm: introduce CONFIG_MEMCG_KMEM as combination of CONFIG_MEMCG && !CONFIG_SLOB Introduce new config option, which is used to replace repeating CONFIG_MEMCG && !CONFIG_SLOB pattern. Next patches add a little more memcg+kmem related code, so let's keep the defines more clearly. Link: http://lkml.kernel.org/r/153063053670.1818.15013136946600481138.stgit@localhost.localdomain Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Tested-by: Shakeel Butt <shakeelb@google.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Guenter Roeck <linux@roeck-us.net> Cc: "Huang, Ying" <ying.huang@intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Josef Bacik <jbacik@fb.com> Cc: Li RongQing <lirongqing@baidu.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Matthias Kaehlcke <mka@chromium.org> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Michal Hocko <mhocko@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Philippe Ombredanne <pombredanne@nexb.com> Cc: Roman Gushchin <guro@fb.com> Cc: Sahitya Tummala <stummala@codeaurora.org> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Waiman Long <longman@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:25 +00:00
#endif /* CONFIG_MEMCG_KMEM */
memcg: kmem controller infrastructure Introduce infrastructure for tracking kernel memory pages to a given memcg. This will happen whenever the caller includes the flag __GFP_KMEMCG flag, and the task belong to a memcg other than the root. In memcontrol.h those functions are wrapped in inline acessors. The idea is to later on, patch those with static branches, so we don't incur any overhead when no mem cgroups with limited kmem are being used. Users of this functionality shall interact with the memcg core code through the following functions: memcg_kmem_newpage_charge: will return true if the group can handle the allocation. At this point, struct page is not yet allocated. memcg_kmem_commit_charge: will either revert the charge, if struct page allocation failed, or embed memcg information into page_cgroup. memcg_kmem_uncharge_page: called at free time, will revert the charge. Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:56 +00:00
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
/*
* Because tail pages are not marked as "used", set it. We're under
* pgdat->lru_lock and migration entries setup in all page mappings.
*/
void mem_cgroup_split_huge_fixup(struct page *head)
{
int i;
if (mem_cgroup_disabled())
return;
for (i = 1; i < HPAGE_PMD_NR; i++)
mm: embed the memcg pointer directly into struct page Memory cgroups used to have 5 per-page pointers. To allow users to disable that amount of overhead during runtime, those pointers were allocated in a separate array, with a translation layer between them and struct page. There is now only one page pointer remaining: the memcg pointer, that indicates which cgroup the page is associated with when charged. The complexity of runtime allocation and the runtime translation overhead is no longer justified to save that *potential* 0.19% of memory. With CONFIG_SLUB, page->mem_cgroup actually sits in the doubleword padding after the page->private member and doesn't even increase struct page, and then this patch actually saves space. Remaining users that care can still compile their kernels without CONFIG_MEMCG. text data bss dec hex filename 8828345 1725264 983040 11536649 b00909 vmlinux.old 8827425 1725264 966656 11519345 afc571 vmlinux.new [mhocko@suse.cz: update Documentation/cgroups/memory.txt] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: David S. Miller <davem@davemloft.net> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Acked-by: Konstantin Khlebnikov <koct9i@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:44:52 +00:00
head[i].mem_cgroup = head->mem_cgroup;
}
#endif /* CONFIG_TRANSPARENT_HUGEPAGE */
#ifdef CONFIG_MEMCG_SWAP
/**
* mem_cgroup_move_swap_account - move swap charge and swap_cgroup's record.
* @entry: swap entry to be moved
* @from: mem_cgroup which the entry is moved from
* @to: mem_cgroup which the entry is moved to
*
* It succeeds only when the swap_cgroup's record for this entry is the same
* as the mem_cgroup's id of @from.
*
* Returns 0 on success, -EINVAL on failure.
*
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
* The caller must have charged to @to, IOW, called page_counter_charge() about
* both res and memsw, and called css_get().
*/
static int mem_cgroup_move_swap_account(swp_entry_t entry,
struct mem_cgroup *from, struct mem_cgroup *to)
{
unsigned short old_id, new_id;
old_id = mem_cgroup_id(from);
new_id = mem_cgroup_id(to);
if (swap_cgroup_cmpxchg(entry, old_id, new_id) == old_id) {
mod_memcg_state(from, MEMCG_SWAP, -1);
mod_memcg_state(to, MEMCG_SWAP, 1);
return 0;
}
return -EINVAL;
}
#else
static inline int mem_cgroup_move_swap_account(swp_entry_t entry,
struct mem_cgroup *from, struct mem_cgroup *to)
{
return -EINVAL;
}
memcg: mem+swap controller core This patch implements per cgroup limit for usage of memory+swap. However there are SwapCache, double counting of swap-cache and swap-entry is avoided. Mem+Swap controller works as following. - memory usage is limited by memory.limit_in_bytes. - memory + swap usage is limited by memory.memsw_limit_in_bytes. This has following benefits. - A user can limit total resource usage of mem+swap. Without this, because memory resource controller doesn't take care of usage of swap, a process can exhaust all the swap (by memory leak.) We can avoid this case. And Swap is shared resource but it cannot be reclaimed (goes back to memory) until it's used. This characteristic can be trouble when the memory is divided into some parts by cpuset or memcg. Assume group A and group B. After some application executes, the system can be.. Group A -- very large free memory space but occupy 99% of swap. Group B -- under memory shortage but cannot use swap...it's nearly full. Ability to set appropriate swap limit for each group is required. Maybe someone wonder "why not swap but mem+swap ?" - The global LRU(kswapd) can swap out arbitrary pages. Swap-out means to move account from memory to swap...there is no change in usage of mem+swap. In other words, when we want to limit the usage of swap without affecting global LRU, mem+swap limit is better than just limiting swap. Accounting target information is stored in swap_cgroup which is per swap entry record. Charge is done as following. map - charge page and memsw. unmap - uncharge page/memsw if not SwapCache. swap-out (__delete_from_swap_cache) - uncharge page - record mem_cgroup information to swap_cgroup. swap-in (do_swap_page) - charged as page and memsw. record in swap_cgroup is cleared. memsw accounting is decremented. swap-free (swap_free()) - if swap entry is freed, memsw is uncharged by PAGE_SIZE. There are people work under never-swap environments and consider swap as something bad. For such people, this mem+swap controller extension is just an overhead. This overhead is avoided by config or boot option. (see Kconfig. detail is not in this patch.) TODO: - maybe more optimization can be don in swap-in path. (but not very safe.) But we just do simple accounting at this stage. [nishimura@mxp.nes.nec.co.jp: make resize limit hold mutex] [hugh@veritas.com: memswap controller core swapcache fixes] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:08:00 +00:00
#endif
memcg: handle swap caches SwapCache support for memory resource controller (memcg) Before mem+swap controller, memcg itself should handle SwapCache in proper way. This is cut-out from it. In current memcg, SwapCache is just leaked and the user can create tons of SwapCache. This is a leak of account and should be handled. SwapCache accounting is done as following. charge (anon) - charged when it's mapped. (because of readahead, charge at add_to_swap_cache() is not sane) uncharge (anon) - uncharged when it's dropped from swapcache and fully unmapped. means it's not uncharged at unmap. Note: delete from swap cache at swap-in is done after rmap information is established. charge (shmem) - charged at swap-in. this prevents charge at add_to_page_cache(). uncharge (shmem) - uncharged when it's dropped from swapcache and not on shmem's radix-tree. at migration, check against 'old page' is modified to handle shmem. Comparing to the old version discussed (and caused troubles), we have advantages of - PCG_USED bit. - simple migrating handling. So, situation is much easier than several months ago, maybe. [hugh@veritas.com: memcg: handle swap caches build fix] Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Tested-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Hugh Dickins <hugh@veritas.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:07:56 +00:00
static DEFINE_MUTEX(memcg_max_mutex);
static int mem_cgroup_resize_max(struct mem_cgroup *memcg,
unsigned long max, bool memsw)
{
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
bool enlarge = false;
bool drained = false;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
int ret;
bool limits_invariant;
struct page_counter *counter = memsw ? &memcg->memsw : &memcg->memory;
memcg: fix shrinking memory to return -EBUSY by fixing retry algorithm As pointed out, shrinking memcg's limit should return -EBUSY after reasonable retries. This patch tries to fix the current behavior of shrink_usage. Before looking into "shrink should return -EBUSY" problem, we should fix hierarchical reclaim code. It compares current usage and current limit, but it only makes sense when the kernel reclaims memory because hit limits. This is also a problem. What this patch does are. 1. add new argument "shrink" to hierarchical reclaim. If "shrink==true", hierarchical reclaim returns immediately and the caller checks the kernel should shrink more or not. (At shrinking memory, usage is always smaller than limit. So check for usage < limit is useless.) 2. For adjusting to above change, 2 changes in "shrink"'s retry path. 2-a. retry_count depends on # of children because the kernel visits the children under hierarchy one by one. 2-b. rather than checking return value of hierarchical_reclaim's progress, compares usage-before-shrink and usage-after-shrink. If usage-before-shrink <= usage-after-shrink, retry_count is decremented. Reported-by: Li Zefan <lizf@cn.fujitsu.com> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Paul Menage <menage@google.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-04-02 23:57:36 +00:00
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
do {
if (signal_pending(current)) {
ret = -EINTR;
break;
}
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
mutex_lock(&memcg_max_mutex);
/*
* Make sure that the new limit (memsw or memory limit) doesn't
* break our basic invariant rule memory.max <= memsw.max.
*/
limits_invariant = memsw ? max >= READ_ONCE(memcg->memory.max) :
max <= memcg->memsw.max;
if (!limits_invariant) {
mutex_unlock(&memcg_max_mutex);
memcg: mem+swap controller core This patch implements per cgroup limit for usage of memory+swap. However there are SwapCache, double counting of swap-cache and swap-entry is avoided. Mem+Swap controller works as following. - memory usage is limited by memory.limit_in_bytes. - memory + swap usage is limited by memory.memsw_limit_in_bytes. This has following benefits. - A user can limit total resource usage of mem+swap. Without this, because memory resource controller doesn't take care of usage of swap, a process can exhaust all the swap (by memory leak.) We can avoid this case. And Swap is shared resource but it cannot be reclaimed (goes back to memory) until it's used. This characteristic can be trouble when the memory is divided into some parts by cpuset or memcg. Assume group A and group B. After some application executes, the system can be.. Group A -- very large free memory space but occupy 99% of swap. Group B -- under memory shortage but cannot use swap...it's nearly full. Ability to set appropriate swap limit for each group is required. Maybe someone wonder "why not swap but mem+swap ?" - The global LRU(kswapd) can swap out arbitrary pages. Swap-out means to move account from memory to swap...there is no change in usage of mem+swap. In other words, when we want to limit the usage of swap without affecting global LRU, mem+swap limit is better than just limiting swap. Accounting target information is stored in swap_cgroup which is per swap entry record. Charge is done as following. map - charge page and memsw. unmap - uncharge page/memsw if not SwapCache. swap-out (__delete_from_swap_cache) - uncharge page - record mem_cgroup information to swap_cgroup. swap-in (do_swap_page) - charged as page and memsw. record in swap_cgroup is cleared. memsw accounting is decremented. swap-free (swap_free()) - if swap entry is freed, memsw is uncharged by PAGE_SIZE. There are people work under never-swap environments and consider swap as something bad. For such people, this mem+swap controller extension is just an overhead. This overhead is avoided by config or boot option. (see Kconfig. detail is not in this patch.) TODO: - maybe more optimization can be don in swap-in path. (but not very safe.) But we just do simple accounting at this stage. [nishimura@mxp.nes.nec.co.jp: make resize limit hold mutex] [hugh@veritas.com: memswap controller core swapcache fixes] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:08:00 +00:00
ret = -EINVAL;
break;
}
if (max > counter->max)
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
enlarge = true;
ret = page_counter_set_max(counter, max);
mutex_unlock(&memcg_max_mutex);
memcg: mem+swap controller core This patch implements per cgroup limit for usage of memory+swap. However there are SwapCache, double counting of swap-cache and swap-entry is avoided. Mem+Swap controller works as following. - memory usage is limited by memory.limit_in_bytes. - memory + swap usage is limited by memory.memsw_limit_in_bytes. This has following benefits. - A user can limit total resource usage of mem+swap. Without this, because memory resource controller doesn't take care of usage of swap, a process can exhaust all the swap (by memory leak.) We can avoid this case. And Swap is shared resource but it cannot be reclaimed (goes back to memory) until it's used. This characteristic can be trouble when the memory is divided into some parts by cpuset or memcg. Assume group A and group B. After some application executes, the system can be.. Group A -- very large free memory space but occupy 99% of swap. Group B -- under memory shortage but cannot use swap...it's nearly full. Ability to set appropriate swap limit for each group is required. Maybe someone wonder "why not swap but mem+swap ?" - The global LRU(kswapd) can swap out arbitrary pages. Swap-out means to move account from memory to swap...there is no change in usage of mem+swap. In other words, when we want to limit the usage of swap without affecting global LRU, mem+swap limit is better than just limiting swap. Accounting target information is stored in swap_cgroup which is per swap entry record. Charge is done as following. map - charge page and memsw. unmap - uncharge page/memsw if not SwapCache. swap-out (__delete_from_swap_cache) - uncharge page - record mem_cgroup information to swap_cgroup. swap-in (do_swap_page) - charged as page and memsw. record in swap_cgroup is cleared. memsw accounting is decremented. swap-free (swap_free()) - if swap entry is freed, memsw is uncharged by PAGE_SIZE. There are people work under never-swap environments and consider swap as something bad. For such people, this mem+swap controller extension is just an overhead. This overhead is avoided by config or boot option. (see Kconfig. detail is not in this patch.) TODO: - maybe more optimization can be don in swap-in path. (but not very safe.) But we just do simple accounting at this stage. [nishimura@mxp.nes.nec.co.jp: make resize limit hold mutex] [hugh@veritas.com: memswap controller core swapcache fixes] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:08:00 +00:00
if (!ret)
break;
if (!drained) {
drain_all_stock(memcg);
drained = true;
continue;
}
if (!try_to_free_mem_cgroup_pages(memcg, 1,
GFP_KERNEL, !memsw)) {
ret = -EBUSY;
break;
}
} while (true);
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
if (!ret && enlarge)
memcg_oom_recover(memcg);
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
return ret;
}
unsigned long mem_cgroup_soft_limit_reclaim(pg_data_t *pgdat, int order,
gfp_t gfp_mask,
unsigned long *total_scanned)
{
unsigned long nr_reclaimed = 0;
struct mem_cgroup_per_node *mz, *next_mz = NULL;
unsigned long reclaimed;
int loop = 0;
struct mem_cgroup_tree_per_node *mctz;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
unsigned long excess;
unsigned long nr_scanned;
if (order > 0)
return 0;
mctz = soft_limit_tree_node(pgdat->node_id);
memcg: put soft limit reclaim out of way if the excess tree is empty We've had a report about soft lockups caused by lock bouncing in the soft reclaim path: BUG: soft lockup - CPU#0 stuck for 22s! [kav4proxy-kavic:3128] RIP: 0010:[<ffffffff81469798>] [<ffffffff81469798>] _raw_spin_lock+0x18/0x20 Call Trace: mem_cgroup_soft_limit_reclaim+0x25a/0x280 shrink_zones+0xed/0x200 do_try_to_free_pages+0x74/0x320 try_to_free_pages+0x112/0x180 __alloc_pages_slowpath+0x3ff/0x820 __alloc_pages_nodemask+0x1e9/0x200 alloc_pages_vma+0xe1/0x290 do_wp_page+0x19f/0x840 handle_pte_fault+0x1cd/0x230 do_page_fault+0x1fd/0x4c0 page_fault+0x25/0x30 There are no memcgs created so there cannot be any in the soft limit excess obviously: [...] memory 0 1 1 so all this just seems to be mem_cgroup_largest_soft_limit_node trying to get spin_lock_irq(&mctz->lock) just to find out that the soft limit excess tree is empty. This is just pointless wasting of cycles and cache line bouncing during heavy parallel reclaim on large machines. The particular machine wasn't very healthy and most probably suffering from a memory leak which just caused the memory reclaim to trash heavily. But bouncing on the lock certainly didn't help... Fix this by optimistic lockless check and bail out early if the tree is empty. This is theoretically racy but that shouldn't matter all that much. First of all soft limit is a best effort feature and it is slowly getting deprecated and its usage should be really scarce. Bouncing on a lock without a good reason is surely much bigger problem, especially on large CPU machines. Link: http://lkml.kernel.org/r/1470073277-1056-1-git-send-email-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-08-02 21:02:37 +00:00
/*
* Do not even bother to check the largest node if the root
* is empty. Do it lockless to prevent lock bouncing. Races
* are acceptable as soft limit is best effort anyway.
*/
mm/cgroup: avoid panic when init with low memory The system may panic when initialisation is done when almost all the memory is assigned to the huge pages using the kernel command line parameter hugepage=xxxx. Panic may occur like this: Unable to handle kernel paging request for data at address 0x00000000 Faulting instruction address: 0xc000000000302b88 Oops: Kernel access of bad area, sig: 11 [#1] SMP NR_CPUS=2048 [ 0.082424] NUMA pSeries Modules linked in: CPU: 0 PID: 1 Comm: swapper/0 Not tainted 4.9.0-15-generic #16-Ubuntu task: c00000021ed01600 task.stack: c00000010d108000 NIP: c000000000302b88 LR: c000000000270e04 CTR: c00000000016cfd0 REGS: c00000010d10b2c0 TRAP: 0300 Not tainted (4.9.0-15-generic) MSR: 8000000002009033 <SF,VEC,EE,ME,IR,DR,RI,LE>[ 0.082770] CR: 28424422 XER: 00000000 CFAR: c0000000003d28b8 DAR: 0000000000000000 DSISR: 40000000 SOFTE: 1 GPR00: c000000000270e04 c00000010d10b540 c00000000141a300 c00000010fff6300 GPR04: 0000000000000000 00000000026012c0 c00000010d10b630 0000000487ab0000 GPR08: 000000010ee90000 c000000001454fd8 0000000000000000 0000000000000000 GPR12: 0000000000004400 c00000000fb80000 00000000026012c0 00000000026012c0 GPR16: 00000000026012c0 0000000000000000 0000000000000000 0000000000000002 GPR20: 000000000000000c 0000000000000000 0000000000000000 00000000024200c0 GPR24: c0000000016eef48 0000000000000000 c00000010fff7d00 00000000026012c0 GPR28: 0000000000000000 c00000010fff7d00 c00000010fff6300 c00000010d10b6d0 NIP mem_cgroup_soft_limit_reclaim+0xf8/0x4f0 LR do_try_to_free_pages+0x1b4/0x450 Call Trace: do_try_to_free_pages+0x1b4/0x450 try_to_free_pages+0xf8/0x270 __alloc_pages_nodemask+0x7a8/0xff0 new_slab+0x104/0x8e0 ___slab_alloc+0x620/0x700 __slab_alloc+0x34/0x60 kmem_cache_alloc_node_trace+0xdc/0x310 mem_cgroup_init+0x158/0x1c8 do_one_initcall+0x68/0x1d0 kernel_init_freeable+0x278/0x360 kernel_init+0x24/0x170 ret_from_kernel_thread+0x5c/0x74 Instruction dump: eb81ffe0 eba1ffe8 ebc1fff0 ebe1fff8 4e800020 3d230001 e9499a42 3d220004 3929acd8 794a1f24 7d295214 eac90100 <e9360000> 2fa90000 419eff74 3b200000 ---[ end trace 342f5208b00d01b6 ]--- This is a chicken and egg issue where the kernel try to get free memory when allocating per node data in mem_cgroup_init(), but in that path mem_cgroup_soft_limit_reclaim() is called which assumes that these data are allocated. As mem_cgroup_soft_limit_reclaim() is best effort, it should return when these data are not yet allocated. This patch also fixes potential null pointer access in mem_cgroup_remove_from_trees() and mem_cgroup_update_tree(). Link: http://lkml.kernel.org/r/1487856999-16581-2-git-send-email-ldufour@linux.vnet.ibm.com Signed-off-by: Laurent Dufour <ldufour@linux.vnet.ibm.com> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Balbir Singh <bsingharora@gmail.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-03-10 00:17:06 +00:00
if (!mctz || RB_EMPTY_ROOT(&mctz->rb_root))
memcg: put soft limit reclaim out of way if the excess tree is empty We've had a report about soft lockups caused by lock bouncing in the soft reclaim path: BUG: soft lockup - CPU#0 stuck for 22s! [kav4proxy-kavic:3128] RIP: 0010:[<ffffffff81469798>] [<ffffffff81469798>] _raw_spin_lock+0x18/0x20 Call Trace: mem_cgroup_soft_limit_reclaim+0x25a/0x280 shrink_zones+0xed/0x200 do_try_to_free_pages+0x74/0x320 try_to_free_pages+0x112/0x180 __alloc_pages_slowpath+0x3ff/0x820 __alloc_pages_nodemask+0x1e9/0x200 alloc_pages_vma+0xe1/0x290 do_wp_page+0x19f/0x840 handle_pte_fault+0x1cd/0x230 do_page_fault+0x1fd/0x4c0 page_fault+0x25/0x30 There are no memcgs created so there cannot be any in the soft limit excess obviously: [...] memory 0 1 1 so all this just seems to be mem_cgroup_largest_soft_limit_node trying to get spin_lock_irq(&mctz->lock) just to find out that the soft limit excess tree is empty. This is just pointless wasting of cycles and cache line bouncing during heavy parallel reclaim on large machines. The particular machine wasn't very healthy and most probably suffering from a memory leak which just caused the memory reclaim to trash heavily. But bouncing on the lock certainly didn't help... Fix this by optimistic lockless check and bail out early if the tree is empty. This is theoretically racy but that shouldn't matter all that much. First of all soft limit is a best effort feature and it is slowly getting deprecated and its usage should be really scarce. Bouncing on a lock without a good reason is surely much bigger problem, especially on large CPU machines. Link: http://lkml.kernel.org/r/1470073277-1056-1-git-send-email-mhocko@kernel.org Signed-off-by: Michal Hocko <mhocko@suse.com> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-08-02 21:02:37 +00:00
return 0;
/*
* This loop can run a while, specially if mem_cgroup's continuously
* keep exceeding their soft limit and putting the system under
* pressure
*/
do {
if (next_mz)
mz = next_mz;
else
mz = mem_cgroup_largest_soft_limit_node(mctz);
if (!mz)
break;
nr_scanned = 0;
reclaimed = mem_cgroup_soft_reclaim(mz->memcg, pgdat,
gfp_mask, &nr_scanned);
nr_reclaimed += reclaimed;
*total_scanned += nr_scanned;
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
spin_lock_irq(&mctz->lock);
__mem_cgroup_remove_exceeded(mz, mctz);
/*
* If we failed to reclaim anything from this memory cgroup
* it is time to move on to the next cgroup
*/
next_mz = NULL;
if (!reclaimed)
next_mz = __mem_cgroup_largest_soft_limit_node(mctz);
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
excess = soft_limit_excess(mz->memcg);
/*
* One school of thought says that we should not add
* back the node to the tree if reclaim returns 0.
* But our reclaim could return 0, simply because due
* to priority we are exposing a smaller subset of
* memory to reclaim from. Consider this as a longer
* term TODO.
*/
/* If excess == 0, no tree ops */
__mem_cgroup_insert_exceeded(mz, mctz, excess);
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
spin_unlock_irq(&mctz->lock);
css_put(&mz->memcg->css);
loop++;
/*
* Could not reclaim anything and there are no more
* mem cgroups to try or we seem to be looping without
* reclaiming anything.
*/
if (!nr_reclaimed &&
(next_mz == NULL ||
loop > MEM_CGROUP_MAX_SOFT_LIMIT_RECLAIM_LOOPS))
break;
} while (!nr_reclaimed);
if (next_mz)
css_put(&next_mz->memcg->css);
return nr_reclaimed;
}
/*
* Test whether @memcg has children, dead or alive. Note that this
* function doesn't care whether @memcg has use_hierarchy enabled and
* returns %true if there are child csses according to the cgroup
* hierarchy. Testing use_hierarchy is the caller's responsibility.
*/
static inline bool memcg_has_children(struct mem_cgroup *memcg)
{
bool ret;
rcu_read_lock();
ret = css_next_child(NULL, &memcg->css);
rcu_read_unlock();
return ret;
}
/*
* Reclaims as many pages from the given memcg as possible.
*
* Caller is responsible for holding css reference for memcg.
*/
static int mem_cgroup_force_empty(struct mem_cgroup *memcg)
{
int nr_retries = MEM_CGROUP_RECLAIM_RETRIES;
/* we call try-to-free pages for make this cgroup empty */
lru_add_drain_all();
drain_all_stock(memcg);
memcg: move all acccounting to parent at rmdir() This patch provides a function to move account information of a page between mem_cgroups and rewrite force_empty to make use of this. This moving of page_cgroup is done under - lru_lock of source/destination mem_cgroup is held. - lock_page_cgroup() is held. Then, a routine which touches pc->mem_cgroup without lock_page_cgroup() should confirm pc->mem_cgroup is still valid or not. Typical code can be following. (while page is not under lock_page()) mem = pc->mem_cgroup; mz = page_cgroup_zoneinfo(pc) spin_lock_irqsave(&mz->lru_lock); if (pc->mem_cgroup == mem) ...../* some list handling */ spin_unlock_irqrestore(&mz->lru_lock); Of course, better way is lock_page_cgroup(pc); .... unlock_page_cgroup(pc); But you should confirm the nest of lock and avoid deadlock. If you treats page_cgroup from mem_cgroup's LRU under mz->lru_lock, you don't have to worry about what pc->mem_cgroup points to. moved pages are added to head of lru, not to tail. Expected users of this routine is: - force_empty (rmdir) - moving tasks between cgroup (for moving account information.) - hierarchy (maybe useful.) force_empty(rmdir) uses this move_account and move pages to its parent. This "move" will not cause OOM (I added "oom" parameter to try_charge().) If the parent is busy (not enough memory), force_empty calls try_to_free_page() and reduce usage. Purpose of this behavior is - Fix "forget all" behavior of force_empty and avoid leak of accounting. - By "moving first, free if necessary", keep pages on memory as much as possible. Adding a switch to change behavior of force_empty to - free first, move if necessary - free all, if there is mlocked/busy pages, return -EBUSY. is under consideration. (I'll add if someone requtests.) This patch also removes memory.force_empty file, a brutal debug-only interface. Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Tested-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Paul Menage <menage@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:07:53 +00:00
/* try to free all pages in this cgroup */
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
while (nr_retries && page_counter_read(&memcg->memory)) {
memcg: move all acccounting to parent at rmdir() This patch provides a function to move account information of a page between mem_cgroups and rewrite force_empty to make use of this. This moving of page_cgroup is done under - lru_lock of source/destination mem_cgroup is held. - lock_page_cgroup() is held. Then, a routine which touches pc->mem_cgroup without lock_page_cgroup() should confirm pc->mem_cgroup is still valid or not. Typical code can be following. (while page is not under lock_page()) mem = pc->mem_cgroup; mz = page_cgroup_zoneinfo(pc) spin_lock_irqsave(&mz->lru_lock); if (pc->mem_cgroup == mem) ...../* some list handling */ spin_unlock_irqrestore(&mz->lru_lock); Of course, better way is lock_page_cgroup(pc); .... unlock_page_cgroup(pc); But you should confirm the nest of lock and avoid deadlock. If you treats page_cgroup from mem_cgroup's LRU under mz->lru_lock, you don't have to worry about what pc->mem_cgroup points to. moved pages are added to head of lru, not to tail. Expected users of this routine is: - force_empty (rmdir) - moving tasks between cgroup (for moving account information.) - hierarchy (maybe useful.) force_empty(rmdir) uses this move_account and move pages to its parent. This "move" will not cause OOM (I added "oom" parameter to try_charge().) If the parent is busy (not enough memory), force_empty calls try_to_free_page() and reduce usage. Purpose of this behavior is - Fix "forget all" behavior of force_empty and avoid leak of accounting. - By "moving first, free if necessary", keep pages on memory as much as possible. Adding a switch to change behavior of force_empty to - free first, move if necessary - free all, if there is mlocked/busy pages, return -EBUSY. is under consideration. (I'll add if someone requtests.) This patch also removes memory.force_empty file, a brutal debug-only interface. Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Tested-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Paul Menage <menage@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:07:53 +00:00
int progress;
if (signal_pending(current))
return -EINTR;
mm: memcontrol: fix transparent huge page allocations under pressure In a memcg with even just moderate cache pressure, success rates for transparent huge page allocations drop to zero, wasting a lot of effort that the allocator puts into assembling these pages. The reason for this is that the memcg reclaim code was never designed for higher-order charges. It reclaims in small batches until there is room for at least one page. Huge page charges only succeed when these batches add up over a series of huge faults, which is unlikely under any significant load involving order-0 allocations in the group. Remove that loop on the memcg side in favor of passing the actual reclaim goal to direct reclaim, which is already set up and optimized to meet higher-order goals efficiently. This brings memcg's THP policy in line with the system policy: if the allocator painstakingly assembles a hugepage, memcg will at least make an honest effort to charge it. As a result, transparent hugepage allocation rates amid cache activity are drastically improved: vanilla patched pgalloc 4717530.80 ( +0.00%) 4451376.40 ( -5.64%) pgfault 491370.60 ( +0.00%) 225477.40 ( -54.11%) pgmajfault 2.00 ( +0.00%) 1.80 ( -6.67%) thp_fault_alloc 0.00 ( +0.00%) 531.60 (+100.00%) thp_fault_fallback 749.00 ( +0.00%) 217.40 ( -70.88%) [ Note: this may in turn increase memory consumption from internal fragmentation, which is an inherent risk of transparent hugepages. Some setups may have to adjust the memcg limits accordingly to accomodate this - or, if the machine is already packed to capacity, disable the transparent huge page feature. ] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Dave Hansen <dave@sr71.net> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-10-09 22:28:56 +00:00
progress = try_to_free_mem_cgroup_pages(memcg, 1,
GFP_KERNEL, true);
if (!progress) {
memcg: move all acccounting to parent at rmdir() This patch provides a function to move account information of a page between mem_cgroups and rewrite force_empty to make use of this. This moving of page_cgroup is done under - lru_lock of source/destination mem_cgroup is held. - lock_page_cgroup() is held. Then, a routine which touches pc->mem_cgroup without lock_page_cgroup() should confirm pc->mem_cgroup is still valid or not. Typical code can be following. (while page is not under lock_page()) mem = pc->mem_cgroup; mz = page_cgroup_zoneinfo(pc) spin_lock_irqsave(&mz->lru_lock); if (pc->mem_cgroup == mem) ...../* some list handling */ spin_unlock_irqrestore(&mz->lru_lock); Of course, better way is lock_page_cgroup(pc); .... unlock_page_cgroup(pc); But you should confirm the nest of lock and avoid deadlock. If you treats page_cgroup from mem_cgroup's LRU under mz->lru_lock, you don't have to worry about what pc->mem_cgroup points to. moved pages are added to head of lru, not to tail. Expected users of this routine is: - force_empty (rmdir) - moving tasks between cgroup (for moving account information.) - hierarchy (maybe useful.) force_empty(rmdir) uses this move_account and move pages to its parent. This "move" will not cause OOM (I added "oom" parameter to try_charge().) If the parent is busy (not enough memory), force_empty calls try_to_free_page() and reduce usage. Purpose of this behavior is - Fix "forget all" behavior of force_empty and avoid leak of accounting. - By "moving first, free if necessary", keep pages on memory as much as possible. Adding a switch to change behavior of force_empty to - free first, move if necessary - free all, if there is mlocked/busy pages, return -EBUSY. is under consideration. (I'll add if someone requtests.) This patch also removes memory.force_empty file, a brutal debug-only interface. Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Tested-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Paul Menage <menage@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:07:53 +00:00
nr_retries--;
/* maybe some writeback is necessary */
congestion_wait(BLK_RW_ASYNC, HZ/10);
}
memcg: move all acccounting to parent at rmdir() This patch provides a function to move account information of a page between mem_cgroups and rewrite force_empty to make use of this. This moving of page_cgroup is done under - lru_lock of source/destination mem_cgroup is held. - lock_page_cgroup() is held. Then, a routine which touches pc->mem_cgroup without lock_page_cgroup() should confirm pc->mem_cgroup is still valid or not. Typical code can be following. (while page is not under lock_page()) mem = pc->mem_cgroup; mz = page_cgroup_zoneinfo(pc) spin_lock_irqsave(&mz->lru_lock); if (pc->mem_cgroup == mem) ...../* some list handling */ spin_unlock_irqrestore(&mz->lru_lock); Of course, better way is lock_page_cgroup(pc); .... unlock_page_cgroup(pc); But you should confirm the nest of lock and avoid deadlock. If you treats page_cgroup from mem_cgroup's LRU under mz->lru_lock, you don't have to worry about what pc->mem_cgroup points to. moved pages are added to head of lru, not to tail. Expected users of this routine is: - force_empty (rmdir) - moving tasks between cgroup (for moving account information.) - hierarchy (maybe useful.) force_empty(rmdir) uses this move_account and move pages to its parent. This "move" will not cause OOM (I added "oom" parameter to try_charge().) If the parent is busy (not enough memory), force_empty calls try_to_free_page() and reduce usage. Purpose of this behavior is - Fix "forget all" behavior of force_empty and avoid leak of accounting. - By "moving first, free if necessary", keep pages on memory as much as possible. Adding a switch to change behavior of force_empty to - free first, move if necessary - free all, if there is mlocked/busy pages, return -EBUSY. is under consideration. (I'll add if someone requtests.) This patch also removes memory.force_empty file, a brutal debug-only interface. Reviewed-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Tested-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Paul Menage <menage@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:07:53 +00:00
}
return 0;
}
static ssize_t mem_cgroup_force_empty_write(struct kernfs_open_file *of,
char *buf, size_t nbytes,
loff_t off)
{
struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of));
if (mem_cgroup_is_root(memcg))
return -EINVAL;
return mem_cgroup_force_empty(memcg) ?: nbytes;
}
cgroup: pass around cgroup_subsys_state instead of cgroup in file methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup. Please see the previous commit which converts the subsystem methods for rationale. This patch converts all cftype file operations to take @css instead of @cgroup. cftypes for the cgroup core files don't have their subsytem pointer set. These will automatically use the dummy_css added by the previous patch and can be converted the same way. Most subsystem conversions are straight forwards but there are some interesting ones. * freezer: update_if_frozen() is also converted to take @css instead of @cgroup for consistency. This will make the code look simpler too once iterators are converted to use css. * memory/vmpressure: mem_cgroup_from_css() needs to be exported to vmpressure while mem_cgroup_from_cont() can be made static. Updated accordingly. * cpu: cgroup_tg() doesn't have any user left. Removed. * cpuacct: cgroup_ca() doesn't have any user left. Removed. * hugetlb: hugetlb_cgroup_form_cgroup() doesn't have any user left. Removed. * net_cls: cgrp_cls_state() doesn't have any user left. Removed. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:24 +00:00
static u64 mem_cgroup_hierarchy_read(struct cgroup_subsys_state *css,
struct cftype *cft)
{
cgroup: pass around cgroup_subsys_state instead of cgroup in file methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup. Please see the previous commit which converts the subsystem methods for rationale. This patch converts all cftype file operations to take @css instead of @cgroup. cftypes for the cgroup core files don't have their subsytem pointer set. These will automatically use the dummy_css added by the previous patch and can be converted the same way. Most subsystem conversions are straight forwards but there are some interesting ones. * freezer: update_if_frozen() is also converted to take @css instead of @cgroup for consistency. This will make the code look simpler too once iterators are converted to use css. * memory/vmpressure: mem_cgroup_from_css() needs to be exported to vmpressure while mem_cgroup_from_cont() can be made static. Updated accordingly. * cpu: cgroup_tg() doesn't have any user left. Removed. * cpuacct: cgroup_ca() doesn't have any user left. Removed. * hugetlb: hugetlb_cgroup_form_cgroup() doesn't have any user left. Removed. * net_cls: cgrp_cls_state() doesn't have any user left. Removed. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:24 +00:00
return mem_cgroup_from_css(css)->use_hierarchy;
}
cgroup: pass around cgroup_subsys_state instead of cgroup in file methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup. Please see the previous commit which converts the subsystem methods for rationale. This patch converts all cftype file operations to take @css instead of @cgroup. cftypes for the cgroup core files don't have their subsytem pointer set. These will automatically use the dummy_css added by the previous patch and can be converted the same way. Most subsystem conversions are straight forwards but there are some interesting ones. * freezer: update_if_frozen() is also converted to take @css instead of @cgroup for consistency. This will make the code look simpler too once iterators are converted to use css. * memory/vmpressure: mem_cgroup_from_css() needs to be exported to vmpressure while mem_cgroup_from_cont() can be made static. Updated accordingly. * cpu: cgroup_tg() doesn't have any user left. Removed. * cpuacct: cgroup_ca() doesn't have any user left. Removed. * hugetlb: hugetlb_cgroup_form_cgroup() doesn't have any user left. Removed. * net_cls: cgrp_cls_state() doesn't have any user left. Removed. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:24 +00:00
static int mem_cgroup_hierarchy_write(struct cgroup_subsys_state *css,
struct cftype *cft, u64 val)
{
int retval = 0;
cgroup: pass around cgroup_subsys_state instead of cgroup in file methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup. Please see the previous commit which converts the subsystem methods for rationale. This patch converts all cftype file operations to take @css instead of @cgroup. cftypes for the cgroup core files don't have their subsytem pointer set. These will automatically use the dummy_css added by the previous patch and can be converted the same way. Most subsystem conversions are straight forwards but there are some interesting ones. * freezer: update_if_frozen() is also converted to take @css instead of @cgroup for consistency. This will make the code look simpler too once iterators are converted to use css. * memory/vmpressure: mem_cgroup_from_css() needs to be exported to vmpressure while mem_cgroup_from_cont() can be made static. Updated accordingly. * cpu: cgroup_tg() doesn't have any user left. Removed. * cpuacct: cgroup_ca() doesn't have any user left. Removed. * hugetlb: hugetlb_cgroup_form_cgroup() doesn't have any user left. Removed. * net_cls: cgrp_cls_state() doesn't have any user left. Removed. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:24 +00:00
struct mem_cgroup *memcg = mem_cgroup_from_css(css);
struct mem_cgroup *parent_memcg = mem_cgroup_from_css(memcg->css.parent);
if (memcg->use_hierarchy == val)
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
return 0;
/*
* If parent's use_hierarchy is set, we can't make any modifications
* in the child subtrees. If it is unset, then the change can
* occur, provided the current cgroup has no children.
*
* For the root cgroup, parent_mem is NULL, we allow value to be
* set if there are no children.
*/
if ((!parent_memcg || !parent_memcg->use_hierarchy) &&
(val == 1 || val == 0)) {
if (!memcg_has_children(memcg))
memcg->use_hierarchy = val;
else
retval = -EBUSY;
} else
retval = -EINVAL;
return retval;
}
static unsigned long mem_cgroup_usage(struct mem_cgroup *memcg, bool swap)
{
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
unsigned long val;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
if (mem_cgroup_is_root(memcg)) {
val = memcg_page_state(memcg, NR_FILE_PAGES) +
mm: memcontrol: switch to native NR_ANON_MAPPED counter Memcg maintains a private MEMCG_RSS counter. This divergence from the generic VM accounting means unnecessary code overhead, and creates a dependency for memcg that page->mapping is set up at the time of charging, so that page types can be told apart. Convert the generic accounting sites to mod_lruvec_page_state and friends to maintain the per-cgroup vmstat counter of NR_ANON_MAPPED. We use lock_page_memcg() to stabilize page->mem_cgroup during rmap changes, the same way we do for NR_FILE_MAPPED. With the previous patch removing MEMCG_CACHE and the private NR_SHMEM counter, this patch finally eliminates the need to have page->mapping set up at charge time. However, we need to have page->mem_cgroup set up by the time rmap runs and does the accounting, so switch the commit and the rmap callbacks around. v2: fix temporary accounting bug by switching rmap<->commit (Joonsoo) Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Cc: Alex Shi <alex.shi@linux.alibaba.com> Cc: Hugh Dickins <hughd@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.com> Cc: Roman Gushchin <guro@fb.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Balbir Singh <bsingharora@gmail.com> Link: http://lkml.kernel.org/r/20200508183105.225460-11-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:01:57 +00:00
memcg_page_state(memcg, NR_ANON_MAPPED);
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
if (swap)
val += memcg_page_state(memcg, MEMCG_SWAP);
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
} else {
if (!swap)
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
val = page_counter_read(&memcg->memory);
else
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
val = page_counter_read(&memcg->memsw);
}
memcg: fix thresholds for 32b architectures. Commit 424cdc141380 ("memcg: convert threshold to bytes") has fixed a regression introduced by 3e32cb2e0a12 ("mm: memcontrol: lockless page counters") where thresholds were silently converted to use page units rather than bytes when interpreting the user input. The fix is not complete, though, as properly pointed out by Ben Hutchings during stable backport review. The page count is converted to bytes but unsigned long is used to hold the value which would be obviously not sufficient for 32b systems with more than 4G thresholds. The same applies to usage as taken from mem_cgroup_usage which might overflow. Let's remove this bytes vs. pages internal tracking differences and handle thresholds in page units internally. Chage mem_cgroup_usage() to return the value in page units and revert 424cdc141380 because this should be sufficient for the consistent handling. mem_cgroup_read_u64 as the only users of mem_cgroup_usage outside of the threshold handling code is converted to give the proper in bytes result. It is doing that already for page_counter output so this is more consistent as well. The value presented to the userspace is still in bytes units. Fixes: 424cdc141380 ("memcg: convert threshold to bytes") Fixes: 3e32cb2e0a12 ("mm: memcontrol: lockless page counters") Signed-off-by: Michal Hocko <mhocko@suse.com> Reported-by: Ben Hutchings <ben@decadent.org.uk> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> From: Michal Hocko <mhocko@kernel.org> Subject: memcg-fix-thresholds-for-32b-architectures-fix Cc: Ben Hutchings <ben@decadent.org.uk> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Johannes Weiner <hannes@cmpxchg.org> From: Andrew Morton <akpm@linux-foundation.org> Subject: memcg-fix-thresholds-for-32b-architectures-fix-fix don't attempt to inline mem_cgroup_usage() The compiler ignores the inline anwyay. And __always_inlining it adds 600 bytes of goop to the .o file. Cc: Ben Hutchings <ben@decadent.org.uk> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:50:29 +00:00
return val;
}
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
enum {
RES_USAGE,
RES_LIMIT,
RES_MAX_USAGE,
RES_FAILCNT,
RES_SOFT_LIMIT,
};
static u64 mem_cgroup_read_u64(struct cgroup_subsys_state *css,
struct cftype *cft)
{
cgroup: pass around cgroup_subsys_state instead of cgroup in file methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup. Please see the previous commit which converts the subsystem methods for rationale. This patch converts all cftype file operations to take @css instead of @cgroup. cftypes for the cgroup core files don't have their subsytem pointer set. These will automatically use the dummy_css added by the previous patch and can be converted the same way. Most subsystem conversions are straight forwards but there are some interesting ones. * freezer: update_if_frozen() is also converted to take @css instead of @cgroup for consistency. This will make the code look simpler too once iterators are converted to use css. * memory/vmpressure: mem_cgroup_from_css() needs to be exported to vmpressure while mem_cgroup_from_cont() can be made static. Updated accordingly. * cpu: cgroup_tg() doesn't have any user left. Removed. * cpuacct: cgroup_ca() doesn't have any user left. Removed. * hugetlb: hugetlb_cgroup_form_cgroup() doesn't have any user left. Removed. * net_cls: cgrp_cls_state() doesn't have any user left. Removed. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:24 +00:00
struct mem_cgroup *memcg = mem_cgroup_from_css(css);
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
struct page_counter *counter;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
switch (MEMFILE_TYPE(cft->private)) {
memcg: mem+swap controller core This patch implements per cgroup limit for usage of memory+swap. However there are SwapCache, double counting of swap-cache and swap-entry is avoided. Mem+Swap controller works as following. - memory usage is limited by memory.limit_in_bytes. - memory + swap usage is limited by memory.memsw_limit_in_bytes. This has following benefits. - A user can limit total resource usage of mem+swap. Without this, because memory resource controller doesn't take care of usage of swap, a process can exhaust all the swap (by memory leak.) We can avoid this case. And Swap is shared resource but it cannot be reclaimed (goes back to memory) until it's used. This characteristic can be trouble when the memory is divided into some parts by cpuset or memcg. Assume group A and group B. After some application executes, the system can be.. Group A -- very large free memory space but occupy 99% of swap. Group B -- under memory shortage but cannot use swap...it's nearly full. Ability to set appropriate swap limit for each group is required. Maybe someone wonder "why not swap but mem+swap ?" - The global LRU(kswapd) can swap out arbitrary pages. Swap-out means to move account from memory to swap...there is no change in usage of mem+swap. In other words, when we want to limit the usage of swap without affecting global LRU, mem+swap limit is better than just limiting swap. Accounting target information is stored in swap_cgroup which is per swap entry record. Charge is done as following. map - charge page and memsw. unmap - uncharge page/memsw if not SwapCache. swap-out (__delete_from_swap_cache) - uncharge page - record mem_cgroup information to swap_cgroup. swap-in (do_swap_page) - charged as page and memsw. record in swap_cgroup is cleared. memsw accounting is decremented. swap-free (swap_free()) - if swap entry is freed, memsw is uncharged by PAGE_SIZE. There are people work under never-swap environments and consider swap as something bad. For such people, this mem+swap controller extension is just an overhead. This overhead is avoided by config or boot option. (see Kconfig. detail is not in this patch.) TODO: - maybe more optimization can be don in swap-in path. (but not very safe.) But we just do simple accounting at this stage. [nishimura@mxp.nes.nec.co.jp: make resize limit hold mutex] [hugh@veritas.com: memswap controller core swapcache fixes] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:08:00 +00:00
case _MEM:
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
counter = &memcg->memory;
break;
memcg: mem+swap controller core This patch implements per cgroup limit for usage of memory+swap. However there are SwapCache, double counting of swap-cache and swap-entry is avoided. Mem+Swap controller works as following. - memory usage is limited by memory.limit_in_bytes. - memory + swap usage is limited by memory.memsw_limit_in_bytes. This has following benefits. - A user can limit total resource usage of mem+swap. Without this, because memory resource controller doesn't take care of usage of swap, a process can exhaust all the swap (by memory leak.) We can avoid this case. And Swap is shared resource but it cannot be reclaimed (goes back to memory) until it's used. This characteristic can be trouble when the memory is divided into some parts by cpuset or memcg. Assume group A and group B. After some application executes, the system can be.. Group A -- very large free memory space but occupy 99% of swap. Group B -- under memory shortage but cannot use swap...it's nearly full. Ability to set appropriate swap limit for each group is required. Maybe someone wonder "why not swap but mem+swap ?" - The global LRU(kswapd) can swap out arbitrary pages. Swap-out means to move account from memory to swap...there is no change in usage of mem+swap. In other words, when we want to limit the usage of swap without affecting global LRU, mem+swap limit is better than just limiting swap. Accounting target information is stored in swap_cgroup which is per swap entry record. Charge is done as following. map - charge page and memsw. unmap - uncharge page/memsw if not SwapCache. swap-out (__delete_from_swap_cache) - uncharge page - record mem_cgroup information to swap_cgroup. swap-in (do_swap_page) - charged as page and memsw. record in swap_cgroup is cleared. memsw accounting is decremented. swap-free (swap_free()) - if swap entry is freed, memsw is uncharged by PAGE_SIZE. There are people work under never-swap environments and consider swap as something bad. For such people, this mem+swap controller extension is just an overhead. This overhead is avoided by config or boot option. (see Kconfig. detail is not in this patch.) TODO: - maybe more optimization can be don in swap-in path. (but not very safe.) But we just do simple accounting at this stage. [nishimura@mxp.nes.nec.co.jp: make resize limit hold mutex] [hugh@veritas.com: memswap controller core swapcache fixes] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:08:00 +00:00
case _MEMSWAP:
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
counter = &memcg->memsw;
break;
memcg: kmem accounting basic infrastructure Add the basic infrastructure for the accounting of kernel memory. To control that, the following files are created: * memory.kmem.usage_in_bytes * memory.kmem.limit_in_bytes * memory.kmem.failcnt * memory.kmem.max_usage_in_bytes They have the same meaning of their user memory counterparts. They reflect the state of the "kmem" res_counter. Per cgroup kmem memory accounting is not enabled until a limit is set for the group. Once the limit is set the accounting cannot be disabled for that group. This means that after the patch is applied, no behavioral changes exists for whoever is still using memcg to control their memory usage, until memory.kmem.limit_in_bytes is set for the first time. We always account to both user and kernel resource_counters. This effectively means that an independent kernel limit is in place when the limit is set to a lower value than the user memory. A equal or higher value means that the user limit will always hit first, meaning that kmem is effectively unlimited. People who want to track kernel memory but not limit it, can set this limit to a very high number (like RESOURCE_MAX - 1page - that no one will ever hit, or equal to the user memory) [akpm@linux-foundation.org: MEMCG_MMEM only works with slab and slub] Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:47 +00:00
case _KMEM:
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
counter = &memcg->kmem;
memcg: kmem accounting basic infrastructure Add the basic infrastructure for the accounting of kernel memory. To control that, the following files are created: * memory.kmem.usage_in_bytes * memory.kmem.limit_in_bytes * memory.kmem.failcnt * memory.kmem.max_usage_in_bytes They have the same meaning of their user memory counterparts. They reflect the state of the "kmem" res_counter. Per cgroup kmem memory accounting is not enabled until a limit is set for the group. Once the limit is set the accounting cannot be disabled for that group. This means that after the patch is applied, no behavioral changes exists for whoever is still using memcg to control their memory usage, until memory.kmem.limit_in_bytes is set for the first time. We always account to both user and kernel resource_counters. This effectively means that an independent kernel limit is in place when the limit is set to a lower value than the user memory. A equal or higher value means that the user limit will always hit first, meaning that kmem is effectively unlimited. People who want to track kernel memory but not limit it, can set this limit to a very high number (like RESOURCE_MAX - 1page - that no one will ever hit, or equal to the user memory) [akpm@linux-foundation.org: MEMCG_MMEM only works with slab and slub] Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:47 +00:00
break;
case _TCP:
counter = &memcg->tcpmem;
break;
memcg: mem+swap controller core This patch implements per cgroup limit for usage of memory+swap. However there are SwapCache, double counting of swap-cache and swap-entry is avoided. Mem+Swap controller works as following. - memory usage is limited by memory.limit_in_bytes. - memory + swap usage is limited by memory.memsw_limit_in_bytes. This has following benefits. - A user can limit total resource usage of mem+swap. Without this, because memory resource controller doesn't take care of usage of swap, a process can exhaust all the swap (by memory leak.) We can avoid this case. And Swap is shared resource but it cannot be reclaimed (goes back to memory) until it's used. This characteristic can be trouble when the memory is divided into some parts by cpuset or memcg. Assume group A and group B. After some application executes, the system can be.. Group A -- very large free memory space but occupy 99% of swap. Group B -- under memory shortage but cannot use swap...it's nearly full. Ability to set appropriate swap limit for each group is required. Maybe someone wonder "why not swap but mem+swap ?" - The global LRU(kswapd) can swap out arbitrary pages. Swap-out means to move account from memory to swap...there is no change in usage of mem+swap. In other words, when we want to limit the usage of swap without affecting global LRU, mem+swap limit is better than just limiting swap. Accounting target information is stored in swap_cgroup which is per swap entry record. Charge is done as following. map - charge page and memsw. unmap - uncharge page/memsw if not SwapCache. swap-out (__delete_from_swap_cache) - uncharge page - record mem_cgroup information to swap_cgroup. swap-in (do_swap_page) - charged as page and memsw. record in swap_cgroup is cleared. memsw accounting is decremented. swap-free (swap_free()) - if swap entry is freed, memsw is uncharged by PAGE_SIZE. There are people work under never-swap environments and consider swap as something bad. For such people, this mem+swap controller extension is just an overhead. This overhead is avoided by config or boot option. (see Kconfig. detail is not in this patch.) TODO: - maybe more optimization can be don in swap-in path. (but not very safe.) But we just do simple accounting at this stage. [nishimura@mxp.nes.nec.co.jp: make resize limit hold mutex] [hugh@veritas.com: memswap controller core swapcache fixes] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:08:00 +00:00
default:
BUG();
}
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
switch (MEMFILE_ATTR(cft->private)) {
case RES_USAGE:
if (counter == &memcg->memory)
memcg: fix thresholds for 32b architectures. Commit 424cdc141380 ("memcg: convert threshold to bytes") has fixed a regression introduced by 3e32cb2e0a12 ("mm: memcontrol: lockless page counters") where thresholds were silently converted to use page units rather than bytes when interpreting the user input. The fix is not complete, though, as properly pointed out by Ben Hutchings during stable backport review. The page count is converted to bytes but unsigned long is used to hold the value which would be obviously not sufficient for 32b systems with more than 4G thresholds. The same applies to usage as taken from mem_cgroup_usage which might overflow. Let's remove this bytes vs. pages internal tracking differences and handle thresholds in page units internally. Chage mem_cgroup_usage() to return the value in page units and revert 424cdc141380 because this should be sufficient for the consistent handling. mem_cgroup_read_u64 as the only users of mem_cgroup_usage outside of the threshold handling code is converted to give the proper in bytes result. It is doing that already for page_counter output so this is more consistent as well. The value presented to the userspace is still in bytes units. Fixes: 424cdc141380 ("memcg: convert threshold to bytes") Fixes: 3e32cb2e0a12 ("mm: memcontrol: lockless page counters") Signed-off-by: Michal Hocko <mhocko@suse.com> Reported-by: Ben Hutchings <ben@decadent.org.uk> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> From: Michal Hocko <mhocko@kernel.org> Subject: memcg-fix-thresholds-for-32b-architectures-fix Cc: Ben Hutchings <ben@decadent.org.uk> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Johannes Weiner <hannes@cmpxchg.org> From: Andrew Morton <akpm@linux-foundation.org> Subject: memcg-fix-thresholds-for-32b-architectures-fix-fix don't attempt to inline mem_cgroup_usage() The compiler ignores the inline anwyay. And __always_inlining it adds 600 bytes of goop to the .o file. Cc: Ben Hutchings <ben@decadent.org.uk> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:50:29 +00:00
return (u64)mem_cgroup_usage(memcg, false) * PAGE_SIZE;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
if (counter == &memcg->memsw)
memcg: fix thresholds for 32b architectures. Commit 424cdc141380 ("memcg: convert threshold to bytes") has fixed a regression introduced by 3e32cb2e0a12 ("mm: memcontrol: lockless page counters") where thresholds were silently converted to use page units rather than bytes when interpreting the user input. The fix is not complete, though, as properly pointed out by Ben Hutchings during stable backport review. The page count is converted to bytes but unsigned long is used to hold the value which would be obviously not sufficient for 32b systems with more than 4G thresholds. The same applies to usage as taken from mem_cgroup_usage which might overflow. Let's remove this bytes vs. pages internal tracking differences and handle thresholds in page units internally. Chage mem_cgroup_usage() to return the value in page units and revert 424cdc141380 because this should be sufficient for the consistent handling. mem_cgroup_read_u64 as the only users of mem_cgroup_usage outside of the threshold handling code is converted to give the proper in bytes result. It is doing that already for page_counter output so this is more consistent as well. The value presented to the userspace is still in bytes units. Fixes: 424cdc141380 ("memcg: convert threshold to bytes") Fixes: 3e32cb2e0a12 ("mm: memcontrol: lockless page counters") Signed-off-by: Michal Hocko <mhocko@suse.com> Reported-by: Ben Hutchings <ben@decadent.org.uk> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> From: Michal Hocko <mhocko@kernel.org> Subject: memcg-fix-thresholds-for-32b-architectures-fix Cc: Ben Hutchings <ben@decadent.org.uk> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Johannes Weiner <hannes@cmpxchg.org> From: Andrew Morton <akpm@linux-foundation.org> Subject: memcg-fix-thresholds-for-32b-architectures-fix-fix don't attempt to inline mem_cgroup_usage() The compiler ignores the inline anwyay. And __always_inlining it adds 600 bytes of goop to the .o file. Cc: Ben Hutchings <ben@decadent.org.uk> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-06 02:50:29 +00:00
return (u64)mem_cgroup_usage(memcg, true) * PAGE_SIZE;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
return (u64)page_counter_read(counter) * PAGE_SIZE;
case RES_LIMIT:
return (u64)counter->max * PAGE_SIZE;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
case RES_MAX_USAGE:
return (u64)counter->watermark * PAGE_SIZE;
case RES_FAILCNT:
return counter->failcnt;
case RES_SOFT_LIMIT:
return (u64)memcg->soft_limit * PAGE_SIZE;
default:
BUG();
}
}
memcg: kmem accounting basic infrastructure Add the basic infrastructure for the accounting of kernel memory. To control that, the following files are created: * memory.kmem.usage_in_bytes * memory.kmem.limit_in_bytes * memory.kmem.failcnt * memory.kmem.max_usage_in_bytes They have the same meaning of their user memory counterparts. They reflect the state of the "kmem" res_counter. Per cgroup kmem memory accounting is not enabled until a limit is set for the group. Once the limit is set the accounting cannot be disabled for that group. This means that after the patch is applied, no behavioral changes exists for whoever is still using memcg to control their memory usage, until memory.kmem.limit_in_bytes is set for the first time. We always account to both user and kernel resource_counters. This effectively means that an independent kernel limit is in place when the limit is set to a lower value than the user memory. A equal or higher value means that the user limit will always hit first, meaning that kmem is effectively unlimited. People who want to track kernel memory but not limit it, can set this limit to a very high number (like RESOURCE_MAX - 1page - that no one will ever hit, or equal to the user memory) [akpm@linux-foundation.org: MEMCG_MMEM only works with slab and slub] Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:47 +00:00
mm: memcg/slab: fix percpu slab vmstats flushing Currently slab percpu vmstats are flushed twice: during the memcg offlining and just before freeing the memcg structure. Each time percpu counters are summed, added to the atomic counterparts and propagated up by the cgroup tree. The second flushing is required due to how recursive vmstats are implemented: counters are batched in percpu variables on a local level, and once a percpu value is crossing some predefined threshold, it spills over to atomic values on the local and each ascendant levels. It means that without flushing some numbers cached in percpu variables will be dropped on floor each time a cgroup is destroyed. And with uptime the error on upper levels might become noticeable. The first flushing aims to make counters on ancestor levels more precise. Dying cgroups may resume in the dying state for a long time. After kmem_cache reparenting which is performed during the offlining slab counters of the dying cgroup don't have any chances to be updated, because any slab operations will be performed on the parent level. It means that the inaccuracy caused by percpu batching will not decrease up to the final destruction of the cgroup. By the original idea flushing slab counters during the offlining should minimize the visible inaccuracy of slab counters on the parent level. The problem is that percpu counters are not zeroed after the first flushing. So every cached percpu value is summed twice. It creates a small error (up to 32 pages per cpu, but usually less) which accumulates on parent cgroup level. After creating and destroying of thousands of child cgroups, slab counter on parent level can be way off the real value. For now, let's just stop flushing slab counters on memcg offlining. It can't be done correctly without scheduling a work on each cpu: reading and zeroing it during css offlining can race with an asynchronous update, which doesn't expect values to be changed underneath. With this change, slab counters on parent level will become eventually consistent. Once all dying children are gone, values are correct. And if not, the error is capped by 32 * NR_CPUS pages per dying cgroup. It's not perfect, as slab are reparented, so any updates after the reparenting will happen on the parent level. It means that if a slab page was allocated, a counter on child level was bumped, then the page was reparented and freed, the annihilation of positive and negative counter values will not happen until the child cgroup is released. It makes slab counters different from others, and it might want us to implement flushing in a correct form again. But it's also a question of performance: scheduling a work on each cpu isn't free, and it's an open question if the benefit of having more accurate counters is worth it. We might also consider flushing all counters on offlining, not only slab counters. So let's fix the main problem now: make the slab counters eventually consistent, so at least the error won't grow with uptime (or more precisely the number of created and destroyed cgroups). And think about the accuracy of counters separately. Link: http://lkml.kernel.org/r/20191220042728.1045881-1-guro@fb.com Fixes: bee07b33db78 ("mm: memcontrol: flush percpu slab vmstats on kmem offlining") Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-01-14 00:29:16 +00:00
static void memcg_flush_percpu_vmstats(struct mem_cgroup *memcg)
mm: memcontrol: flush percpu vmstats before releasing memcg Percpu caching of local vmstats with the conditional propagation by the cgroup tree leads to an accumulation of errors on non-leaf levels. Let's imagine two nested memory cgroups A and A/B. Say, a process belonging to A/B allocates 100 pagecache pages on the CPU 0. The percpu cache will spill 3 times, so that 32*3=96 pages will be accounted to A/B and A atomic vmstat counters, 4 pages will remain in the percpu cache. Imagine A/B is nearby memory.max, so that every following allocation triggers a direct reclaim on the local CPU. Say, each such attempt will free 16 pages on a new cpu. That means every percpu cache will have -16 pages, except the first one, which will have 4 - 16 = -12. A/B and A atomic counters will not be touched at all. Now a user removes A/B. All percpu caches are freed and corresponding vmstat numbers are forgotten. A has 96 pages more than expected. As memory cgroups are created and destroyed, errors do accumulate. Even 1-2 pages differences can accumulate into large numbers. To fix this issue let's accumulate and propagate percpu vmstat values before releasing the memory cgroup. At this point these numbers are stable and cannot be changed. Since on cpu hotplug we do flush percpu vmstats anyway, we can iterate only over online cpus. Link: http://lkml.kernel.org/r/20190819202338.363363-2-guro@fb.com Fixes: 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty") Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-08-25 00:54:47 +00:00
{
mm: memcg/slab: fix percpu slab vmstats flushing Currently slab percpu vmstats are flushed twice: during the memcg offlining and just before freeing the memcg structure. Each time percpu counters are summed, added to the atomic counterparts and propagated up by the cgroup tree. The second flushing is required due to how recursive vmstats are implemented: counters are batched in percpu variables on a local level, and once a percpu value is crossing some predefined threshold, it spills over to atomic values on the local and each ascendant levels. It means that without flushing some numbers cached in percpu variables will be dropped on floor each time a cgroup is destroyed. And with uptime the error on upper levels might become noticeable. The first flushing aims to make counters on ancestor levels more precise. Dying cgroups may resume in the dying state for a long time. After kmem_cache reparenting which is performed during the offlining slab counters of the dying cgroup don't have any chances to be updated, because any slab operations will be performed on the parent level. It means that the inaccuracy caused by percpu batching will not decrease up to the final destruction of the cgroup. By the original idea flushing slab counters during the offlining should minimize the visible inaccuracy of slab counters on the parent level. The problem is that percpu counters are not zeroed after the first flushing. So every cached percpu value is summed twice. It creates a small error (up to 32 pages per cpu, but usually less) which accumulates on parent cgroup level. After creating and destroying of thousands of child cgroups, slab counter on parent level can be way off the real value. For now, let's just stop flushing slab counters on memcg offlining. It can't be done correctly without scheduling a work on each cpu: reading and zeroing it during css offlining can race with an asynchronous update, which doesn't expect values to be changed underneath. With this change, slab counters on parent level will become eventually consistent. Once all dying children are gone, values are correct. And if not, the error is capped by 32 * NR_CPUS pages per dying cgroup. It's not perfect, as slab are reparented, so any updates after the reparenting will happen on the parent level. It means that if a slab page was allocated, a counter on child level was bumped, then the page was reparented and freed, the annihilation of positive and negative counter values will not happen until the child cgroup is released. It makes slab counters different from others, and it might want us to implement flushing in a correct form again. But it's also a question of performance: scheduling a work on each cpu isn't free, and it's an open question if the benefit of having more accurate counters is worth it. We might also consider flushing all counters on offlining, not only slab counters. So let's fix the main problem now: make the slab counters eventually consistent, so at least the error won't grow with uptime (or more precisely the number of created and destroyed cgroups). And think about the accuracy of counters separately. Link: http://lkml.kernel.org/r/20191220042728.1045881-1-guro@fb.com Fixes: bee07b33db78 ("mm: memcontrol: flush percpu slab vmstats on kmem offlining") Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-01-14 00:29:16 +00:00
unsigned long stat[MEMCG_NR_STAT] = {0};
mm: memcontrol: flush percpu vmstats before releasing memcg Percpu caching of local vmstats with the conditional propagation by the cgroup tree leads to an accumulation of errors on non-leaf levels. Let's imagine two nested memory cgroups A and A/B. Say, a process belonging to A/B allocates 100 pagecache pages on the CPU 0. The percpu cache will spill 3 times, so that 32*3=96 pages will be accounted to A/B and A atomic vmstat counters, 4 pages will remain in the percpu cache. Imagine A/B is nearby memory.max, so that every following allocation triggers a direct reclaim on the local CPU. Say, each such attempt will free 16 pages on a new cpu. That means every percpu cache will have -16 pages, except the first one, which will have 4 - 16 = -12. A/B and A atomic counters will not be touched at all. Now a user removes A/B. All percpu caches are freed and corresponding vmstat numbers are forgotten. A has 96 pages more than expected. As memory cgroups are created and destroyed, errors do accumulate. Even 1-2 pages differences can accumulate into large numbers. To fix this issue let's accumulate and propagate percpu vmstat values before releasing the memory cgroup. At this point these numbers are stable and cannot be changed. Since on cpu hotplug we do flush percpu vmstats anyway, we can iterate only over online cpus. Link: http://lkml.kernel.org/r/20190819202338.363363-2-guro@fb.com Fixes: 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty") Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-08-25 00:54:47 +00:00
struct mem_cgroup *mi;
int node, cpu, i;
for_each_online_cpu(cpu)
mm: memcg/slab: fix percpu slab vmstats flushing Currently slab percpu vmstats are flushed twice: during the memcg offlining and just before freeing the memcg structure. Each time percpu counters are summed, added to the atomic counterparts and propagated up by the cgroup tree. The second flushing is required due to how recursive vmstats are implemented: counters are batched in percpu variables on a local level, and once a percpu value is crossing some predefined threshold, it spills over to atomic values on the local and each ascendant levels. It means that without flushing some numbers cached in percpu variables will be dropped on floor each time a cgroup is destroyed. And with uptime the error on upper levels might become noticeable. The first flushing aims to make counters on ancestor levels more precise. Dying cgroups may resume in the dying state for a long time. After kmem_cache reparenting which is performed during the offlining slab counters of the dying cgroup don't have any chances to be updated, because any slab operations will be performed on the parent level. It means that the inaccuracy caused by percpu batching will not decrease up to the final destruction of the cgroup. By the original idea flushing slab counters during the offlining should minimize the visible inaccuracy of slab counters on the parent level. The problem is that percpu counters are not zeroed after the first flushing. So every cached percpu value is summed twice. It creates a small error (up to 32 pages per cpu, but usually less) which accumulates on parent cgroup level. After creating and destroying of thousands of child cgroups, slab counter on parent level can be way off the real value. For now, let's just stop flushing slab counters on memcg offlining. It can't be done correctly without scheduling a work on each cpu: reading and zeroing it during css offlining can race with an asynchronous update, which doesn't expect values to be changed underneath. With this change, slab counters on parent level will become eventually consistent. Once all dying children are gone, values are correct. And if not, the error is capped by 32 * NR_CPUS pages per dying cgroup. It's not perfect, as slab are reparented, so any updates after the reparenting will happen on the parent level. It means that if a slab page was allocated, a counter on child level was bumped, then the page was reparented and freed, the annihilation of positive and negative counter values will not happen until the child cgroup is released. It makes slab counters different from others, and it might want us to implement flushing in a correct form again. But it's also a question of performance: scheduling a work on each cpu isn't free, and it's an open question if the benefit of having more accurate counters is worth it. We might also consider flushing all counters on offlining, not only slab counters. So let's fix the main problem now: make the slab counters eventually consistent, so at least the error won't grow with uptime (or more precisely the number of created and destroyed cgroups). And think about the accuracy of counters separately. Link: http://lkml.kernel.org/r/20191220042728.1045881-1-guro@fb.com Fixes: bee07b33db78 ("mm: memcontrol: flush percpu slab vmstats on kmem offlining") Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-01-14 00:29:16 +00:00
for (i = 0; i < MEMCG_NR_STAT; i++)
stat[i] += per_cpu(memcg->vmstats_percpu->stat[i], cpu);
mm: memcontrol: flush percpu vmstats before releasing memcg Percpu caching of local vmstats with the conditional propagation by the cgroup tree leads to an accumulation of errors on non-leaf levels. Let's imagine two nested memory cgroups A and A/B. Say, a process belonging to A/B allocates 100 pagecache pages on the CPU 0. The percpu cache will spill 3 times, so that 32*3=96 pages will be accounted to A/B and A atomic vmstat counters, 4 pages will remain in the percpu cache. Imagine A/B is nearby memory.max, so that every following allocation triggers a direct reclaim on the local CPU. Say, each such attempt will free 16 pages on a new cpu. That means every percpu cache will have -16 pages, except the first one, which will have 4 - 16 = -12. A/B and A atomic counters will not be touched at all. Now a user removes A/B. All percpu caches are freed and corresponding vmstat numbers are forgotten. A has 96 pages more than expected. As memory cgroups are created and destroyed, errors do accumulate. Even 1-2 pages differences can accumulate into large numbers. To fix this issue let's accumulate and propagate percpu vmstat values before releasing the memory cgroup. At this point these numbers are stable and cannot be changed. Since on cpu hotplug we do flush percpu vmstats anyway, we can iterate only over online cpus. Link: http://lkml.kernel.org/r/20190819202338.363363-2-guro@fb.com Fixes: 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty") Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-08-25 00:54:47 +00:00
for (mi = memcg; mi; mi = parent_mem_cgroup(mi))
mm: memcg/slab: fix percpu slab vmstats flushing Currently slab percpu vmstats are flushed twice: during the memcg offlining and just before freeing the memcg structure. Each time percpu counters are summed, added to the atomic counterparts and propagated up by the cgroup tree. The second flushing is required due to how recursive vmstats are implemented: counters are batched in percpu variables on a local level, and once a percpu value is crossing some predefined threshold, it spills over to atomic values on the local and each ascendant levels. It means that without flushing some numbers cached in percpu variables will be dropped on floor each time a cgroup is destroyed. And with uptime the error on upper levels might become noticeable. The first flushing aims to make counters on ancestor levels more precise. Dying cgroups may resume in the dying state for a long time. After kmem_cache reparenting which is performed during the offlining slab counters of the dying cgroup don't have any chances to be updated, because any slab operations will be performed on the parent level. It means that the inaccuracy caused by percpu batching will not decrease up to the final destruction of the cgroup. By the original idea flushing slab counters during the offlining should minimize the visible inaccuracy of slab counters on the parent level. The problem is that percpu counters are not zeroed after the first flushing. So every cached percpu value is summed twice. It creates a small error (up to 32 pages per cpu, but usually less) which accumulates on parent cgroup level. After creating and destroying of thousands of child cgroups, slab counter on parent level can be way off the real value. For now, let's just stop flushing slab counters on memcg offlining. It can't be done correctly without scheduling a work on each cpu: reading and zeroing it during css offlining can race with an asynchronous update, which doesn't expect values to be changed underneath. With this change, slab counters on parent level will become eventually consistent. Once all dying children are gone, values are correct. And if not, the error is capped by 32 * NR_CPUS pages per dying cgroup. It's not perfect, as slab are reparented, so any updates after the reparenting will happen on the parent level. It means that if a slab page was allocated, a counter on child level was bumped, then the page was reparented and freed, the annihilation of positive and negative counter values will not happen until the child cgroup is released. It makes slab counters different from others, and it might want us to implement flushing in a correct form again. But it's also a question of performance: scheduling a work on each cpu isn't free, and it's an open question if the benefit of having more accurate counters is worth it. We might also consider flushing all counters on offlining, not only slab counters. So let's fix the main problem now: make the slab counters eventually consistent, so at least the error won't grow with uptime (or more precisely the number of created and destroyed cgroups). And think about the accuracy of counters separately. Link: http://lkml.kernel.org/r/20191220042728.1045881-1-guro@fb.com Fixes: bee07b33db78 ("mm: memcontrol: flush percpu slab vmstats on kmem offlining") Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-01-14 00:29:16 +00:00
for (i = 0; i < MEMCG_NR_STAT; i++)
mm: memcontrol: flush percpu vmstats before releasing memcg Percpu caching of local vmstats with the conditional propagation by the cgroup tree leads to an accumulation of errors on non-leaf levels. Let's imagine two nested memory cgroups A and A/B. Say, a process belonging to A/B allocates 100 pagecache pages on the CPU 0. The percpu cache will spill 3 times, so that 32*3=96 pages will be accounted to A/B and A atomic vmstat counters, 4 pages will remain in the percpu cache. Imagine A/B is nearby memory.max, so that every following allocation triggers a direct reclaim on the local CPU. Say, each such attempt will free 16 pages on a new cpu. That means every percpu cache will have -16 pages, except the first one, which will have 4 - 16 = -12. A/B and A atomic counters will not be touched at all. Now a user removes A/B. All percpu caches are freed and corresponding vmstat numbers are forgotten. A has 96 pages more than expected. As memory cgroups are created and destroyed, errors do accumulate. Even 1-2 pages differences can accumulate into large numbers. To fix this issue let's accumulate and propagate percpu vmstat values before releasing the memory cgroup. At this point these numbers are stable and cannot be changed. Since on cpu hotplug we do flush percpu vmstats anyway, we can iterate only over online cpus. Link: http://lkml.kernel.org/r/20190819202338.363363-2-guro@fb.com Fixes: 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty") Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-08-25 00:54:47 +00:00
atomic_long_add(stat[i], &mi->vmstats[i]);
for_each_node(node) {
struct mem_cgroup_per_node *pn = memcg->nodeinfo[node];
struct mem_cgroup_per_node *pi;
mm: memcg/slab: fix percpu slab vmstats flushing Currently slab percpu vmstats are flushed twice: during the memcg offlining and just before freeing the memcg structure. Each time percpu counters are summed, added to the atomic counterparts and propagated up by the cgroup tree. The second flushing is required due to how recursive vmstats are implemented: counters are batched in percpu variables on a local level, and once a percpu value is crossing some predefined threshold, it spills over to atomic values on the local and each ascendant levels. It means that without flushing some numbers cached in percpu variables will be dropped on floor each time a cgroup is destroyed. And with uptime the error on upper levels might become noticeable. The first flushing aims to make counters on ancestor levels more precise. Dying cgroups may resume in the dying state for a long time. After kmem_cache reparenting which is performed during the offlining slab counters of the dying cgroup don't have any chances to be updated, because any slab operations will be performed on the parent level. It means that the inaccuracy caused by percpu batching will not decrease up to the final destruction of the cgroup. By the original idea flushing slab counters during the offlining should minimize the visible inaccuracy of slab counters on the parent level. The problem is that percpu counters are not zeroed after the first flushing. So every cached percpu value is summed twice. It creates a small error (up to 32 pages per cpu, but usually less) which accumulates on parent cgroup level. After creating and destroying of thousands of child cgroups, slab counter on parent level can be way off the real value. For now, let's just stop flushing slab counters on memcg offlining. It can't be done correctly without scheduling a work on each cpu: reading and zeroing it during css offlining can race with an asynchronous update, which doesn't expect values to be changed underneath. With this change, slab counters on parent level will become eventually consistent. Once all dying children are gone, values are correct. And if not, the error is capped by 32 * NR_CPUS pages per dying cgroup. It's not perfect, as slab are reparented, so any updates after the reparenting will happen on the parent level. It means that if a slab page was allocated, a counter on child level was bumped, then the page was reparented and freed, the annihilation of positive and negative counter values will not happen until the child cgroup is released. It makes slab counters different from others, and it might want us to implement flushing in a correct form again. But it's also a question of performance: scheduling a work on each cpu isn't free, and it's an open question if the benefit of having more accurate counters is worth it. We might also consider flushing all counters on offlining, not only slab counters. So let's fix the main problem now: make the slab counters eventually consistent, so at least the error won't grow with uptime (or more precisely the number of created and destroyed cgroups). And think about the accuracy of counters separately. Link: http://lkml.kernel.org/r/20191220042728.1045881-1-guro@fb.com Fixes: bee07b33db78 ("mm: memcontrol: flush percpu slab vmstats on kmem offlining") Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-01-14 00:29:16 +00:00
for (i = 0; i < NR_VM_NODE_STAT_ITEMS; i++)
mm: memcontrol: flush percpu vmstats before releasing memcg Percpu caching of local vmstats with the conditional propagation by the cgroup tree leads to an accumulation of errors on non-leaf levels. Let's imagine two nested memory cgroups A and A/B. Say, a process belonging to A/B allocates 100 pagecache pages on the CPU 0. The percpu cache will spill 3 times, so that 32*3=96 pages will be accounted to A/B and A atomic vmstat counters, 4 pages will remain in the percpu cache. Imagine A/B is nearby memory.max, so that every following allocation triggers a direct reclaim on the local CPU. Say, each such attempt will free 16 pages on a new cpu. That means every percpu cache will have -16 pages, except the first one, which will have 4 - 16 = -12. A/B and A atomic counters will not be touched at all. Now a user removes A/B. All percpu caches are freed and corresponding vmstat numbers are forgotten. A has 96 pages more than expected. As memory cgroups are created and destroyed, errors do accumulate. Even 1-2 pages differences can accumulate into large numbers. To fix this issue let's accumulate and propagate percpu vmstat values before releasing the memory cgroup. At this point these numbers are stable and cannot be changed. Since on cpu hotplug we do flush percpu vmstats anyway, we can iterate only over online cpus. Link: http://lkml.kernel.org/r/20190819202338.363363-2-guro@fb.com Fixes: 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty") Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-08-25 00:54:47 +00:00
stat[i] = 0;
for_each_online_cpu(cpu)
mm: memcg/slab: fix percpu slab vmstats flushing Currently slab percpu vmstats are flushed twice: during the memcg offlining and just before freeing the memcg structure. Each time percpu counters are summed, added to the atomic counterparts and propagated up by the cgroup tree. The second flushing is required due to how recursive vmstats are implemented: counters are batched in percpu variables on a local level, and once a percpu value is crossing some predefined threshold, it spills over to atomic values on the local and each ascendant levels. It means that without flushing some numbers cached in percpu variables will be dropped on floor each time a cgroup is destroyed. And with uptime the error on upper levels might become noticeable. The first flushing aims to make counters on ancestor levels more precise. Dying cgroups may resume in the dying state for a long time. After kmem_cache reparenting which is performed during the offlining slab counters of the dying cgroup don't have any chances to be updated, because any slab operations will be performed on the parent level. It means that the inaccuracy caused by percpu batching will not decrease up to the final destruction of the cgroup. By the original idea flushing slab counters during the offlining should minimize the visible inaccuracy of slab counters on the parent level. The problem is that percpu counters are not zeroed after the first flushing. So every cached percpu value is summed twice. It creates a small error (up to 32 pages per cpu, but usually less) which accumulates on parent cgroup level. After creating and destroying of thousands of child cgroups, slab counter on parent level can be way off the real value. For now, let's just stop flushing slab counters on memcg offlining. It can't be done correctly without scheduling a work on each cpu: reading and zeroing it during css offlining can race with an asynchronous update, which doesn't expect values to be changed underneath. With this change, slab counters on parent level will become eventually consistent. Once all dying children are gone, values are correct. And if not, the error is capped by 32 * NR_CPUS pages per dying cgroup. It's not perfect, as slab are reparented, so any updates after the reparenting will happen on the parent level. It means that if a slab page was allocated, a counter on child level was bumped, then the page was reparented and freed, the annihilation of positive and negative counter values will not happen until the child cgroup is released. It makes slab counters different from others, and it might want us to implement flushing in a correct form again. But it's also a question of performance: scheduling a work on each cpu isn't free, and it's an open question if the benefit of having more accurate counters is worth it. We might also consider flushing all counters on offlining, not only slab counters. So let's fix the main problem now: make the slab counters eventually consistent, so at least the error won't grow with uptime (or more precisely the number of created and destroyed cgroups). And think about the accuracy of counters separately. Link: http://lkml.kernel.org/r/20191220042728.1045881-1-guro@fb.com Fixes: bee07b33db78 ("mm: memcontrol: flush percpu slab vmstats on kmem offlining") Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-01-14 00:29:16 +00:00
for (i = 0; i < NR_VM_NODE_STAT_ITEMS; i++)
stat[i] += per_cpu(
pn->lruvec_stat_cpu->count[i], cpu);
mm: memcontrol: flush percpu vmstats before releasing memcg Percpu caching of local vmstats with the conditional propagation by the cgroup tree leads to an accumulation of errors on non-leaf levels. Let's imagine two nested memory cgroups A and A/B. Say, a process belonging to A/B allocates 100 pagecache pages on the CPU 0. The percpu cache will spill 3 times, so that 32*3=96 pages will be accounted to A/B and A atomic vmstat counters, 4 pages will remain in the percpu cache. Imagine A/B is nearby memory.max, so that every following allocation triggers a direct reclaim on the local CPU. Say, each such attempt will free 16 pages on a new cpu. That means every percpu cache will have -16 pages, except the first one, which will have 4 - 16 = -12. A/B and A atomic counters will not be touched at all. Now a user removes A/B. All percpu caches are freed and corresponding vmstat numbers are forgotten. A has 96 pages more than expected. As memory cgroups are created and destroyed, errors do accumulate. Even 1-2 pages differences can accumulate into large numbers. To fix this issue let's accumulate and propagate percpu vmstat values before releasing the memory cgroup. At this point these numbers are stable and cannot be changed. Since on cpu hotplug we do flush percpu vmstats anyway, we can iterate only over online cpus. Link: http://lkml.kernel.org/r/20190819202338.363363-2-guro@fb.com Fixes: 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty") Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-08-25 00:54:47 +00:00
for (pi = pn; pi; pi = parent_nodeinfo(pi, node))
mm: memcg/slab: fix percpu slab vmstats flushing Currently slab percpu vmstats are flushed twice: during the memcg offlining and just before freeing the memcg structure. Each time percpu counters are summed, added to the atomic counterparts and propagated up by the cgroup tree. The second flushing is required due to how recursive vmstats are implemented: counters are batched in percpu variables on a local level, and once a percpu value is crossing some predefined threshold, it spills over to atomic values on the local and each ascendant levels. It means that without flushing some numbers cached in percpu variables will be dropped on floor each time a cgroup is destroyed. And with uptime the error on upper levels might become noticeable. The first flushing aims to make counters on ancestor levels more precise. Dying cgroups may resume in the dying state for a long time. After kmem_cache reparenting which is performed during the offlining slab counters of the dying cgroup don't have any chances to be updated, because any slab operations will be performed on the parent level. It means that the inaccuracy caused by percpu batching will not decrease up to the final destruction of the cgroup. By the original idea flushing slab counters during the offlining should minimize the visible inaccuracy of slab counters on the parent level. The problem is that percpu counters are not zeroed after the first flushing. So every cached percpu value is summed twice. It creates a small error (up to 32 pages per cpu, but usually less) which accumulates on parent cgroup level. After creating and destroying of thousands of child cgroups, slab counter on parent level can be way off the real value. For now, let's just stop flushing slab counters on memcg offlining. It can't be done correctly without scheduling a work on each cpu: reading and zeroing it during css offlining can race with an asynchronous update, which doesn't expect values to be changed underneath. With this change, slab counters on parent level will become eventually consistent. Once all dying children are gone, values are correct. And if not, the error is capped by 32 * NR_CPUS pages per dying cgroup. It's not perfect, as slab are reparented, so any updates after the reparenting will happen on the parent level. It means that if a slab page was allocated, a counter on child level was bumped, then the page was reparented and freed, the annihilation of positive and negative counter values will not happen until the child cgroup is released. It makes slab counters different from others, and it might want us to implement flushing in a correct form again. But it's also a question of performance: scheduling a work on each cpu isn't free, and it's an open question if the benefit of having more accurate counters is worth it. We might also consider flushing all counters on offlining, not only slab counters. So let's fix the main problem now: make the slab counters eventually consistent, so at least the error won't grow with uptime (or more precisely the number of created and destroyed cgroups). And think about the accuracy of counters separately. Link: http://lkml.kernel.org/r/20191220042728.1045881-1-guro@fb.com Fixes: bee07b33db78 ("mm: memcontrol: flush percpu slab vmstats on kmem offlining") Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-01-14 00:29:16 +00:00
for (i = 0; i < NR_VM_NODE_STAT_ITEMS; i++)
mm: memcontrol: flush percpu vmstats before releasing memcg Percpu caching of local vmstats with the conditional propagation by the cgroup tree leads to an accumulation of errors on non-leaf levels. Let's imagine two nested memory cgroups A and A/B. Say, a process belonging to A/B allocates 100 pagecache pages on the CPU 0. The percpu cache will spill 3 times, so that 32*3=96 pages will be accounted to A/B and A atomic vmstat counters, 4 pages will remain in the percpu cache. Imagine A/B is nearby memory.max, so that every following allocation triggers a direct reclaim on the local CPU. Say, each such attempt will free 16 pages on a new cpu. That means every percpu cache will have -16 pages, except the first one, which will have 4 - 16 = -12. A/B and A atomic counters will not be touched at all. Now a user removes A/B. All percpu caches are freed and corresponding vmstat numbers are forgotten. A has 96 pages more than expected. As memory cgroups are created and destroyed, errors do accumulate. Even 1-2 pages differences can accumulate into large numbers. To fix this issue let's accumulate and propagate percpu vmstat values before releasing the memory cgroup. At this point these numbers are stable and cannot be changed. Since on cpu hotplug we do flush percpu vmstats anyway, we can iterate only over online cpus. Link: http://lkml.kernel.org/r/20190819202338.363363-2-guro@fb.com Fixes: 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty") Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-08-25 00:54:47 +00:00
atomic_long_add(stat[i], &pi->lruvec_stat[i]);
}
}
static void memcg_flush_percpu_vmevents(struct mem_cgroup *memcg)
{
unsigned long events[NR_VM_EVENT_ITEMS];
struct mem_cgroup *mi;
int cpu, i;
for (i = 0; i < NR_VM_EVENT_ITEMS; i++)
events[i] = 0;
for_each_online_cpu(cpu)
for (i = 0; i < NR_VM_EVENT_ITEMS; i++)
events[i] += per_cpu(memcg->vmstats_percpu->events[i],
cpu);
for (mi = memcg; mi; mi = parent_mem_cgroup(mi))
for (i = 0; i < NR_VM_EVENT_ITEMS; i++)
atomic_long_add(events[i], &mi->vmevents[i]);
}
mm: introduce CONFIG_MEMCG_KMEM as combination of CONFIG_MEMCG && !CONFIG_SLOB Introduce new config option, which is used to replace repeating CONFIG_MEMCG && !CONFIG_SLOB pattern. Next patches add a little more memcg+kmem related code, so let's keep the defines more clearly. Link: http://lkml.kernel.org/r/153063053670.1818.15013136946600481138.stgit@localhost.localdomain Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Tested-by: Shakeel Butt <shakeelb@google.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Guenter Roeck <linux@roeck-us.net> Cc: "Huang, Ying" <ying.huang@intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Josef Bacik <jbacik@fb.com> Cc: Li RongQing <lirongqing@baidu.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Matthias Kaehlcke <mka@chromium.org> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Michal Hocko <mhocko@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Philippe Ombredanne <pombredanne@nexb.com> Cc: Roman Gushchin <guro@fb.com> Cc: Sahitya Tummala <stummala@codeaurora.org> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Waiman Long <longman@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:25 +00:00
#ifdef CONFIG_MEMCG_KMEM
static int memcg_online_kmem(struct mem_cgroup *memcg)
memcg: rework memcg_update_kmem_limit synchronization Currently we take both the memcg_create_mutex and the set_limit_mutex when we enable kmem accounting for a memory cgroup, which makes kmem activation events serialize with both memcg creations and other memcg limit updates (memory.limit, memory.memsw.limit). However, there is no point in such strict synchronization rules there. First, the set_limit_mutex was introduced to keep the memory.limit and memory.memsw.limit values in sync. Since memory.kmem.limit can be set independently of them, it is better to introduce a separate mutex to synchronize against concurrent kmem limit updates. Second, we take the memcg_create_mutex in order to make sure all children of this memcg will be kmem-active as well. For achieving that, it is enough to hold this mutex only while checking if memcg_has_children() though. This guarantees that if a child is added after we checked that the memcg has no children, the newly added cgroup will see its parent kmem-active (of course if the latter succeeded), and call kmem activation for itself. This patch simplifies the locking rules of memcg_update_kmem_limit() according to these considerations. [vdavydov@parallels.com: fix unintialized var warning] Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Glauber Costa <glommer@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-23 23:53:09 +00:00
{
int memcg_id;
mm: memcontrol: enable kmem accounting for all cgroups in the legacy hierarchy Workingset code was recently made memcg aware, but shadow node shrinker is still global. As a result, one small cgroup can consume all memory available for shadow nodes, possibly hurting other cgroups by reclaiming their shadow nodes, even though reclaim distances stored in its shadow nodes have no effect. To avoid this, we need to make shadow node shrinker memcg aware. The actual work is done in patch 6 of the series. Patches 1 and 2 prepare memcg/shrinker infrastructure for the change. Patch 3 is just a collateral cleanup. Patch 4 makes radix_tree_node accounted, which is necessary for making shadow node shrinker memcg aware. Patch 5 reduces shadow nodes overhead in case workload mostly uses anonymous pages. This patch: Currently, in the legacy hierarchy kmem accounting is off for all cgroups by default and must be enabled explicitly by writing something to memory.kmem.limit_in_bytes. Since we don't support reclaim on hitting kmem limit, nor do we have any plans to implement it, this is likely to be -1, just to enable kmem accounting and limit kernel memory consumption by the memory.limit_in_bytes along with user memory. This user API was introduced when the implementation of kmem accounting lacked slab shrinker support and hence was useless in practice. Things have changed since then - slab shrinkers were made memcg aware, the accounting overhead seems to be negligible, and a failure to charge a kmem allocation should not have critical consequences, because we only account those kernel objects that should be safe to fail. That's why kmem accounting is enabled by default for all cgroups in the default hierarchy, which will eventually replace the legacy one. The ability to enable kmem accounting for some cgroups while keeping it disabled for others is getting difficult to maintain. E.g. to make shadow node shrinker memcg aware (see mm/workingset.c), we need to know the relationship between the number of shadow nodes allocated for a cgroup and the size of its lru list. If kmem accounting is enabled for all cgroups there is no problem, but what should we do if kmem accounting is enabled only for half of cgroups? We've no other choice but use global lru stats while scanning root cgroup's shadow nodes, but that would be wrong if kmem accounting was enabled for all cgroups (which is the case if the unified hierarchy is used), in which case we should use lru stats of the root cgroup's lruvec. That being said, let's enable kmem accounting for all memory cgroups by default. If one finds it unstable or too costly, it can always be disabled system-wide by passing cgroup.memory=nokmem to the kernel at boot time. Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-03-17 21:18:27 +00:00
if (cgroup_memory_nokmem)
return 0;
BUG_ON(memcg->kmemcg_id >= 0);
BUG_ON(memcg->kmem_state);
memcg: rework memcg_update_kmem_limit synchronization Currently we take both the memcg_create_mutex and the set_limit_mutex when we enable kmem accounting for a memory cgroup, which makes kmem activation events serialize with both memcg creations and other memcg limit updates (memory.limit, memory.memsw.limit). However, there is no point in such strict synchronization rules there. First, the set_limit_mutex was introduced to keep the memory.limit and memory.memsw.limit values in sync. Since memory.kmem.limit can be set independently of them, it is better to introduce a separate mutex to synchronize against concurrent kmem limit updates. Second, we take the memcg_create_mutex in order to make sure all children of this memcg will be kmem-active as well. For achieving that, it is enough to hold this mutex only while checking if memcg_has_children() though. This guarantees that if a child is added after we checked that the memcg has no children, the newly added cgroup will see its parent kmem-active (of course if the latter succeeded), and call kmem activation for itself. This patch simplifies the locking rules of memcg_update_kmem_limit() according to these considerations. [vdavydov@parallels.com: fix unintialized var warning] Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Glauber Costa <glommer@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-23 23:53:09 +00:00
memcg_id = memcg_alloc_cache_id();
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
if (memcg_id < 0)
return memcg_id;
memcg: rework memcg_update_kmem_limit synchronization Currently we take both the memcg_create_mutex and the set_limit_mutex when we enable kmem accounting for a memory cgroup, which makes kmem activation events serialize with both memcg creations and other memcg limit updates (memory.limit, memory.memsw.limit). However, there is no point in such strict synchronization rules there. First, the set_limit_mutex was introduced to keep the memory.limit and memory.memsw.limit values in sync. Since memory.kmem.limit can be set independently of them, it is better to introduce a separate mutex to synchronize against concurrent kmem limit updates. Second, we take the memcg_create_mutex in order to make sure all children of this memcg will be kmem-active as well. For achieving that, it is enough to hold this mutex only while checking if memcg_has_children() though. This guarantees that if a child is added after we checked that the memcg has no children, the newly added cgroup will see its parent kmem-active (of course if the latter succeeded), and call kmem activation for itself. This patch simplifies the locking rules of memcg_update_kmem_limit() according to these considerations. [vdavydov@parallels.com: fix unintialized var warning] Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Glauber Costa <glommer@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-23 23:53:09 +00:00
static_branch_inc(&memcg_kmem_enabled_key);
memcg: rework memcg_update_kmem_limit synchronization Currently we take both the memcg_create_mutex and the set_limit_mutex when we enable kmem accounting for a memory cgroup, which makes kmem activation events serialize with both memcg creations and other memcg limit updates (memory.limit, memory.memsw.limit). However, there is no point in such strict synchronization rules there. First, the set_limit_mutex was introduced to keep the memory.limit and memory.memsw.limit values in sync. Since memory.kmem.limit can be set independently of them, it is better to introduce a separate mutex to synchronize against concurrent kmem limit updates. Second, we take the memcg_create_mutex in order to make sure all children of this memcg will be kmem-active as well. For achieving that, it is enough to hold this mutex only while checking if memcg_has_children() though. This guarantees that if a child is added after we checked that the memcg has no children, the newly added cgroup will see its parent kmem-active (of course if the latter succeeded), and call kmem activation for itself. This patch simplifies the locking rules of memcg_update_kmem_limit() according to these considerations. [vdavydov@parallels.com: fix unintialized var warning] Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Glauber Costa <glommer@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-23 23:53:09 +00:00
/*
* A memory cgroup is considered kmem-online as soon as it gets
* kmemcg_id. Setting the id after enabling static branching will
memcg: rework memcg_update_kmem_limit synchronization Currently we take both the memcg_create_mutex and the set_limit_mutex when we enable kmem accounting for a memory cgroup, which makes kmem activation events serialize with both memcg creations and other memcg limit updates (memory.limit, memory.memsw.limit). However, there is no point in such strict synchronization rules there. First, the set_limit_mutex was introduced to keep the memory.limit and memory.memsw.limit values in sync. Since memory.kmem.limit can be set independently of them, it is better to introduce a separate mutex to synchronize against concurrent kmem limit updates. Second, we take the memcg_create_mutex in order to make sure all children of this memcg will be kmem-active as well. For achieving that, it is enough to hold this mutex only while checking if memcg_has_children() though. This guarantees that if a child is added after we checked that the memcg has no children, the newly added cgroup will see its parent kmem-active (of course if the latter succeeded), and call kmem activation for itself. This patch simplifies the locking rules of memcg_update_kmem_limit() according to these considerations. [vdavydov@parallels.com: fix unintialized var warning] Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Glauber Costa <glommer@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-23 23:53:09 +00:00
* guarantee no one starts accounting before all call sites are
* patched.
*/
memcg->kmemcg_id = memcg_id;
memcg->kmem_state = KMEM_ONLINE;
slab: link memcg kmem_caches on their associated memory cgroup With kmem cgroup support enabled, kmem_caches can be created and destroyed frequently and a great number of near empty kmem_caches can accumulate if there are a lot of transient cgroups and the system is not under memory pressure. When memory reclaim starts under such conditions, it can lead to consecutive deactivation and destruction of many kmem_caches, easily hundreds of thousands on moderately large systems, exposing scalability issues in the current slab management code. This is one of the patches to address the issue. While a memcg kmem_cache is listed on its root cache's ->children list, there is no direct way to iterate all kmem_caches which are assocaited with a memory cgroup. The only way to iterate them is walking all caches while filtering out caches which don't match, which would be most of them. This makes memcg destruction operations O(N^2) where N is the total number of slab caches which can be huge. This combined with the synchronous RCU operations can tie up a CPU and affect the whole machine for many hours when memory reclaim triggers offlining and destruction of the stale memcgs. This patch adds mem_cgroup->kmem_caches list which goes through memcg_cache_params->kmem_caches_node of all kmem_caches which are associated with the memcg. All memcg specific iterations, including stat file access, are updated to use the new list instead. Link: http://lkml.kernel.org/r/20170117235411.9408-6-tj@kernel.org Signed-off-by: Tejun Heo <tj@kernel.org> Reported-by: Jay Vana <jsvana@fb.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-02-22 23:41:21 +00:00
INIT_LIST_HEAD(&memcg->kmem_caches);
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
return 0;
memcg: rework memcg_update_kmem_limit synchronization Currently we take both the memcg_create_mutex and the set_limit_mutex when we enable kmem accounting for a memory cgroup, which makes kmem activation events serialize with both memcg creations and other memcg limit updates (memory.limit, memory.memsw.limit). However, there is no point in such strict synchronization rules there. First, the set_limit_mutex was introduced to keep the memory.limit and memory.memsw.limit values in sync. Since memory.kmem.limit can be set independently of them, it is better to introduce a separate mutex to synchronize against concurrent kmem limit updates. Second, we take the memcg_create_mutex in order to make sure all children of this memcg will be kmem-active as well. For achieving that, it is enough to hold this mutex only while checking if memcg_has_children() though. This guarantees that if a child is added after we checked that the memcg has no children, the newly added cgroup will see its parent kmem-active (of course if the latter succeeded), and call kmem activation for itself. This patch simplifies the locking rules of memcg_update_kmem_limit() according to these considerations. [vdavydov@parallels.com: fix unintialized var warning] Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Glauber Costa <glommer@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-23 23:53:09 +00:00
}
static void memcg_offline_kmem(struct mem_cgroup *memcg)
{
struct cgroup_subsys_state *css;
struct mem_cgroup *parent, *child;
int kmemcg_id;
if (memcg->kmem_state != KMEM_ONLINE)
return;
/*
* Clear the online state before clearing memcg_caches array
* entries. The slab_mutex in memcg_deactivate_kmem_caches()
* guarantees that no cache will be created for this cgroup
* after we are done (see memcg_create_kmem_cache()).
*/
memcg->kmem_state = KMEM_ALLOCATED;
parent = parent_mem_cgroup(memcg);
if (!parent)
parent = root_mem_cgroup;
/*
mm: memcg/slab: fix percpu slab vmstats flushing Currently slab percpu vmstats are flushed twice: during the memcg offlining and just before freeing the memcg structure. Each time percpu counters are summed, added to the atomic counterparts and propagated up by the cgroup tree. The second flushing is required due to how recursive vmstats are implemented: counters are batched in percpu variables on a local level, and once a percpu value is crossing some predefined threshold, it spills over to atomic values on the local and each ascendant levels. It means that without flushing some numbers cached in percpu variables will be dropped on floor each time a cgroup is destroyed. And with uptime the error on upper levels might become noticeable. The first flushing aims to make counters on ancestor levels more precise. Dying cgroups may resume in the dying state for a long time. After kmem_cache reparenting which is performed during the offlining slab counters of the dying cgroup don't have any chances to be updated, because any slab operations will be performed on the parent level. It means that the inaccuracy caused by percpu batching will not decrease up to the final destruction of the cgroup. By the original idea flushing slab counters during the offlining should minimize the visible inaccuracy of slab counters on the parent level. The problem is that percpu counters are not zeroed after the first flushing. So every cached percpu value is summed twice. It creates a small error (up to 32 pages per cpu, but usually less) which accumulates on parent cgroup level. After creating and destroying of thousands of child cgroups, slab counter on parent level can be way off the real value. For now, let's just stop flushing slab counters on memcg offlining. It can't be done correctly without scheduling a work on each cpu: reading and zeroing it during css offlining can race with an asynchronous update, which doesn't expect values to be changed underneath. With this change, slab counters on parent level will become eventually consistent. Once all dying children are gone, values are correct. And if not, the error is capped by 32 * NR_CPUS pages per dying cgroup. It's not perfect, as slab are reparented, so any updates after the reparenting will happen on the parent level. It means that if a slab page was allocated, a counter on child level was bumped, then the page was reparented and freed, the annihilation of positive and negative counter values will not happen until the child cgroup is released. It makes slab counters different from others, and it might want us to implement flushing in a correct form again. But it's also a question of performance: scheduling a work on each cpu isn't free, and it's an open question if the benefit of having more accurate counters is worth it. We might also consider flushing all counters on offlining, not only slab counters. So let's fix the main problem now: make the slab counters eventually consistent, so at least the error won't grow with uptime (or more precisely the number of created and destroyed cgroups). And think about the accuracy of counters separately. Link: http://lkml.kernel.org/r/20191220042728.1045881-1-guro@fb.com Fixes: bee07b33db78 ("mm: memcontrol: flush percpu slab vmstats on kmem offlining") Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-01-14 00:29:16 +00:00
* Deactivate and reparent kmem_caches.
*/
mm: memcg/slab: reparent memcg kmem_caches on cgroup removal Let's reparent non-root kmem_caches on memcg offlining. This allows us to release the memory cgroup without waiting for the last outstanding kernel object (e.g. dentry used by another application). Since the parent cgroup is already charged, everything we need to do is to splice the list of kmem_caches to the parent's kmem_caches list, swap the memcg pointer, drop the css refcounter for each kmem_cache and adjust the parent's css refcounter. Please, note that kmem_cache->memcg_params.memcg isn't a stable pointer anymore. It's safe to read it under rcu_read_lock(), cgroup_mutex held, or any other way that protects the memory cgroup from being released. We can race with the slab allocation and deallocation paths. It's not a big problem: parent's charge and slab global stats are always correct, and we don't care anymore about the child usage and global stats. The child cgroup is already offline, so we don't use or show it anywhere. Local slab stats (NR_SLAB_RECLAIMABLE and NR_SLAB_UNRECLAIMABLE) aren't used anywhere except count_shadow_nodes(). But even there it won't break anything: after reparenting "nodes" will be 0 on child level (because we're already reparenting shrinker lists), and on parent level page stats always were 0, and this patch won't change anything. [guro@fb.com: properly handle kmem_caches reparented to root_mem_cgroup] Link: http://lkml.kernel.org/r/20190620213427.1691847-1-guro@fb.com Link: http://lkml.kernel.org/r/20190611231813.3148843-11-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Reviewed-by: Shakeel Butt <shakeelb@google.com> Acked-by: David Rientjes <rientjes@google.com> Cc: Christoph Lameter <cl@linux.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Waiman Long <longman@redhat.com> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: Andrei Vagin <avagin@gmail.com> Cc: Qian Cai <cai@lca.pw> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:56:34 +00:00
memcg_deactivate_kmem_caches(memcg, parent);
kmemcg_id = memcg->kmemcg_id;
BUG_ON(kmemcg_id < 0);
/*
* Change kmemcg_id of this cgroup and all its descendants to the
* parent's id, and then move all entries from this cgroup's list_lrus
* to ones of the parent. After we have finished, all list_lrus
* corresponding to this cgroup are guaranteed to remain empty. The
* ordering is imposed by list_lru_node->lock taken by
* memcg_drain_all_list_lrus().
*/
memcg: add RCU locking around css_for_each_descendant_pre() in memcg_offline_kmem() memcg_offline_kmem() may be called from memcg_free_kmem() after a css init failure. memcg_free_kmem() is a ->css_free callback which is called without cgroup_mutex and memcg_offline_kmem() ends up using css_for_each_descendant_pre() without any locking. Fix it by adding rcu read locking around it. mkdir: cannot create directory `65530': No space left on device =============================== [ INFO: suspicious RCU usage. ] 4.6.0-work+ #321 Not tainted ------------------------------- kernel/cgroup.c:4008 cgroup_mutex or RCU read lock required! [ 527.243970] other info that might help us debug this: [ 527.244715] rcu_scheduler_active = 1, debug_locks = 0 2 locks held by kworker/0:5/1664: #0: ("cgroup_destroy"){.+.+..}, at: [<ffffffff81060ab5>] process_one_work+0x165/0x4a0 #1: ((&css->destroy_work)#3){+.+...}, at: [<ffffffff81060ab5>] process_one_work+0x165/0x4a0 [ 527.248098] stack backtrace: CPU: 0 PID: 1664 Comm: kworker/0:5 Not tainted 4.6.0-work+ #321 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.9.1-1.fc24 04/01/2014 Workqueue: cgroup_destroy css_free_work_fn Call Trace: dump_stack+0x68/0xa1 lockdep_rcu_suspicious+0xd7/0x110 css_next_descendant_pre+0x7d/0xb0 memcg_offline_kmem.part.44+0x4a/0xc0 mem_cgroup_css_free+0x1ec/0x200 css_free_work_fn+0x49/0x5e0 process_one_work+0x1c5/0x4a0 worker_thread+0x49/0x490 kthread+0xea/0x100 ret_from_fork+0x1f/0x40 Link: http://lkml.kernel.org/r/20160526203018.GG23194@mtj.duckdns.org Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: <stable@vger.kernel.org> [4.5+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-06-03 21:55:44 +00:00
rcu_read_lock(); /* can be called from css_free w/o cgroup_mutex */
css_for_each_descendant_pre(css, &memcg->css) {
child = mem_cgroup_from_css(css);
BUG_ON(child->kmemcg_id != kmemcg_id);
child->kmemcg_id = parent->kmemcg_id;
if (!memcg->use_hierarchy)
break;
}
memcg: add RCU locking around css_for_each_descendant_pre() in memcg_offline_kmem() memcg_offline_kmem() may be called from memcg_free_kmem() after a css init failure. memcg_free_kmem() is a ->css_free callback which is called without cgroup_mutex and memcg_offline_kmem() ends up using css_for_each_descendant_pre() without any locking. Fix it by adding rcu read locking around it. mkdir: cannot create directory `65530': No space left on device =============================== [ INFO: suspicious RCU usage. ] 4.6.0-work+ #321 Not tainted ------------------------------- kernel/cgroup.c:4008 cgroup_mutex or RCU read lock required! [ 527.243970] other info that might help us debug this: [ 527.244715] rcu_scheduler_active = 1, debug_locks = 0 2 locks held by kworker/0:5/1664: #0: ("cgroup_destroy"){.+.+..}, at: [<ffffffff81060ab5>] process_one_work+0x165/0x4a0 #1: ((&css->destroy_work)#3){+.+...}, at: [<ffffffff81060ab5>] process_one_work+0x165/0x4a0 [ 527.248098] stack backtrace: CPU: 0 PID: 1664 Comm: kworker/0:5 Not tainted 4.6.0-work+ #321 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.9.1-1.fc24 04/01/2014 Workqueue: cgroup_destroy css_free_work_fn Call Trace: dump_stack+0x68/0xa1 lockdep_rcu_suspicious+0xd7/0x110 css_next_descendant_pre+0x7d/0xb0 memcg_offline_kmem.part.44+0x4a/0xc0 mem_cgroup_css_free+0x1ec/0x200 css_free_work_fn+0x49/0x5e0 process_one_work+0x1c5/0x4a0 worker_thread+0x49/0x490 kthread+0xea/0x100 ret_from_fork+0x1f/0x40 Link: http://lkml.kernel.org/r/20160526203018.GG23194@mtj.duckdns.org Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: <stable@vger.kernel.org> [4.5+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-06-03 21:55:44 +00:00
rcu_read_unlock();
memcg_drain_all_list_lrus(kmemcg_id, parent);
memcg_free_cache_id(kmemcg_id);
}
static void memcg_free_kmem(struct mem_cgroup *memcg)
{
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
/* css_alloc() failed, offlining didn't happen */
if (unlikely(memcg->kmem_state == KMEM_ONLINE))
memcg_offline_kmem(memcg);
if (memcg->kmem_state == KMEM_ALLOCATED) {
mm: memcg/slab: rework non-root kmem_cache lifecycle management Currently each charged slab page holds a reference to the cgroup to which it's charged. Kmem_caches are held by the memcg and are released all together with the memory cgroup. It means that none of kmem_caches are released unless at least one reference to the memcg exists, which is very far from optimal. Let's rework it in a way that allows releasing individual kmem_caches as soon as the cgroup is offline, the kmem_cache is empty and there are no pending allocations. To make it possible, let's introduce a new percpu refcounter for non-root kmem caches. The counter is initialized to the percpu mode, and is switched to the atomic mode during kmem_cache deactivation. The counter is bumped for every charged page and also for every running allocation. So the kmem_cache can't be released unless all allocations complete. To shutdown non-active empty kmem_caches, let's reuse the work queue, previously used for the kmem_cache deactivation. Once the reference counter reaches 0, let's schedule an asynchronous kmem_cache release. * I used the following simple approach to test the performance (stolen from another patchset by T. Harding): time find / -name fname-no-exist echo 2 > /proc/sys/vm/drop_caches repeat 10 times Results: orig patched real 0m1.455s real 0m1.355s user 0m0.206s user 0m0.219s sys 0m0.855s sys 0m0.807s real 0m1.487s real 0m1.699s user 0m0.221s user 0m0.256s sys 0m0.806s sys 0m0.948s real 0m1.515s real 0m1.505s user 0m0.183s user 0m0.215s sys 0m0.876s sys 0m0.858s real 0m1.291s real 0m1.380s user 0m0.193s user 0m0.198s sys 0m0.843s sys 0m0.786s real 0m1.364s real 0m1.374s user 0m0.180s user 0m0.182s sys 0m0.868s sys 0m0.806s real 0m1.352s real 0m1.312s user 0m0.201s user 0m0.212s sys 0m0.820s sys 0m0.761s real 0m1.302s real 0m1.349s user 0m0.205s user 0m0.203s sys 0m0.803s sys 0m0.792s real 0m1.334s real 0m1.301s user 0m0.194s user 0m0.201s sys 0m0.806s sys 0m0.779s real 0m1.426s real 0m1.434s user 0m0.216s user 0m0.181s sys 0m0.824s sys 0m0.864s real 0m1.350s real 0m1.295s user 0m0.200s user 0m0.190s sys 0m0.842s sys 0m0.811s So it looks like the difference is not noticeable in this test. [cai@lca.pw: fix an use-after-free in kmemcg_workfn()] Link: http://lkml.kernel.org/r/1560977573-10715-1-git-send-email-cai@lca.pw Link: http://lkml.kernel.org/r/20190611231813.3148843-9-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Signed-off-by: Qian Cai <cai@lca.pw> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Christoph Lameter <cl@linux.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Waiman Long <longman@redhat.com> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: Andrei Vagin <avagin@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:56:27 +00:00
WARN_ON(!list_empty(&memcg->kmem_caches));
static_branch_dec(&memcg_kmem_enabled_key);
}
}
memcg: rework memcg_update_kmem_limit synchronization Currently we take both the memcg_create_mutex and the set_limit_mutex when we enable kmem accounting for a memory cgroup, which makes kmem activation events serialize with both memcg creations and other memcg limit updates (memory.limit, memory.memsw.limit). However, there is no point in such strict synchronization rules there. First, the set_limit_mutex was introduced to keep the memory.limit and memory.memsw.limit values in sync. Since memory.kmem.limit can be set independently of them, it is better to introduce a separate mutex to synchronize against concurrent kmem limit updates. Second, we take the memcg_create_mutex in order to make sure all children of this memcg will be kmem-active as well. For achieving that, it is enough to hold this mutex only while checking if memcg_has_children() though. This guarantees that if a child is added after we checked that the memcg has no children, the newly added cgroup will see its parent kmem-active (of course if the latter succeeded), and call kmem activation for itself. This patch simplifies the locking rules of memcg_update_kmem_limit() according to these considerations. [vdavydov@parallels.com: fix unintialized var warning] Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Glauber Costa <glommer@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-23 23:53:09 +00:00
#else
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
static int memcg_online_kmem(struct mem_cgroup *memcg)
{
return 0;
}
static void memcg_offline_kmem(struct mem_cgroup *memcg)
{
}
static void memcg_free_kmem(struct mem_cgroup *memcg)
{
}
mm: introduce CONFIG_MEMCG_KMEM as combination of CONFIG_MEMCG && !CONFIG_SLOB Introduce new config option, which is used to replace repeating CONFIG_MEMCG && !CONFIG_SLOB pattern. Next patches add a little more memcg+kmem related code, so let's keep the defines more clearly. Link: http://lkml.kernel.org/r/153063053670.1818.15013136946600481138.stgit@localhost.localdomain Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Tested-by: Shakeel Butt <shakeelb@google.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Guenter Roeck <linux@roeck-us.net> Cc: "Huang, Ying" <ying.huang@intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Josef Bacik <jbacik@fb.com> Cc: Li RongQing <lirongqing@baidu.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Matthias Kaehlcke <mka@chromium.org> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Michal Hocko <mhocko@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Philippe Ombredanne <pombredanne@nexb.com> Cc: Roman Gushchin <guro@fb.com> Cc: Sahitya Tummala <stummala@codeaurora.org> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Waiman Long <longman@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:25 +00:00
#endif /* CONFIG_MEMCG_KMEM */
static int memcg_update_kmem_max(struct mem_cgroup *memcg,
unsigned long max)
memcg: rework memcg_update_kmem_limit synchronization Currently we take both the memcg_create_mutex and the set_limit_mutex when we enable kmem accounting for a memory cgroup, which makes kmem activation events serialize with both memcg creations and other memcg limit updates (memory.limit, memory.memsw.limit). However, there is no point in such strict synchronization rules there. First, the set_limit_mutex was introduced to keep the memory.limit and memory.memsw.limit values in sync. Since memory.kmem.limit can be set independently of them, it is better to introduce a separate mutex to synchronize against concurrent kmem limit updates. Second, we take the memcg_create_mutex in order to make sure all children of this memcg will be kmem-active as well. For achieving that, it is enough to hold this mutex only while checking if memcg_has_children() though. This guarantees that if a child is added after we checked that the memcg has no children, the newly added cgroup will see its parent kmem-active (of course if the latter succeeded), and call kmem activation for itself. This patch simplifies the locking rules of memcg_update_kmem_limit() according to these considerations. [vdavydov@parallels.com: fix unintialized var warning] Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Glauber Costa <glommer@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-23 23:53:09 +00:00
{
mm: memcontrol: enable kmem accounting for all cgroups in the legacy hierarchy Workingset code was recently made memcg aware, but shadow node shrinker is still global. As a result, one small cgroup can consume all memory available for shadow nodes, possibly hurting other cgroups by reclaiming their shadow nodes, even though reclaim distances stored in its shadow nodes have no effect. To avoid this, we need to make shadow node shrinker memcg aware. The actual work is done in patch 6 of the series. Patches 1 and 2 prepare memcg/shrinker infrastructure for the change. Patch 3 is just a collateral cleanup. Patch 4 makes radix_tree_node accounted, which is necessary for making shadow node shrinker memcg aware. Patch 5 reduces shadow nodes overhead in case workload mostly uses anonymous pages. This patch: Currently, in the legacy hierarchy kmem accounting is off for all cgroups by default and must be enabled explicitly by writing something to memory.kmem.limit_in_bytes. Since we don't support reclaim on hitting kmem limit, nor do we have any plans to implement it, this is likely to be -1, just to enable kmem accounting and limit kernel memory consumption by the memory.limit_in_bytes along with user memory. This user API was introduced when the implementation of kmem accounting lacked slab shrinker support and hence was useless in practice. Things have changed since then - slab shrinkers were made memcg aware, the accounting overhead seems to be negligible, and a failure to charge a kmem allocation should not have critical consequences, because we only account those kernel objects that should be safe to fail. That's why kmem accounting is enabled by default for all cgroups in the default hierarchy, which will eventually replace the legacy one. The ability to enable kmem accounting for some cgroups while keeping it disabled for others is getting difficult to maintain. E.g. to make shadow node shrinker memcg aware (see mm/workingset.c), we need to know the relationship between the number of shadow nodes allocated for a cgroup and the size of its lru list. If kmem accounting is enabled for all cgroups there is no problem, but what should we do if kmem accounting is enabled only for half of cgroups? We've no other choice but use global lru stats while scanning root cgroup's shadow nodes, but that would be wrong if kmem accounting was enabled for all cgroups (which is the case if the unified hierarchy is used), in which case we should use lru stats of the root cgroup's lruvec. That being said, let's enable kmem accounting for all memory cgroups by default. If one finds it unstable or too costly, it can always be disabled system-wide by passing cgroup.memory=nokmem to the kernel at boot time. Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-03-17 21:18:27 +00:00
int ret;
mutex_lock(&memcg_max_mutex);
ret = page_counter_set_max(&memcg->kmem, max);
mutex_unlock(&memcg_max_mutex);
return ret;
memcg: rework memcg_update_kmem_limit synchronization Currently we take both the memcg_create_mutex and the set_limit_mutex when we enable kmem accounting for a memory cgroup, which makes kmem activation events serialize with both memcg creations and other memcg limit updates (memory.limit, memory.memsw.limit). However, there is no point in such strict synchronization rules there. First, the set_limit_mutex was introduced to keep the memory.limit and memory.memsw.limit values in sync. Since memory.kmem.limit can be set independently of them, it is better to introduce a separate mutex to synchronize against concurrent kmem limit updates. Second, we take the memcg_create_mutex in order to make sure all children of this memcg will be kmem-active as well. For achieving that, it is enough to hold this mutex only while checking if memcg_has_children() though. This guarantees that if a child is added after we checked that the memcg has no children, the newly added cgroup will see its parent kmem-active (of course if the latter succeeded), and call kmem activation for itself. This patch simplifies the locking rules of memcg_update_kmem_limit() according to these considerations. [vdavydov@parallels.com: fix unintialized var warning] Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Glauber Costa <glommer@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-23 23:53:09 +00:00
}
memcg: kmem accounting basic infrastructure Add the basic infrastructure for the accounting of kernel memory. To control that, the following files are created: * memory.kmem.usage_in_bytes * memory.kmem.limit_in_bytes * memory.kmem.failcnt * memory.kmem.max_usage_in_bytes They have the same meaning of their user memory counterparts. They reflect the state of the "kmem" res_counter. Per cgroup kmem memory accounting is not enabled until a limit is set for the group. Once the limit is set the accounting cannot be disabled for that group. This means that after the patch is applied, no behavioral changes exists for whoever is still using memcg to control their memory usage, until memory.kmem.limit_in_bytes is set for the first time. We always account to both user and kernel resource_counters. This effectively means that an independent kernel limit is in place when the limit is set to a lower value than the user memory. A equal or higher value means that the user limit will always hit first, meaning that kmem is effectively unlimited. People who want to track kernel memory but not limit it, can set this limit to a very high number (like RESOURCE_MAX - 1page - that no one will ever hit, or equal to the user memory) [akpm@linux-foundation.org: MEMCG_MMEM only works with slab and slub] Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:47 +00:00
static int memcg_update_tcp_max(struct mem_cgroup *memcg, unsigned long max)
{
int ret;
mutex_lock(&memcg_max_mutex);
ret = page_counter_set_max(&memcg->tcpmem, max);
if (ret)
goto out;
if (!memcg->tcpmem_active) {
/*
* The active flag needs to be written after the static_key
* update. This is what guarantees that the socket activation
* function is the last one to run. See mem_cgroup_sk_alloc()
* for details, and note that we don't mark any socket as
* belonging to this memcg until that flag is up.
*
* We need to do this, because static_keys will span multiple
* sites, but we can't control their order. If we mark a socket
* as accounted, but the accounting functions are not patched in
* yet, we'll lose accounting.
*
* We never race with the readers in mem_cgroup_sk_alloc(),
* because when this value change, the code to process it is not
* patched in yet.
*/
static_branch_inc(&memcg_sockets_enabled_key);
memcg->tcpmem_active = true;
}
out:
mutex_unlock(&memcg_max_mutex);
return ret;
}
/*
* The user of this function is...
* RES_LIMIT.
*/
static ssize_t mem_cgroup_write(struct kernfs_open_file *of,
char *buf, size_t nbytes, loff_t off)
{
struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of));
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
unsigned long nr_pages;
int ret;
buf = strstrip(buf);
ret = page_counter_memparse(buf, "-1", &nr_pages);
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
if (ret)
return ret;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
switch (MEMFILE_ATTR(of_cft(of)->private)) {
case RES_LIMIT:
if (mem_cgroup_is_root(memcg)) { /* Can't set limit on root */
ret = -EINVAL;
break;
}
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
switch (MEMFILE_TYPE(of_cft(of)->private)) {
case _MEM:
ret = mem_cgroup_resize_max(memcg, nr_pages, false);
memcg: mem+swap controller core This patch implements per cgroup limit for usage of memory+swap. However there are SwapCache, double counting of swap-cache and swap-entry is avoided. Mem+Swap controller works as following. - memory usage is limited by memory.limit_in_bytes. - memory + swap usage is limited by memory.memsw_limit_in_bytes. This has following benefits. - A user can limit total resource usage of mem+swap. Without this, because memory resource controller doesn't take care of usage of swap, a process can exhaust all the swap (by memory leak.) We can avoid this case. And Swap is shared resource but it cannot be reclaimed (goes back to memory) until it's used. This characteristic can be trouble when the memory is divided into some parts by cpuset or memcg. Assume group A and group B. After some application executes, the system can be.. Group A -- very large free memory space but occupy 99% of swap. Group B -- under memory shortage but cannot use swap...it's nearly full. Ability to set appropriate swap limit for each group is required. Maybe someone wonder "why not swap but mem+swap ?" - The global LRU(kswapd) can swap out arbitrary pages. Swap-out means to move account from memory to swap...there is no change in usage of mem+swap. In other words, when we want to limit the usage of swap without affecting global LRU, mem+swap limit is better than just limiting swap. Accounting target information is stored in swap_cgroup which is per swap entry record. Charge is done as following. map - charge page and memsw. unmap - uncharge page/memsw if not SwapCache. swap-out (__delete_from_swap_cache) - uncharge page - record mem_cgroup information to swap_cgroup. swap-in (do_swap_page) - charged as page and memsw. record in swap_cgroup is cleared. memsw accounting is decremented. swap-free (swap_free()) - if swap entry is freed, memsw is uncharged by PAGE_SIZE. There are people work under never-swap environments and consider swap as something bad. For such people, this mem+swap controller extension is just an overhead. This overhead is avoided by config or boot option. (see Kconfig. detail is not in this patch.) TODO: - maybe more optimization can be don in swap-in path. (but not very safe.) But we just do simple accounting at this stage. [nishimura@mxp.nes.nec.co.jp: make resize limit hold mutex] [hugh@veritas.com: memswap controller core swapcache fixes] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:08:00 +00:00
break;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
case _MEMSWAP:
ret = mem_cgroup_resize_max(memcg, nr_pages, true);
break;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
case _KMEM:
memcg, kmem: deprecate kmem.limit_in_bytes Cgroup v1 memcg controller has exposed a dedicated kmem limit to users which turned out to be really a bad idea because there are paths which cannot shrink the kernel memory usage enough to get below the limit (e.g. because the accounted memory is not reclaimable). There are cases when the failure is even not allowed (e.g. __GFP_NOFAIL). This means that the kmem limit is in excess to the hard limit without any way to shrink and thus completely useless. OOM killer cannot be invoked to handle the situation because that would lead to a premature oom killing. As a result many places might see ENOMEM returning from kmalloc and result in unexpected errors. E.g. a global OOM killer when there is a lot of free memory because ENOMEM is translated into VM_FAULT_OOM in #PF path and therefore pagefault_out_of_memory would result in OOM killer. Please note that the kernel memory is still accounted to the overall limit along with the user memory so removing the kmem specific limit should still allow to contain kernel memory consumption. Unlike the kmem one, though, it invokes memory reclaim and targeted memcg oom killing if necessary. Start the deprecation process by crying to the kernel log. Let's see whether there are relevant usecases and simply return to EINVAL in the second stage if nobody complains in few releases. [akpm@linux-foundation.org: tweak documentation text] Link: http://lkml.kernel.org/r/20190911151612.GI4023@dhcp22.suse.cz Signed-off-by: Michal Hocko <mhocko@suse.com> Reviewed-by: Shakeel Butt <shakeelb@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Thomas Lindroth <thomas.lindroth@gmail.com> Cc: Tetsuo Handa <penguin-kernel@i-love.sakura.ne.jp> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-09-23 22:37:22 +00:00
pr_warn_once("kmem.limit_in_bytes is deprecated and will be removed. "
"Please report your usecase to linux-mm@kvack.org if you "
"depend on this functionality.\n");
ret = memcg_update_kmem_max(memcg, nr_pages);
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
break;
case _TCP:
ret = memcg_update_tcp_max(memcg, nr_pages);
break;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
}
break;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
case RES_SOFT_LIMIT:
memcg->soft_limit = nr_pages;
ret = 0;
break;
}
return ret ?: nbytes;
}
static ssize_t mem_cgroup_reset(struct kernfs_open_file *of, char *buf,
size_t nbytes, loff_t off)
{
struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of));
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
struct page_counter *counter;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
switch (MEMFILE_TYPE(of_cft(of)->private)) {
case _MEM:
counter = &memcg->memory;
break;
case _MEMSWAP:
counter = &memcg->memsw;
break;
case _KMEM:
counter = &memcg->kmem;
break;
case _TCP:
counter = &memcg->tcpmem;
break;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
default:
BUG();
}
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
switch (MEMFILE_ATTR(of_cft(of)->private)) {
case RES_MAX_USAGE:
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
page_counter_reset_watermark(counter);
break;
case RES_FAILCNT:
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
counter->failcnt = 0;
break;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
default:
BUG();
}
return nbytes;
}
cgroup: pass around cgroup_subsys_state instead of cgroup in file methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup. Please see the previous commit which converts the subsystem methods for rationale. This patch converts all cftype file operations to take @css instead of @cgroup. cftypes for the cgroup core files don't have their subsytem pointer set. These will automatically use the dummy_css added by the previous patch and can be converted the same way. Most subsystem conversions are straight forwards but there are some interesting ones. * freezer: update_if_frozen() is also converted to take @css instead of @cgroup for consistency. This will make the code look simpler too once iterators are converted to use css. * memory/vmpressure: mem_cgroup_from_css() needs to be exported to vmpressure while mem_cgroup_from_cont() can be made static. Updated accordingly. * cpu: cgroup_tg() doesn't have any user left. Removed. * cpuacct: cgroup_ca() doesn't have any user left. Removed. * hugetlb: hugetlb_cgroup_form_cgroup() doesn't have any user left. Removed. * net_cls: cgrp_cls_state() doesn't have any user left. Removed. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:24 +00:00
static u64 mem_cgroup_move_charge_read(struct cgroup_subsys_state *css,
struct cftype *cft)
{
cgroup: pass around cgroup_subsys_state instead of cgroup in file methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup. Please see the previous commit which converts the subsystem methods for rationale. This patch converts all cftype file operations to take @css instead of @cgroup. cftypes for the cgroup core files don't have their subsytem pointer set. These will automatically use the dummy_css added by the previous patch and can be converted the same way. Most subsystem conversions are straight forwards but there are some interesting ones. * freezer: update_if_frozen() is also converted to take @css instead of @cgroup for consistency. This will make the code look simpler too once iterators are converted to use css. * memory/vmpressure: mem_cgroup_from_css() needs to be exported to vmpressure while mem_cgroup_from_cont() can be made static. Updated accordingly. * cpu: cgroup_tg() doesn't have any user left. Removed. * cpuacct: cgroup_ca() doesn't have any user left. Removed. * hugetlb: hugetlb_cgroup_form_cgroup() doesn't have any user left. Removed. * net_cls: cgrp_cls_state() doesn't have any user left. Removed. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:24 +00:00
return mem_cgroup_from_css(css)->move_charge_at_immigrate;
}
#ifdef CONFIG_MMU
cgroup: pass around cgroup_subsys_state instead of cgroup in file methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup. Please see the previous commit which converts the subsystem methods for rationale. This patch converts all cftype file operations to take @css instead of @cgroup. cftypes for the cgroup core files don't have their subsytem pointer set. These will automatically use the dummy_css added by the previous patch and can be converted the same way. Most subsystem conversions are straight forwards but there are some interesting ones. * freezer: update_if_frozen() is also converted to take @css instead of @cgroup for consistency. This will make the code look simpler too once iterators are converted to use css. * memory/vmpressure: mem_cgroup_from_css() needs to be exported to vmpressure while mem_cgroup_from_cont() can be made static. Updated accordingly. * cpu: cgroup_tg() doesn't have any user left. Removed. * cpuacct: cgroup_ca() doesn't have any user left. Removed. * hugetlb: hugetlb_cgroup_form_cgroup() doesn't have any user left. Removed. * net_cls: cgrp_cls_state() doesn't have any user left. Removed. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:24 +00:00
static int mem_cgroup_move_charge_write(struct cgroup_subsys_state *css,
struct cftype *cft, u64 val)
{
cgroup: pass around cgroup_subsys_state instead of cgroup in file methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup. Please see the previous commit which converts the subsystem methods for rationale. This patch converts all cftype file operations to take @css instead of @cgroup. cftypes for the cgroup core files don't have their subsytem pointer set. These will automatically use the dummy_css added by the previous patch and can be converted the same way. Most subsystem conversions are straight forwards but there are some interesting ones. * freezer: update_if_frozen() is also converted to take @css instead of @cgroup for consistency. This will make the code look simpler too once iterators are converted to use css. * memory/vmpressure: mem_cgroup_from_css() needs to be exported to vmpressure while mem_cgroup_from_cont() can be made static. Updated accordingly. * cpu: cgroup_tg() doesn't have any user left. Removed. * cpuacct: cgroup_ca() doesn't have any user left. Removed. * hugetlb: hugetlb_cgroup_form_cgroup() doesn't have any user left. Removed. * net_cls: cgrp_cls_state() doesn't have any user left. Removed. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:24 +00:00
struct mem_cgroup *memcg = mem_cgroup_from_css(css);
if (val & ~MOVE_MASK)
return -EINVAL;
memcg: prevent changes to move_charge_at_immigrate during task attach In memcg, we use the cgroup_lock basically to synchronize against attaching new children to a cgroup. We do this because we rely on cgroup core to provide us with this information. We need to guarantee that upon child creation, our tunables are consistent. For those, the calls to cgroup_lock() all live in handlers like mem_cgroup_hierarchy_write(), where we change a tunable in the group that is hierarchy-related. For instance, the use_hierarchy flag cannot be changed if the cgroup already have children. Furthermore, those values are propagated from the parent to the child when a new child is created. So if we don't lock like this, we can end up with the following situation: A B memcg_css_alloc() mem_cgroup_hierarchy_write() copy use hierarchy from parent change use hierarchy in parent finish creation. This is mainly because during create, we are still not fully connected to the css tree. So all iterators and the such that we could use, will fail to show that the group has children. My observation is that all of creation can proceed in parallel with those tasks, except value assignment. So what this patch series does is to first move all value assignment that is dependent on parent values from css_alloc to css_online, where the iterators all work, and then we lock only the value assignment. This will guarantee that parent and children always have consistent values. Together with an online test, that can be derived from the observation that the refcount of an online memcg can be made to be always positive, we should be able to synchronize our side without the cgroup lock. This patch: Currently, we rely on the cgroup_lock() to prevent changes to move_charge_at_immigrate during task migration. However, this is only needed because the current strategy keeps checking this value throughout the whole process. Since all we need is serialization, one needs only to guarantee that whatever decision we made in the beginning of a specific migration is respected throughout the process. We can achieve this by just saving it in mc. By doing this, no kind of locking is needed. Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Hiroyuki Kamezawa <kamezawa.hiroyuki@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-23 00:34:50 +00:00
/*
memcg: prevent changes to move_charge_at_immigrate during task attach In memcg, we use the cgroup_lock basically to synchronize against attaching new children to a cgroup. We do this because we rely on cgroup core to provide us with this information. We need to guarantee that upon child creation, our tunables are consistent. For those, the calls to cgroup_lock() all live in handlers like mem_cgroup_hierarchy_write(), where we change a tunable in the group that is hierarchy-related. For instance, the use_hierarchy flag cannot be changed if the cgroup already have children. Furthermore, those values are propagated from the parent to the child when a new child is created. So if we don't lock like this, we can end up with the following situation: A B memcg_css_alloc() mem_cgroup_hierarchy_write() copy use hierarchy from parent change use hierarchy in parent finish creation. This is mainly because during create, we are still not fully connected to the css tree. So all iterators and the such that we could use, will fail to show that the group has children. My observation is that all of creation can proceed in parallel with those tasks, except value assignment. So what this patch series does is to first move all value assignment that is dependent on parent values from css_alloc to css_online, where the iterators all work, and then we lock only the value assignment. This will guarantee that parent and children always have consistent values. Together with an online test, that can be derived from the observation that the refcount of an online memcg can be made to be always positive, we should be able to synchronize our side without the cgroup lock. This patch: Currently, we rely on the cgroup_lock() to prevent changes to move_charge_at_immigrate during task migration. However, this is only needed because the current strategy keeps checking this value throughout the whole process. Since all we need is serialization, one needs only to guarantee that whatever decision we made in the beginning of a specific migration is respected throughout the process. We can achieve this by just saving it in mc. By doing this, no kind of locking is needed. Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Hiroyuki Kamezawa <kamezawa.hiroyuki@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-02-23 00:34:50 +00:00
* No kind of locking is needed in here, because ->can_attach() will
* check this value once in the beginning of the process, and then carry
* on with stale data. This means that changes to this value will only
* affect task migrations starting after the change.
*/
memcg->move_charge_at_immigrate = val;
return 0;
}
#else
cgroup: pass around cgroup_subsys_state instead of cgroup in file methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup. Please see the previous commit which converts the subsystem methods for rationale. This patch converts all cftype file operations to take @css instead of @cgroup. cftypes for the cgroup core files don't have their subsytem pointer set. These will automatically use the dummy_css added by the previous patch and can be converted the same way. Most subsystem conversions are straight forwards but there are some interesting ones. * freezer: update_if_frozen() is also converted to take @css instead of @cgroup for consistency. This will make the code look simpler too once iterators are converted to use css. * memory/vmpressure: mem_cgroup_from_css() needs to be exported to vmpressure while mem_cgroup_from_cont() can be made static. Updated accordingly. * cpu: cgroup_tg() doesn't have any user left. Removed. * cpuacct: cgroup_ca() doesn't have any user left. Removed. * hugetlb: hugetlb_cgroup_form_cgroup() doesn't have any user left. Removed. * net_cls: cgrp_cls_state() doesn't have any user left. Removed. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:24 +00:00
static int mem_cgroup_move_charge_write(struct cgroup_subsys_state *css,
struct cftype *cft, u64 val)
{
return -ENOSYS;
}
#endif
#ifdef CONFIG_NUMA
#define LRU_ALL_FILE (BIT(LRU_INACTIVE_FILE) | BIT(LRU_ACTIVE_FILE))
#define LRU_ALL_ANON (BIT(LRU_INACTIVE_ANON) | BIT(LRU_ACTIVE_ANON))
#define LRU_ALL ((1 << NR_LRU_LISTS) - 1)
static unsigned long mem_cgroup_node_nr_lru_pages(struct mem_cgroup *memcg,
int nid, unsigned int lru_mask, bool tree)
{
struct lruvec *lruvec = mem_cgroup_lruvec(memcg, NODE_DATA(nid));
unsigned long nr = 0;
enum lru_list lru;
VM_BUG_ON((unsigned)nid >= nr_node_ids);
for_each_lru(lru) {
if (!(BIT(lru) & lru_mask))
continue;
if (tree)
nr += lruvec_page_state(lruvec, NR_LRU_BASE + lru);
else
nr += lruvec_page_state_local(lruvec, NR_LRU_BASE + lru);
}
return nr;
}
static unsigned long mem_cgroup_nr_lru_pages(struct mem_cgroup *memcg,
unsigned int lru_mask,
bool tree)
{
unsigned long nr = 0;
enum lru_list lru;
for_each_lru(lru) {
if (!(BIT(lru) & lru_mask))
continue;
if (tree)
nr += memcg_page_state(memcg, NR_LRU_BASE + lru);
else
nr += memcg_page_state_local(memcg, NR_LRU_BASE + lru);
}
return nr;
}
static int memcg_numa_stat_show(struct seq_file *m, void *v)
{
struct numa_stat {
const char *name;
unsigned int lru_mask;
};
static const struct numa_stat stats[] = {
{ "total", LRU_ALL },
{ "file", LRU_ALL_FILE },
{ "anon", LRU_ALL_ANON },
{ "unevictable", BIT(LRU_UNEVICTABLE) },
};
const struct numa_stat *stat;
int nid;
struct mem_cgroup *memcg = mem_cgroup_from_seq(m);
for (stat = stats; stat < stats + ARRAY_SIZE(stats); stat++) {
seq_printf(m, "%s=%lu", stat->name,
mem_cgroup_nr_lru_pages(memcg, stat->lru_mask,
false));
for_each_node_state(nid, N_MEMORY)
seq_printf(m, " N%d=%lu", nid,
mem_cgroup_node_nr_lru_pages(memcg, nid,
stat->lru_mask, false));
seq_putc(m, '\n');
}
memcg: support hierarchical memory.numa_stats The memory.numa_stat file was not hierarchical. Memory charged to the children was not shown in parent's numa_stat. This change adds the "hierarchical_" stats to the existing stats. The new hierarchical stats include the sum of all children's values in addition to the value of the memcg. Tested: Create cgroup a, a/b and run workload under b. The values of b are included in the "hierarchical_*" under a. $ cd /sys/fs/cgroup $ echo 1 > memory.use_hierarchy $ mkdir a a/b Run workload in a/b: $ (echo $BASHPID >> a/b/cgroup.procs && cat /some/file && bash) & The hierarchical_ fields in parent (a) show use of workload in a/b: $ cat a/memory.numa_stat total=0 N0=0 N1=0 N2=0 N3=0 file=0 N0=0 N1=0 N2=0 N3=0 anon=0 N0=0 N1=0 N2=0 N3=0 unevictable=0 N0=0 N1=0 N2=0 N3=0 hierarchical_total=908 N0=552 N1=317 N2=39 N3=0 hierarchical_file=850 N0=549 N1=301 N2=0 N3=0 hierarchical_anon=58 N0=3 N1=16 N2=39 N3=0 hierarchical_unevictable=0 N0=0 N1=0 N2=0 N3=0 $ cat a/b/memory.numa_stat total=908 N0=552 N1=317 N2=39 N3=0 file=850 N0=549 N1=301 N2=0 N3=0 anon=58 N0=3 N1=16 N2=39 N3=0 unevictable=0 N0=0 N1=0 N2=0 N3=0 hierarchical_total=908 N0=552 N1=317 N2=39 N3=0 hierarchical_file=850 N0=549 N1=301 N2=0 N3=0 hierarchical_anon=58 N0=3 N1=16 N2=39 N3=0 hierarchical_unevictable=0 N0=0 N1=0 N2=0 N3=0 Signed-off-by: Ying Han <yinghan@google.com> Signed-off-by: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-11-12 23:07:41 +00:00
for (stat = stats; stat < stats + ARRAY_SIZE(stats); stat++) {
seq_printf(m, "hierarchical_%s=%lu", stat->name,
mem_cgroup_nr_lru_pages(memcg, stat->lru_mask,
true));
for_each_node_state(nid, N_MEMORY)
seq_printf(m, " N%d=%lu", nid,
mem_cgroup_node_nr_lru_pages(memcg, nid,
stat->lru_mask, true));
memcg: support hierarchical memory.numa_stats The memory.numa_stat file was not hierarchical. Memory charged to the children was not shown in parent's numa_stat. This change adds the "hierarchical_" stats to the existing stats. The new hierarchical stats include the sum of all children's values in addition to the value of the memcg. Tested: Create cgroup a, a/b and run workload under b. The values of b are included in the "hierarchical_*" under a. $ cd /sys/fs/cgroup $ echo 1 > memory.use_hierarchy $ mkdir a a/b Run workload in a/b: $ (echo $BASHPID >> a/b/cgroup.procs && cat /some/file && bash) & The hierarchical_ fields in parent (a) show use of workload in a/b: $ cat a/memory.numa_stat total=0 N0=0 N1=0 N2=0 N3=0 file=0 N0=0 N1=0 N2=0 N3=0 anon=0 N0=0 N1=0 N2=0 N3=0 unevictable=0 N0=0 N1=0 N2=0 N3=0 hierarchical_total=908 N0=552 N1=317 N2=39 N3=0 hierarchical_file=850 N0=549 N1=301 N2=0 N3=0 hierarchical_anon=58 N0=3 N1=16 N2=39 N3=0 hierarchical_unevictable=0 N0=0 N1=0 N2=0 N3=0 $ cat a/b/memory.numa_stat total=908 N0=552 N1=317 N2=39 N3=0 file=850 N0=549 N1=301 N2=0 N3=0 anon=58 N0=3 N1=16 N2=39 N3=0 unevictable=0 N0=0 N1=0 N2=0 N3=0 hierarchical_total=908 N0=552 N1=317 N2=39 N3=0 hierarchical_file=850 N0=549 N1=301 N2=0 N3=0 hierarchical_anon=58 N0=3 N1=16 N2=39 N3=0 hierarchical_unevictable=0 N0=0 N1=0 N2=0 N3=0 Signed-off-by: Ying Han <yinghan@google.com> Signed-off-by: Greg Thelen <gthelen@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-11-12 23:07:41 +00:00
seq_putc(m, '\n');
}
return 0;
}
#endif /* CONFIG_NUMA */
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
static const unsigned int memcg1_stats[] = {
NR_FILE_PAGES,
mm: memcontrol: switch to native NR_ANON_MAPPED counter Memcg maintains a private MEMCG_RSS counter. This divergence from the generic VM accounting means unnecessary code overhead, and creates a dependency for memcg that page->mapping is set up at the time of charging, so that page types can be told apart. Convert the generic accounting sites to mod_lruvec_page_state and friends to maintain the per-cgroup vmstat counter of NR_ANON_MAPPED. We use lock_page_memcg() to stabilize page->mem_cgroup during rmap changes, the same way we do for NR_FILE_MAPPED. With the previous patch removing MEMCG_CACHE and the private NR_SHMEM counter, this patch finally eliminates the need to have page->mapping set up at charge time. However, we need to have page->mem_cgroup set up by the time rmap runs and does the accounting, so switch the commit and the rmap callbacks around. v2: fix temporary accounting bug by switching rmap<->commit (Joonsoo) Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Cc: Alex Shi <alex.shi@linux.alibaba.com> Cc: Hugh Dickins <hughd@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.com> Cc: Roman Gushchin <guro@fb.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Balbir Singh <bsingharora@gmail.com> Link: http://lkml.kernel.org/r/20200508183105.225460-11-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:01:57 +00:00
NR_ANON_MAPPED,
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
NR_ANON_THPS,
#endif
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
NR_SHMEM,
NR_FILE_MAPPED,
NR_FILE_DIRTY,
NR_WRITEBACK,
MEMCG_SWAP,
};
static const char *const memcg1_stat_names[] = {
"cache",
"rss",
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
"rss_huge",
#endif
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
"shmem",
"mapped_file",
"dirty",
"writeback",
"swap",
};
/* Universal VM events cgroup1 shows, original sort order */
static const unsigned int memcg1_events[] = {
PGPGIN,
PGPGOUT,
PGFAULT,
PGMAJFAULT,
};
static int memcg_stat_show(struct seq_file *m, void *v)
{
struct mem_cgroup *memcg = mem_cgroup_from_seq(m);
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
unsigned long memory, memsw;
struct mem_cgroup *mi;
unsigned int i;
BUILD_BUG_ON(ARRAY_SIZE(memcg1_stat_names) != ARRAY_SIZE(memcg1_stats));
for (i = 0; i < ARRAY_SIZE(memcg1_stats); i++) {
unsigned long nr;
if (memcg1_stats[i] == MEMCG_SWAP && !do_memsw_account())
continue;
nr = memcg_page_state_local(memcg, memcg1_stats[i]);
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
if (memcg1_stats[i] == NR_ANON_THPS)
nr *= HPAGE_PMD_NR;
#endif
seq_printf(m, "%s %lu\n", memcg1_stat_names[i], nr * PAGE_SIZE);
}
for (i = 0; i < ARRAY_SIZE(memcg1_events); i++)
seq_printf(m, "%s %lu\n", vm_event_name(memcg1_events[i]),
mm: memcontrol: make cgroup stats and events query API explicitly local Patch series "mm: memcontrol: memory.stat cost & correctness". The cgroup memory.stat file holds recursive statistics for the entire subtree. The current implementation does this tree walk on-demand whenever the file is read. This is giving us problems in production. 1. The cost of aggregating the statistics on-demand is high. A lot of system service cgroups are mostly idle and their stats don't change between reads, yet we always have to check them. There are also always some lazily-dying cgroups sitting around that are pinned by a handful of remaining page cache; the same applies to them. In an application that periodically monitors memory.stat in our fleet, we have seen the aggregation consume up to 5% CPU time. 2. When cgroups die and disappear from the cgroup tree, so do their accumulated vm events. The result is that the event counters at higher-level cgroups can go backwards and confuse some of our automation, let alone people looking at the graphs over time. To address both issues, this patch series changes the stat implementation to spill counts upwards when the counters change. The upward spilling is batched using the existing per-cpu cache. In a sparse file stress test with 5 level cgroup nesting, the additional cost of the flushing was negligible (a little under 1% of CPU at 100% CPU utilization, compared to the 5% of reading memory.stat during regular operation). This patch (of 4): memcg_page_state(), lruvec_page_state(), memcg_sum_events() are currently returning the state of the local memcg or lruvec, not the recursive state. In practice there is a demand for both versions, although the callers that want the recursive counts currently sum them up by hand. Per default, cgroups are considered recursive entities and generally we expect more users of the recursive counters, with the local counts being special cases. To reflect that in the name, add a _local suffix to the current implementations. The following patch will re-incarnate these functions with recursive semantics, but with an O(1) implementation. [hannes@cmpxchg.org: fix bisection hole] Link: http://lkml.kernel.org/r/20190417160347.GC23013@cmpxchg.org Link: http://lkml.kernel.org/r/20190412151507.2769-2-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:06 +00:00
memcg_events_local(memcg, memcg1_events[i]));
for (i = 0; i < NR_LRU_LISTS; i++)
seq_printf(m, "%s %lu\n", lru_list_name(i),
mm: memcontrol: make cgroup stats and events query API explicitly local Patch series "mm: memcontrol: memory.stat cost & correctness". The cgroup memory.stat file holds recursive statistics for the entire subtree. The current implementation does this tree walk on-demand whenever the file is read. This is giving us problems in production. 1. The cost of aggregating the statistics on-demand is high. A lot of system service cgroups are mostly idle and their stats don't change between reads, yet we always have to check them. There are also always some lazily-dying cgroups sitting around that are pinned by a handful of remaining page cache; the same applies to them. In an application that periodically monitors memory.stat in our fleet, we have seen the aggregation consume up to 5% CPU time. 2. When cgroups die and disappear from the cgroup tree, so do their accumulated vm events. The result is that the event counters at higher-level cgroups can go backwards and confuse some of our automation, let alone people looking at the graphs over time. To address both issues, this patch series changes the stat implementation to spill counts upwards when the counters change. The upward spilling is batched using the existing per-cpu cache. In a sparse file stress test with 5 level cgroup nesting, the additional cost of the flushing was negligible (a little under 1% of CPU at 100% CPU utilization, compared to the 5% of reading memory.stat during regular operation). This patch (of 4): memcg_page_state(), lruvec_page_state(), memcg_sum_events() are currently returning the state of the local memcg or lruvec, not the recursive state. In practice there is a demand for both versions, although the callers that want the recursive counts currently sum them up by hand. Per default, cgroups are considered recursive entities and generally we expect more users of the recursive counters, with the local counts being special cases. To reflect that in the name, add a _local suffix to the current implementations. The following patch will re-incarnate these functions with recursive semantics, but with an O(1) implementation. [hannes@cmpxchg.org: fix bisection hole] Link: http://lkml.kernel.org/r/20190417160347.GC23013@cmpxchg.org Link: http://lkml.kernel.org/r/20190412151507.2769-2-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:06 +00:00
memcg_page_state_local(memcg, NR_LRU_BASE + i) *
PAGE_SIZE);
/* Hierarchical information */
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
memory = memsw = PAGE_COUNTER_MAX;
for (mi = memcg; mi; mi = parent_mem_cgroup(mi)) {
memory = min(memory, READ_ONCE(mi->memory.max));
memsw = min(memsw, READ_ONCE(mi->memsw.max));
}
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
seq_printf(m, "hierarchical_memory_limit %llu\n",
(u64)memory * PAGE_SIZE);
if (do_memsw_account())
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
seq_printf(m, "hierarchical_memsw_limit %llu\n",
(u64)memsw * PAGE_SIZE);
for (i = 0; i < ARRAY_SIZE(memcg1_stats); i++) {
if (memcg1_stats[i] == MEMCG_SWAP && !do_memsw_account())
continue;
seq_printf(m, "total_%s %llu\n", memcg1_stat_names[i],
(u64)memcg_page_state(memcg, memcg1_stats[i]) *
PAGE_SIZE);
}
for (i = 0; i < ARRAY_SIZE(memcg1_events); i++)
seq_printf(m, "total_%s %llu\n",
vm_event_name(memcg1_events[i]),
(u64)memcg_events(memcg, memcg1_events[i]));
for (i = 0; i < NR_LRU_LISTS; i++)
seq_printf(m, "total_%s %llu\n", lru_list_name(i),
mm: memcontrol: fix recursive statistics correctness & scalabilty Right now, when somebody needs to know the recursive memory statistics and events of a cgroup subtree, they need to walk the entire subtree and sum up the counters manually. There are two issues with this: 1. When a cgroup gets deleted, its stats are lost. The state counters should all be 0 at that point, of course, but the events are not. When this happens, the event counters, which are supposed to be monotonic, can go backwards in the parent cgroups. 2. During regular operation, we always have a certain number of lazily freed cgroups sitting around that have been deleted, have no tasks, but have a few cache pages remaining. These groups' statistics do not change until we eventually hit memory pressure, but somebody watching, say, memory.stat on an ancestor has to iterate those every time. This patch addresses both issues by introducing recursive counters at each level that are propagated from the write side when stats change. Upward propagation happens when the per-cpu caches spill over into the local atomic counter. This is the same thing we do during charge and uncharge, except that the latter uses atomic RMWs, which are more expensive; stat changes happen at around the same rate. In a sparse file test (page faults and reclaim at maximum CPU speed) with 5 cgroup nesting levels, perf shows __mod_memcg_page state at ~1%. Link: http://lkml.kernel.org/r/20190412151507.2769-4-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reviewed-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-05-14 22:47:12 +00:00
(u64)memcg_page_state(memcg, NR_LRU_BASE + i) *
PAGE_SIZE);
#ifdef CONFIG_DEBUG_VM
{
pg_data_t *pgdat;
struct mem_cgroup_per_node *mz;
mm: base LRU balancing on an explicit cost model Currently, scan pressure between the anon and file LRU lists is balanced based on a mixture of reclaim efficiency and a somewhat vague notion of "value" of having certain pages in memory over others. That concept of value is problematic, because it has caused us to count any event that remotely makes one LRU list more or less preferrable for reclaim, even when these events are not directly comparable and impose very different costs on the system. One example is referenced file pages that we still deactivate and referenced anonymous pages that we actually rotate back to the head of the list. There is also conceptual overlap with the LRU algorithm itself. By rotating recently used pages instead of reclaiming them, the algorithm already biases the applied scan pressure based on page value. Thus, when rebalancing scan pressure due to rotations, we should think of reclaim cost, and leave assessing the page value to the LRU algorithm. Lastly, considering both value-increasing as well as value-decreasing events can sometimes cause the same type of event to be counted twice, i.e. how rotating a page increases the LRU value, while reclaiming it succesfully decreases the value. In itself this will balance out fine, but it quietly skews the impact of events that are only recorded once. The abstract metric of "value", the murky relationship with the LRU algorithm, and accounting both negative and positive events make the current pressure balancing model hard to reason about and modify. This patch switches to a balancing model of accounting the concrete, actually observed cost of reclaiming one LRU over another. For now, that cost includes pages that are scanned but rotated back to the list head. Subsequent patches will add consideration for IO caused by refaulting of recently evicted pages. Replace struct zone_reclaim_stat with two cost counters in the lruvec, and make everything that affects cost go through a new lru_note_cost() function. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Rik van Riel <riel@surriel.com> Link: http://lkml.kernel.org/r/20200520232525.798933-9-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:02:53 +00:00
unsigned long anon_cost = 0;
unsigned long file_cost = 0;
for_each_online_pgdat(pgdat) {
mz = mem_cgroup_nodeinfo(memcg, pgdat->node_id);
mm: base LRU balancing on an explicit cost model Currently, scan pressure between the anon and file LRU lists is balanced based on a mixture of reclaim efficiency and a somewhat vague notion of "value" of having certain pages in memory over others. That concept of value is problematic, because it has caused us to count any event that remotely makes one LRU list more or less preferrable for reclaim, even when these events are not directly comparable and impose very different costs on the system. One example is referenced file pages that we still deactivate and referenced anonymous pages that we actually rotate back to the head of the list. There is also conceptual overlap with the LRU algorithm itself. By rotating recently used pages instead of reclaiming them, the algorithm already biases the applied scan pressure based on page value. Thus, when rebalancing scan pressure due to rotations, we should think of reclaim cost, and leave assessing the page value to the LRU algorithm. Lastly, considering both value-increasing as well as value-decreasing events can sometimes cause the same type of event to be counted twice, i.e. how rotating a page increases the LRU value, while reclaiming it succesfully decreases the value. In itself this will balance out fine, but it quietly skews the impact of events that are only recorded once. The abstract metric of "value", the murky relationship with the LRU algorithm, and accounting both negative and positive events make the current pressure balancing model hard to reason about and modify. This patch switches to a balancing model of accounting the concrete, actually observed cost of reclaiming one LRU over another. For now, that cost includes pages that are scanned but rotated back to the list head. Subsequent patches will add consideration for IO caused by refaulting of recently evicted pages. Replace struct zone_reclaim_stat with two cost counters in the lruvec, and make everything that affects cost go through a new lru_note_cost() function. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Rik van Riel <riel@surriel.com> Link: http://lkml.kernel.org/r/20200520232525.798933-9-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:02:53 +00:00
anon_cost += mz->lruvec.anon_cost;
file_cost += mz->lruvec.file_cost;
}
mm: base LRU balancing on an explicit cost model Currently, scan pressure between the anon and file LRU lists is balanced based on a mixture of reclaim efficiency and a somewhat vague notion of "value" of having certain pages in memory over others. That concept of value is problematic, because it has caused us to count any event that remotely makes one LRU list more or less preferrable for reclaim, even when these events are not directly comparable and impose very different costs on the system. One example is referenced file pages that we still deactivate and referenced anonymous pages that we actually rotate back to the head of the list. There is also conceptual overlap with the LRU algorithm itself. By rotating recently used pages instead of reclaiming them, the algorithm already biases the applied scan pressure based on page value. Thus, when rebalancing scan pressure due to rotations, we should think of reclaim cost, and leave assessing the page value to the LRU algorithm. Lastly, considering both value-increasing as well as value-decreasing events can sometimes cause the same type of event to be counted twice, i.e. how rotating a page increases the LRU value, while reclaiming it succesfully decreases the value. In itself this will balance out fine, but it quietly skews the impact of events that are only recorded once. The abstract metric of "value", the murky relationship with the LRU algorithm, and accounting both negative and positive events make the current pressure balancing model hard to reason about and modify. This patch switches to a balancing model of accounting the concrete, actually observed cost of reclaiming one LRU over another. For now, that cost includes pages that are scanned but rotated back to the list head. Subsequent patches will add consideration for IO caused by refaulting of recently evicted pages. Replace struct zone_reclaim_stat with two cost counters in the lruvec, and make everything that affects cost go through a new lru_note_cost() function. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Rik van Riel <riel@surriel.com> Link: http://lkml.kernel.org/r/20200520232525.798933-9-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:02:53 +00:00
seq_printf(m, "anon_cost %lu\n", anon_cost);
seq_printf(m, "file_cost %lu\n", file_cost);
}
#endif
return 0;
}
cgroup: pass around cgroup_subsys_state instead of cgroup in file methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup. Please see the previous commit which converts the subsystem methods for rationale. This patch converts all cftype file operations to take @css instead of @cgroup. cftypes for the cgroup core files don't have their subsytem pointer set. These will automatically use the dummy_css added by the previous patch and can be converted the same way. Most subsystem conversions are straight forwards but there are some interesting ones. * freezer: update_if_frozen() is also converted to take @css instead of @cgroup for consistency. This will make the code look simpler too once iterators are converted to use css. * memory/vmpressure: mem_cgroup_from_css() needs to be exported to vmpressure while mem_cgroup_from_cont() can be made static. Updated accordingly. * cpu: cgroup_tg() doesn't have any user left. Removed. * cpuacct: cgroup_ca() doesn't have any user left. Removed. * hugetlb: hugetlb_cgroup_form_cgroup() doesn't have any user left. Removed. * net_cls: cgrp_cls_state() doesn't have any user left. Removed. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:24 +00:00
static u64 mem_cgroup_swappiness_read(struct cgroup_subsys_state *css,
struct cftype *cft)
{
cgroup: pass around cgroup_subsys_state instead of cgroup in file methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup. Please see the previous commit which converts the subsystem methods for rationale. This patch converts all cftype file operations to take @css instead of @cgroup. cftypes for the cgroup core files don't have their subsytem pointer set. These will automatically use the dummy_css added by the previous patch and can be converted the same way. Most subsystem conversions are straight forwards but there are some interesting ones. * freezer: update_if_frozen() is also converted to take @css instead of @cgroup for consistency. This will make the code look simpler too once iterators are converted to use css. * memory/vmpressure: mem_cgroup_from_css() needs to be exported to vmpressure while mem_cgroup_from_cont() can be made static. Updated accordingly. * cpu: cgroup_tg() doesn't have any user left. Removed. * cpuacct: cgroup_ca() doesn't have any user left. Removed. * hugetlb: hugetlb_cgroup_form_cgroup() doesn't have any user left. Removed. * net_cls: cgrp_cls_state() doesn't have any user left. Removed. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:24 +00:00
struct mem_cgroup *memcg = mem_cgroup_from_css(css);
return mem_cgroup_swappiness(memcg);
}
cgroup: pass around cgroup_subsys_state instead of cgroup in file methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup. Please see the previous commit which converts the subsystem methods for rationale. This patch converts all cftype file operations to take @css instead of @cgroup. cftypes for the cgroup core files don't have their subsytem pointer set. These will automatically use the dummy_css added by the previous patch and can be converted the same way. Most subsystem conversions are straight forwards but there are some interesting ones. * freezer: update_if_frozen() is also converted to take @css instead of @cgroup for consistency. This will make the code look simpler too once iterators are converted to use css. * memory/vmpressure: mem_cgroup_from_css() needs to be exported to vmpressure while mem_cgroup_from_cont() can be made static. Updated accordingly. * cpu: cgroup_tg() doesn't have any user left. Removed. * cpuacct: cgroup_ca() doesn't have any user left. Removed. * hugetlb: hugetlb_cgroup_form_cgroup() doesn't have any user left. Removed. * net_cls: cgrp_cls_state() doesn't have any user left. Removed. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:24 +00:00
static int mem_cgroup_swappiness_write(struct cgroup_subsys_state *css,
struct cftype *cft, u64 val)
{
cgroup: pass around cgroup_subsys_state instead of cgroup in file methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup. Please see the previous commit which converts the subsystem methods for rationale. This patch converts all cftype file operations to take @css instead of @cgroup. cftypes for the cgroup core files don't have their subsytem pointer set. These will automatically use the dummy_css added by the previous patch and can be converted the same way. Most subsystem conversions are straight forwards but there are some interesting ones. * freezer: update_if_frozen() is also converted to take @css instead of @cgroup for consistency. This will make the code look simpler too once iterators are converted to use css. * memory/vmpressure: mem_cgroup_from_css() needs to be exported to vmpressure while mem_cgroup_from_cont() can be made static. Updated accordingly. * cpu: cgroup_tg() doesn't have any user left. Removed. * cpuacct: cgroup_ca() doesn't have any user left. Removed. * hugetlb: hugetlb_cgroup_form_cgroup() doesn't have any user left. Removed. * net_cls: cgrp_cls_state() doesn't have any user left. Removed. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:24 +00:00
struct mem_cgroup *memcg = mem_cgroup_from_css(css);
if (val > 100)
return -EINVAL;
Merge branch 'for-3.16' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup Pull cgroup updates from Tejun Heo: "A lot of activities on cgroup side. Heavy restructuring including locking simplification took place to improve the code base and enable implementation of the unified hierarchy, which currently exists behind a __DEVEL__ mount option. The core support is mostly complete but individual controllers need further work. To explain the design and rationales of the the unified hierarchy Documentation/cgroups/unified-hierarchy.txt is added. Another notable change is css (cgroup_subsys_state - what each controller uses to identify and interact with a cgroup) iteration update. This is part of continuing updates on css object lifetime and visibility. cgroup started with reference count draining on removal way back and is now reaching a point where csses behave and are iterated like normal refcnted objects albeit with some complexities to allow distinguishing the state where they're being deleted. The css iteration update isn't taken advantage of yet but is planned to be used to simplify memcg significantly" * 'for-3.16' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup: (77 commits) cgroup: disallow disabled controllers on the default hierarchy cgroup: don't destroy the default root cgroup: disallow debug controller on the default hierarchy cgroup: clean up MAINTAINERS entries cgroup: implement css_tryget() device_cgroup: use css_has_online_children() instead of has_children() cgroup: convert cgroup_has_live_children() into css_has_online_children() cgroup: use CSS_ONLINE instead of CGRP_DEAD cgroup: iterate cgroup_subsys_states directly cgroup: introduce CSS_RELEASED and reduce css iteration fallback window cgroup: move cgroup->serial_nr into cgroup_subsys_state cgroup: link all cgroup_subsys_states in their sibling lists cgroup: move cgroup->sibling and ->children into cgroup_subsys_state cgroup: remove cgroup->parent device_cgroup: remove direct access to cgroup->children memcg: update memcg_has_children() to use css_next_child() memcg: remove tasks/children test from mem_cgroup_force_empty() cgroup: remove css_parent() cgroup: skip refcnting on normal root csses and cgrp_dfl_root self css cgroup: use cgroup->self.refcnt for cgroup refcnting ...
2014-06-09 22:03:33 +00:00
if (css->parent)
memcg->swappiness = val;
else
vm_swappiness = val;
return 0;
}
static void __mem_cgroup_threshold(struct mem_cgroup *memcg, bool swap)
{
struct mem_cgroup_threshold_ary *t;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
unsigned long usage;
int i;
rcu_read_lock();
if (!swap)
t = rcu_dereference(memcg->thresholds.primary);
else
t = rcu_dereference(memcg->memsw_thresholds.primary);
if (!t)
goto unlock;
usage = mem_cgroup_usage(memcg, swap);
/*
* current_threshold points to threshold just below or equal to usage.
* If it's not true, a threshold was crossed after last
* call of __mem_cgroup_threshold().
*/
i = t->current_threshold;
/*
* Iterate backward over array of thresholds starting from
* current_threshold and check if a threshold is crossed.
* If none of thresholds below usage is crossed, we read
* only one element of the array here.
*/
for (; i >= 0 && unlikely(t->entries[i].threshold > usage); i--)
eventfd_signal(t->entries[i].eventfd, 1);
/* i = current_threshold + 1 */
i++;
/*
* Iterate forward over array of thresholds starting from
* current_threshold+1 and check if a threshold is crossed.
* If none of thresholds above usage is crossed, we read
* only one element of the array here.
*/
for (; i < t->size && unlikely(t->entries[i].threshold <= usage); i++)
eventfd_signal(t->entries[i].eventfd, 1);
/* Update current_threshold */
t->current_threshold = i - 1;
unlock:
rcu_read_unlock();
}
static void mem_cgroup_threshold(struct mem_cgroup *memcg)
{
while (memcg) {
__mem_cgroup_threshold(memcg, false);
if (do_memsw_account())
__mem_cgroup_threshold(memcg, true);
memcg = parent_mem_cgroup(memcg);
}
}
static int compare_thresholds(const void *a, const void *b)
{
const struct mem_cgroup_threshold *_a = a;
const struct mem_cgroup_threshold *_b = b;
memcg: fix multiple large threshold notifications A memory cgroup with (1) multiple threshold notifications and (2) at least one threshold >=2G was not reliable. Specifically the notifications would either not fire or would not fire in the proper order. The __mem_cgroup_threshold() signaling logic depends on keeping 64 bit thresholds in sorted order. mem_cgroup_usage_register_event() sorts them with compare_thresholds(), which returns the difference of two 64 bit thresholds as an int. If the difference is positive but has bit[31] set, then sort() treats the difference as negative and breaks sort order. This fix compares the two arbitrary 64 bit thresholds returning the classic -1, 0, 1 result. The test below sets two notifications (at 0x1000 and 0x81001000): cd /sys/fs/cgroup/memory mkdir x for x in 4096 2164264960; do cgroup_event_listener x/memory.usage_in_bytes $x | sed "s/^/$x listener:/" & done echo $$ > x/cgroup.procs anon_leaker 500M v3.11-rc7 fails to signal the 4096 event listener: Leaking... Done leaking pages. Patched v3.11-rc7 properly notifies: Leaking... 4096 listener:2013:8:31:14:13:36 Done leaking pages. The fixed bug is old. It appears to date back to the introduction of memcg threshold notifications in v2.6.34-rc1-116-g2e72b6347c94 "memcg: implement memory thresholds" Signed-off-by: Greg Thelen <gthelen@google.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-09-11 21:23:08 +00:00
if (_a->threshold > _b->threshold)
return 1;
if (_a->threshold < _b->threshold)
return -1;
return 0;
}
static int mem_cgroup_oom_notify_cb(struct mem_cgroup *memcg)
{
struct mem_cgroup_eventfd_list *ev;
memcg: oom_notify use-after-free fix Paul Furtado has reported the following GPF: general protection fault: 0000 [#1] SMP Modules linked in: ipv6 dm_mod xen_netfront coretemp hwmon x86_pkg_temp_thermal crc32_pclmul crc32c_intel ghash_clmulni_intel aesni_intel ablk_helper cryptd lrw gf128mul glue_helper aes_x86_64 microcode pcspkr ext4 jbd2 mbcache raid0 xen_blkfront CPU: 3 PID: 3062 Comm: java Not tainted 3.16.0-rc5 #1 task: ffff8801cfe8f170 ti: ffff8801d2ec4000 task.ti: ffff8801d2ec4000 RIP: e030:mem_cgroup_oom_synchronize+0x140/0x240 RSP: e02b:ffff8801d2ec7d48 EFLAGS: 00010283 RAX: 0000000000000001 RBX: ffff88009d633800 RCX: 000000000000000e RDX: fffffffffffffffe RSI: ffff88009d630200 RDI: ffff88009d630200 RBP: ffff8801d2ec7da8 R08: 0000000000000012 R09: 00000000fffffffe R10: 0000000000000000 R11: 0000000000000000 R12: ffff88009d633800 R13: ffff8801d2ec7d48 R14: dead000000100100 R15: ffff88009d633a30 FS: 00007f1748bb4700(0000) GS:ffff8801def80000(0000) knlGS:0000000000000000 CS: e033 DS: 0000 ES: 0000 CR0: 000000008005003b CR2: 00007f4110300308 CR3: 00000000c05f7000 CR4: 0000000000002660 Call Trace: pagefault_out_of_memory+0x18/0x90 mm_fault_error+0xa9/0x1a0 __do_page_fault+0x478/0x4c0 do_page_fault+0x2c/0x40 page_fault+0x28/0x30 Code: 44 00 00 48 89 df e8 40 ca ff ff 48 85 c0 49 89 c4 74 35 4c 8b b0 30 02 00 00 4c 8d b8 30 02 00 00 4d 39 fe 74 1b 0f 1f 44 00 00 <49> 8b 7e 10 be 01 00 00 00 e8 42 d2 04 00 4d 8b 36 4d 39 fe 75 RIP mem_cgroup_oom_synchronize+0x140/0x240 Commit fb2a6fc56be6 ("mm: memcg: rework and document OOM waiting and wakeup") has moved mem_cgroup_oom_notify outside of memcg_oom_lock assuming it is protected by the hierarchical OOM-lock. Although this is true for the notification part the protection doesn't cover unregistration of event which can happen in parallel now so mem_cgroup_oom_notify can see already unlinked and/or freed mem_cgroup_eventfd_list. Fix this by using memcg_oom_lock also in mem_cgroup_oom_notify. Addresses https://bugzilla.kernel.org/show_bug.cgi?id=80881 Fixes: fb2a6fc56be6 (mm: memcg: rework and document OOM waiting and wakeup) Signed-off-by: Michal Hocko <mhocko@suse.cz> Reported-by: Paul Furtado <paulfurtado91@gmail.com> Tested-by: Paul Furtado <paulfurtado91@gmail.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> [3.12+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-07-30 23:08:33 +00:00
spin_lock(&memcg_oom_lock);
list_for_each_entry(ev, &memcg->oom_notify, list)
eventfd_signal(ev->eventfd, 1);
memcg: oom_notify use-after-free fix Paul Furtado has reported the following GPF: general protection fault: 0000 [#1] SMP Modules linked in: ipv6 dm_mod xen_netfront coretemp hwmon x86_pkg_temp_thermal crc32_pclmul crc32c_intel ghash_clmulni_intel aesni_intel ablk_helper cryptd lrw gf128mul glue_helper aes_x86_64 microcode pcspkr ext4 jbd2 mbcache raid0 xen_blkfront CPU: 3 PID: 3062 Comm: java Not tainted 3.16.0-rc5 #1 task: ffff8801cfe8f170 ti: ffff8801d2ec4000 task.ti: ffff8801d2ec4000 RIP: e030:mem_cgroup_oom_synchronize+0x140/0x240 RSP: e02b:ffff8801d2ec7d48 EFLAGS: 00010283 RAX: 0000000000000001 RBX: ffff88009d633800 RCX: 000000000000000e RDX: fffffffffffffffe RSI: ffff88009d630200 RDI: ffff88009d630200 RBP: ffff8801d2ec7da8 R08: 0000000000000012 R09: 00000000fffffffe R10: 0000000000000000 R11: 0000000000000000 R12: ffff88009d633800 R13: ffff8801d2ec7d48 R14: dead000000100100 R15: ffff88009d633a30 FS: 00007f1748bb4700(0000) GS:ffff8801def80000(0000) knlGS:0000000000000000 CS: e033 DS: 0000 ES: 0000 CR0: 000000008005003b CR2: 00007f4110300308 CR3: 00000000c05f7000 CR4: 0000000000002660 Call Trace: pagefault_out_of_memory+0x18/0x90 mm_fault_error+0xa9/0x1a0 __do_page_fault+0x478/0x4c0 do_page_fault+0x2c/0x40 page_fault+0x28/0x30 Code: 44 00 00 48 89 df e8 40 ca ff ff 48 85 c0 49 89 c4 74 35 4c 8b b0 30 02 00 00 4c 8d b8 30 02 00 00 4d 39 fe 74 1b 0f 1f 44 00 00 <49> 8b 7e 10 be 01 00 00 00 e8 42 d2 04 00 4d 8b 36 4d 39 fe 75 RIP mem_cgroup_oom_synchronize+0x140/0x240 Commit fb2a6fc56be6 ("mm: memcg: rework and document OOM waiting and wakeup") has moved mem_cgroup_oom_notify outside of memcg_oom_lock assuming it is protected by the hierarchical OOM-lock. Although this is true for the notification part the protection doesn't cover unregistration of event which can happen in parallel now so mem_cgroup_oom_notify can see already unlinked and/or freed mem_cgroup_eventfd_list. Fix this by using memcg_oom_lock also in mem_cgroup_oom_notify. Addresses https://bugzilla.kernel.org/show_bug.cgi?id=80881 Fixes: fb2a6fc56be6 (mm: memcg: rework and document OOM waiting and wakeup) Signed-off-by: Michal Hocko <mhocko@suse.cz> Reported-by: Paul Furtado <paulfurtado91@gmail.com> Tested-by: Paul Furtado <paulfurtado91@gmail.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> [3.12+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-07-30 23:08:33 +00:00
spin_unlock(&memcg_oom_lock);
return 0;
}
static void mem_cgroup_oom_notify(struct mem_cgroup *memcg)
{
struct mem_cgroup *iter;
for_each_mem_cgroup_tree(iter, memcg)
mem_cgroup_oom_notify_cb(iter);
}
static int __mem_cgroup_usage_register_event(struct mem_cgroup *memcg,
struct eventfd_ctx *eventfd, const char *args, enum res_type type)
{
struct mem_cgroup_thresholds *thresholds;
struct mem_cgroup_threshold_ary *new;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
unsigned long threshold;
unsigned long usage;
int i, size, ret;
ret = page_counter_memparse(args, "-1", &threshold);
if (ret)
return ret;
mutex_lock(&memcg->thresholds_lock);
if (type == _MEM) {
thresholds = &memcg->thresholds;
usage = mem_cgroup_usage(memcg, false);
} else if (type == _MEMSWAP) {
thresholds = &memcg->memsw_thresholds;
usage = mem_cgroup_usage(memcg, true);
} else
BUG();
/* Check if a threshold crossed before adding a new one */
if (thresholds->primary)
__mem_cgroup_threshold(memcg, type == _MEMSWAP);
size = thresholds->primary ? thresholds->primary->size + 1 : 1;
/* Allocate memory for new array of thresholds */
new = kmalloc(struct_size(new, entries, size), GFP_KERNEL);
if (!new) {
ret = -ENOMEM;
goto unlock;
}
new->size = size;
/* Copy thresholds (if any) to new array */
if (thresholds->primary) {
memcpy(new->entries, thresholds->primary->entries, (size - 1) *
sizeof(struct mem_cgroup_threshold));
}
/* Add new threshold */
new->entries[size - 1].eventfd = eventfd;
new->entries[size - 1].threshold = threshold;
/* Sort thresholds. Registering of new threshold isn't time-critical */
sort(new->entries, size, sizeof(struct mem_cgroup_threshold),
compare_thresholds, NULL);
/* Find current threshold */
new->current_threshold = -1;
for (i = 0; i < size; i++) {
if (new->entries[i].threshold <= usage) {
/*
* new->current_threshold will not be used until
* rcu_assign_pointer(), so it's safe to increment
* it here.
*/
++new->current_threshold;
} else
break;
}
/* Free old spare buffer and save old primary buffer as spare */
kfree(thresholds->spare);
thresholds->spare = thresholds->primary;
rcu_assign_pointer(thresholds->primary, new);
/* To be sure that nobody uses thresholds */
synchronize_rcu();
unlock:
mutex_unlock(&memcg->thresholds_lock);
return ret;
}
static int mem_cgroup_usage_register_event(struct mem_cgroup *memcg,
struct eventfd_ctx *eventfd, const char *args)
{
return __mem_cgroup_usage_register_event(memcg, eventfd, args, _MEM);
}
static int memsw_cgroup_usage_register_event(struct mem_cgroup *memcg,
struct eventfd_ctx *eventfd, const char *args)
{
return __mem_cgroup_usage_register_event(memcg, eventfd, args, _MEMSWAP);
}
static void __mem_cgroup_usage_unregister_event(struct mem_cgroup *memcg,
struct eventfd_ctx *eventfd, enum res_type type)
{
struct mem_cgroup_thresholds *thresholds;
struct mem_cgroup_threshold_ary *new;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
unsigned long usage;
memcg: fix NULL pointer dereference in __mem_cgroup_usage_unregister_event An eventfd monitors multiple memory thresholds of the cgroup, closes them, the kernel deletes all events related to this eventfd. Before all events are deleted, another eventfd monitors the memory threshold of this cgroup, leading to a crash: BUG: kernel NULL pointer dereference, address: 0000000000000004 #PF: supervisor write access in kernel mode #PF: error_code(0x0002) - not-present page PGD 800000033058e067 P4D 800000033058e067 PUD 3355ce067 PMD 0 Oops: 0002 [#1] SMP PTI CPU: 2 PID: 14012 Comm: kworker/2:6 Kdump: loaded Not tainted 5.6.0-rc4 #3 Hardware name: LENOVO 20AWS01K00/20AWS01K00, BIOS GLET70WW (2.24 ) 05/21/2014 Workqueue: events memcg_event_remove RIP: 0010:__mem_cgroup_usage_unregister_event+0xb3/0x190 RSP: 0018:ffffb47e01c4fe18 EFLAGS: 00010202 RAX: 0000000000000001 RBX: ffff8bb223a8a000 RCX: 0000000000000001 RDX: 0000000000000001 RSI: ffff8bb22fb83540 RDI: 0000000000000001 RBP: ffffb47e01c4fe48 R08: 0000000000000000 R09: 0000000000000010 R10: 000000000000000c R11: 071c71c71c71c71c R12: ffff8bb226aba880 R13: ffff8bb223a8a480 R14: 0000000000000000 R15: 0000000000000000 FS:  0000000000000000(0000) GS:ffff8bb242680000(0000) knlGS:0000000000000000 CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000004 CR3: 000000032c29c003 CR4: 00000000001606e0 Call Trace: memcg_event_remove+0x32/0x90 process_one_work+0x172/0x380 worker_thread+0x49/0x3f0 kthread+0xf8/0x130 ret_from_fork+0x35/0x40 CR2: 0000000000000004 We can reproduce this problem in the following ways: 1. We create a new cgroup subdirectory and a new eventfd, and then we monitor multiple memory thresholds of the cgroup through this eventfd. 2. closing this eventfd, and __mem_cgroup_usage_unregister_event () will be called multiple times to delete all events related to this eventfd. The first time __mem_cgroup_usage_unregister_event() is called, the kernel will clear all items related to this eventfd in thresholds-> primary. Since there is currently only one eventfd, thresholds-> primary becomes empty, so the kernel will set thresholds-> primary and hresholds-> spare to NULL. If at this time, the user creates a new eventfd and monitor the memory threshold of this cgroup, kernel will re-initialize thresholds-> primary. Then when __mem_cgroup_usage_unregister_event () is called for the second time, because thresholds-> primary is not empty, the system will access thresholds-> spare, but thresholds-> spare is NULL, which will trigger a crash. In general, the longer it takes to delete all events related to this eventfd, the easier it is to trigger this problem. The solution is to check whether the thresholds associated with the eventfd has been cleared when deleting the event. If so, we do nothing. [akpm@linux-foundation.org: fix comment, per Kirill] Fixes: 907860ed381a ("cgroups: make cftype.unregister_event() void-returning") Signed-off-by: Chunguang Xu <brookxu@tencent.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: <stable@vger.kernel.org> Link: http://lkml.kernel.org/r/077a6f67-aefa-4591-efec-f2f3af2b0b02@gmail.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-03-22 01:22:10 +00:00
int i, j, size, entries;
mutex_lock(&memcg->thresholds_lock);
if (type == _MEM) {
thresholds = &memcg->thresholds;
usage = mem_cgroup_usage(memcg, false);
} else if (type == _MEMSWAP) {
thresholds = &memcg->memsw_thresholds;
usage = mem_cgroup_usage(memcg, true);
} else
BUG();
mm: memcg: Correct unregistring of events attached to the same eventfd There is an issue when memcg unregisters events that were attached to the same eventfd: - On the first call mem_cgroup_usage_unregister_event() removes all events attached to a given eventfd, and if there were no events left, thresholds->primary would become NULL; - Since there were several events registered, cgroups core will call mem_cgroup_usage_unregister_event() again, but now kernel will oops, as the function doesn't expect that threshold->primary may be NULL. That's a good question whether mem_cgroup_usage_unregister_event() should actually remove all events in one go, but nowadays it can't do any better as cftype->unregister_event callback doesn't pass any private event-associated cookie. So, let's fix the issue by simply checking for threshold->primary. FWIW, w/o the patch the following oops may be observed: BUG: unable to handle kernel NULL pointer dereference at 0000000000000004 IP: [<ffffffff810be32c>] mem_cgroup_usage_unregister_event+0x9c/0x1f0 Pid: 574, comm: kworker/0:2 Not tainted 3.3.0-rc4+ #9 Bochs Bochs RIP: 0010:[<ffffffff810be32c>] [<ffffffff810be32c>] mem_cgroup_usage_unregister_event+0x9c/0x1f0 RSP: 0018:ffff88001d0b9d60 EFLAGS: 00010246 Process kworker/0:2 (pid: 574, threadinfo ffff88001d0b8000, task ffff88001de91cc0) Call Trace: [<ffffffff8107092b>] cgroup_event_remove+0x2b/0x60 [<ffffffff8103db94>] process_one_work+0x174/0x450 [<ffffffff8103e413>] worker_thread+0x123/0x2d0 Cc: stable <stable@vger.kernel.org> Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kirill A. Shutemov <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-02-24 01:14:46 +00:00
if (!thresholds->primary)
goto unlock;
/* Check if a threshold crossed before removing */
__mem_cgroup_threshold(memcg, type == _MEMSWAP);
/* Calculate new number of threshold */
memcg: fix NULL pointer dereference in __mem_cgroup_usage_unregister_event An eventfd monitors multiple memory thresholds of the cgroup, closes them, the kernel deletes all events related to this eventfd. Before all events are deleted, another eventfd monitors the memory threshold of this cgroup, leading to a crash: BUG: kernel NULL pointer dereference, address: 0000000000000004 #PF: supervisor write access in kernel mode #PF: error_code(0x0002) - not-present page PGD 800000033058e067 P4D 800000033058e067 PUD 3355ce067 PMD 0 Oops: 0002 [#1] SMP PTI CPU: 2 PID: 14012 Comm: kworker/2:6 Kdump: loaded Not tainted 5.6.0-rc4 #3 Hardware name: LENOVO 20AWS01K00/20AWS01K00, BIOS GLET70WW (2.24 ) 05/21/2014 Workqueue: events memcg_event_remove RIP: 0010:__mem_cgroup_usage_unregister_event+0xb3/0x190 RSP: 0018:ffffb47e01c4fe18 EFLAGS: 00010202 RAX: 0000000000000001 RBX: ffff8bb223a8a000 RCX: 0000000000000001 RDX: 0000000000000001 RSI: ffff8bb22fb83540 RDI: 0000000000000001 RBP: ffffb47e01c4fe48 R08: 0000000000000000 R09: 0000000000000010 R10: 000000000000000c R11: 071c71c71c71c71c R12: ffff8bb226aba880 R13: ffff8bb223a8a480 R14: 0000000000000000 R15: 0000000000000000 FS:  0000000000000000(0000) GS:ffff8bb242680000(0000) knlGS:0000000000000000 CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000004 CR3: 000000032c29c003 CR4: 00000000001606e0 Call Trace: memcg_event_remove+0x32/0x90 process_one_work+0x172/0x380 worker_thread+0x49/0x3f0 kthread+0xf8/0x130 ret_from_fork+0x35/0x40 CR2: 0000000000000004 We can reproduce this problem in the following ways: 1. We create a new cgroup subdirectory and a new eventfd, and then we monitor multiple memory thresholds of the cgroup through this eventfd. 2. closing this eventfd, and __mem_cgroup_usage_unregister_event () will be called multiple times to delete all events related to this eventfd. The first time __mem_cgroup_usage_unregister_event() is called, the kernel will clear all items related to this eventfd in thresholds-> primary. Since there is currently only one eventfd, thresholds-> primary becomes empty, so the kernel will set thresholds-> primary and hresholds-> spare to NULL. If at this time, the user creates a new eventfd and monitor the memory threshold of this cgroup, kernel will re-initialize thresholds-> primary. Then when __mem_cgroup_usage_unregister_event () is called for the second time, because thresholds-> primary is not empty, the system will access thresholds-> spare, but thresholds-> spare is NULL, which will trigger a crash. In general, the longer it takes to delete all events related to this eventfd, the easier it is to trigger this problem. The solution is to check whether the thresholds associated with the eventfd has been cleared when deleting the event. If so, we do nothing. [akpm@linux-foundation.org: fix comment, per Kirill] Fixes: 907860ed381a ("cgroups: make cftype.unregister_event() void-returning") Signed-off-by: Chunguang Xu <brookxu@tencent.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: <stable@vger.kernel.org> Link: http://lkml.kernel.org/r/077a6f67-aefa-4591-efec-f2f3af2b0b02@gmail.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-03-22 01:22:10 +00:00
size = entries = 0;
for (i = 0; i < thresholds->primary->size; i++) {
if (thresholds->primary->entries[i].eventfd != eventfd)
size++;
memcg: fix NULL pointer dereference in __mem_cgroup_usage_unregister_event An eventfd monitors multiple memory thresholds of the cgroup, closes them, the kernel deletes all events related to this eventfd. Before all events are deleted, another eventfd monitors the memory threshold of this cgroup, leading to a crash: BUG: kernel NULL pointer dereference, address: 0000000000000004 #PF: supervisor write access in kernel mode #PF: error_code(0x0002) - not-present page PGD 800000033058e067 P4D 800000033058e067 PUD 3355ce067 PMD 0 Oops: 0002 [#1] SMP PTI CPU: 2 PID: 14012 Comm: kworker/2:6 Kdump: loaded Not tainted 5.6.0-rc4 #3 Hardware name: LENOVO 20AWS01K00/20AWS01K00, BIOS GLET70WW (2.24 ) 05/21/2014 Workqueue: events memcg_event_remove RIP: 0010:__mem_cgroup_usage_unregister_event+0xb3/0x190 RSP: 0018:ffffb47e01c4fe18 EFLAGS: 00010202 RAX: 0000000000000001 RBX: ffff8bb223a8a000 RCX: 0000000000000001 RDX: 0000000000000001 RSI: ffff8bb22fb83540 RDI: 0000000000000001 RBP: ffffb47e01c4fe48 R08: 0000000000000000 R09: 0000000000000010 R10: 000000000000000c R11: 071c71c71c71c71c R12: ffff8bb226aba880 R13: ffff8bb223a8a480 R14: 0000000000000000 R15: 0000000000000000 FS:  0000000000000000(0000) GS:ffff8bb242680000(0000) knlGS:0000000000000000 CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000004 CR3: 000000032c29c003 CR4: 00000000001606e0 Call Trace: memcg_event_remove+0x32/0x90 process_one_work+0x172/0x380 worker_thread+0x49/0x3f0 kthread+0xf8/0x130 ret_from_fork+0x35/0x40 CR2: 0000000000000004 We can reproduce this problem in the following ways: 1. We create a new cgroup subdirectory and a new eventfd, and then we monitor multiple memory thresholds of the cgroup through this eventfd. 2. closing this eventfd, and __mem_cgroup_usage_unregister_event () will be called multiple times to delete all events related to this eventfd. The first time __mem_cgroup_usage_unregister_event() is called, the kernel will clear all items related to this eventfd in thresholds-> primary. Since there is currently only one eventfd, thresholds-> primary becomes empty, so the kernel will set thresholds-> primary and hresholds-> spare to NULL. If at this time, the user creates a new eventfd and monitor the memory threshold of this cgroup, kernel will re-initialize thresholds-> primary. Then when __mem_cgroup_usage_unregister_event () is called for the second time, because thresholds-> primary is not empty, the system will access thresholds-> spare, but thresholds-> spare is NULL, which will trigger a crash. In general, the longer it takes to delete all events related to this eventfd, the easier it is to trigger this problem. The solution is to check whether the thresholds associated with the eventfd has been cleared when deleting the event. If so, we do nothing. [akpm@linux-foundation.org: fix comment, per Kirill] Fixes: 907860ed381a ("cgroups: make cftype.unregister_event() void-returning") Signed-off-by: Chunguang Xu <brookxu@tencent.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: <stable@vger.kernel.org> Link: http://lkml.kernel.org/r/077a6f67-aefa-4591-efec-f2f3af2b0b02@gmail.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-03-22 01:22:10 +00:00
else
entries++;
}
new = thresholds->spare;
memcg: fix NULL pointer dereference in __mem_cgroup_usage_unregister_event An eventfd monitors multiple memory thresholds of the cgroup, closes them, the kernel deletes all events related to this eventfd. Before all events are deleted, another eventfd monitors the memory threshold of this cgroup, leading to a crash: BUG: kernel NULL pointer dereference, address: 0000000000000004 #PF: supervisor write access in kernel mode #PF: error_code(0x0002) - not-present page PGD 800000033058e067 P4D 800000033058e067 PUD 3355ce067 PMD 0 Oops: 0002 [#1] SMP PTI CPU: 2 PID: 14012 Comm: kworker/2:6 Kdump: loaded Not tainted 5.6.0-rc4 #3 Hardware name: LENOVO 20AWS01K00/20AWS01K00, BIOS GLET70WW (2.24 ) 05/21/2014 Workqueue: events memcg_event_remove RIP: 0010:__mem_cgroup_usage_unregister_event+0xb3/0x190 RSP: 0018:ffffb47e01c4fe18 EFLAGS: 00010202 RAX: 0000000000000001 RBX: ffff8bb223a8a000 RCX: 0000000000000001 RDX: 0000000000000001 RSI: ffff8bb22fb83540 RDI: 0000000000000001 RBP: ffffb47e01c4fe48 R08: 0000000000000000 R09: 0000000000000010 R10: 000000000000000c R11: 071c71c71c71c71c R12: ffff8bb226aba880 R13: ffff8bb223a8a480 R14: 0000000000000000 R15: 0000000000000000 FS:  0000000000000000(0000) GS:ffff8bb242680000(0000) knlGS:0000000000000000 CS:  0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000004 CR3: 000000032c29c003 CR4: 00000000001606e0 Call Trace: memcg_event_remove+0x32/0x90 process_one_work+0x172/0x380 worker_thread+0x49/0x3f0 kthread+0xf8/0x130 ret_from_fork+0x35/0x40 CR2: 0000000000000004 We can reproduce this problem in the following ways: 1. We create a new cgroup subdirectory and a new eventfd, and then we monitor multiple memory thresholds of the cgroup through this eventfd. 2. closing this eventfd, and __mem_cgroup_usage_unregister_event () will be called multiple times to delete all events related to this eventfd. The first time __mem_cgroup_usage_unregister_event() is called, the kernel will clear all items related to this eventfd in thresholds-> primary. Since there is currently only one eventfd, thresholds-> primary becomes empty, so the kernel will set thresholds-> primary and hresholds-> spare to NULL. If at this time, the user creates a new eventfd and monitor the memory threshold of this cgroup, kernel will re-initialize thresholds-> primary. Then when __mem_cgroup_usage_unregister_event () is called for the second time, because thresholds-> primary is not empty, the system will access thresholds-> spare, but thresholds-> spare is NULL, which will trigger a crash. In general, the longer it takes to delete all events related to this eventfd, the easier it is to trigger this problem. The solution is to check whether the thresholds associated with the eventfd has been cleared when deleting the event. If so, we do nothing. [akpm@linux-foundation.org: fix comment, per Kirill] Fixes: 907860ed381a ("cgroups: make cftype.unregister_event() void-returning") Signed-off-by: Chunguang Xu <brookxu@tencent.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: <stable@vger.kernel.org> Link: http://lkml.kernel.org/r/077a6f67-aefa-4591-efec-f2f3af2b0b02@gmail.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-03-22 01:22:10 +00:00
/* If no items related to eventfd have been cleared, nothing to do */
if (!entries)
goto unlock;
/* Set thresholds array to NULL if we don't have thresholds */
if (!size) {
kfree(new);
new = NULL;
goto swap_buffers;
}
new->size = size;
/* Copy thresholds and find current threshold */
new->current_threshold = -1;
for (i = 0, j = 0; i < thresholds->primary->size; i++) {
if (thresholds->primary->entries[i].eventfd == eventfd)
continue;
new->entries[j] = thresholds->primary->entries[i];
if (new->entries[j].threshold <= usage) {
/*
* new->current_threshold will not be used
* until rcu_assign_pointer(), so it's safe to increment
* it here.
*/
++new->current_threshold;
}
j++;
}
swap_buffers:
/* Swap primary and spare array */
thresholds->spare = thresholds->primary;
rcu_assign_pointer(thresholds->primary, new);
/* To be sure that nobody uses thresholds */
synchronize_rcu();
/* If all events are unregistered, free the spare array */
if (!new) {
kfree(thresholds->spare);
thresholds->spare = NULL;
}
mm: memcg: Correct unregistring of events attached to the same eventfd There is an issue when memcg unregisters events that were attached to the same eventfd: - On the first call mem_cgroup_usage_unregister_event() removes all events attached to a given eventfd, and if there were no events left, thresholds->primary would become NULL; - Since there were several events registered, cgroups core will call mem_cgroup_usage_unregister_event() again, but now kernel will oops, as the function doesn't expect that threshold->primary may be NULL. That's a good question whether mem_cgroup_usage_unregister_event() should actually remove all events in one go, but nowadays it can't do any better as cftype->unregister_event callback doesn't pass any private event-associated cookie. So, let's fix the issue by simply checking for threshold->primary. FWIW, w/o the patch the following oops may be observed: BUG: unable to handle kernel NULL pointer dereference at 0000000000000004 IP: [<ffffffff810be32c>] mem_cgroup_usage_unregister_event+0x9c/0x1f0 Pid: 574, comm: kworker/0:2 Not tainted 3.3.0-rc4+ #9 Bochs Bochs RIP: 0010:[<ffffffff810be32c>] [<ffffffff810be32c>] mem_cgroup_usage_unregister_event+0x9c/0x1f0 RSP: 0018:ffff88001d0b9d60 EFLAGS: 00010246 Process kworker/0:2 (pid: 574, threadinfo ffff88001d0b8000, task ffff88001de91cc0) Call Trace: [<ffffffff8107092b>] cgroup_event_remove+0x2b/0x60 [<ffffffff8103db94>] process_one_work+0x174/0x450 [<ffffffff8103e413>] worker_thread+0x123/0x2d0 Cc: stable <stable@vger.kernel.org> Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Kirill A. Shutemov <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-02-24 01:14:46 +00:00
unlock:
mutex_unlock(&memcg->thresholds_lock);
}
static void mem_cgroup_usage_unregister_event(struct mem_cgroup *memcg,
struct eventfd_ctx *eventfd)
{
return __mem_cgroup_usage_unregister_event(memcg, eventfd, _MEM);
}
static void memsw_cgroup_usage_unregister_event(struct mem_cgroup *memcg,
struct eventfd_ctx *eventfd)
{
return __mem_cgroup_usage_unregister_event(memcg, eventfd, _MEMSWAP);
}
static int mem_cgroup_oom_register_event(struct mem_cgroup *memcg,
struct eventfd_ctx *eventfd, const char *args)
{
struct mem_cgroup_eventfd_list *event;
event = kmalloc(sizeof(*event), GFP_KERNEL);
if (!event)
return -ENOMEM;
spin_lock(&memcg_oom_lock);
event->eventfd = eventfd;
list_add(&event->list, &memcg->oom_notify);
/* already in OOM ? */
if (memcg->under_oom)
eventfd_signal(eventfd, 1);
spin_unlock(&memcg_oom_lock);
return 0;
}
static void mem_cgroup_oom_unregister_event(struct mem_cgroup *memcg,
struct eventfd_ctx *eventfd)
{
struct mem_cgroup_eventfd_list *ev, *tmp;
spin_lock(&memcg_oom_lock);
list_for_each_entry_safe(ev, tmp, &memcg->oom_notify, list) {
if (ev->eventfd == eventfd) {
list_del(&ev->list);
kfree(ev);
}
}
spin_unlock(&memcg_oom_lock);
}
static int mem_cgroup_oom_control_read(struct seq_file *sf, void *v)
{
struct mem_cgroup *memcg = mem_cgroup_from_seq(sf);
seq_printf(sf, "oom_kill_disable %d\n", memcg->oom_kill_disable);
seq_printf(sf, "under_oom %d\n", (bool)memcg->under_oom);
seq_printf(sf, "oom_kill %lu\n",
atomic_long_read(&memcg->memory_events[MEMCG_OOM_KILL]));
return 0;
}
cgroup: pass around cgroup_subsys_state instead of cgroup in file methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup. Please see the previous commit which converts the subsystem methods for rationale. This patch converts all cftype file operations to take @css instead of @cgroup. cftypes for the cgroup core files don't have their subsytem pointer set. These will automatically use the dummy_css added by the previous patch and can be converted the same way. Most subsystem conversions are straight forwards but there are some interesting ones. * freezer: update_if_frozen() is also converted to take @css instead of @cgroup for consistency. This will make the code look simpler too once iterators are converted to use css. * memory/vmpressure: mem_cgroup_from_css() needs to be exported to vmpressure while mem_cgroup_from_cont() can be made static. Updated accordingly. * cpu: cgroup_tg() doesn't have any user left. Removed. * cpuacct: cgroup_ca() doesn't have any user left. Removed. * hugetlb: hugetlb_cgroup_form_cgroup() doesn't have any user left. Removed. * net_cls: cgrp_cls_state() doesn't have any user left. Removed. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:24 +00:00
static int mem_cgroup_oom_control_write(struct cgroup_subsys_state *css,
struct cftype *cft, u64 val)
{
cgroup: pass around cgroup_subsys_state instead of cgroup in file methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup. Please see the previous commit which converts the subsystem methods for rationale. This patch converts all cftype file operations to take @css instead of @cgroup. cftypes for the cgroup core files don't have their subsytem pointer set. These will automatically use the dummy_css added by the previous patch and can be converted the same way. Most subsystem conversions are straight forwards but there are some interesting ones. * freezer: update_if_frozen() is also converted to take @css instead of @cgroup for consistency. This will make the code look simpler too once iterators are converted to use css. * memory/vmpressure: mem_cgroup_from_css() needs to be exported to vmpressure while mem_cgroup_from_cont() can be made static. Updated accordingly. * cpu: cgroup_tg() doesn't have any user left. Removed. * cpuacct: cgroup_ca() doesn't have any user left. Removed. * hugetlb: hugetlb_cgroup_form_cgroup() doesn't have any user left. Removed. * net_cls: cgrp_cls_state() doesn't have any user left. Removed. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:24 +00:00
struct mem_cgroup *memcg = mem_cgroup_from_css(css);
/* cannot set to root cgroup and only 0 and 1 are allowed */
Merge branch 'for-3.16' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup Pull cgroup updates from Tejun Heo: "A lot of activities on cgroup side. Heavy restructuring including locking simplification took place to improve the code base and enable implementation of the unified hierarchy, which currently exists behind a __DEVEL__ mount option. The core support is mostly complete but individual controllers need further work. To explain the design and rationales of the the unified hierarchy Documentation/cgroups/unified-hierarchy.txt is added. Another notable change is css (cgroup_subsys_state - what each controller uses to identify and interact with a cgroup) iteration update. This is part of continuing updates on css object lifetime and visibility. cgroup started with reference count draining on removal way back and is now reaching a point where csses behave and are iterated like normal refcnted objects albeit with some complexities to allow distinguishing the state where they're being deleted. The css iteration update isn't taken advantage of yet but is planned to be used to simplify memcg significantly" * 'for-3.16' of git://git.kernel.org/pub/scm/linux/kernel/git/tj/cgroup: (77 commits) cgroup: disallow disabled controllers on the default hierarchy cgroup: don't destroy the default root cgroup: disallow debug controller on the default hierarchy cgroup: clean up MAINTAINERS entries cgroup: implement css_tryget() device_cgroup: use css_has_online_children() instead of has_children() cgroup: convert cgroup_has_live_children() into css_has_online_children() cgroup: use CSS_ONLINE instead of CGRP_DEAD cgroup: iterate cgroup_subsys_states directly cgroup: introduce CSS_RELEASED and reduce css iteration fallback window cgroup: move cgroup->serial_nr into cgroup_subsys_state cgroup: link all cgroup_subsys_states in their sibling lists cgroup: move cgroup->sibling and ->children into cgroup_subsys_state cgroup: remove cgroup->parent device_cgroup: remove direct access to cgroup->children memcg: update memcg_has_children() to use css_next_child() memcg: remove tasks/children test from mem_cgroup_force_empty() cgroup: remove css_parent() cgroup: skip refcnting on normal root csses and cgrp_dfl_root self css cgroup: use cgroup->self.refcnt for cgroup refcnting ...
2014-06-09 22:03:33 +00:00
if (!css->parent || !((val == 0) || (val == 1)))
return -EINVAL;
memcg->oom_kill_disable = val;
if (!val)
memcg_oom_recover(memcg);
return 0;
}
writeback: make backing_dev_info host cgroup-specific bdi_writebacks For the planned cgroup writeback support, on each bdi (backing_dev_info), each memcg will be served by a separate wb (bdi_writeback). This patch updates bdi so that a bdi can host multiple wbs (bdi_writebacks). On the default hierarchy, blkcg implicitly enables memcg. This allows using memcg's page ownership for attributing writeback IOs, and every memcg - blkcg combination can be served by its own wb by assigning a dedicated wb to each memcg. This means that there may be multiple wb's of a bdi mapped to the same blkcg. As congested state is per blkcg - bdi combination, those wb's should share the same congested state. This is achieved by tracking congested state via bdi_writeback_congested structs which are keyed by blkcg. bdi->wb remains unchanged and will keep serving the root cgroup. cgwb's (cgroup wb's) for non-root cgroups are created on-demand or looked up while dirtying an inode according to the memcg of the page being dirtied or current task. Each cgwb is indexed on bdi->cgwb_tree by its memcg id. Once an inode is associated with its wb, it can be retrieved using inode_to_wb(). Currently, none of the filesystems has FS_CGROUP_WRITEBACK and all pages will keep being associated with bdi->wb. v3: inode_attach_wb() in account_page_dirtied() moved inside mapping_cap_account_dirty() block where it's known to be !NULL. Also, an unnecessary NULL check before kfree() removed. Both detected by the kbuild bot. v2: Updated so that wb association is per inode and wb is per memcg rather than blkcg. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: kbuild test robot <fengguang.wu@intel.com> Cc: Dan Carpenter <dan.carpenter@oracle.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-22 21:13:37 +00:00
#ifdef CONFIG_CGROUP_WRITEBACK
#include <trace/events/writeback.h>
static int memcg_wb_domain_init(struct mem_cgroup *memcg, gfp_t gfp)
{
return wb_domain_init(&memcg->cgwb_domain, gfp);
}
static void memcg_wb_domain_exit(struct mem_cgroup *memcg)
{
wb_domain_exit(&memcg->cgwb_domain);
}
static void memcg_wb_domain_size_changed(struct mem_cgroup *memcg)
{
wb_domain_size_changed(&memcg->cgwb_domain);
}
struct wb_domain *mem_cgroup_wb_domain(struct bdi_writeback *wb)
{
struct mem_cgroup *memcg = mem_cgroup_from_css(wb->memcg_css);
if (!memcg->css.parent)
return NULL;
return &memcg->cgwb_domain;
}
mm: writeback: use exact memcg dirty counts Since commit a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting") memcg dirty and writeback counters are managed as: 1) per-memcg per-cpu values in range of [-32..32] 2) per-memcg atomic counter When a per-cpu counter cannot fit in [-32..32] it's flushed to the atomic. Stat readers only check the atomic. Thus readers such as balance_dirty_pages() may see a nontrivial error margin: 32 pages per cpu. Assuming 100 cpus: 4k x86 page_size: 13 MiB error per memcg 64k ppc page_size: 200 MiB error per memcg Considering that dirty+writeback are used together for some decisions the errors double. This inaccuracy can lead to undeserved oom kills. One nasty case is when all per-cpu counters hold positive values offsetting an atomic negative value (i.e. per_cpu[*]=32, atomic=n_cpu*-32). balance_dirty_pages() only consults the atomic and does not consider throttling the next n_cpu*32 dirty pages. If the file_lru is in the 13..200 MiB range then there's absolutely no dirty throttling, which burdens vmscan with only dirty+writeback pages thus resorting to oom kill. It could be argued that tiny containers are not supported, but it's more subtle. It's the amount the space available for file lru that matters. If a container has memory.max-200MiB of non reclaimable memory, then it will also suffer such oom kills on a 100 cpu machine. The following test reliably ooms without this patch. This patch avoids oom kills. $ cat test mount -t cgroup2 none /dev/cgroup cd /dev/cgroup echo +io +memory > cgroup.subtree_control mkdir test cd test echo 10M > memory.max (echo $BASHPID > cgroup.procs && exec /memcg-writeback-stress /foo) (echo $BASHPID > cgroup.procs && exec dd if=/dev/zero of=/foo bs=2M count=100) $ cat memcg-writeback-stress.c /* * Dirty pages from all but one cpu. * Clean pages from the non dirtying cpu. * This is to stress per cpu counter imbalance. * On a 100 cpu machine: * - per memcg per cpu dirty count is 32 pages for each of 99 cpus * - per memcg atomic is -99*32 pages * - thus the complete dirty limit: sum of all counters 0 * - balance_dirty_pages() only sees atomic count -99*32 pages, which * it max()s to 0. * - So a workload can dirty -99*32 pages before balance_dirty_pages() * cares. */ #define _GNU_SOURCE #include <err.h> #include <fcntl.h> #include <sched.h> #include <stdlib.h> #include <stdio.h> #include <sys/stat.h> #include <sys/sysinfo.h> #include <sys/types.h> #include <unistd.h> static char *buf; static int bufSize; static void set_affinity(int cpu) { cpu_set_t affinity; CPU_ZERO(&affinity); CPU_SET(cpu, &affinity); if (sched_setaffinity(0, sizeof(affinity), &affinity)) err(1, "sched_setaffinity"); } static void dirty_on(int output_fd, int cpu) { int i, wrote; set_affinity(cpu); for (i = 0; i < 32; i++) { for (wrote = 0; wrote < bufSize; ) { int ret = write(output_fd, buf+wrote, bufSize-wrote); if (ret == -1) err(1, "write"); wrote += ret; } } } int main(int argc, char **argv) { int cpu, flush_cpu = 1, output_fd; const char *output; if (argc != 2) errx(1, "usage: output_file"); output = argv[1]; bufSize = getpagesize(); buf = malloc(getpagesize()); if (buf == NULL) errx(1, "malloc failed"); output_fd = open(output, O_CREAT|O_RDWR); if (output_fd == -1) err(1, "open(%s)", output); for (cpu = 0; cpu < get_nprocs(); cpu++) { if (cpu != flush_cpu) dirty_on(output_fd, cpu); } set_affinity(flush_cpu); if (fsync(output_fd)) err(1, "fsync(%s)", output); if (close(output_fd)) err(1, "close(%s)", output); free(buf); } Make balance_dirty_pages() and wb_over_bg_thresh() work harder to collect exact per memcg counters. This avoids the aforementioned oom kills. This does not affect the overhead of memory.stat, which still reads the single atomic counter. Why not use percpu_counter? memcg already handles cpus going offline, so no need for that overhead from percpu_counter. And the percpu_counter spinlocks are more heavyweight than is required. It probably also makes sense to use exact dirty and writeback counters in memcg oom reports. But that is saved for later. Link: http://lkml.kernel.org/r/20190329174609.164344-1-gthelen@google.com Signed-off-by: Greg Thelen <gthelen@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Cc: <stable@vger.kernel.org> [4.16+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-04-06 01:39:18 +00:00
/*
* idx can be of type enum memcg_stat_item or node_stat_item.
* Keep in sync with memcg_exact_page().
*/
static unsigned long memcg_exact_page_state(struct mem_cgroup *memcg, int idx)
{
long x = atomic_long_read(&memcg->vmstats[idx]);
mm: writeback: use exact memcg dirty counts Since commit a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting") memcg dirty and writeback counters are managed as: 1) per-memcg per-cpu values in range of [-32..32] 2) per-memcg atomic counter When a per-cpu counter cannot fit in [-32..32] it's flushed to the atomic. Stat readers only check the atomic. Thus readers such as balance_dirty_pages() may see a nontrivial error margin: 32 pages per cpu. Assuming 100 cpus: 4k x86 page_size: 13 MiB error per memcg 64k ppc page_size: 200 MiB error per memcg Considering that dirty+writeback are used together for some decisions the errors double. This inaccuracy can lead to undeserved oom kills. One nasty case is when all per-cpu counters hold positive values offsetting an atomic negative value (i.e. per_cpu[*]=32, atomic=n_cpu*-32). balance_dirty_pages() only consults the atomic and does not consider throttling the next n_cpu*32 dirty pages. If the file_lru is in the 13..200 MiB range then there's absolutely no dirty throttling, which burdens vmscan with only dirty+writeback pages thus resorting to oom kill. It could be argued that tiny containers are not supported, but it's more subtle. It's the amount the space available for file lru that matters. If a container has memory.max-200MiB of non reclaimable memory, then it will also suffer such oom kills on a 100 cpu machine. The following test reliably ooms without this patch. This patch avoids oom kills. $ cat test mount -t cgroup2 none /dev/cgroup cd /dev/cgroup echo +io +memory > cgroup.subtree_control mkdir test cd test echo 10M > memory.max (echo $BASHPID > cgroup.procs && exec /memcg-writeback-stress /foo) (echo $BASHPID > cgroup.procs && exec dd if=/dev/zero of=/foo bs=2M count=100) $ cat memcg-writeback-stress.c /* * Dirty pages from all but one cpu. * Clean pages from the non dirtying cpu. * This is to stress per cpu counter imbalance. * On a 100 cpu machine: * - per memcg per cpu dirty count is 32 pages for each of 99 cpus * - per memcg atomic is -99*32 pages * - thus the complete dirty limit: sum of all counters 0 * - balance_dirty_pages() only sees atomic count -99*32 pages, which * it max()s to 0. * - So a workload can dirty -99*32 pages before balance_dirty_pages() * cares. */ #define _GNU_SOURCE #include <err.h> #include <fcntl.h> #include <sched.h> #include <stdlib.h> #include <stdio.h> #include <sys/stat.h> #include <sys/sysinfo.h> #include <sys/types.h> #include <unistd.h> static char *buf; static int bufSize; static void set_affinity(int cpu) { cpu_set_t affinity; CPU_ZERO(&affinity); CPU_SET(cpu, &affinity); if (sched_setaffinity(0, sizeof(affinity), &affinity)) err(1, "sched_setaffinity"); } static void dirty_on(int output_fd, int cpu) { int i, wrote; set_affinity(cpu); for (i = 0; i < 32; i++) { for (wrote = 0; wrote < bufSize; ) { int ret = write(output_fd, buf+wrote, bufSize-wrote); if (ret == -1) err(1, "write"); wrote += ret; } } } int main(int argc, char **argv) { int cpu, flush_cpu = 1, output_fd; const char *output; if (argc != 2) errx(1, "usage: output_file"); output = argv[1]; bufSize = getpagesize(); buf = malloc(getpagesize()); if (buf == NULL) errx(1, "malloc failed"); output_fd = open(output, O_CREAT|O_RDWR); if (output_fd == -1) err(1, "open(%s)", output); for (cpu = 0; cpu < get_nprocs(); cpu++) { if (cpu != flush_cpu) dirty_on(output_fd, cpu); } set_affinity(flush_cpu); if (fsync(output_fd)) err(1, "fsync(%s)", output); if (close(output_fd)) err(1, "close(%s)", output); free(buf); } Make balance_dirty_pages() and wb_over_bg_thresh() work harder to collect exact per memcg counters. This avoids the aforementioned oom kills. This does not affect the overhead of memory.stat, which still reads the single atomic counter. Why not use percpu_counter? memcg already handles cpus going offline, so no need for that overhead from percpu_counter. And the percpu_counter spinlocks are more heavyweight than is required. It probably also makes sense to use exact dirty and writeback counters in memcg oom reports. But that is saved for later. Link: http://lkml.kernel.org/r/20190329174609.164344-1-gthelen@google.com Signed-off-by: Greg Thelen <gthelen@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Cc: <stable@vger.kernel.org> [4.16+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-04-06 01:39:18 +00:00
int cpu;
for_each_online_cpu(cpu)
x += per_cpu_ptr(memcg->vmstats_percpu, cpu)->stat[idx];
mm: writeback: use exact memcg dirty counts Since commit a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting") memcg dirty and writeback counters are managed as: 1) per-memcg per-cpu values in range of [-32..32] 2) per-memcg atomic counter When a per-cpu counter cannot fit in [-32..32] it's flushed to the atomic. Stat readers only check the atomic. Thus readers such as balance_dirty_pages() may see a nontrivial error margin: 32 pages per cpu. Assuming 100 cpus: 4k x86 page_size: 13 MiB error per memcg 64k ppc page_size: 200 MiB error per memcg Considering that dirty+writeback are used together for some decisions the errors double. This inaccuracy can lead to undeserved oom kills. One nasty case is when all per-cpu counters hold positive values offsetting an atomic negative value (i.e. per_cpu[*]=32, atomic=n_cpu*-32). balance_dirty_pages() only consults the atomic and does not consider throttling the next n_cpu*32 dirty pages. If the file_lru is in the 13..200 MiB range then there's absolutely no dirty throttling, which burdens vmscan with only dirty+writeback pages thus resorting to oom kill. It could be argued that tiny containers are not supported, but it's more subtle. It's the amount the space available for file lru that matters. If a container has memory.max-200MiB of non reclaimable memory, then it will also suffer such oom kills on a 100 cpu machine. The following test reliably ooms without this patch. This patch avoids oom kills. $ cat test mount -t cgroup2 none /dev/cgroup cd /dev/cgroup echo +io +memory > cgroup.subtree_control mkdir test cd test echo 10M > memory.max (echo $BASHPID > cgroup.procs && exec /memcg-writeback-stress /foo) (echo $BASHPID > cgroup.procs && exec dd if=/dev/zero of=/foo bs=2M count=100) $ cat memcg-writeback-stress.c /* * Dirty pages from all but one cpu. * Clean pages from the non dirtying cpu. * This is to stress per cpu counter imbalance. * On a 100 cpu machine: * - per memcg per cpu dirty count is 32 pages for each of 99 cpus * - per memcg atomic is -99*32 pages * - thus the complete dirty limit: sum of all counters 0 * - balance_dirty_pages() only sees atomic count -99*32 pages, which * it max()s to 0. * - So a workload can dirty -99*32 pages before balance_dirty_pages() * cares. */ #define _GNU_SOURCE #include <err.h> #include <fcntl.h> #include <sched.h> #include <stdlib.h> #include <stdio.h> #include <sys/stat.h> #include <sys/sysinfo.h> #include <sys/types.h> #include <unistd.h> static char *buf; static int bufSize; static void set_affinity(int cpu) { cpu_set_t affinity; CPU_ZERO(&affinity); CPU_SET(cpu, &affinity); if (sched_setaffinity(0, sizeof(affinity), &affinity)) err(1, "sched_setaffinity"); } static void dirty_on(int output_fd, int cpu) { int i, wrote; set_affinity(cpu); for (i = 0; i < 32; i++) { for (wrote = 0; wrote < bufSize; ) { int ret = write(output_fd, buf+wrote, bufSize-wrote); if (ret == -1) err(1, "write"); wrote += ret; } } } int main(int argc, char **argv) { int cpu, flush_cpu = 1, output_fd; const char *output; if (argc != 2) errx(1, "usage: output_file"); output = argv[1]; bufSize = getpagesize(); buf = malloc(getpagesize()); if (buf == NULL) errx(1, "malloc failed"); output_fd = open(output, O_CREAT|O_RDWR); if (output_fd == -1) err(1, "open(%s)", output); for (cpu = 0; cpu < get_nprocs(); cpu++) { if (cpu != flush_cpu) dirty_on(output_fd, cpu); } set_affinity(flush_cpu); if (fsync(output_fd)) err(1, "fsync(%s)", output); if (close(output_fd)) err(1, "close(%s)", output); free(buf); } Make balance_dirty_pages() and wb_over_bg_thresh() work harder to collect exact per memcg counters. This avoids the aforementioned oom kills. This does not affect the overhead of memory.stat, which still reads the single atomic counter. Why not use percpu_counter? memcg already handles cpus going offline, so no need for that overhead from percpu_counter. And the percpu_counter spinlocks are more heavyweight than is required. It probably also makes sense to use exact dirty and writeback counters in memcg oom reports. But that is saved for later. Link: http://lkml.kernel.org/r/20190329174609.164344-1-gthelen@google.com Signed-off-by: Greg Thelen <gthelen@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Cc: <stable@vger.kernel.org> [4.16+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-04-06 01:39:18 +00:00
if (x < 0)
x = 0;
return x;
}
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-22 22:23:35 +00:00
/**
* mem_cgroup_wb_stats - retrieve writeback related stats from its memcg
* @wb: bdi_writeback in question
2015-09-29 17:04:26 +00:00
* @pfilepages: out parameter for number of file pages
* @pheadroom: out parameter for number of allocatable pages according to memcg
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-22 22:23:35 +00:00
* @pdirty: out parameter for number of dirty pages
* @pwriteback: out parameter for number of pages under writeback
*
2015-09-29 17:04:26 +00:00
* Determine the numbers of file, headroom, dirty, and writeback pages in
* @wb's memcg. File, dirty and writeback are self-explanatory. Headroom
* is a bit more involved.
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-22 22:23:35 +00:00
*
2015-09-29 17:04:26 +00:00
* A memcg's headroom is "min(max, high) - used". In the hierarchy, the
* headroom is calculated as the lowest headroom of itself and the
* ancestors. Note that this doesn't consider the actual amount of
* available memory in the system. The caller should further cap
* *@pheadroom accordingly.
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-22 22:23:35 +00:00
*/
2015-09-29 17:04:26 +00:00
void mem_cgroup_wb_stats(struct bdi_writeback *wb, unsigned long *pfilepages,
unsigned long *pheadroom, unsigned long *pdirty,
unsigned long *pwriteback)
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-22 22:23:35 +00:00
{
struct mem_cgroup *memcg = mem_cgroup_from_css(wb->memcg_css);
struct mem_cgroup *parent;
mm: writeback: use exact memcg dirty counts Since commit a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting") memcg dirty and writeback counters are managed as: 1) per-memcg per-cpu values in range of [-32..32] 2) per-memcg atomic counter When a per-cpu counter cannot fit in [-32..32] it's flushed to the atomic. Stat readers only check the atomic. Thus readers such as balance_dirty_pages() may see a nontrivial error margin: 32 pages per cpu. Assuming 100 cpus: 4k x86 page_size: 13 MiB error per memcg 64k ppc page_size: 200 MiB error per memcg Considering that dirty+writeback are used together for some decisions the errors double. This inaccuracy can lead to undeserved oom kills. One nasty case is when all per-cpu counters hold positive values offsetting an atomic negative value (i.e. per_cpu[*]=32, atomic=n_cpu*-32). balance_dirty_pages() only consults the atomic and does not consider throttling the next n_cpu*32 dirty pages. If the file_lru is in the 13..200 MiB range then there's absolutely no dirty throttling, which burdens vmscan with only dirty+writeback pages thus resorting to oom kill. It could be argued that tiny containers are not supported, but it's more subtle. It's the amount the space available for file lru that matters. If a container has memory.max-200MiB of non reclaimable memory, then it will also suffer such oom kills on a 100 cpu machine. The following test reliably ooms without this patch. This patch avoids oom kills. $ cat test mount -t cgroup2 none /dev/cgroup cd /dev/cgroup echo +io +memory > cgroup.subtree_control mkdir test cd test echo 10M > memory.max (echo $BASHPID > cgroup.procs && exec /memcg-writeback-stress /foo) (echo $BASHPID > cgroup.procs && exec dd if=/dev/zero of=/foo bs=2M count=100) $ cat memcg-writeback-stress.c /* * Dirty pages from all but one cpu. * Clean pages from the non dirtying cpu. * This is to stress per cpu counter imbalance. * On a 100 cpu machine: * - per memcg per cpu dirty count is 32 pages for each of 99 cpus * - per memcg atomic is -99*32 pages * - thus the complete dirty limit: sum of all counters 0 * - balance_dirty_pages() only sees atomic count -99*32 pages, which * it max()s to 0. * - So a workload can dirty -99*32 pages before balance_dirty_pages() * cares. */ #define _GNU_SOURCE #include <err.h> #include <fcntl.h> #include <sched.h> #include <stdlib.h> #include <stdio.h> #include <sys/stat.h> #include <sys/sysinfo.h> #include <sys/types.h> #include <unistd.h> static char *buf; static int bufSize; static void set_affinity(int cpu) { cpu_set_t affinity; CPU_ZERO(&affinity); CPU_SET(cpu, &affinity); if (sched_setaffinity(0, sizeof(affinity), &affinity)) err(1, "sched_setaffinity"); } static void dirty_on(int output_fd, int cpu) { int i, wrote; set_affinity(cpu); for (i = 0; i < 32; i++) { for (wrote = 0; wrote < bufSize; ) { int ret = write(output_fd, buf+wrote, bufSize-wrote); if (ret == -1) err(1, "write"); wrote += ret; } } } int main(int argc, char **argv) { int cpu, flush_cpu = 1, output_fd; const char *output; if (argc != 2) errx(1, "usage: output_file"); output = argv[1]; bufSize = getpagesize(); buf = malloc(getpagesize()); if (buf == NULL) errx(1, "malloc failed"); output_fd = open(output, O_CREAT|O_RDWR); if (output_fd == -1) err(1, "open(%s)", output); for (cpu = 0; cpu < get_nprocs(); cpu++) { if (cpu != flush_cpu) dirty_on(output_fd, cpu); } set_affinity(flush_cpu); if (fsync(output_fd)) err(1, "fsync(%s)", output); if (close(output_fd)) err(1, "close(%s)", output); free(buf); } Make balance_dirty_pages() and wb_over_bg_thresh() work harder to collect exact per memcg counters. This avoids the aforementioned oom kills. This does not affect the overhead of memory.stat, which still reads the single atomic counter. Why not use percpu_counter? memcg already handles cpus going offline, so no need for that overhead from percpu_counter. And the percpu_counter spinlocks are more heavyweight than is required. It probably also makes sense to use exact dirty and writeback counters in memcg oom reports. But that is saved for later. Link: http://lkml.kernel.org/r/20190329174609.164344-1-gthelen@google.com Signed-off-by: Greg Thelen <gthelen@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Cc: <stable@vger.kernel.org> [4.16+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-04-06 01:39:18 +00:00
*pdirty = memcg_exact_page_state(memcg, NR_FILE_DIRTY);
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-22 22:23:35 +00:00
mm: writeback: use exact memcg dirty counts Since commit a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting") memcg dirty and writeback counters are managed as: 1) per-memcg per-cpu values in range of [-32..32] 2) per-memcg atomic counter When a per-cpu counter cannot fit in [-32..32] it's flushed to the atomic. Stat readers only check the atomic. Thus readers such as balance_dirty_pages() may see a nontrivial error margin: 32 pages per cpu. Assuming 100 cpus: 4k x86 page_size: 13 MiB error per memcg 64k ppc page_size: 200 MiB error per memcg Considering that dirty+writeback are used together for some decisions the errors double. This inaccuracy can lead to undeserved oom kills. One nasty case is when all per-cpu counters hold positive values offsetting an atomic negative value (i.e. per_cpu[*]=32, atomic=n_cpu*-32). balance_dirty_pages() only consults the atomic and does not consider throttling the next n_cpu*32 dirty pages. If the file_lru is in the 13..200 MiB range then there's absolutely no dirty throttling, which burdens vmscan with only dirty+writeback pages thus resorting to oom kill. It could be argued that tiny containers are not supported, but it's more subtle. It's the amount the space available for file lru that matters. If a container has memory.max-200MiB of non reclaimable memory, then it will also suffer such oom kills on a 100 cpu machine. The following test reliably ooms without this patch. This patch avoids oom kills. $ cat test mount -t cgroup2 none /dev/cgroup cd /dev/cgroup echo +io +memory > cgroup.subtree_control mkdir test cd test echo 10M > memory.max (echo $BASHPID > cgroup.procs && exec /memcg-writeback-stress /foo) (echo $BASHPID > cgroup.procs && exec dd if=/dev/zero of=/foo bs=2M count=100) $ cat memcg-writeback-stress.c /* * Dirty pages from all but one cpu. * Clean pages from the non dirtying cpu. * This is to stress per cpu counter imbalance. * On a 100 cpu machine: * - per memcg per cpu dirty count is 32 pages for each of 99 cpus * - per memcg atomic is -99*32 pages * - thus the complete dirty limit: sum of all counters 0 * - balance_dirty_pages() only sees atomic count -99*32 pages, which * it max()s to 0. * - So a workload can dirty -99*32 pages before balance_dirty_pages() * cares. */ #define _GNU_SOURCE #include <err.h> #include <fcntl.h> #include <sched.h> #include <stdlib.h> #include <stdio.h> #include <sys/stat.h> #include <sys/sysinfo.h> #include <sys/types.h> #include <unistd.h> static char *buf; static int bufSize; static void set_affinity(int cpu) { cpu_set_t affinity; CPU_ZERO(&affinity); CPU_SET(cpu, &affinity); if (sched_setaffinity(0, sizeof(affinity), &affinity)) err(1, "sched_setaffinity"); } static void dirty_on(int output_fd, int cpu) { int i, wrote; set_affinity(cpu); for (i = 0; i < 32; i++) { for (wrote = 0; wrote < bufSize; ) { int ret = write(output_fd, buf+wrote, bufSize-wrote); if (ret == -1) err(1, "write"); wrote += ret; } } } int main(int argc, char **argv) { int cpu, flush_cpu = 1, output_fd; const char *output; if (argc != 2) errx(1, "usage: output_file"); output = argv[1]; bufSize = getpagesize(); buf = malloc(getpagesize()); if (buf == NULL) errx(1, "malloc failed"); output_fd = open(output, O_CREAT|O_RDWR); if (output_fd == -1) err(1, "open(%s)", output); for (cpu = 0; cpu < get_nprocs(); cpu++) { if (cpu != flush_cpu) dirty_on(output_fd, cpu); } set_affinity(flush_cpu); if (fsync(output_fd)) err(1, "fsync(%s)", output); if (close(output_fd)) err(1, "close(%s)", output); free(buf); } Make balance_dirty_pages() and wb_over_bg_thresh() work harder to collect exact per memcg counters. This avoids the aforementioned oom kills. This does not affect the overhead of memory.stat, which still reads the single atomic counter. Why not use percpu_counter? memcg already handles cpus going offline, so no need for that overhead from percpu_counter. And the percpu_counter spinlocks are more heavyweight than is required. It probably also makes sense to use exact dirty and writeback counters in memcg oom reports. But that is saved for later. Link: http://lkml.kernel.org/r/20190329174609.164344-1-gthelen@google.com Signed-off-by: Greg Thelen <gthelen@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Cc: <stable@vger.kernel.org> [4.16+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-04-06 01:39:18 +00:00
*pwriteback = memcg_exact_page_state(memcg, NR_WRITEBACK);
*pfilepages = memcg_exact_page_state(memcg, NR_INACTIVE_FILE) +
memcg_exact_page_state(memcg, NR_ACTIVE_FILE);
2015-09-29 17:04:26 +00:00
*pheadroom = PAGE_COUNTER_MAX;
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-22 22:23:35 +00:00
while ((parent = parent_mem_cgroup(memcg))) {
unsigned long ceiling = min(READ_ONCE(memcg->memory.max),
READ_ONCE(memcg->memory.high));
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-22 22:23:35 +00:00
unsigned long used = page_counter_read(&memcg->memory);
2015-09-29 17:04:26 +00:00
*pheadroom = min(*pheadroom, ceiling - min(ceiling, used));
writeback: implement memcg writeback domain based throttling While cgroup writeback support now connects memcg and blkcg so that writeback IOs are properly attributed and controlled, the IO back pressure propagation mechanism implemented in balance_dirty_pages() and its subroutines wasn't aware of cgroup writeback. Processes belonging to a memcg may have access to only subset of total memory available in the system and not factoring this into dirty throttling rendered it completely ineffective for processes under memcg limits and memcg ended up building a separate ad-hoc degenerate mechanism directly into vmscan code to limit page dirtying. The previous patches updated balance_dirty_pages() and its subroutines so that they can deal with multiple wb_domain's (writeback domains) and defined per-memcg wb_domain. Processes belonging to a non-root memcg are bound to two wb_domains, global wb_domain and memcg wb_domain, and should be throttled according to IO pressures from both domains. This patch updates dirty throttling code so that it repeats similar calculations for the two domains - the differences between the two are few and minor - and applies the lower of the two sets of resulting constraints. wb_over_bg_thresh(), which controls when background writeback terminates, is also updated to consider both global and memcg wb_domains. It returns true if dirty is over bg_thresh for either domain. This makes the dirty throttling mechanism operational for memcg domains including writeback-bandwidth-proportional dirty page distribution inside them but the ad-hoc memcg throttling mechanism in vmscan is still in place. The next patch will rip it out. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Cc: Wu Fengguang <fengguang.wu@intel.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-22 22:23:35 +00:00
memcg = parent;
}
}
writeback, memcg: Implement foreign dirty flushing There's an inherent mismatch between memcg and writeback. The former trackes ownership per-page while the latter per-inode. This was a deliberate design decision because honoring per-page ownership in the writeback path is complicated, may lead to higher CPU and IO overheads and deemed unnecessary given that write-sharing an inode across different cgroups isn't a common use-case. Combined with inode majority-writer ownership switching, this works well enough in most cases but there are some pathological cases. For example, let's say there are two cgroups A and B which keep writing to different but confined parts of the same inode. B owns the inode and A's memory is limited far below B's. A's dirty ratio can rise enough to trigger balance_dirty_pages() sleeps but B's can be low enough to avoid triggering background writeback. A will be slowed down without a way to make writeback of the dirty pages happen. This patch implements foreign dirty recording and foreign mechanism so that when a memcg encounters a condition as above it can trigger flushes on bdi_writebacks which can clean its pages. Please see the comment on top of mem_cgroup_track_foreign_dirty_slowpath() for details. A reproducer follows. write-range.c:: #include <stdio.h> #include <stdlib.h> #include <unistd.h> #include <fcntl.h> #include <sys/types.h> static const char *usage = "write-range FILE START SIZE\n"; int main(int argc, char **argv) { int fd; unsigned long start, size, end, pos; char *endp; char buf[4096]; if (argc < 4) { fprintf(stderr, usage); return 1; } fd = open(argv[1], O_WRONLY); if (fd < 0) { perror("open"); return 1; } start = strtoul(argv[2], &endp, 0); if (*endp != '\0') { fprintf(stderr, usage); return 1; } size = strtoul(argv[3], &endp, 0); if (*endp != '\0') { fprintf(stderr, usage); return 1; } end = start + size; while (1) { for (pos = start; pos < end; ) { long bread, bwritten = 0; if (lseek(fd, pos, SEEK_SET) < 0) { perror("lseek"); return 1; } bread = read(0, buf, sizeof(buf) < end - pos ? sizeof(buf) : end - pos); if (bread < 0) { perror("read"); return 1; } if (bread == 0) return 0; while (bwritten < bread) { long this; this = write(fd, buf + bwritten, bread - bwritten); if (this < 0) { perror("write"); return 1; } bwritten += this; pos += bwritten; } } } } repro.sh:: #!/bin/bash set -e set -x sysctl -w vm.dirty_expire_centisecs=300000 sysctl -w vm.dirty_writeback_centisecs=300000 sysctl -w vm.dirtytime_expire_seconds=300000 echo 3 > /proc/sys/vm/drop_caches TEST=/sys/fs/cgroup/test A=$TEST/A B=$TEST/B mkdir -p $A $B echo "+memory +io" > $TEST/cgroup.subtree_control echo $((1<<30)) > $A/memory.high echo $((32<<30)) > $B/memory.high rm -f testfile touch testfile fallocate -l 4G testfile echo "Starting B" (echo $BASHPID > $B/cgroup.procs pv -q --rate-limit 70M < /dev/urandom | ./write-range testfile $((2<<30)) $((2<<30))) & echo "Waiting 10s to ensure B claims the testfile inode" sleep 5 sync sleep 5 sync echo "Starting A" (echo $BASHPID > $A/cgroup.procs pv < /dev/urandom | ./write-range testfile 0 $((2<<30))) v2: Added comments explaining why the specific intervals are being used. v3: Use 0 @nr when calling cgroup_writeback_by_id() to use best-effort flushing while avoding possible livelocks. v4: Use get_jiffies_64() and time_before/after64() instead of raw jiffies_64 and arthimetic comparisons as suggested by Jan. Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-08-26 16:06:56 +00:00
/*
* Foreign dirty flushing
*
* There's an inherent mismatch between memcg and writeback. The former
* trackes ownership per-page while the latter per-inode. This was a
* deliberate design decision because honoring per-page ownership in the
* writeback path is complicated, may lead to higher CPU and IO overheads
* and deemed unnecessary given that write-sharing an inode across
* different cgroups isn't a common use-case.
*
* Combined with inode majority-writer ownership switching, this works well
* enough in most cases but there are some pathological cases. For
* example, let's say there are two cgroups A and B which keep writing to
* different but confined parts of the same inode. B owns the inode and
* A's memory is limited far below B's. A's dirty ratio can rise enough to
* trigger balance_dirty_pages() sleeps but B's can be low enough to avoid
* triggering background writeback. A will be slowed down without a way to
* make writeback of the dirty pages happen.
*
* Conditions like the above can lead to a cgroup getting repatedly and
* severely throttled after making some progress after each
* dirty_expire_interval while the underyling IO device is almost
* completely idle.
*
* Solving this problem completely requires matching the ownership tracking
* granularities between memcg and writeback in either direction. However,
* the more egregious behaviors can be avoided by simply remembering the
* most recent foreign dirtying events and initiating remote flushes on
* them when local writeback isn't enough to keep the memory clean enough.
*
* The following two functions implement such mechanism. When a foreign
* page - a page whose memcg and writeback ownerships don't match - is
* dirtied, mem_cgroup_track_foreign_dirty() records the inode owning
* bdi_writeback on the page owning memcg. When balance_dirty_pages()
* decides that the memcg needs to sleep due to high dirty ratio, it calls
* mem_cgroup_flush_foreign() which queues writeback on the recorded
* foreign bdi_writebacks which haven't expired. Both the numbers of
* recorded bdi_writebacks and concurrent in-flight foreign writebacks are
* limited to MEMCG_CGWB_FRN_CNT.
*
* The mechanism only remembers IDs and doesn't hold any object references.
* As being wrong occasionally doesn't matter, updates and accesses to the
* records are lockless and racy.
*/
void mem_cgroup_track_foreign_dirty_slowpath(struct page *page,
struct bdi_writeback *wb)
{
struct mem_cgroup *memcg = page->mem_cgroup;
struct memcg_cgwb_frn *frn;
u64 now = get_jiffies_64();
u64 oldest_at = now;
int oldest = -1;
int i;
trace_track_foreign_dirty(page, wb);
writeback, memcg: Implement foreign dirty flushing There's an inherent mismatch between memcg and writeback. The former trackes ownership per-page while the latter per-inode. This was a deliberate design decision because honoring per-page ownership in the writeback path is complicated, may lead to higher CPU and IO overheads and deemed unnecessary given that write-sharing an inode across different cgroups isn't a common use-case. Combined with inode majority-writer ownership switching, this works well enough in most cases but there are some pathological cases. For example, let's say there are two cgroups A and B which keep writing to different but confined parts of the same inode. B owns the inode and A's memory is limited far below B's. A's dirty ratio can rise enough to trigger balance_dirty_pages() sleeps but B's can be low enough to avoid triggering background writeback. A will be slowed down without a way to make writeback of the dirty pages happen. This patch implements foreign dirty recording and foreign mechanism so that when a memcg encounters a condition as above it can trigger flushes on bdi_writebacks which can clean its pages. Please see the comment on top of mem_cgroup_track_foreign_dirty_slowpath() for details. A reproducer follows. write-range.c:: #include <stdio.h> #include <stdlib.h> #include <unistd.h> #include <fcntl.h> #include <sys/types.h> static const char *usage = "write-range FILE START SIZE\n"; int main(int argc, char **argv) { int fd; unsigned long start, size, end, pos; char *endp; char buf[4096]; if (argc < 4) { fprintf(stderr, usage); return 1; } fd = open(argv[1], O_WRONLY); if (fd < 0) { perror("open"); return 1; } start = strtoul(argv[2], &endp, 0); if (*endp != '\0') { fprintf(stderr, usage); return 1; } size = strtoul(argv[3], &endp, 0); if (*endp != '\0') { fprintf(stderr, usage); return 1; } end = start + size; while (1) { for (pos = start; pos < end; ) { long bread, bwritten = 0; if (lseek(fd, pos, SEEK_SET) < 0) { perror("lseek"); return 1; } bread = read(0, buf, sizeof(buf) < end - pos ? sizeof(buf) : end - pos); if (bread < 0) { perror("read"); return 1; } if (bread == 0) return 0; while (bwritten < bread) { long this; this = write(fd, buf + bwritten, bread - bwritten); if (this < 0) { perror("write"); return 1; } bwritten += this; pos += bwritten; } } } } repro.sh:: #!/bin/bash set -e set -x sysctl -w vm.dirty_expire_centisecs=300000 sysctl -w vm.dirty_writeback_centisecs=300000 sysctl -w vm.dirtytime_expire_seconds=300000 echo 3 > /proc/sys/vm/drop_caches TEST=/sys/fs/cgroup/test A=$TEST/A B=$TEST/B mkdir -p $A $B echo "+memory +io" > $TEST/cgroup.subtree_control echo $((1<<30)) > $A/memory.high echo $((32<<30)) > $B/memory.high rm -f testfile touch testfile fallocate -l 4G testfile echo "Starting B" (echo $BASHPID > $B/cgroup.procs pv -q --rate-limit 70M < /dev/urandom | ./write-range testfile $((2<<30)) $((2<<30))) & echo "Waiting 10s to ensure B claims the testfile inode" sleep 5 sync sleep 5 sync echo "Starting A" (echo $BASHPID > $A/cgroup.procs pv < /dev/urandom | ./write-range testfile 0 $((2<<30))) v2: Added comments explaining why the specific intervals are being used. v3: Use 0 @nr when calling cgroup_writeback_by_id() to use best-effort flushing while avoding possible livelocks. v4: Use get_jiffies_64() and time_before/after64() instead of raw jiffies_64 and arthimetic comparisons as suggested by Jan. Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-08-26 16:06:56 +00:00
/*
* Pick the slot to use. If there is already a slot for @wb, keep
* using it. If not replace the oldest one which isn't being
* written out.
*/
for (i = 0; i < MEMCG_CGWB_FRN_CNT; i++) {
frn = &memcg->cgwb_frn[i];
if (frn->bdi_id == wb->bdi->id &&
frn->memcg_id == wb->memcg_css->id)
break;
if (time_before64(frn->at, oldest_at) &&
atomic_read(&frn->done.cnt) == 1) {
oldest = i;
oldest_at = frn->at;
}
}
if (i < MEMCG_CGWB_FRN_CNT) {
/*
* Re-using an existing one. Update timestamp lazily to
* avoid making the cacheline hot. We want them to be
* reasonably up-to-date and significantly shorter than
* dirty_expire_interval as that's what expires the record.
* Use the shorter of 1s and dirty_expire_interval / 8.
*/
unsigned long update_intv =
min_t(unsigned long, HZ,
msecs_to_jiffies(dirty_expire_interval * 10) / 8);
if (time_before64(frn->at, now - update_intv))
frn->at = now;
} else if (oldest >= 0) {
/* replace the oldest free one */
frn = &memcg->cgwb_frn[oldest];
frn->bdi_id = wb->bdi->id;
frn->memcg_id = wb->memcg_css->id;
frn->at = now;
}
}
/* issue foreign writeback flushes for recorded foreign dirtying events */
void mem_cgroup_flush_foreign(struct bdi_writeback *wb)
{
struct mem_cgroup *memcg = mem_cgroup_from_css(wb->memcg_css);
unsigned long intv = msecs_to_jiffies(dirty_expire_interval * 10);
u64 now = jiffies_64;
int i;
for (i = 0; i < MEMCG_CGWB_FRN_CNT; i++) {
struct memcg_cgwb_frn *frn = &memcg->cgwb_frn[i];
/*
* If the record is older than dirty_expire_interval,
* writeback on it has already started. No need to kick it
* off again. Also, don't start a new one if there's
* already one in flight.
*/
if (time_after64(frn->at, now - intv) &&
atomic_read(&frn->done.cnt) == 1) {
frn->at = 0;
trace_flush_foreign(wb, frn->bdi_id, frn->memcg_id);
writeback, memcg: Implement foreign dirty flushing There's an inherent mismatch between memcg and writeback. The former trackes ownership per-page while the latter per-inode. This was a deliberate design decision because honoring per-page ownership in the writeback path is complicated, may lead to higher CPU and IO overheads and deemed unnecessary given that write-sharing an inode across different cgroups isn't a common use-case. Combined with inode majority-writer ownership switching, this works well enough in most cases but there are some pathological cases. For example, let's say there are two cgroups A and B which keep writing to different but confined parts of the same inode. B owns the inode and A's memory is limited far below B's. A's dirty ratio can rise enough to trigger balance_dirty_pages() sleeps but B's can be low enough to avoid triggering background writeback. A will be slowed down without a way to make writeback of the dirty pages happen. This patch implements foreign dirty recording and foreign mechanism so that when a memcg encounters a condition as above it can trigger flushes on bdi_writebacks which can clean its pages. Please see the comment on top of mem_cgroup_track_foreign_dirty_slowpath() for details. A reproducer follows. write-range.c:: #include <stdio.h> #include <stdlib.h> #include <unistd.h> #include <fcntl.h> #include <sys/types.h> static const char *usage = "write-range FILE START SIZE\n"; int main(int argc, char **argv) { int fd; unsigned long start, size, end, pos; char *endp; char buf[4096]; if (argc < 4) { fprintf(stderr, usage); return 1; } fd = open(argv[1], O_WRONLY); if (fd < 0) { perror("open"); return 1; } start = strtoul(argv[2], &endp, 0); if (*endp != '\0') { fprintf(stderr, usage); return 1; } size = strtoul(argv[3], &endp, 0); if (*endp != '\0') { fprintf(stderr, usage); return 1; } end = start + size; while (1) { for (pos = start; pos < end; ) { long bread, bwritten = 0; if (lseek(fd, pos, SEEK_SET) < 0) { perror("lseek"); return 1; } bread = read(0, buf, sizeof(buf) < end - pos ? sizeof(buf) : end - pos); if (bread < 0) { perror("read"); return 1; } if (bread == 0) return 0; while (bwritten < bread) { long this; this = write(fd, buf + bwritten, bread - bwritten); if (this < 0) { perror("write"); return 1; } bwritten += this; pos += bwritten; } } } } repro.sh:: #!/bin/bash set -e set -x sysctl -w vm.dirty_expire_centisecs=300000 sysctl -w vm.dirty_writeback_centisecs=300000 sysctl -w vm.dirtytime_expire_seconds=300000 echo 3 > /proc/sys/vm/drop_caches TEST=/sys/fs/cgroup/test A=$TEST/A B=$TEST/B mkdir -p $A $B echo "+memory +io" > $TEST/cgroup.subtree_control echo $((1<<30)) > $A/memory.high echo $((32<<30)) > $B/memory.high rm -f testfile touch testfile fallocate -l 4G testfile echo "Starting B" (echo $BASHPID > $B/cgroup.procs pv -q --rate-limit 70M < /dev/urandom | ./write-range testfile $((2<<30)) $((2<<30))) & echo "Waiting 10s to ensure B claims the testfile inode" sleep 5 sync sleep 5 sync echo "Starting A" (echo $BASHPID > $A/cgroup.procs pv < /dev/urandom | ./write-range testfile 0 $((2<<30))) v2: Added comments explaining why the specific intervals are being used. v3: Use 0 @nr when calling cgroup_writeback_by_id() to use best-effort flushing while avoding possible livelocks. v4: Use get_jiffies_64() and time_before/after64() instead of raw jiffies_64 and arthimetic comparisons as suggested by Jan. Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-08-26 16:06:56 +00:00
cgroup_writeback_by_id(frn->bdi_id, frn->memcg_id, 0,
WB_REASON_FOREIGN_FLUSH,
&frn->done);
}
}
}
#else /* CONFIG_CGROUP_WRITEBACK */
static int memcg_wb_domain_init(struct mem_cgroup *memcg, gfp_t gfp)
{
return 0;
}
static void memcg_wb_domain_exit(struct mem_cgroup *memcg)
{
}
static void memcg_wb_domain_size_changed(struct mem_cgroup *memcg)
{
}
writeback: make backing_dev_info host cgroup-specific bdi_writebacks For the planned cgroup writeback support, on each bdi (backing_dev_info), each memcg will be served by a separate wb (bdi_writeback). This patch updates bdi so that a bdi can host multiple wbs (bdi_writebacks). On the default hierarchy, blkcg implicitly enables memcg. This allows using memcg's page ownership for attributing writeback IOs, and every memcg - blkcg combination can be served by its own wb by assigning a dedicated wb to each memcg. This means that there may be multiple wb's of a bdi mapped to the same blkcg. As congested state is per blkcg - bdi combination, those wb's should share the same congested state. This is achieved by tracking congested state via bdi_writeback_congested structs which are keyed by blkcg. bdi->wb remains unchanged and will keep serving the root cgroup. cgwb's (cgroup wb's) for non-root cgroups are created on-demand or looked up while dirtying an inode according to the memcg of the page being dirtied or current task. Each cgwb is indexed on bdi->cgwb_tree by its memcg id. Once an inode is associated with its wb, it can be retrieved using inode_to_wb(). Currently, none of the filesystems has FS_CGROUP_WRITEBACK and all pages will keep being associated with bdi->wb. v3: inode_attach_wb() in account_page_dirtied() moved inside mapping_cap_account_dirty() block where it's known to be !NULL. Also, an unnecessary NULL check before kfree() removed. Both detected by the kbuild bot. v2: Updated so that wb association is per inode and wb is per memcg rather than blkcg. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: kbuild test robot <fengguang.wu@intel.com> Cc: Dan Carpenter <dan.carpenter@oracle.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-22 21:13:37 +00:00
#endif /* CONFIG_CGROUP_WRITEBACK */
/*
* DO NOT USE IN NEW FILES.
*
* "cgroup.event_control" implementation.
*
* This is way over-engineered. It tries to support fully configurable
* events for each user. Such level of flexibility is completely
* unnecessary especially in the light of the planned unified hierarchy.
*
* Please deprecate this and replace with something simpler if at all
* possible.
*/
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
/*
* Unregister event and free resources.
*
* Gets called from workqueue.
*/
static void memcg_event_remove(struct work_struct *work)
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
{
struct mem_cgroup_event *event =
container_of(work, struct mem_cgroup_event, remove);
struct mem_cgroup *memcg = event->memcg;
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
remove_wait_queue(event->wqh, &event->wait);
event->unregister_event(memcg, event->eventfd);
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
/* Notify userspace the event is going away. */
eventfd_signal(event->eventfd, 1);
eventfd_ctx_put(event->eventfd);
kfree(event);
css_put(&memcg->css);
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
}
/*
* Gets called on EPOLLHUP on eventfd when user closes it.
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
*
* Called with wqh->lock held and interrupts disabled.
*/
static int memcg_event_wake(wait_queue_entry_t *wait, unsigned mode,
int sync, void *key)
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
{
struct mem_cgroup_event *event =
container_of(wait, struct mem_cgroup_event, wait);
struct mem_cgroup *memcg = event->memcg;
__poll_t flags = key_to_poll(key);
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
if (flags & EPOLLHUP) {
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
/*
* If the event has been detached at cgroup removal, we
* can simply return knowing the other side will cleanup
* for us.
*
* We can't race against event freeing since the other
* side will require wqh->lock via remove_wait_queue(),
* which we hold.
*/
spin_lock(&memcg->event_list_lock);
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
if (!list_empty(&event->list)) {
list_del_init(&event->list);
/*
* We are in atomic context, but cgroup_event_remove()
* may sleep, so we have to call it in workqueue.
*/
schedule_work(&event->remove);
}
spin_unlock(&memcg->event_list_lock);
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
}
return 0;
}
static void memcg_event_ptable_queue_proc(struct file *file,
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
wait_queue_head_t *wqh, poll_table *pt)
{
struct mem_cgroup_event *event =
container_of(pt, struct mem_cgroup_event, pt);
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
event->wqh = wqh;
add_wait_queue(wqh, &event->wait);
}
/*
* DO NOT USE IN NEW FILES.
*
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
* Parse input and register new cgroup event handler.
*
* Input must be in format '<event_fd> <control_fd> <args>'.
* Interpretation of args is defined by control file implementation.
*/
static ssize_t memcg_write_event_control(struct kernfs_open_file *of,
char *buf, size_t nbytes, loff_t off)
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
{
struct cgroup_subsys_state *css = of_css(of);
struct mem_cgroup *memcg = mem_cgroup_from_css(css);
struct mem_cgroup_event *event;
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
struct cgroup_subsys_state *cfile_css;
unsigned int efd, cfd;
struct fd efile;
struct fd cfile;
const char *name;
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
char *endp;
int ret;
buf = strstrip(buf);
efd = simple_strtoul(buf, &endp, 10);
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
if (*endp != ' ')
return -EINVAL;
buf = endp + 1;
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
cfd = simple_strtoul(buf, &endp, 10);
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
if ((*endp != ' ') && (*endp != '\0'))
return -EINVAL;
buf = endp + 1;
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
event = kzalloc(sizeof(*event), GFP_KERNEL);
if (!event)
return -ENOMEM;
event->memcg = memcg;
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
INIT_LIST_HEAD(&event->list);
init_poll_funcptr(&event->pt, memcg_event_ptable_queue_proc);
init_waitqueue_func_entry(&event->wait, memcg_event_wake);
INIT_WORK(&event->remove, memcg_event_remove);
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
efile = fdget(efd);
if (!efile.file) {
ret = -EBADF;
goto out_kfree;
}
event->eventfd = eventfd_ctx_fileget(efile.file);
if (IS_ERR(event->eventfd)) {
ret = PTR_ERR(event->eventfd);
goto out_put_efile;
}
cfile = fdget(cfd);
if (!cfile.file) {
ret = -EBADF;
goto out_put_eventfd;
}
/* the process need read permission on control file */
/* AV: shouldn't we check that it's been opened for read instead? */
ret = inode_permission(file_inode(cfile.file), MAY_READ);
if (ret < 0)
goto out_put_cfile;
/*
* Determine the event callbacks and set them in @event. This used
* to be done via struct cftype but cgroup core no longer knows
* about these events. The following is crude but the whole thing
* is for compatibility anyway.
*
* DO NOT ADD NEW FILES.
*/
name = cfile.file->f_path.dentry->d_name.name;
if (!strcmp(name, "memory.usage_in_bytes")) {
event->register_event = mem_cgroup_usage_register_event;
event->unregister_event = mem_cgroup_usage_unregister_event;
} else if (!strcmp(name, "memory.oom_control")) {
event->register_event = mem_cgroup_oom_register_event;
event->unregister_event = mem_cgroup_oom_unregister_event;
} else if (!strcmp(name, "memory.pressure_level")) {
event->register_event = vmpressure_register_event;
event->unregister_event = vmpressure_unregister_event;
} else if (!strcmp(name, "memory.memsw.usage_in_bytes")) {
event->register_event = memsw_cgroup_usage_register_event;
event->unregister_event = memsw_cgroup_usage_unregister_event;
} else {
ret = -EINVAL;
goto out_put_cfile;
}
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
/*
* Verify @cfile should belong to @css. Also, remaining events are
* automatically removed on cgroup destruction but the removal is
* asynchronous, so take an extra ref on @css.
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
*/
cfile_css = css_tryget_online_from_dir(cfile.file->f_path.dentry->d_parent,
&memory_cgrp_subsys);
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
ret = -EINVAL;
if (IS_ERR(cfile_css))
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
goto out_put_cfile;
if (cfile_css != css) {
css_put(cfile_css);
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
goto out_put_cfile;
}
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
ret = event->register_event(memcg, event->eventfd, buf);
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
if (ret)
goto out_put_css;
vfs_poll(efile.file, &event->pt);
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
spin_lock(&memcg->event_list_lock);
list_add(&event->list, &memcg->event_list);
spin_unlock(&memcg->event_list_lock);
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
fdput(cfile);
fdput(efile);
return nbytes;
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
out_put_css:
css_put(css);
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
out_put_cfile:
fdput(cfile);
out_put_eventfd:
eventfd_ctx_put(event->eventfd);
out_put_efile:
fdput(efile);
out_kfree:
kfree(event);
return ret;
}
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
static struct cftype mem_cgroup_legacy_files[] = {
{
.name = "usage_in_bytes",
memcg: mem+swap controller core This patch implements per cgroup limit for usage of memory+swap. However there are SwapCache, double counting of swap-cache and swap-entry is avoided. Mem+Swap controller works as following. - memory usage is limited by memory.limit_in_bytes. - memory + swap usage is limited by memory.memsw_limit_in_bytes. This has following benefits. - A user can limit total resource usage of mem+swap. Without this, because memory resource controller doesn't take care of usage of swap, a process can exhaust all the swap (by memory leak.) We can avoid this case. And Swap is shared resource but it cannot be reclaimed (goes back to memory) until it's used. This characteristic can be trouble when the memory is divided into some parts by cpuset or memcg. Assume group A and group B. After some application executes, the system can be.. Group A -- very large free memory space but occupy 99% of swap. Group B -- under memory shortage but cannot use swap...it's nearly full. Ability to set appropriate swap limit for each group is required. Maybe someone wonder "why not swap but mem+swap ?" - The global LRU(kswapd) can swap out arbitrary pages. Swap-out means to move account from memory to swap...there is no change in usage of mem+swap. In other words, when we want to limit the usage of swap without affecting global LRU, mem+swap limit is better than just limiting swap. Accounting target information is stored in swap_cgroup which is per swap entry record. Charge is done as following. map - charge page and memsw. unmap - uncharge page/memsw if not SwapCache. swap-out (__delete_from_swap_cache) - uncharge page - record mem_cgroup information to swap_cgroup. swap-in (do_swap_page) - charged as page and memsw. record in swap_cgroup is cleared. memsw accounting is decremented. swap-free (swap_free()) - if swap entry is freed, memsw is uncharged by PAGE_SIZE. There are people work under never-swap environments and consider swap as something bad. For such people, this mem+swap controller extension is just an overhead. This overhead is avoided by config or boot option. (see Kconfig. detail is not in this patch.) TODO: - maybe more optimization can be don in swap-in path. (but not very safe.) But we just do simple accounting at this stage. [nishimura@mxp.nes.nec.co.jp: make resize limit hold mutex] [hugh@veritas.com: memswap controller core swapcache fixes] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:08:00 +00:00
.private = MEMFILE_PRIVATE(_MEM, RES_USAGE),
.read_u64 = mem_cgroup_read_u64,
},
{
.name = "max_usage_in_bytes",
memcg: mem+swap controller core This patch implements per cgroup limit for usage of memory+swap. However there are SwapCache, double counting of swap-cache and swap-entry is avoided. Mem+Swap controller works as following. - memory usage is limited by memory.limit_in_bytes. - memory + swap usage is limited by memory.memsw_limit_in_bytes. This has following benefits. - A user can limit total resource usage of mem+swap. Without this, because memory resource controller doesn't take care of usage of swap, a process can exhaust all the swap (by memory leak.) We can avoid this case. And Swap is shared resource but it cannot be reclaimed (goes back to memory) until it's used. This characteristic can be trouble when the memory is divided into some parts by cpuset or memcg. Assume group A and group B. After some application executes, the system can be.. Group A -- very large free memory space but occupy 99% of swap. Group B -- under memory shortage but cannot use swap...it's nearly full. Ability to set appropriate swap limit for each group is required. Maybe someone wonder "why not swap but mem+swap ?" - The global LRU(kswapd) can swap out arbitrary pages. Swap-out means to move account from memory to swap...there is no change in usage of mem+swap. In other words, when we want to limit the usage of swap without affecting global LRU, mem+swap limit is better than just limiting swap. Accounting target information is stored in swap_cgroup which is per swap entry record. Charge is done as following. map - charge page and memsw. unmap - uncharge page/memsw if not SwapCache. swap-out (__delete_from_swap_cache) - uncharge page - record mem_cgroup information to swap_cgroup. swap-in (do_swap_page) - charged as page and memsw. record in swap_cgroup is cleared. memsw accounting is decremented. swap-free (swap_free()) - if swap entry is freed, memsw is uncharged by PAGE_SIZE. There are people work under never-swap environments and consider swap as something bad. For such people, this mem+swap controller extension is just an overhead. This overhead is avoided by config or boot option. (see Kconfig. detail is not in this patch.) TODO: - maybe more optimization can be don in swap-in path. (but not very safe.) But we just do simple accounting at this stage. [nishimura@mxp.nes.nec.co.jp: make resize limit hold mutex] [hugh@veritas.com: memswap controller core swapcache fixes] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:08:00 +00:00
.private = MEMFILE_PRIVATE(_MEM, RES_MAX_USAGE),
.write = mem_cgroup_reset,
.read_u64 = mem_cgroup_read_u64,
},
{
.name = "limit_in_bytes",
memcg: mem+swap controller core This patch implements per cgroup limit for usage of memory+swap. However there are SwapCache, double counting of swap-cache and swap-entry is avoided. Mem+Swap controller works as following. - memory usage is limited by memory.limit_in_bytes. - memory + swap usage is limited by memory.memsw_limit_in_bytes. This has following benefits. - A user can limit total resource usage of mem+swap. Without this, because memory resource controller doesn't take care of usage of swap, a process can exhaust all the swap (by memory leak.) We can avoid this case. And Swap is shared resource but it cannot be reclaimed (goes back to memory) until it's used. This characteristic can be trouble when the memory is divided into some parts by cpuset or memcg. Assume group A and group B. After some application executes, the system can be.. Group A -- very large free memory space but occupy 99% of swap. Group B -- under memory shortage but cannot use swap...it's nearly full. Ability to set appropriate swap limit for each group is required. Maybe someone wonder "why not swap but mem+swap ?" - The global LRU(kswapd) can swap out arbitrary pages. Swap-out means to move account from memory to swap...there is no change in usage of mem+swap. In other words, when we want to limit the usage of swap without affecting global LRU, mem+swap limit is better than just limiting swap. Accounting target information is stored in swap_cgroup which is per swap entry record. Charge is done as following. map - charge page and memsw. unmap - uncharge page/memsw if not SwapCache. swap-out (__delete_from_swap_cache) - uncharge page - record mem_cgroup information to swap_cgroup. swap-in (do_swap_page) - charged as page and memsw. record in swap_cgroup is cleared. memsw accounting is decremented. swap-free (swap_free()) - if swap entry is freed, memsw is uncharged by PAGE_SIZE. There are people work under never-swap environments and consider swap as something bad. For such people, this mem+swap controller extension is just an overhead. This overhead is avoided by config or boot option. (see Kconfig. detail is not in this patch.) TODO: - maybe more optimization can be don in swap-in path. (but not very safe.) But we just do simple accounting at this stage. [nishimura@mxp.nes.nec.co.jp: make resize limit hold mutex] [hugh@veritas.com: memswap controller core swapcache fixes] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:08:00 +00:00
.private = MEMFILE_PRIVATE(_MEM, RES_LIMIT),
.write = mem_cgroup_write,
.read_u64 = mem_cgroup_read_u64,
},
{
.name = "soft_limit_in_bytes",
.private = MEMFILE_PRIVATE(_MEM, RES_SOFT_LIMIT),
.write = mem_cgroup_write,
.read_u64 = mem_cgroup_read_u64,
},
{
.name = "failcnt",
memcg: mem+swap controller core This patch implements per cgroup limit for usage of memory+swap. However there are SwapCache, double counting of swap-cache and swap-entry is avoided. Mem+Swap controller works as following. - memory usage is limited by memory.limit_in_bytes. - memory + swap usage is limited by memory.memsw_limit_in_bytes. This has following benefits. - A user can limit total resource usage of mem+swap. Without this, because memory resource controller doesn't take care of usage of swap, a process can exhaust all the swap (by memory leak.) We can avoid this case. And Swap is shared resource but it cannot be reclaimed (goes back to memory) until it's used. This characteristic can be trouble when the memory is divided into some parts by cpuset or memcg. Assume group A and group B. After some application executes, the system can be.. Group A -- very large free memory space but occupy 99% of swap. Group B -- under memory shortage but cannot use swap...it's nearly full. Ability to set appropriate swap limit for each group is required. Maybe someone wonder "why not swap but mem+swap ?" - The global LRU(kswapd) can swap out arbitrary pages. Swap-out means to move account from memory to swap...there is no change in usage of mem+swap. In other words, when we want to limit the usage of swap without affecting global LRU, mem+swap limit is better than just limiting swap. Accounting target information is stored in swap_cgroup which is per swap entry record. Charge is done as following. map - charge page and memsw. unmap - uncharge page/memsw if not SwapCache. swap-out (__delete_from_swap_cache) - uncharge page - record mem_cgroup information to swap_cgroup. swap-in (do_swap_page) - charged as page and memsw. record in swap_cgroup is cleared. memsw accounting is decremented. swap-free (swap_free()) - if swap entry is freed, memsw is uncharged by PAGE_SIZE. There are people work under never-swap environments and consider swap as something bad. For such people, this mem+swap controller extension is just an overhead. This overhead is avoided by config or boot option. (see Kconfig. detail is not in this patch.) TODO: - maybe more optimization can be don in swap-in path. (but not very safe.) But we just do simple accounting at this stage. [nishimura@mxp.nes.nec.co.jp: make resize limit hold mutex] [hugh@veritas.com: memswap controller core swapcache fixes] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:08:00 +00:00
.private = MEMFILE_PRIVATE(_MEM, RES_FAILCNT),
.write = mem_cgroup_reset,
.read_u64 = mem_cgroup_read_u64,
},
{
.name = "stat",
.seq_show = memcg_stat_show,
},
{
.name = "force_empty",
.write = mem_cgroup_force_empty_write,
},
{
.name = "use_hierarchy",
.write_u64 = mem_cgroup_hierarchy_write,
.read_u64 = mem_cgroup_hierarchy_read,
},
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
{
.name = "cgroup.event_control", /* XXX: for compat */
.write = memcg_write_event_control,
.flags = CFTYPE_NO_PREFIX | CFTYPE_WORLD_WRITABLE,
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
},
{
.name = "swappiness",
.read_u64 = mem_cgroup_swappiness_read,
.write_u64 = mem_cgroup_swappiness_write,
},
{
.name = "move_charge_at_immigrate",
.read_u64 = mem_cgroup_move_charge_read,
.write_u64 = mem_cgroup_move_charge_write,
},
{
.name = "oom_control",
.seq_show = mem_cgroup_oom_control_read,
.write_u64 = mem_cgroup_oom_control_write,
.private = MEMFILE_PRIVATE(_OOM_TYPE, OOM_CONTROL),
},
memcg: add memory.pressure_level events With this patch userland applications that want to maintain the interactivity/memory allocation cost can use the pressure level notifications. The levels are defined like this: The "low" level means that the system is reclaiming memory for new allocations. Monitoring this reclaiming activity might be useful for maintaining cache level. Upon notification, the program (typically "Activity Manager") might analyze vmstat and act in advance (i.e. prematurely shutdown unimportant services). The "medium" level means that the system is experiencing medium memory pressure, the system might be making swap, paging out active file caches, etc. Upon this event applications may decide to further analyze vmstat/zoneinfo/memcg or internal memory usage statistics and free any resources that can be easily reconstructed or re-read from a disk. The "critical" level means that the system is actively thrashing, it is about to out of memory (OOM) or even the in-kernel OOM killer is on its way to trigger. Applications should do whatever they can to help the system. It might be too late to consult with vmstat or any other statistics, so it's advisable to take an immediate action. The events are propagated upward until the event is handled, i.e. the events are not pass-through. Here is what this means: for example you have three cgroups: A->B->C. Now you set up an event listener on cgroups A, B and C, and suppose group C experiences some pressure. In this situation, only group C will receive the notification, i.e. groups A and B will not receive it. This is done to avoid excessive "broadcasting" of messages, which disturbs the system and which is especially bad if we are low on memory or thrashing. So, organize the cgroups wisely, or propagate the events manually (or, ask us to implement the pass-through events, explaining why would you need them.) Performance wise, the memory pressure notifications feature itself is lightweight and does not require much of bookkeeping, in contrast to the rest of memcg features. Unfortunately, as of current memcg implementation, pages accounting is an inseparable part and cannot be turned off. The good news is that there are some efforts[1] to improve the situation; plus, implementing the same, fully API-compatible[2] interface for CONFIG_MEMCG=n case (e.g. embedded) is also a viable option, so it will not require any changes on the userland side. [1] http://permalink.gmane.org/gmane.linux.kernel.cgroups/6291 [2] http://lkml.org/lkml/2013/2/21/454 [akpm@linux-foundation.org: coding-style fixes] [akpm@linux-foundation.org: fix CONFIG_CGROPUPS=n warnings] Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Acked-by: Kirill A. Shutemov <kirill@shutemov.name> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: Mel Gorman <mgorman@suse.de> Cc: Glauber Costa <glommer@parallels.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Luiz Capitulino <lcapitulino@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Leonid Moiseichuk <leonid.moiseichuk@nokia.com> Cc: KOSAKI Motohiro <kosaki.motohiro@gmail.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> Cc: John Stultz <john.stultz@linaro.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-04-29 22:08:31 +00:00
{
.name = "pressure_level",
},
#ifdef CONFIG_NUMA
{
.name = "numa_stat",
.seq_show = memcg_numa_stat_show,
},
#endif
memcg: kmem accounting basic infrastructure Add the basic infrastructure for the accounting of kernel memory. To control that, the following files are created: * memory.kmem.usage_in_bytes * memory.kmem.limit_in_bytes * memory.kmem.failcnt * memory.kmem.max_usage_in_bytes They have the same meaning of their user memory counterparts. They reflect the state of the "kmem" res_counter. Per cgroup kmem memory accounting is not enabled until a limit is set for the group. Once the limit is set the accounting cannot be disabled for that group. This means that after the patch is applied, no behavioral changes exists for whoever is still using memcg to control their memory usage, until memory.kmem.limit_in_bytes is set for the first time. We always account to both user and kernel resource_counters. This effectively means that an independent kernel limit is in place when the limit is set to a lower value than the user memory. A equal or higher value means that the user limit will always hit first, meaning that kmem is effectively unlimited. People who want to track kernel memory but not limit it, can set this limit to a very high number (like RESOURCE_MAX - 1page - that no one will ever hit, or equal to the user memory) [akpm@linux-foundation.org: MEMCG_MMEM only works with slab and slub] Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:47 +00:00
{
.name = "kmem.limit_in_bytes",
.private = MEMFILE_PRIVATE(_KMEM, RES_LIMIT),
.write = mem_cgroup_write,
.read_u64 = mem_cgroup_read_u64,
memcg: kmem accounting basic infrastructure Add the basic infrastructure for the accounting of kernel memory. To control that, the following files are created: * memory.kmem.usage_in_bytes * memory.kmem.limit_in_bytes * memory.kmem.failcnt * memory.kmem.max_usage_in_bytes They have the same meaning of their user memory counterparts. They reflect the state of the "kmem" res_counter. Per cgroup kmem memory accounting is not enabled until a limit is set for the group. Once the limit is set the accounting cannot be disabled for that group. This means that after the patch is applied, no behavioral changes exists for whoever is still using memcg to control their memory usage, until memory.kmem.limit_in_bytes is set for the first time. We always account to both user and kernel resource_counters. This effectively means that an independent kernel limit is in place when the limit is set to a lower value than the user memory. A equal or higher value means that the user limit will always hit first, meaning that kmem is effectively unlimited. People who want to track kernel memory but not limit it, can set this limit to a very high number (like RESOURCE_MAX - 1page - that no one will ever hit, or equal to the user memory) [akpm@linux-foundation.org: MEMCG_MMEM only works with slab and slub] Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:47 +00:00
},
{
.name = "kmem.usage_in_bytes",
.private = MEMFILE_PRIVATE(_KMEM, RES_USAGE),
.read_u64 = mem_cgroup_read_u64,
memcg: kmem accounting basic infrastructure Add the basic infrastructure for the accounting of kernel memory. To control that, the following files are created: * memory.kmem.usage_in_bytes * memory.kmem.limit_in_bytes * memory.kmem.failcnt * memory.kmem.max_usage_in_bytes They have the same meaning of their user memory counterparts. They reflect the state of the "kmem" res_counter. Per cgroup kmem memory accounting is not enabled until a limit is set for the group. Once the limit is set the accounting cannot be disabled for that group. This means that after the patch is applied, no behavioral changes exists for whoever is still using memcg to control their memory usage, until memory.kmem.limit_in_bytes is set for the first time. We always account to both user and kernel resource_counters. This effectively means that an independent kernel limit is in place when the limit is set to a lower value than the user memory. A equal or higher value means that the user limit will always hit first, meaning that kmem is effectively unlimited. People who want to track kernel memory but not limit it, can set this limit to a very high number (like RESOURCE_MAX - 1page - that no one will ever hit, or equal to the user memory) [akpm@linux-foundation.org: MEMCG_MMEM only works with slab and slub] Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:47 +00:00
},
{
.name = "kmem.failcnt",
.private = MEMFILE_PRIVATE(_KMEM, RES_FAILCNT),
.write = mem_cgroup_reset,
.read_u64 = mem_cgroup_read_u64,
memcg: kmem accounting basic infrastructure Add the basic infrastructure for the accounting of kernel memory. To control that, the following files are created: * memory.kmem.usage_in_bytes * memory.kmem.limit_in_bytes * memory.kmem.failcnt * memory.kmem.max_usage_in_bytes They have the same meaning of their user memory counterparts. They reflect the state of the "kmem" res_counter. Per cgroup kmem memory accounting is not enabled until a limit is set for the group. Once the limit is set the accounting cannot be disabled for that group. This means that after the patch is applied, no behavioral changes exists for whoever is still using memcg to control their memory usage, until memory.kmem.limit_in_bytes is set for the first time. We always account to both user and kernel resource_counters. This effectively means that an independent kernel limit is in place when the limit is set to a lower value than the user memory. A equal or higher value means that the user limit will always hit first, meaning that kmem is effectively unlimited. People who want to track kernel memory but not limit it, can set this limit to a very high number (like RESOURCE_MAX - 1page - that no one will ever hit, or equal to the user memory) [akpm@linux-foundation.org: MEMCG_MMEM only works with slab and slub] Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:47 +00:00
},
{
.name = "kmem.max_usage_in_bytes",
.private = MEMFILE_PRIVATE(_KMEM, RES_MAX_USAGE),
.write = mem_cgroup_reset,
.read_u64 = mem_cgroup_read_u64,
memcg: kmem accounting basic infrastructure Add the basic infrastructure for the accounting of kernel memory. To control that, the following files are created: * memory.kmem.usage_in_bytes * memory.kmem.limit_in_bytes * memory.kmem.failcnt * memory.kmem.max_usage_in_bytes They have the same meaning of their user memory counterparts. They reflect the state of the "kmem" res_counter. Per cgroup kmem memory accounting is not enabled until a limit is set for the group. Once the limit is set the accounting cannot be disabled for that group. This means that after the patch is applied, no behavioral changes exists for whoever is still using memcg to control their memory usage, until memory.kmem.limit_in_bytes is set for the first time. We always account to both user and kernel resource_counters. This effectively means that an independent kernel limit is in place when the limit is set to a lower value than the user memory. A equal or higher value means that the user limit will always hit first, meaning that kmem is effectively unlimited. People who want to track kernel memory but not limit it, can set this limit to a very high number (like RESOURCE_MAX - 1page - that no one will ever hit, or equal to the user memory) [akpm@linux-foundation.org: MEMCG_MMEM only works with slab and slub] Signed-off-by: Glauber Costa <glommer@parallels.com> Acked-by: Kamezawa Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:21:47 +00:00
},
mm, memcg: fix build error around the usage of kmem_caches When I manually set default n to MEMCG_KMEM in init/Kconfig, bellow error occurs, mm/slab_common.c: In function 'memcg_slab_start': mm/slab_common.c:1530:30: error: 'struct mem_cgroup' has no member named 'kmem_caches' return seq_list_start(&memcg->kmem_caches, *pos); ^ mm/slab_common.c: In function 'memcg_slab_next': mm/slab_common.c:1537:32: error: 'struct mem_cgroup' has no member named 'kmem_caches' return seq_list_next(p, &memcg->kmem_caches, pos); ^ mm/slab_common.c: In function 'memcg_slab_show': mm/slab_common.c:1551:16: error: 'struct mem_cgroup' has no member named 'kmem_caches' if (p == memcg->kmem_caches.next) ^ CC arch/x86/xen/smp.o mm/slab_common.c: In function 'memcg_slab_start': mm/slab_common.c:1531:1: warning: control reaches end of non-void function [-Wreturn-type] } ^ mm/slab_common.c: In function 'memcg_slab_next': mm/slab_common.c:1538:1: warning: control reaches end of non-void function [-Wreturn-type] } ^ That's because kmem_caches is defined only when CONFIG_MEMCG_KMEM is set, while memcg_slab_start() will use it no matter CONFIG_MEMCG_KMEM is defined or not. By the way, the reason I mannuly undefined CONFIG_MEMCG_KMEM is to verify whether my some other code change is still stable when CONFIG_MEMCG_KMEM is not set. Unfortunately, the existing code has been already unstable since v4.11. Fixes: bc2791f857e1 ("slab: link memcg kmem_caches on their associated memory cgroup") Signed-off-by: Yafang Shao <laoar.shao@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Andrew Morton <akpm@linux-foundation.org> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Link: http://lkml.kernel.org/r/1580970260-2045-1-git-send-email-laoar.shao@gmail.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-04-02 04:06:30 +00:00
#if defined(CONFIG_MEMCG_KMEM) && \
(defined(CONFIG_SLAB) || defined(CONFIG_SLUB_DEBUG))
{
.name = "kmem.slabinfo",
slab: link memcg kmem_caches on their associated memory cgroup With kmem cgroup support enabled, kmem_caches can be created and destroyed frequently and a great number of near empty kmem_caches can accumulate if there are a lot of transient cgroups and the system is not under memory pressure. When memory reclaim starts under such conditions, it can lead to consecutive deactivation and destruction of many kmem_caches, easily hundreds of thousands on moderately large systems, exposing scalability issues in the current slab management code. This is one of the patches to address the issue. While a memcg kmem_cache is listed on its root cache's ->children list, there is no direct way to iterate all kmem_caches which are assocaited with a memory cgroup. The only way to iterate them is walking all caches while filtering out caches which don't match, which would be most of them. This makes memcg destruction operations O(N^2) where N is the total number of slab caches which can be huge. This combined with the synchronous RCU operations can tie up a CPU and affect the whole machine for many hours when memory reclaim triggers offlining and destruction of the stale memcgs. This patch adds mem_cgroup->kmem_caches list which goes through memcg_cache_params->kmem_caches_node of all kmem_caches which are associated with the memcg. All memcg specific iterations, including stat file access, are updated to use the new list instead. Link: http://lkml.kernel.org/r/20170117235411.9408-6-tj@kernel.org Signed-off-by: Tejun Heo <tj@kernel.org> Reported-by: Jay Vana <jsvana@fb.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Christoph Lameter <cl@linux.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-02-22 23:41:21 +00:00
.seq_start = memcg_slab_start,
.seq_next = memcg_slab_next,
.seq_stop = memcg_slab_stop,
.seq_show = memcg_slab_show,
},
#endif
{
.name = "kmem.tcp.limit_in_bytes",
.private = MEMFILE_PRIVATE(_TCP, RES_LIMIT),
.write = mem_cgroup_write,
.read_u64 = mem_cgroup_read_u64,
},
{
.name = "kmem.tcp.usage_in_bytes",
.private = MEMFILE_PRIVATE(_TCP, RES_USAGE),
.read_u64 = mem_cgroup_read_u64,
},
{
.name = "kmem.tcp.failcnt",
.private = MEMFILE_PRIVATE(_TCP, RES_FAILCNT),
.write = mem_cgroup_reset,
.read_u64 = mem_cgroup_read_u64,
},
{
.name = "kmem.tcp.max_usage_in_bytes",
.private = MEMFILE_PRIVATE(_TCP, RES_MAX_USAGE),
.write = mem_cgroup_reset,
.read_u64 = mem_cgroup_read_u64,
},
{ }, /* terminate */
};
memcg: mem+swap controller core This patch implements per cgroup limit for usage of memory+swap. However there are SwapCache, double counting of swap-cache and swap-entry is avoided. Mem+Swap controller works as following. - memory usage is limited by memory.limit_in_bytes. - memory + swap usage is limited by memory.memsw_limit_in_bytes. This has following benefits. - A user can limit total resource usage of mem+swap. Without this, because memory resource controller doesn't take care of usage of swap, a process can exhaust all the swap (by memory leak.) We can avoid this case. And Swap is shared resource but it cannot be reclaimed (goes back to memory) until it's used. This characteristic can be trouble when the memory is divided into some parts by cpuset or memcg. Assume group A and group B. After some application executes, the system can be.. Group A -- very large free memory space but occupy 99% of swap. Group B -- under memory shortage but cannot use swap...it's nearly full. Ability to set appropriate swap limit for each group is required. Maybe someone wonder "why not swap but mem+swap ?" - The global LRU(kswapd) can swap out arbitrary pages. Swap-out means to move account from memory to swap...there is no change in usage of mem+swap. In other words, when we want to limit the usage of swap without affecting global LRU, mem+swap limit is better than just limiting swap. Accounting target information is stored in swap_cgroup which is per swap entry record. Charge is done as following. map - charge page and memsw. unmap - uncharge page/memsw if not SwapCache. swap-out (__delete_from_swap_cache) - uncharge page - record mem_cgroup information to swap_cgroup. swap-in (do_swap_page) - charged as page and memsw. record in swap_cgroup is cleared. memsw accounting is decremented. swap-free (swap_free()) - if swap entry is freed, memsw is uncharged by PAGE_SIZE. There are people work under never-swap environments and consider swap as something bad. For such people, this mem+swap controller extension is just an overhead. This overhead is avoided by config or boot option. (see Kconfig. detail is not in this patch.) TODO: - maybe more optimization can be don in swap-in path. (but not very safe.) But we just do simple accounting at this stage. [nishimura@mxp.nes.nec.co.jp: make resize limit hold mutex] [hugh@veritas.com: memswap controller core swapcache fixes] Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Li Zefan <lizf@cn.fujitsu.com> Cc: Balbir Singh <balbir@in.ibm.com> Cc: Pavel Emelyanov <xemul@openvz.org> Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Signed-off-by: Hugh Dickins <hugh@veritas.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2009-01-08 02:08:00 +00:00
mm: memcontrol: fix cgroup creation failure after many small jobs The memory controller has quite a bit of state that usually outlives the cgroup and pins its CSS until said state disappears. At the same time it imposes a 16-bit limit on the CSS ID space to economically store IDs in the wild. Consequently, when we use cgroups to contain frequent but small and short-lived jobs that leave behind some page cache, we quickly run into the 64k limitations of outstanding CSSs. Creating a new cgroup fails with -ENOSPC while there are only a few, or even no user-visible cgroups in existence. Although pinning CSSs past cgroup removal is common, there are only two instances that actually need an ID after a cgroup is deleted: cache shadow entries and swapout records. Cache shadow entries reference the ID weakly and can deal with the CSS having disappeared when it's looked up later. They pose no hurdle. Swap-out records do need to pin the css to hierarchically attribute swapins after the cgroup has been deleted; though the only pages that remain swapped out after offlining are tmpfs/shmem pages. And those references are under the user's control, so they are manageable. This patch introduces a private 16-bit memcg ID and switches swap and cache shadow entries over to using that. This ID can then be recycled after offlining when the CSS remains pinned only by objects that don't specifically need it. This script demonstrates the problem by faulting one cache page in a new cgroup and deleting it again: set -e mkdir -p pages for x in `seq 128000`; do [ $((x % 1000)) -eq 0 ] && echo $x mkdir /cgroup/foo echo $$ >/cgroup/foo/cgroup.procs echo trex >pages/$x echo $$ >/cgroup/cgroup.procs rmdir /cgroup/foo done When run on an unpatched kernel, we eventually run out of possible IDs even though there are no visible cgroups: [root@ham ~]# ./cssidstress.sh [...] 65000 mkdir: cannot create directory '/cgroup/foo': No space left on device After this patch, the IDs get released upon cgroup destruction and the cache and css objects get released once memory reclaim kicks in. [hannes@cmpxchg.org: init the IDR] Link: http://lkml.kernel.org/r/20160621154601.GA22431@cmpxchg.org Fixes: b2052564e66d ("mm: memcontrol: continue cache reclaim from offlined groups") Link: http://lkml.kernel.org/r/20160617162516.GD19084@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: John Garcia <john.garcia@mesosphere.io> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: Nikolay Borisov <kernel@kyup.com> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-20 22:44:57 +00:00
/*
* Private memory cgroup IDR
*
* Swap-out records and page cache shadow entries need to store memcg
* references in constrained space, so we maintain an ID space that is
* limited to 16 bit (MEM_CGROUP_ID_MAX), limiting the total number of
* memory-controlled cgroups to 64k.
*
* However, there usually are many references to the offline CSS after
mm: memcontrol: fix cgroup creation failure after many small jobs The memory controller has quite a bit of state that usually outlives the cgroup and pins its CSS until said state disappears. At the same time it imposes a 16-bit limit on the CSS ID space to economically store IDs in the wild. Consequently, when we use cgroups to contain frequent but small and short-lived jobs that leave behind some page cache, we quickly run into the 64k limitations of outstanding CSSs. Creating a new cgroup fails with -ENOSPC while there are only a few, or even no user-visible cgroups in existence. Although pinning CSSs past cgroup removal is common, there are only two instances that actually need an ID after a cgroup is deleted: cache shadow entries and swapout records. Cache shadow entries reference the ID weakly and can deal with the CSS having disappeared when it's looked up later. They pose no hurdle. Swap-out records do need to pin the css to hierarchically attribute swapins after the cgroup has been deleted; though the only pages that remain swapped out after offlining are tmpfs/shmem pages. And those references are under the user's control, so they are manageable. This patch introduces a private 16-bit memcg ID and switches swap and cache shadow entries over to using that. This ID can then be recycled after offlining when the CSS remains pinned only by objects that don't specifically need it. This script demonstrates the problem by faulting one cache page in a new cgroup and deleting it again: set -e mkdir -p pages for x in `seq 128000`; do [ $((x % 1000)) -eq 0 ] && echo $x mkdir /cgroup/foo echo $$ >/cgroup/foo/cgroup.procs echo trex >pages/$x echo $$ >/cgroup/cgroup.procs rmdir /cgroup/foo done When run on an unpatched kernel, we eventually run out of possible IDs even though there are no visible cgroups: [root@ham ~]# ./cssidstress.sh [...] 65000 mkdir: cannot create directory '/cgroup/foo': No space left on device After this patch, the IDs get released upon cgroup destruction and the cache and css objects get released once memory reclaim kicks in. [hannes@cmpxchg.org: init the IDR] Link: http://lkml.kernel.org/r/20160621154601.GA22431@cmpxchg.org Fixes: b2052564e66d ("mm: memcontrol: continue cache reclaim from offlined groups") Link: http://lkml.kernel.org/r/20160617162516.GD19084@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: John Garcia <john.garcia@mesosphere.io> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: Nikolay Borisov <kernel@kyup.com> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-20 22:44:57 +00:00
* the cgroup has been destroyed, such as page cache or reclaimable
* slab objects, that don't need to hang on to the ID. We want to keep
* those dead CSS from occupying IDs, or we might quickly exhaust the
* relatively small ID space and prevent the creation of new cgroups
* even when there are much fewer than 64k cgroups - possibly none.
*
* Maintain a private 16-bit ID space for memcg, and allow the ID to
* be freed and recycled when it's no longer needed, which is usually
* when the CSS is offlined.
*
* The only exception to that are records of swapped out tmpfs/shmem
* pages that need to be attributed to live ancestors on swapin. But
* those references are manageable from userspace.
*/
static DEFINE_IDR(mem_cgroup_idr);
static void mem_cgroup_id_remove(struct mem_cgroup *memcg)
{
if (memcg->id.id > 0) {
idr_remove(&mem_cgroup_idr, memcg->id.id);
memcg->id.id = 0;
}
}
static void __maybe_unused mem_cgroup_id_get_many(struct mem_cgroup *memcg,
unsigned int n)
mm: memcontrol: fix cgroup creation failure after many small jobs The memory controller has quite a bit of state that usually outlives the cgroup and pins its CSS until said state disappears. At the same time it imposes a 16-bit limit on the CSS ID space to economically store IDs in the wild. Consequently, when we use cgroups to contain frequent but small and short-lived jobs that leave behind some page cache, we quickly run into the 64k limitations of outstanding CSSs. Creating a new cgroup fails with -ENOSPC while there are only a few, or even no user-visible cgroups in existence. Although pinning CSSs past cgroup removal is common, there are only two instances that actually need an ID after a cgroup is deleted: cache shadow entries and swapout records. Cache shadow entries reference the ID weakly and can deal with the CSS having disappeared when it's looked up later. They pose no hurdle. Swap-out records do need to pin the css to hierarchically attribute swapins after the cgroup has been deleted; though the only pages that remain swapped out after offlining are tmpfs/shmem pages. And those references are under the user's control, so they are manageable. This patch introduces a private 16-bit memcg ID and switches swap and cache shadow entries over to using that. This ID can then be recycled after offlining when the CSS remains pinned only by objects that don't specifically need it. This script demonstrates the problem by faulting one cache page in a new cgroup and deleting it again: set -e mkdir -p pages for x in `seq 128000`; do [ $((x % 1000)) -eq 0 ] && echo $x mkdir /cgroup/foo echo $$ >/cgroup/foo/cgroup.procs echo trex >pages/$x echo $$ >/cgroup/cgroup.procs rmdir /cgroup/foo done When run on an unpatched kernel, we eventually run out of possible IDs even though there are no visible cgroups: [root@ham ~]# ./cssidstress.sh [...] 65000 mkdir: cannot create directory '/cgroup/foo': No space left on device After this patch, the IDs get released upon cgroup destruction and the cache and css objects get released once memory reclaim kicks in. [hannes@cmpxchg.org: init the IDR] Link: http://lkml.kernel.org/r/20160621154601.GA22431@cmpxchg.org Fixes: b2052564e66d ("mm: memcontrol: continue cache reclaim from offlined groups") Link: http://lkml.kernel.org/r/20160617162516.GD19084@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: John Garcia <john.garcia@mesosphere.io> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: Nikolay Borisov <kernel@kyup.com> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-20 22:44:57 +00:00
{
refcount_add(n, &memcg->id.ref);
mm: memcontrol: fix cgroup creation failure after many small jobs The memory controller has quite a bit of state that usually outlives the cgroup and pins its CSS until said state disappears. At the same time it imposes a 16-bit limit on the CSS ID space to economically store IDs in the wild. Consequently, when we use cgroups to contain frequent but small and short-lived jobs that leave behind some page cache, we quickly run into the 64k limitations of outstanding CSSs. Creating a new cgroup fails with -ENOSPC while there are only a few, or even no user-visible cgroups in existence. Although pinning CSSs past cgroup removal is common, there are only two instances that actually need an ID after a cgroup is deleted: cache shadow entries and swapout records. Cache shadow entries reference the ID weakly and can deal with the CSS having disappeared when it's looked up later. They pose no hurdle. Swap-out records do need to pin the css to hierarchically attribute swapins after the cgroup has been deleted; though the only pages that remain swapped out after offlining are tmpfs/shmem pages. And those references are under the user's control, so they are manageable. This patch introduces a private 16-bit memcg ID and switches swap and cache shadow entries over to using that. This ID can then be recycled after offlining when the CSS remains pinned only by objects that don't specifically need it. This script demonstrates the problem by faulting one cache page in a new cgroup and deleting it again: set -e mkdir -p pages for x in `seq 128000`; do [ $((x % 1000)) -eq 0 ] && echo $x mkdir /cgroup/foo echo $$ >/cgroup/foo/cgroup.procs echo trex >pages/$x echo $$ >/cgroup/cgroup.procs rmdir /cgroup/foo done When run on an unpatched kernel, we eventually run out of possible IDs even though there are no visible cgroups: [root@ham ~]# ./cssidstress.sh [...] 65000 mkdir: cannot create directory '/cgroup/foo': No space left on device After this patch, the IDs get released upon cgroup destruction and the cache and css objects get released once memory reclaim kicks in. [hannes@cmpxchg.org: init the IDR] Link: http://lkml.kernel.org/r/20160621154601.GA22431@cmpxchg.org Fixes: b2052564e66d ("mm: memcontrol: continue cache reclaim from offlined groups") Link: http://lkml.kernel.org/r/20160617162516.GD19084@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: John Garcia <john.garcia@mesosphere.io> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: Nikolay Borisov <kernel@kyup.com> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-20 22:44:57 +00:00
}
static void mem_cgroup_id_put_many(struct mem_cgroup *memcg, unsigned int n)
mm: memcontrol: fix cgroup creation failure after many small jobs The memory controller has quite a bit of state that usually outlives the cgroup and pins its CSS until said state disappears. At the same time it imposes a 16-bit limit on the CSS ID space to economically store IDs in the wild. Consequently, when we use cgroups to contain frequent but small and short-lived jobs that leave behind some page cache, we quickly run into the 64k limitations of outstanding CSSs. Creating a new cgroup fails with -ENOSPC while there are only a few, or even no user-visible cgroups in existence. Although pinning CSSs past cgroup removal is common, there are only two instances that actually need an ID after a cgroup is deleted: cache shadow entries and swapout records. Cache shadow entries reference the ID weakly and can deal with the CSS having disappeared when it's looked up later. They pose no hurdle. Swap-out records do need to pin the css to hierarchically attribute swapins after the cgroup has been deleted; though the only pages that remain swapped out after offlining are tmpfs/shmem pages. And those references are under the user's control, so they are manageable. This patch introduces a private 16-bit memcg ID and switches swap and cache shadow entries over to using that. This ID can then be recycled after offlining when the CSS remains pinned only by objects that don't specifically need it. This script demonstrates the problem by faulting one cache page in a new cgroup and deleting it again: set -e mkdir -p pages for x in `seq 128000`; do [ $((x % 1000)) -eq 0 ] && echo $x mkdir /cgroup/foo echo $$ >/cgroup/foo/cgroup.procs echo trex >pages/$x echo $$ >/cgroup/cgroup.procs rmdir /cgroup/foo done When run on an unpatched kernel, we eventually run out of possible IDs even though there are no visible cgroups: [root@ham ~]# ./cssidstress.sh [...] 65000 mkdir: cannot create directory '/cgroup/foo': No space left on device After this patch, the IDs get released upon cgroup destruction and the cache and css objects get released once memory reclaim kicks in. [hannes@cmpxchg.org: init the IDR] Link: http://lkml.kernel.org/r/20160621154601.GA22431@cmpxchg.org Fixes: b2052564e66d ("mm: memcontrol: continue cache reclaim from offlined groups") Link: http://lkml.kernel.org/r/20160617162516.GD19084@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: John Garcia <john.garcia@mesosphere.io> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: Nikolay Borisov <kernel@kyup.com> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-20 22:44:57 +00:00
{
if (refcount_sub_and_test(n, &memcg->id.ref)) {
mem_cgroup_id_remove(memcg);
mm: memcontrol: fix cgroup creation failure after many small jobs The memory controller has quite a bit of state that usually outlives the cgroup and pins its CSS until said state disappears. At the same time it imposes a 16-bit limit on the CSS ID space to economically store IDs in the wild. Consequently, when we use cgroups to contain frequent but small and short-lived jobs that leave behind some page cache, we quickly run into the 64k limitations of outstanding CSSs. Creating a new cgroup fails with -ENOSPC while there are only a few, or even no user-visible cgroups in existence. Although pinning CSSs past cgroup removal is common, there are only two instances that actually need an ID after a cgroup is deleted: cache shadow entries and swapout records. Cache shadow entries reference the ID weakly and can deal with the CSS having disappeared when it's looked up later. They pose no hurdle. Swap-out records do need to pin the css to hierarchically attribute swapins after the cgroup has been deleted; though the only pages that remain swapped out after offlining are tmpfs/shmem pages. And those references are under the user's control, so they are manageable. This patch introduces a private 16-bit memcg ID and switches swap and cache shadow entries over to using that. This ID can then be recycled after offlining when the CSS remains pinned only by objects that don't specifically need it. This script demonstrates the problem by faulting one cache page in a new cgroup and deleting it again: set -e mkdir -p pages for x in `seq 128000`; do [ $((x % 1000)) -eq 0 ] && echo $x mkdir /cgroup/foo echo $$ >/cgroup/foo/cgroup.procs echo trex >pages/$x echo $$ >/cgroup/cgroup.procs rmdir /cgroup/foo done When run on an unpatched kernel, we eventually run out of possible IDs even though there are no visible cgroups: [root@ham ~]# ./cssidstress.sh [...] 65000 mkdir: cannot create directory '/cgroup/foo': No space left on device After this patch, the IDs get released upon cgroup destruction and the cache and css objects get released once memory reclaim kicks in. [hannes@cmpxchg.org: init the IDR] Link: http://lkml.kernel.org/r/20160621154601.GA22431@cmpxchg.org Fixes: b2052564e66d ("mm: memcontrol: continue cache reclaim from offlined groups") Link: http://lkml.kernel.org/r/20160617162516.GD19084@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: John Garcia <john.garcia@mesosphere.io> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: Nikolay Borisov <kernel@kyup.com> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-20 22:44:57 +00:00
/* Memcg ID pins CSS */
css_put(&memcg->css);
}
}
static inline void mem_cgroup_id_put(struct mem_cgroup *memcg)
{
mem_cgroup_id_put_many(memcg, 1);
}
mm: memcontrol: fix cgroup creation failure after many small jobs The memory controller has quite a bit of state that usually outlives the cgroup and pins its CSS until said state disappears. At the same time it imposes a 16-bit limit on the CSS ID space to economically store IDs in the wild. Consequently, when we use cgroups to contain frequent but small and short-lived jobs that leave behind some page cache, we quickly run into the 64k limitations of outstanding CSSs. Creating a new cgroup fails with -ENOSPC while there are only a few, or even no user-visible cgroups in existence. Although pinning CSSs past cgroup removal is common, there are only two instances that actually need an ID after a cgroup is deleted: cache shadow entries and swapout records. Cache shadow entries reference the ID weakly and can deal with the CSS having disappeared when it's looked up later. They pose no hurdle. Swap-out records do need to pin the css to hierarchically attribute swapins after the cgroup has been deleted; though the only pages that remain swapped out after offlining are tmpfs/shmem pages. And those references are under the user's control, so they are manageable. This patch introduces a private 16-bit memcg ID and switches swap and cache shadow entries over to using that. This ID can then be recycled after offlining when the CSS remains pinned only by objects that don't specifically need it. This script demonstrates the problem by faulting one cache page in a new cgroup and deleting it again: set -e mkdir -p pages for x in `seq 128000`; do [ $((x % 1000)) -eq 0 ] && echo $x mkdir /cgroup/foo echo $$ >/cgroup/foo/cgroup.procs echo trex >pages/$x echo $$ >/cgroup/cgroup.procs rmdir /cgroup/foo done When run on an unpatched kernel, we eventually run out of possible IDs even though there are no visible cgroups: [root@ham ~]# ./cssidstress.sh [...] 65000 mkdir: cannot create directory '/cgroup/foo': No space left on device After this patch, the IDs get released upon cgroup destruction and the cache and css objects get released once memory reclaim kicks in. [hannes@cmpxchg.org: init the IDR] Link: http://lkml.kernel.org/r/20160621154601.GA22431@cmpxchg.org Fixes: b2052564e66d ("mm: memcontrol: continue cache reclaim from offlined groups") Link: http://lkml.kernel.org/r/20160617162516.GD19084@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: John Garcia <john.garcia@mesosphere.io> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: Nikolay Borisov <kernel@kyup.com> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-20 22:44:57 +00:00
/**
* mem_cgroup_from_id - look up a memcg from a memcg id
* @id: the memcg id to look up
*
* Caller must hold rcu_read_lock().
*/
struct mem_cgroup *mem_cgroup_from_id(unsigned short id)
{
WARN_ON_ONCE(!rcu_read_lock_held());
return idr_find(&mem_cgroup_idr, id);
}
static int alloc_mem_cgroup_per_node_info(struct mem_cgroup *memcg, int node)
{
struct mem_cgroup_per_node *pn;
int tmp = node;
/*
* This routine is called against possible nodes.
* But it's BUG to call kmalloc() against offline node.
*
* TODO: this routine can waste much memory for nodes which will
* never be onlined. It's better to use memory hotplug callback
* function.
*/
if (!node_state(node, N_NORMAL_MEMORY))
tmp = -1;
pn = kzalloc_node(sizeof(*pn), GFP_KERNEL, tmp);
if (!pn)
return 1;
mm: memcontrol: don't batch updates of local VM stats and events The kernel test robot noticed a 26% will-it-scale pagefault regression from commit 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty"). This appears to be caused by bouncing the additional cachelines from the new hierarchical statistics counters. We can fix this by getting rid of the batched local counters instead. Originally, there were *only* group-local counters, and they were fully maintained per cpu. A reader of a stats file high up in the cgroup tree would have to walk the entire subtree and collect each level's per-cpu counters to get the recursive view. This was prohibitively expensive, and so we switched to per-cpu batched updates of the local counters during a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting"), reducing the complexity from nr_subgroups * nr_cpus to nr_subgroups. With growing machines and cgroup trees, the tree walk itself became too expensive for monitoring top-level groups, and this is when the culprit patch added hierarchy counters on each cgroup level. When the per-cpu batch size would be reached, both the local and the hierarchy counters would get batch-updated from the per-cpu delta simultaneously. This makes local and hierarchical counter reads blazingly fast, but it unfortunately makes the write-side too cache line intense. Since local counter reads were never a problem - we only centralized them to accelerate the hierarchy walk - and use of the local counters are becoming rarer due to replacement with hierarchical views (ongoing rework in the page reclaim and workingset code), we can make those local counters unbatched per-cpu counters again. The scheme will then be as such: when a memcg statistic changes, the writer will: - update the local counter (per-cpu) - update the batch counter (per-cpu). If the batch is full: - spill the batch into the group's atomic_t - spill the batch into all ancestors' atomic_ts - empty out the batch counter (per-cpu) when a local memcg counter is read, the reader will: - collect the local counter from all cpus when a hiearchy memcg counter is read, the reader will: - read the atomic_t We might be able to simplify this further and make the recursive counters unbatched per-cpu counters as well (batch upward propagation, but leave per-cpu collection to the readers), but that will require a more in-depth analysis and testing of all the callsites. Deal with the immediate regression for now. Link: http://lkml.kernel.org/r/20190521151647.GB2870@cmpxchg.org Fixes: 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: kernel test robot <rong.a.chen@intel.com> Tested-by: kernel test robot <rong.a.chen@intel.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Shakeel Butt <shakeelb@google.com> Cc: Roman Gushchin <guro@fb.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-06-13 22:55:46 +00:00
pn->lruvec_stat_local = alloc_percpu(struct lruvec_stat);
if (!pn->lruvec_stat_local) {
kfree(pn);
return 1;
}
mm: memcontrol: fix excessive complexity in memory.stat reporting We've seen memory.stat reads in top-level cgroups take up to fourteen seconds during a userspace bug that created tens of thousands of ghost cgroups pinned by lingering page cache. Even with a more reasonable number of cgroups, aggregating memory.stat is unnecessarily heavy. The complexity is this: nr_cgroups * nr_stat_items * nr_possible_cpus where the stat items are ~70 at this point. With 128 cgroups and 128 CPUs - decent, not enormous setups - reading the top-level memory.stat has to aggregate over a million per-cpu counters. This doesn't scale. Instead of spreading the source of truth across all CPUs, use the per-cpu counters merely to batch updates to shared atomic counters. This is the same as the per-cpu stocks we use for charging memory to the shared atomic page_counters, and also the way the global vmstat counters are implemented. Vmstat has elaborate spilling thresholds that depend on the number of CPUs, amount of memory, and memory pressure - carefully balancing the cost of counter updates with the amount of per-cpu error. That's because the vmstat counters are system-wide, but also used for decisions inside the kernel (e.g. NR_FREE_PAGES in the allocator). Neither is true for the memory controller. Use the same static batch size we already use for page_counter updates during charging. The per-cpu error in the stats will be 128k, which is an acceptable ratio of cores to memory accounting granularity. [hannes@cmpxchg.org: fix warning in __this_cpu_xchg() calls] Link: http://lkml.kernel.org/r/20171201135750.GB8097@cmpxchg.org Link: http://lkml.kernel.org/r/20171103153336.24044-3-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Michal Hocko <mhocko@suse.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-02-01 00:16:45 +00:00
pn->lruvec_stat_cpu = alloc_percpu(struct lruvec_stat);
if (!pn->lruvec_stat_cpu) {
mm: memcontrol: don't batch updates of local VM stats and events The kernel test robot noticed a 26% will-it-scale pagefault regression from commit 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty"). This appears to be caused by bouncing the additional cachelines from the new hierarchical statistics counters. We can fix this by getting rid of the batched local counters instead. Originally, there were *only* group-local counters, and they were fully maintained per cpu. A reader of a stats file high up in the cgroup tree would have to walk the entire subtree and collect each level's per-cpu counters to get the recursive view. This was prohibitively expensive, and so we switched to per-cpu batched updates of the local counters during a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting"), reducing the complexity from nr_subgroups * nr_cpus to nr_subgroups. With growing machines and cgroup trees, the tree walk itself became too expensive for monitoring top-level groups, and this is when the culprit patch added hierarchy counters on each cgroup level. When the per-cpu batch size would be reached, both the local and the hierarchy counters would get batch-updated from the per-cpu delta simultaneously. This makes local and hierarchical counter reads blazingly fast, but it unfortunately makes the write-side too cache line intense. Since local counter reads were never a problem - we only centralized them to accelerate the hierarchy walk - and use of the local counters are becoming rarer due to replacement with hierarchical views (ongoing rework in the page reclaim and workingset code), we can make those local counters unbatched per-cpu counters again. The scheme will then be as such: when a memcg statistic changes, the writer will: - update the local counter (per-cpu) - update the batch counter (per-cpu). If the batch is full: - spill the batch into the group's atomic_t - spill the batch into all ancestors' atomic_ts - empty out the batch counter (per-cpu) when a local memcg counter is read, the reader will: - collect the local counter from all cpus when a hiearchy memcg counter is read, the reader will: - read the atomic_t We might be able to simplify this further and make the recursive counters unbatched per-cpu counters as well (batch upward propagation, but leave per-cpu collection to the readers), but that will require a more in-depth analysis and testing of all the callsites. Deal with the immediate regression for now. Link: http://lkml.kernel.org/r/20190521151647.GB2870@cmpxchg.org Fixes: 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: kernel test robot <rong.a.chen@intel.com> Tested-by: kernel test robot <rong.a.chen@intel.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Shakeel Butt <shakeelb@google.com> Cc: Roman Gushchin <guro@fb.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-06-13 22:55:46 +00:00
free_percpu(pn->lruvec_stat_local);
kfree(pn);
return 1;
}
lruvec_init(&pn->lruvec);
pn->usage_in_excess = 0;
pn->on_tree = false;
pn->memcg = memcg;
memcg->nodeinfo[node] = pn;
return 0;
}
static void free_mem_cgroup_per_node_info(struct mem_cgroup *memcg, int node)
{
struct mem_cgroup_per_node *pn = memcg->nodeinfo[node];
if (!pn)
return;
mm: memcontrol: fix excessive complexity in memory.stat reporting We've seen memory.stat reads in top-level cgroups take up to fourteen seconds during a userspace bug that created tens of thousands of ghost cgroups pinned by lingering page cache. Even with a more reasonable number of cgroups, aggregating memory.stat is unnecessarily heavy. The complexity is this: nr_cgroups * nr_stat_items * nr_possible_cpus where the stat items are ~70 at this point. With 128 cgroups and 128 CPUs - decent, not enormous setups - reading the top-level memory.stat has to aggregate over a million per-cpu counters. This doesn't scale. Instead of spreading the source of truth across all CPUs, use the per-cpu counters merely to batch updates to shared atomic counters. This is the same as the per-cpu stocks we use for charging memory to the shared atomic page_counters, and also the way the global vmstat counters are implemented. Vmstat has elaborate spilling thresholds that depend on the number of CPUs, amount of memory, and memory pressure - carefully balancing the cost of counter updates with the amount of per-cpu error. That's because the vmstat counters are system-wide, but also used for decisions inside the kernel (e.g. NR_FREE_PAGES in the allocator). Neither is true for the memory controller. Use the same static batch size we already use for page_counter updates during charging. The per-cpu error in the stats will be 128k, which is an acceptable ratio of cores to memory accounting granularity. [hannes@cmpxchg.org: fix warning in __this_cpu_xchg() calls] Link: http://lkml.kernel.org/r/20171201135750.GB8097@cmpxchg.org Link: http://lkml.kernel.org/r/20171103153336.24044-3-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Michal Hocko <mhocko@suse.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-02-01 00:16:45 +00:00
free_percpu(pn->lruvec_stat_cpu);
mm: memcontrol: don't batch updates of local VM stats and events The kernel test robot noticed a 26% will-it-scale pagefault regression from commit 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty"). This appears to be caused by bouncing the additional cachelines from the new hierarchical statistics counters. We can fix this by getting rid of the batched local counters instead. Originally, there were *only* group-local counters, and they were fully maintained per cpu. A reader of a stats file high up in the cgroup tree would have to walk the entire subtree and collect each level's per-cpu counters to get the recursive view. This was prohibitively expensive, and so we switched to per-cpu batched updates of the local counters during a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting"), reducing the complexity from nr_subgroups * nr_cpus to nr_subgroups. With growing machines and cgroup trees, the tree walk itself became too expensive for monitoring top-level groups, and this is when the culprit patch added hierarchy counters on each cgroup level. When the per-cpu batch size would be reached, both the local and the hierarchy counters would get batch-updated from the per-cpu delta simultaneously. This makes local and hierarchical counter reads blazingly fast, but it unfortunately makes the write-side too cache line intense. Since local counter reads were never a problem - we only centralized them to accelerate the hierarchy walk - and use of the local counters are becoming rarer due to replacement with hierarchical views (ongoing rework in the page reclaim and workingset code), we can make those local counters unbatched per-cpu counters again. The scheme will then be as such: when a memcg statistic changes, the writer will: - update the local counter (per-cpu) - update the batch counter (per-cpu). If the batch is full: - spill the batch into the group's atomic_t - spill the batch into all ancestors' atomic_ts - empty out the batch counter (per-cpu) when a local memcg counter is read, the reader will: - collect the local counter from all cpus when a hiearchy memcg counter is read, the reader will: - read the atomic_t We might be able to simplify this further and make the recursive counters unbatched per-cpu counters as well (batch upward propagation, but leave per-cpu collection to the readers), but that will require a more in-depth analysis and testing of all the callsites. Deal with the immediate regression for now. Link: http://lkml.kernel.org/r/20190521151647.GB2870@cmpxchg.org Fixes: 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: kernel test robot <rong.a.chen@intel.com> Tested-by: kernel test robot <rong.a.chen@intel.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Shakeel Butt <shakeelb@google.com> Cc: Roman Gushchin <guro@fb.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-06-13 22:55:46 +00:00
free_percpu(pn->lruvec_stat_local);
kfree(pn);
}
mm: do not call mem_cgroup_free() from within mem_cgroup_alloc() mem_cgroup_free() indirectly calls wb_domain_exit() which is not prepared to deal with a struct wb_domain object that hasn't executed wb_domain_init(). For instance, the following warning message is printed by lockdep if alloc_percpu() fails in mem_cgroup_alloc(): INFO: trying to register non-static key. the code is fine but needs lockdep annotation. turning off the locking correctness validator. CPU: 1 PID: 1950 Comm: mkdir Not tainted 4.10.0+ #151 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011 Call Trace: dump_stack+0x67/0x99 register_lock_class+0x36d/0x540 __lock_acquire+0x7f/0x1a30 lock_acquire+0xcc/0x200 del_timer_sync+0x3c/0xc0 wb_domain_exit+0x14/0x20 mem_cgroup_free+0x14/0x40 mem_cgroup_css_alloc+0x3f9/0x620 cgroup_apply_control_enable+0x190/0x390 cgroup_mkdir+0x290/0x3d0 kernfs_iop_mkdir+0x58/0x80 vfs_mkdir+0x10e/0x1a0 SyS_mkdirat+0xa8/0xd0 SyS_mkdir+0x14/0x20 entry_SYSCALL_64_fastpath+0x18/0xad Add __mem_cgroup_free() which skips wb_domain_exit(). This is used by both mem_cgroup_free() and mem_cgroup_alloc() clean up. Fixes: 0b8f73e104285 ("mm: memcontrol: clean up alloc, online, offline, free functions") Link: http://lkml.kernel.org/r/20170306192122.24262-1-tahsin@google.com Signed-off-by: Tahsin Erdogan <tahsin@google.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-03-10 00:17:26 +00:00
static void __mem_cgroup_free(struct mem_cgroup *memcg)
memcg: free mem_cgroup by RCU to fix oops After fixing the GPF in mem_cgroup_lru_del_list(), three times one machine running a similar load (moving and removing memcgs while swapping) has oopsed in mem_cgroup_zone_nr_lru_pages(), when retrieving memcg zone numbers for get_scan_count() for shrink_mem_cgroup_zone(): this is where a struct mem_cgroup is first accessed after being chosen by mem_cgroup_iter(). Just what protects a struct mem_cgroup from being freed, in between mem_cgroup_iter()'s css_get_next() and its css_tryget()? css_tryget() fails once css->refcnt is zero with CSS_REMOVED set in flags, yes: but what if that memory is freed and reused for something else, which sets "refcnt" non-zero? Hmm, and scope for an indefinite freeze if refcnt is left at zero but flags are cleared. It's tempting to move the css_tryget() into css_get_next(), to make it really "get" the css, but I don't think that actually solves anything: the same difficulty in moving from css_id found to stable css remains. But we already have rcu_read_lock() around the two, so it's easily fixed if __mem_cgroup_free() just uses kfree_rcu() to free mem_cgroup. However, a big struct mem_cgroup is allocated with vzalloc() instead of kzalloc(), and we're not allowed to vfree() at interrupt time: there doesn't appear to be a general vfree_rcu() to help with this, so roll our own using schedule_work(). The compiler decently removes vfree_work() and vfree_rcu() when the config doesn't need them. Signed-off-by: Hugh Dickins <hughd@google.com> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Konstantin Khlebnikov <khlebnikov@openvz.org> Cc: Tejun Heo <tj@kernel.org> Cc: Ying Han <yinghan@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-15 22:17:07 +00:00
{
memcg: execute the whole memcg freeing in free_worker() A lot of the initialization we do in mem_cgroup_create() is done with softirqs enabled. This include grabbing a css id, which holds &ss->id_lock->rlock, and the per-zone trees, which holds rtpz->lock->rlock. All of those signal to the lockdep mechanism that those locks can be used in SOFTIRQ-ON-W context. This means that the freeing of memcg structure must happen in a compatible context, otherwise we'll get a deadlock, like the one below, caught by lockdep: free_accounted_pages+0x47/0x4c free_task+0x31/0x5c __put_task_struct+0xc2/0xdb put_task_struct+0x1e/0x22 delayed_put_task_struct+0x7a/0x98 __rcu_process_callbacks+0x269/0x3df rcu_process_callbacks+0x31/0x5b __do_softirq+0x122/0x277 This usage pattern could not be triggered before kmem came into play. With the introduction of kmem stack handling, it is possible that we call the last mem_cgroup_put() from the task destructor, which is run in an rcu callback. Such callbacks are run with softirqs disabled, leading to the offensive usage pattern. In general, we have little, if any, means to guarantee in which context the last memcg_put will happen. The best we can do is test it and try to make sure no invalid context releases are happening. But as we add more code to memcg, the possible interactions grow in number and expose more ways to get context conflicts. One thing to keep in mind, is that part of the freeing process is already deferred to a worker, such as vfree(), that can only be called from process context. For the moment, the only two functions we really need moved away are: * free_css_id(), and * mem_cgroup_remove_from_trees(). But because the later accesses per-zone info, free_mem_cgroup_per_zone_info() needs to be moved as well. With that, we are left with the per_cpu stats only. Better move it all. Signed-off-by: Glauber Costa <glommer@parallels.com> Tested-by: Greg Thelen <gthelen@google.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:13 +00:00
int node;
memcg: free mem_cgroup by RCU to fix oops After fixing the GPF in mem_cgroup_lru_del_list(), three times one machine running a similar load (moving and removing memcgs while swapping) has oopsed in mem_cgroup_zone_nr_lru_pages(), when retrieving memcg zone numbers for get_scan_count() for shrink_mem_cgroup_zone(): this is where a struct mem_cgroup is first accessed after being chosen by mem_cgroup_iter(). Just what protects a struct mem_cgroup from being freed, in between mem_cgroup_iter()'s css_get_next() and its css_tryget()? css_tryget() fails once css->refcnt is zero with CSS_REMOVED set in flags, yes: but what if that memory is freed and reused for something else, which sets "refcnt" non-zero? Hmm, and scope for an indefinite freeze if refcnt is left at zero but flags are cleared. It's tempting to move the css_tryget() into css_get_next(), to make it really "get" the css, but I don't think that actually solves anything: the same difficulty in moving from css_id found to stable css remains. But we already have rcu_read_lock() around the two, so it's easily fixed if __mem_cgroup_free() just uses kfree_rcu() to free mem_cgroup. However, a big struct mem_cgroup is allocated with vzalloc() instead of kzalloc(), and we're not allowed to vfree() at interrupt time: there doesn't appear to be a general vfree_rcu() to help with this, so roll our own using schedule_work(). The compiler decently removes vfree_work() and vfree_rcu() when the config doesn't need them. Signed-off-by: Hugh Dickins <hughd@google.com> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Konstantin Khlebnikov <khlebnikov@openvz.org> Cc: Tejun Heo <tj@kernel.org> Cc: Ying Han <yinghan@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-15 22:17:07 +00:00
memcg: execute the whole memcg freeing in free_worker() A lot of the initialization we do in mem_cgroup_create() is done with softirqs enabled. This include grabbing a css id, which holds &ss->id_lock->rlock, and the per-zone trees, which holds rtpz->lock->rlock. All of those signal to the lockdep mechanism that those locks can be used in SOFTIRQ-ON-W context. This means that the freeing of memcg structure must happen in a compatible context, otherwise we'll get a deadlock, like the one below, caught by lockdep: free_accounted_pages+0x47/0x4c free_task+0x31/0x5c __put_task_struct+0xc2/0xdb put_task_struct+0x1e/0x22 delayed_put_task_struct+0x7a/0x98 __rcu_process_callbacks+0x269/0x3df rcu_process_callbacks+0x31/0x5b __do_softirq+0x122/0x277 This usage pattern could not be triggered before kmem came into play. With the introduction of kmem stack handling, it is possible that we call the last mem_cgroup_put() from the task destructor, which is run in an rcu callback. Such callbacks are run with softirqs disabled, leading to the offensive usage pattern. In general, we have little, if any, means to guarantee in which context the last memcg_put will happen. The best we can do is test it and try to make sure no invalid context releases are happening. But as we add more code to memcg, the possible interactions grow in number and expose more ways to get context conflicts. One thing to keep in mind, is that part of the freeing process is already deferred to a worker, such as vfree(), that can only be called from process context. For the moment, the only two functions we really need moved away are: * free_css_id(), and * mem_cgroup_remove_from_trees(). But because the later accesses per-zone info, free_mem_cgroup_per_zone_info() needs to be moved as well. With that, we are left with the per_cpu stats only. Better move it all. Signed-off-by: Glauber Costa <glommer@parallels.com> Tested-by: Greg Thelen <gthelen@google.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Frederic Weisbecker <fweisbec@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: JoonSoo Kim <js1304@gmail.com> Cc: Mel Gorman <mel@csn.ul.ie> Cc: Pekka Enberg <penberg@cs.helsinki.fi> Cc: Rik van Riel <riel@redhat.com> Cc: Suleiman Souhlal <suleiman@google.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-12-18 22:22:13 +00:00
for_each_node(node)
free_mem_cgroup_per_node_info(memcg, node);
free_percpu(memcg->vmstats_percpu);
mm: memcontrol: don't batch updates of local VM stats and events The kernel test robot noticed a 26% will-it-scale pagefault regression from commit 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty"). This appears to be caused by bouncing the additional cachelines from the new hierarchical statistics counters. We can fix this by getting rid of the batched local counters instead. Originally, there were *only* group-local counters, and they were fully maintained per cpu. A reader of a stats file high up in the cgroup tree would have to walk the entire subtree and collect each level's per-cpu counters to get the recursive view. This was prohibitively expensive, and so we switched to per-cpu batched updates of the local counters during a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting"), reducing the complexity from nr_subgroups * nr_cpus to nr_subgroups. With growing machines and cgroup trees, the tree walk itself became too expensive for monitoring top-level groups, and this is when the culprit patch added hierarchy counters on each cgroup level. When the per-cpu batch size would be reached, both the local and the hierarchy counters would get batch-updated from the per-cpu delta simultaneously. This makes local and hierarchical counter reads blazingly fast, but it unfortunately makes the write-side too cache line intense. Since local counter reads were never a problem - we only centralized them to accelerate the hierarchy walk - and use of the local counters are becoming rarer due to replacement with hierarchical views (ongoing rework in the page reclaim and workingset code), we can make those local counters unbatched per-cpu counters again. The scheme will then be as such: when a memcg statistic changes, the writer will: - update the local counter (per-cpu) - update the batch counter (per-cpu). If the batch is full: - spill the batch into the group's atomic_t - spill the batch into all ancestors' atomic_ts - empty out the batch counter (per-cpu) when a local memcg counter is read, the reader will: - collect the local counter from all cpus when a hiearchy memcg counter is read, the reader will: - read the atomic_t We might be able to simplify this further and make the recursive counters unbatched per-cpu counters as well (batch upward propagation, but leave per-cpu collection to the readers), but that will require a more in-depth analysis and testing of all the callsites. Deal with the immediate regression for now. Link: http://lkml.kernel.org/r/20190521151647.GB2870@cmpxchg.org Fixes: 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: kernel test robot <rong.a.chen@intel.com> Tested-by: kernel test robot <rong.a.chen@intel.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Shakeel Butt <shakeelb@google.com> Cc: Roman Gushchin <guro@fb.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-06-13 22:55:46 +00:00
free_percpu(memcg->vmstats_local);
kfree(memcg);
memcg: free mem_cgroup by RCU to fix oops After fixing the GPF in mem_cgroup_lru_del_list(), three times one machine running a similar load (moving and removing memcgs while swapping) has oopsed in mem_cgroup_zone_nr_lru_pages(), when retrieving memcg zone numbers for get_scan_count() for shrink_mem_cgroup_zone(): this is where a struct mem_cgroup is first accessed after being chosen by mem_cgroup_iter(). Just what protects a struct mem_cgroup from being freed, in between mem_cgroup_iter()'s css_get_next() and its css_tryget()? css_tryget() fails once css->refcnt is zero with CSS_REMOVED set in flags, yes: but what if that memory is freed and reused for something else, which sets "refcnt" non-zero? Hmm, and scope for an indefinite freeze if refcnt is left at zero but flags are cleared. It's tempting to move the css_tryget() into css_get_next(), to make it really "get" the css, but I don't think that actually solves anything: the same difficulty in moving from css_id found to stable css remains. But we already have rcu_read_lock() around the two, so it's easily fixed if __mem_cgroup_free() just uses kfree_rcu() to free mem_cgroup. However, a big struct mem_cgroup is allocated with vzalloc() instead of kzalloc(), and we're not allowed to vfree() at interrupt time: there doesn't appear to be a general vfree_rcu() to help with this, so roll our own using schedule_work(). The compiler decently removes vfree_work() and vfree_rcu() when the config doesn't need them. Signed-off-by: Hugh Dickins <hughd@google.com> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Konstantin Khlebnikov <khlebnikov@openvz.org> Cc: Tejun Heo <tj@kernel.org> Cc: Ying Han <yinghan@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-15 22:17:07 +00:00
}
mm: do not call mem_cgroup_free() from within mem_cgroup_alloc() mem_cgroup_free() indirectly calls wb_domain_exit() which is not prepared to deal with a struct wb_domain object that hasn't executed wb_domain_init(). For instance, the following warning message is printed by lockdep if alloc_percpu() fails in mem_cgroup_alloc(): INFO: trying to register non-static key. the code is fine but needs lockdep annotation. turning off the locking correctness validator. CPU: 1 PID: 1950 Comm: mkdir Not tainted 4.10.0+ #151 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011 Call Trace: dump_stack+0x67/0x99 register_lock_class+0x36d/0x540 __lock_acquire+0x7f/0x1a30 lock_acquire+0xcc/0x200 del_timer_sync+0x3c/0xc0 wb_domain_exit+0x14/0x20 mem_cgroup_free+0x14/0x40 mem_cgroup_css_alloc+0x3f9/0x620 cgroup_apply_control_enable+0x190/0x390 cgroup_mkdir+0x290/0x3d0 kernfs_iop_mkdir+0x58/0x80 vfs_mkdir+0x10e/0x1a0 SyS_mkdirat+0xa8/0xd0 SyS_mkdir+0x14/0x20 entry_SYSCALL_64_fastpath+0x18/0xad Add __mem_cgroup_free() which skips wb_domain_exit(). This is used by both mem_cgroup_free() and mem_cgroup_alloc() clean up. Fixes: 0b8f73e104285 ("mm: memcontrol: clean up alloc, online, offline, free functions") Link: http://lkml.kernel.org/r/20170306192122.24262-1-tahsin@google.com Signed-off-by: Tahsin Erdogan <tahsin@google.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-03-10 00:17:26 +00:00
static void mem_cgroup_free(struct mem_cgroup *memcg)
{
memcg_wb_domain_exit(memcg);
mm: memcontrol: fix NULL-ptr deref in percpu stats flush __mem_cgroup_free() can be called on the failure path in mem_cgroup_alloc(). However memcg_flush_percpu_vmstats() and memcg_flush_percpu_vmevents() which are called from __mem_cgroup_free() access the fields of memcg which can potentially be null if called from failure path from mem_cgroup_alloc(). Indeed syzbot has reported the following crash: kasan: CONFIG_KASAN_INLINE enabled kasan: GPF could be caused by NULL-ptr deref or user memory access general protection fault: 0000 [#1] PREEMPT SMP KASAN CPU: 0 PID: 30393 Comm: syz-executor.1 Not tainted 5.4.0-rc2+ #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 RIP: 0010:memcg_flush_percpu_vmstats+0x4ae/0x930 mm/memcontrol.c:3436 Code: 05 41 89 c0 41 0f b6 04 24 41 38 c7 7c 08 84 c0 0f 85 5d 03 00 00 44 3b 05 33 d5 12 08 0f 83 e2 00 00 00 4c 89 f0 48 c1 e8 03 <42> 80 3c 28 00 0f 85 91 03 00 00 48 8b 85 10 fe ff ff 48 8b b0 90 RSP: 0018:ffff888095c27980 EFLAGS: 00010206 RAX: 0000000000000012 RBX: ffff888095c27b28 RCX: ffffc90008192000 RDX: 0000000000040000 RSI: ffffffff8340fae7 RDI: 0000000000000007 RBP: ffff888095c27be0 R08: 0000000000000000 R09: ffffed1013f0da33 R10: ffffed1013f0da32 R11: ffff88809f86d197 R12: fffffbfff138b760 R13: dffffc0000000000 R14: 0000000000000090 R15: 0000000000000007 FS: 00007f5027170700(0000) GS:ffff8880ae800000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000710158 CR3: 00000000a7b18000 CR4: 00000000001406f0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: __mem_cgroup_free+0x1a/0x190 mm/memcontrol.c:5021 mem_cgroup_free mm/memcontrol.c:5033 [inline] mem_cgroup_css_alloc+0x3a1/0x1ae0 mm/memcontrol.c:5160 css_create kernel/cgroup/cgroup.c:5156 [inline] cgroup_apply_control_enable+0x44d/0xc40 kernel/cgroup/cgroup.c:3119 cgroup_mkdir+0x899/0x11b0 kernel/cgroup/cgroup.c:5401 kernfs_iop_mkdir+0x14d/0x1d0 fs/kernfs/dir.c:1124 vfs_mkdir+0x42e/0x670 fs/namei.c:3807 do_mkdirat+0x234/0x2a0 fs/namei.c:3830 __do_sys_mkdir fs/namei.c:3846 [inline] __se_sys_mkdir fs/namei.c:3844 [inline] __x64_sys_mkdir+0x5c/0x80 fs/namei.c:3844 do_syscall_64+0xfa/0x760 arch/x86/entry/common.c:290 entry_SYSCALL_64_after_hwframe+0x49/0xbe Fixing this by moving the flush to mem_cgroup_free as there is no need to flush anything if we see failure in mem_cgroup_alloc(). Link: http://lkml.kernel.org/r/20191018165231.249872-1-shakeelb@google.com Fixes: bb65f89b7d3d ("mm: memcontrol: flush percpu vmevents before releasing memcg") Fixes: c350a99ea2b1 ("mm: memcontrol: flush percpu vmstats before releasing memcg") Signed-off-by: Shakeel Butt <shakeelb@google.com> Reported-by: syzbot+515d5bcfe179cdf049b2@syzkaller.appspotmail.com Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-11-06 05:16:21 +00:00
/*
* Flush percpu vmstats and vmevents to guarantee the value correctness
* on parent's and all ancestor levels.
*/
mm: memcg/slab: fix percpu slab vmstats flushing Currently slab percpu vmstats are flushed twice: during the memcg offlining and just before freeing the memcg structure. Each time percpu counters are summed, added to the atomic counterparts and propagated up by the cgroup tree. The second flushing is required due to how recursive vmstats are implemented: counters are batched in percpu variables on a local level, and once a percpu value is crossing some predefined threshold, it spills over to atomic values on the local and each ascendant levels. It means that without flushing some numbers cached in percpu variables will be dropped on floor each time a cgroup is destroyed. And with uptime the error on upper levels might become noticeable. The first flushing aims to make counters on ancestor levels more precise. Dying cgroups may resume in the dying state for a long time. After kmem_cache reparenting which is performed during the offlining slab counters of the dying cgroup don't have any chances to be updated, because any slab operations will be performed on the parent level. It means that the inaccuracy caused by percpu batching will not decrease up to the final destruction of the cgroup. By the original idea flushing slab counters during the offlining should minimize the visible inaccuracy of slab counters on the parent level. The problem is that percpu counters are not zeroed after the first flushing. So every cached percpu value is summed twice. It creates a small error (up to 32 pages per cpu, but usually less) which accumulates on parent cgroup level. After creating and destroying of thousands of child cgroups, slab counter on parent level can be way off the real value. For now, let's just stop flushing slab counters on memcg offlining. It can't be done correctly without scheduling a work on each cpu: reading and zeroing it during css offlining can race with an asynchronous update, which doesn't expect values to be changed underneath. With this change, slab counters on parent level will become eventually consistent. Once all dying children are gone, values are correct. And if not, the error is capped by 32 * NR_CPUS pages per dying cgroup. It's not perfect, as slab are reparented, so any updates after the reparenting will happen on the parent level. It means that if a slab page was allocated, a counter on child level was bumped, then the page was reparented and freed, the annihilation of positive and negative counter values will not happen until the child cgroup is released. It makes slab counters different from others, and it might want us to implement flushing in a correct form again. But it's also a question of performance: scheduling a work on each cpu isn't free, and it's an open question if the benefit of having more accurate counters is worth it. We might also consider flushing all counters on offlining, not only slab counters. So let's fix the main problem now: make the slab counters eventually consistent, so at least the error won't grow with uptime (or more precisely the number of created and destroyed cgroups). And think about the accuracy of counters separately. Link: http://lkml.kernel.org/r/20191220042728.1045881-1-guro@fb.com Fixes: bee07b33db78 ("mm: memcontrol: flush percpu slab vmstats on kmem offlining") Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-01-14 00:29:16 +00:00
memcg_flush_percpu_vmstats(memcg);
mm: memcontrol: fix NULL-ptr deref in percpu stats flush __mem_cgroup_free() can be called on the failure path in mem_cgroup_alloc(). However memcg_flush_percpu_vmstats() and memcg_flush_percpu_vmevents() which are called from __mem_cgroup_free() access the fields of memcg which can potentially be null if called from failure path from mem_cgroup_alloc(). Indeed syzbot has reported the following crash: kasan: CONFIG_KASAN_INLINE enabled kasan: GPF could be caused by NULL-ptr deref or user memory access general protection fault: 0000 [#1] PREEMPT SMP KASAN CPU: 0 PID: 30393 Comm: syz-executor.1 Not tainted 5.4.0-rc2+ #0 Hardware name: Google Google Compute Engine/Google Compute Engine, BIOS Google 01/01/2011 RIP: 0010:memcg_flush_percpu_vmstats+0x4ae/0x930 mm/memcontrol.c:3436 Code: 05 41 89 c0 41 0f b6 04 24 41 38 c7 7c 08 84 c0 0f 85 5d 03 00 00 44 3b 05 33 d5 12 08 0f 83 e2 00 00 00 4c 89 f0 48 c1 e8 03 <42> 80 3c 28 00 0f 85 91 03 00 00 48 8b 85 10 fe ff ff 48 8b b0 90 RSP: 0018:ffff888095c27980 EFLAGS: 00010206 RAX: 0000000000000012 RBX: ffff888095c27b28 RCX: ffffc90008192000 RDX: 0000000000040000 RSI: ffffffff8340fae7 RDI: 0000000000000007 RBP: ffff888095c27be0 R08: 0000000000000000 R09: ffffed1013f0da33 R10: ffffed1013f0da32 R11: ffff88809f86d197 R12: fffffbfff138b760 R13: dffffc0000000000 R14: 0000000000000090 R15: 0000000000000007 FS: 00007f5027170700(0000) GS:ffff8880ae800000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000710158 CR3: 00000000a7b18000 CR4: 00000000001406f0 DR0: 0000000000000000 DR1: 0000000000000000 DR2: 0000000000000000 DR3: 0000000000000000 DR6: 00000000fffe0ff0 DR7: 0000000000000400 Call Trace: __mem_cgroup_free+0x1a/0x190 mm/memcontrol.c:5021 mem_cgroup_free mm/memcontrol.c:5033 [inline] mem_cgroup_css_alloc+0x3a1/0x1ae0 mm/memcontrol.c:5160 css_create kernel/cgroup/cgroup.c:5156 [inline] cgroup_apply_control_enable+0x44d/0xc40 kernel/cgroup/cgroup.c:3119 cgroup_mkdir+0x899/0x11b0 kernel/cgroup/cgroup.c:5401 kernfs_iop_mkdir+0x14d/0x1d0 fs/kernfs/dir.c:1124 vfs_mkdir+0x42e/0x670 fs/namei.c:3807 do_mkdirat+0x234/0x2a0 fs/namei.c:3830 __do_sys_mkdir fs/namei.c:3846 [inline] __se_sys_mkdir fs/namei.c:3844 [inline] __x64_sys_mkdir+0x5c/0x80 fs/namei.c:3844 do_syscall_64+0xfa/0x760 arch/x86/entry/common.c:290 entry_SYSCALL_64_after_hwframe+0x49/0xbe Fixing this by moving the flush to mem_cgroup_free as there is no need to flush anything if we see failure in mem_cgroup_alloc(). Link: http://lkml.kernel.org/r/20191018165231.249872-1-shakeelb@google.com Fixes: bb65f89b7d3d ("mm: memcontrol: flush percpu vmevents before releasing memcg") Fixes: c350a99ea2b1 ("mm: memcontrol: flush percpu vmstats before releasing memcg") Signed-off-by: Shakeel Butt <shakeelb@google.com> Reported-by: syzbot+515d5bcfe179cdf049b2@syzkaller.appspotmail.com Reviewed-by: Roman Gushchin <guro@fb.com> Cc: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-11-06 05:16:21 +00:00
memcg_flush_percpu_vmevents(memcg);
mm: do not call mem_cgroup_free() from within mem_cgroup_alloc() mem_cgroup_free() indirectly calls wb_domain_exit() which is not prepared to deal with a struct wb_domain object that hasn't executed wb_domain_init(). For instance, the following warning message is printed by lockdep if alloc_percpu() fails in mem_cgroup_alloc(): INFO: trying to register non-static key. the code is fine but needs lockdep annotation. turning off the locking correctness validator. CPU: 1 PID: 1950 Comm: mkdir Not tainted 4.10.0+ #151 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011 Call Trace: dump_stack+0x67/0x99 register_lock_class+0x36d/0x540 __lock_acquire+0x7f/0x1a30 lock_acquire+0xcc/0x200 del_timer_sync+0x3c/0xc0 wb_domain_exit+0x14/0x20 mem_cgroup_free+0x14/0x40 mem_cgroup_css_alloc+0x3f9/0x620 cgroup_apply_control_enable+0x190/0x390 cgroup_mkdir+0x290/0x3d0 kernfs_iop_mkdir+0x58/0x80 vfs_mkdir+0x10e/0x1a0 SyS_mkdirat+0xa8/0xd0 SyS_mkdir+0x14/0x20 entry_SYSCALL_64_fastpath+0x18/0xad Add __mem_cgroup_free() which skips wb_domain_exit(). This is used by both mem_cgroup_free() and mem_cgroup_alloc() clean up. Fixes: 0b8f73e104285 ("mm: memcontrol: clean up alloc, online, offline, free functions") Link: http://lkml.kernel.org/r/20170306192122.24262-1-tahsin@google.com Signed-off-by: Tahsin Erdogan <tahsin@google.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-03-10 00:17:26 +00:00
__mem_cgroup_free(memcg);
}
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
static struct mem_cgroup *mem_cgroup_alloc(void)
{
struct mem_cgroup *memcg;
unsigned int size;
int node;
writeback, memcg: Implement foreign dirty flushing There's an inherent mismatch between memcg and writeback. The former trackes ownership per-page while the latter per-inode. This was a deliberate design decision because honoring per-page ownership in the writeback path is complicated, may lead to higher CPU and IO overheads and deemed unnecessary given that write-sharing an inode across different cgroups isn't a common use-case. Combined with inode majority-writer ownership switching, this works well enough in most cases but there are some pathological cases. For example, let's say there are two cgroups A and B which keep writing to different but confined parts of the same inode. B owns the inode and A's memory is limited far below B's. A's dirty ratio can rise enough to trigger balance_dirty_pages() sleeps but B's can be low enough to avoid triggering background writeback. A will be slowed down without a way to make writeback of the dirty pages happen. This patch implements foreign dirty recording and foreign mechanism so that when a memcg encounters a condition as above it can trigger flushes on bdi_writebacks which can clean its pages. Please see the comment on top of mem_cgroup_track_foreign_dirty_slowpath() for details. A reproducer follows. write-range.c:: #include <stdio.h> #include <stdlib.h> #include <unistd.h> #include <fcntl.h> #include <sys/types.h> static const char *usage = "write-range FILE START SIZE\n"; int main(int argc, char **argv) { int fd; unsigned long start, size, end, pos; char *endp; char buf[4096]; if (argc < 4) { fprintf(stderr, usage); return 1; } fd = open(argv[1], O_WRONLY); if (fd < 0) { perror("open"); return 1; } start = strtoul(argv[2], &endp, 0); if (*endp != '\0') { fprintf(stderr, usage); return 1; } size = strtoul(argv[3], &endp, 0); if (*endp != '\0') { fprintf(stderr, usage); return 1; } end = start + size; while (1) { for (pos = start; pos < end; ) { long bread, bwritten = 0; if (lseek(fd, pos, SEEK_SET) < 0) { perror("lseek"); return 1; } bread = read(0, buf, sizeof(buf) < end - pos ? sizeof(buf) : end - pos); if (bread < 0) { perror("read"); return 1; } if (bread == 0) return 0; while (bwritten < bread) { long this; this = write(fd, buf + bwritten, bread - bwritten); if (this < 0) { perror("write"); return 1; } bwritten += this; pos += bwritten; } } } } repro.sh:: #!/bin/bash set -e set -x sysctl -w vm.dirty_expire_centisecs=300000 sysctl -w vm.dirty_writeback_centisecs=300000 sysctl -w vm.dirtytime_expire_seconds=300000 echo 3 > /proc/sys/vm/drop_caches TEST=/sys/fs/cgroup/test A=$TEST/A B=$TEST/B mkdir -p $A $B echo "+memory +io" > $TEST/cgroup.subtree_control echo $((1<<30)) > $A/memory.high echo $((32<<30)) > $B/memory.high rm -f testfile touch testfile fallocate -l 4G testfile echo "Starting B" (echo $BASHPID > $B/cgroup.procs pv -q --rate-limit 70M < /dev/urandom | ./write-range testfile $((2<<30)) $((2<<30))) & echo "Waiting 10s to ensure B claims the testfile inode" sleep 5 sync sleep 5 sync echo "Starting A" (echo $BASHPID > $A/cgroup.procs pv < /dev/urandom | ./write-range testfile 0 $((2<<30))) v2: Added comments explaining why the specific intervals are being used. v3: Use 0 @nr when calling cgroup_writeback_by_id() to use best-effort flushing while avoding possible livelocks. v4: Use get_jiffies_64() and time_before/after64() instead of raw jiffies_64 and arthimetic comparisons as suggested by Jan. Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-08-26 16:06:56 +00:00
int __maybe_unused i;
long error = -ENOMEM;
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
size = sizeof(struct mem_cgroup);
size += nr_node_ids * sizeof(struct mem_cgroup_per_node *);
memcg = kzalloc(size, GFP_KERNEL);
if (!memcg)
return ERR_PTR(error);
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
mm: memcontrol: fix cgroup creation failure after many small jobs The memory controller has quite a bit of state that usually outlives the cgroup and pins its CSS until said state disappears. At the same time it imposes a 16-bit limit on the CSS ID space to economically store IDs in the wild. Consequently, when we use cgroups to contain frequent but small and short-lived jobs that leave behind some page cache, we quickly run into the 64k limitations of outstanding CSSs. Creating a new cgroup fails with -ENOSPC while there are only a few, or even no user-visible cgroups in existence. Although pinning CSSs past cgroup removal is common, there are only two instances that actually need an ID after a cgroup is deleted: cache shadow entries and swapout records. Cache shadow entries reference the ID weakly and can deal with the CSS having disappeared when it's looked up later. They pose no hurdle. Swap-out records do need to pin the css to hierarchically attribute swapins after the cgroup has been deleted; though the only pages that remain swapped out after offlining are tmpfs/shmem pages. And those references are under the user's control, so they are manageable. This patch introduces a private 16-bit memcg ID and switches swap and cache shadow entries over to using that. This ID can then be recycled after offlining when the CSS remains pinned only by objects that don't specifically need it. This script demonstrates the problem by faulting one cache page in a new cgroup and deleting it again: set -e mkdir -p pages for x in `seq 128000`; do [ $((x % 1000)) -eq 0 ] && echo $x mkdir /cgroup/foo echo $$ >/cgroup/foo/cgroup.procs echo trex >pages/$x echo $$ >/cgroup/cgroup.procs rmdir /cgroup/foo done When run on an unpatched kernel, we eventually run out of possible IDs even though there are no visible cgroups: [root@ham ~]# ./cssidstress.sh [...] 65000 mkdir: cannot create directory '/cgroup/foo': No space left on device After this patch, the IDs get released upon cgroup destruction and the cache and css objects get released once memory reclaim kicks in. [hannes@cmpxchg.org: init the IDR] Link: http://lkml.kernel.org/r/20160621154601.GA22431@cmpxchg.org Fixes: b2052564e66d ("mm: memcontrol: continue cache reclaim from offlined groups") Link: http://lkml.kernel.org/r/20160617162516.GD19084@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: John Garcia <john.garcia@mesosphere.io> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: Nikolay Borisov <kernel@kyup.com> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-20 22:44:57 +00:00
memcg->id.id = idr_alloc(&mem_cgroup_idr, NULL,
1, MEM_CGROUP_ID_MAX,
GFP_KERNEL);
if (memcg->id.id < 0) {
error = memcg->id.id;
mm: memcontrol: fix cgroup creation failure after many small jobs The memory controller has quite a bit of state that usually outlives the cgroup and pins its CSS until said state disappears. At the same time it imposes a 16-bit limit on the CSS ID space to economically store IDs in the wild. Consequently, when we use cgroups to contain frequent but small and short-lived jobs that leave behind some page cache, we quickly run into the 64k limitations of outstanding CSSs. Creating a new cgroup fails with -ENOSPC while there are only a few, or even no user-visible cgroups in existence. Although pinning CSSs past cgroup removal is common, there are only two instances that actually need an ID after a cgroup is deleted: cache shadow entries and swapout records. Cache shadow entries reference the ID weakly and can deal with the CSS having disappeared when it's looked up later. They pose no hurdle. Swap-out records do need to pin the css to hierarchically attribute swapins after the cgroup has been deleted; though the only pages that remain swapped out after offlining are tmpfs/shmem pages. And those references are under the user's control, so they are manageable. This patch introduces a private 16-bit memcg ID and switches swap and cache shadow entries over to using that. This ID can then be recycled after offlining when the CSS remains pinned only by objects that don't specifically need it. This script demonstrates the problem by faulting one cache page in a new cgroup and deleting it again: set -e mkdir -p pages for x in `seq 128000`; do [ $((x % 1000)) -eq 0 ] && echo $x mkdir /cgroup/foo echo $$ >/cgroup/foo/cgroup.procs echo trex >pages/$x echo $$ >/cgroup/cgroup.procs rmdir /cgroup/foo done When run on an unpatched kernel, we eventually run out of possible IDs even though there are no visible cgroups: [root@ham ~]# ./cssidstress.sh [...] 65000 mkdir: cannot create directory '/cgroup/foo': No space left on device After this patch, the IDs get released upon cgroup destruction and the cache and css objects get released once memory reclaim kicks in. [hannes@cmpxchg.org: init the IDR] Link: http://lkml.kernel.org/r/20160621154601.GA22431@cmpxchg.org Fixes: b2052564e66d ("mm: memcontrol: continue cache reclaim from offlined groups") Link: http://lkml.kernel.org/r/20160617162516.GD19084@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: John Garcia <john.garcia@mesosphere.io> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: Nikolay Borisov <kernel@kyup.com> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-20 22:44:57 +00:00
goto fail;
}
mm: memcontrol: fix cgroup creation failure after many small jobs The memory controller has quite a bit of state that usually outlives the cgroup and pins its CSS until said state disappears. At the same time it imposes a 16-bit limit on the CSS ID space to economically store IDs in the wild. Consequently, when we use cgroups to contain frequent but small and short-lived jobs that leave behind some page cache, we quickly run into the 64k limitations of outstanding CSSs. Creating a new cgroup fails with -ENOSPC while there are only a few, or even no user-visible cgroups in existence. Although pinning CSSs past cgroup removal is common, there are only two instances that actually need an ID after a cgroup is deleted: cache shadow entries and swapout records. Cache shadow entries reference the ID weakly and can deal with the CSS having disappeared when it's looked up later. They pose no hurdle. Swap-out records do need to pin the css to hierarchically attribute swapins after the cgroup has been deleted; though the only pages that remain swapped out after offlining are tmpfs/shmem pages. And those references are under the user's control, so they are manageable. This patch introduces a private 16-bit memcg ID and switches swap and cache shadow entries over to using that. This ID can then be recycled after offlining when the CSS remains pinned only by objects that don't specifically need it. This script demonstrates the problem by faulting one cache page in a new cgroup and deleting it again: set -e mkdir -p pages for x in `seq 128000`; do [ $((x % 1000)) -eq 0 ] && echo $x mkdir /cgroup/foo echo $$ >/cgroup/foo/cgroup.procs echo trex >pages/$x echo $$ >/cgroup/cgroup.procs rmdir /cgroup/foo done When run on an unpatched kernel, we eventually run out of possible IDs even though there are no visible cgroups: [root@ham ~]# ./cssidstress.sh [...] 65000 mkdir: cannot create directory '/cgroup/foo': No space left on device After this patch, the IDs get released upon cgroup destruction and the cache and css objects get released once memory reclaim kicks in. [hannes@cmpxchg.org: init the IDR] Link: http://lkml.kernel.org/r/20160621154601.GA22431@cmpxchg.org Fixes: b2052564e66d ("mm: memcontrol: continue cache reclaim from offlined groups") Link: http://lkml.kernel.org/r/20160617162516.GD19084@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: John Garcia <john.garcia@mesosphere.io> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: Nikolay Borisov <kernel@kyup.com> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-20 22:44:57 +00:00
mm: memcontrol: don't batch updates of local VM stats and events The kernel test robot noticed a 26% will-it-scale pagefault regression from commit 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty"). This appears to be caused by bouncing the additional cachelines from the new hierarchical statistics counters. We can fix this by getting rid of the batched local counters instead. Originally, there were *only* group-local counters, and they were fully maintained per cpu. A reader of a stats file high up in the cgroup tree would have to walk the entire subtree and collect each level's per-cpu counters to get the recursive view. This was prohibitively expensive, and so we switched to per-cpu batched updates of the local counters during a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting"), reducing the complexity from nr_subgroups * nr_cpus to nr_subgroups. With growing machines and cgroup trees, the tree walk itself became too expensive for monitoring top-level groups, and this is when the culprit patch added hierarchy counters on each cgroup level. When the per-cpu batch size would be reached, both the local and the hierarchy counters would get batch-updated from the per-cpu delta simultaneously. This makes local and hierarchical counter reads blazingly fast, but it unfortunately makes the write-side too cache line intense. Since local counter reads were never a problem - we only centralized them to accelerate the hierarchy walk - and use of the local counters are becoming rarer due to replacement with hierarchical views (ongoing rework in the page reclaim and workingset code), we can make those local counters unbatched per-cpu counters again. The scheme will then be as such: when a memcg statistic changes, the writer will: - update the local counter (per-cpu) - update the batch counter (per-cpu). If the batch is full: - spill the batch into the group's atomic_t - spill the batch into all ancestors' atomic_ts - empty out the batch counter (per-cpu) when a local memcg counter is read, the reader will: - collect the local counter from all cpus when a hiearchy memcg counter is read, the reader will: - read the atomic_t We might be able to simplify this further and make the recursive counters unbatched per-cpu counters as well (batch upward propagation, but leave per-cpu collection to the readers), but that will require a more in-depth analysis and testing of all the callsites. Deal with the immediate regression for now. Link: http://lkml.kernel.org/r/20190521151647.GB2870@cmpxchg.org Fixes: 42a300353577 ("mm: memcontrol: fix recursive statistics correctness & scalabilty") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: kernel test robot <rong.a.chen@intel.com> Tested-by: kernel test robot <rong.a.chen@intel.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Shakeel Butt <shakeelb@google.com> Cc: Roman Gushchin <guro@fb.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-06-13 22:55:46 +00:00
memcg->vmstats_local = alloc_percpu(struct memcg_vmstats_percpu);
if (!memcg->vmstats_local)
goto fail;
memcg->vmstats_percpu = alloc_percpu(struct memcg_vmstats_percpu);
if (!memcg->vmstats_percpu)
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
goto fail;
for_each_node(node)
if (alloc_mem_cgroup_per_node_info(memcg, node))
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
goto fail;
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
if (memcg_wb_domain_init(memcg, GFP_KERNEL))
goto fail;
INIT_WORK(&memcg->high_work, high_work_func);
INIT_LIST_HEAD(&memcg->oom_notify);
mutex_init(&memcg->thresholds_lock);
spin_lock_init(&memcg->move_lock);
memcg: add memory.pressure_level events With this patch userland applications that want to maintain the interactivity/memory allocation cost can use the pressure level notifications. The levels are defined like this: The "low" level means that the system is reclaiming memory for new allocations. Monitoring this reclaiming activity might be useful for maintaining cache level. Upon notification, the program (typically "Activity Manager") might analyze vmstat and act in advance (i.e. prematurely shutdown unimportant services). The "medium" level means that the system is experiencing medium memory pressure, the system might be making swap, paging out active file caches, etc. Upon this event applications may decide to further analyze vmstat/zoneinfo/memcg or internal memory usage statistics and free any resources that can be easily reconstructed or re-read from a disk. The "critical" level means that the system is actively thrashing, it is about to out of memory (OOM) or even the in-kernel OOM killer is on its way to trigger. Applications should do whatever they can to help the system. It might be too late to consult with vmstat or any other statistics, so it's advisable to take an immediate action. The events are propagated upward until the event is handled, i.e. the events are not pass-through. Here is what this means: for example you have three cgroups: A->B->C. Now you set up an event listener on cgroups A, B and C, and suppose group C experiences some pressure. In this situation, only group C will receive the notification, i.e. groups A and B will not receive it. This is done to avoid excessive "broadcasting" of messages, which disturbs the system and which is especially bad if we are low on memory or thrashing. So, organize the cgroups wisely, or propagate the events manually (or, ask us to implement the pass-through events, explaining why would you need them.) Performance wise, the memory pressure notifications feature itself is lightweight and does not require much of bookkeeping, in contrast to the rest of memcg features. Unfortunately, as of current memcg implementation, pages accounting is an inseparable part and cannot be turned off. The good news is that there are some efforts[1] to improve the situation; plus, implementing the same, fully API-compatible[2] interface for CONFIG_MEMCG=n case (e.g. embedded) is also a viable option, so it will not require any changes on the userland side. [1] http://permalink.gmane.org/gmane.linux.kernel.cgroups/6291 [2] http://lkml.org/lkml/2013/2/21/454 [akpm@linux-foundation.org: coding-style fixes] [akpm@linux-foundation.org: fix CONFIG_CGROPUPS=n warnings] Signed-off-by: Anton Vorontsov <anton.vorontsov@linaro.org> Acked-by: Kirill A. Shutemov <kirill@shutemov.name> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: Mel Gorman <mgorman@suse.de> Cc: Glauber Costa <glommer@parallels.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Luiz Capitulino <lcapitulino@redhat.com> Cc: Greg Thelen <gthelen@google.com> Cc: Leonid Moiseichuk <leonid.moiseichuk@nokia.com> Cc: KOSAKI Motohiro <kosaki.motohiro@gmail.com> Cc: Minchan Kim <minchan@kernel.org> Cc: Bartlomiej Zolnierkiewicz <b.zolnierkie@samsung.com> Cc: John Stultz <john.stultz@linaro.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-04-29 22:08:31 +00:00
vmpressure_init(&memcg->vmpressure);
INIT_LIST_HEAD(&memcg->event_list);
spin_lock_init(&memcg->event_list_lock);
memcg->socket_pressure = jiffies;
mm: introduce CONFIG_MEMCG_KMEM as combination of CONFIG_MEMCG && !CONFIG_SLOB Introduce new config option, which is used to replace repeating CONFIG_MEMCG && !CONFIG_SLOB pattern. Next patches add a little more memcg+kmem related code, so let's keep the defines more clearly. Link: http://lkml.kernel.org/r/153063053670.1818.15013136946600481138.stgit@localhost.localdomain Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Tested-by: Shakeel Butt <shakeelb@google.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Guenter Roeck <linux@roeck-us.net> Cc: "Huang, Ying" <ying.huang@intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Josef Bacik <jbacik@fb.com> Cc: Li RongQing <lirongqing@baidu.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Matthias Kaehlcke <mka@chromium.org> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Michal Hocko <mhocko@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Philippe Ombredanne <pombredanne@nexb.com> Cc: Roman Gushchin <guro@fb.com> Cc: Sahitya Tummala <stummala@codeaurora.org> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Waiman Long <longman@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:25 +00:00
#ifdef CONFIG_MEMCG_KMEM
memcg->kmemcg_id = -1;
#endif
writeback: make backing_dev_info host cgroup-specific bdi_writebacks For the planned cgroup writeback support, on each bdi (backing_dev_info), each memcg will be served by a separate wb (bdi_writeback). This patch updates bdi so that a bdi can host multiple wbs (bdi_writebacks). On the default hierarchy, blkcg implicitly enables memcg. This allows using memcg's page ownership for attributing writeback IOs, and every memcg - blkcg combination can be served by its own wb by assigning a dedicated wb to each memcg. This means that there may be multiple wb's of a bdi mapped to the same blkcg. As congested state is per blkcg - bdi combination, those wb's should share the same congested state. This is achieved by tracking congested state via bdi_writeback_congested structs which are keyed by blkcg. bdi->wb remains unchanged and will keep serving the root cgroup. cgwb's (cgroup wb's) for non-root cgroups are created on-demand or looked up while dirtying an inode according to the memcg of the page being dirtied or current task. Each cgwb is indexed on bdi->cgwb_tree by its memcg id. Once an inode is associated with its wb, it can be retrieved using inode_to_wb(). Currently, none of the filesystems has FS_CGROUP_WRITEBACK and all pages will keep being associated with bdi->wb. v3: inode_attach_wb() in account_page_dirtied() moved inside mapping_cap_account_dirty() block where it's known to be !NULL. Also, an unnecessary NULL check before kfree() removed. Both detected by the kbuild bot. v2: Updated so that wb association is per inode and wb is per memcg rather than blkcg. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: kbuild test robot <fengguang.wu@intel.com> Cc: Dan Carpenter <dan.carpenter@oracle.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-22 21:13:37 +00:00
#ifdef CONFIG_CGROUP_WRITEBACK
INIT_LIST_HEAD(&memcg->cgwb_list);
writeback, memcg: Implement foreign dirty flushing There's an inherent mismatch between memcg and writeback. The former trackes ownership per-page while the latter per-inode. This was a deliberate design decision because honoring per-page ownership in the writeback path is complicated, may lead to higher CPU and IO overheads and deemed unnecessary given that write-sharing an inode across different cgroups isn't a common use-case. Combined with inode majority-writer ownership switching, this works well enough in most cases but there are some pathological cases. For example, let's say there are two cgroups A and B which keep writing to different but confined parts of the same inode. B owns the inode and A's memory is limited far below B's. A's dirty ratio can rise enough to trigger balance_dirty_pages() sleeps but B's can be low enough to avoid triggering background writeback. A will be slowed down without a way to make writeback of the dirty pages happen. This patch implements foreign dirty recording and foreign mechanism so that when a memcg encounters a condition as above it can trigger flushes on bdi_writebacks which can clean its pages. Please see the comment on top of mem_cgroup_track_foreign_dirty_slowpath() for details. A reproducer follows. write-range.c:: #include <stdio.h> #include <stdlib.h> #include <unistd.h> #include <fcntl.h> #include <sys/types.h> static const char *usage = "write-range FILE START SIZE\n"; int main(int argc, char **argv) { int fd; unsigned long start, size, end, pos; char *endp; char buf[4096]; if (argc < 4) { fprintf(stderr, usage); return 1; } fd = open(argv[1], O_WRONLY); if (fd < 0) { perror("open"); return 1; } start = strtoul(argv[2], &endp, 0); if (*endp != '\0') { fprintf(stderr, usage); return 1; } size = strtoul(argv[3], &endp, 0); if (*endp != '\0') { fprintf(stderr, usage); return 1; } end = start + size; while (1) { for (pos = start; pos < end; ) { long bread, bwritten = 0; if (lseek(fd, pos, SEEK_SET) < 0) { perror("lseek"); return 1; } bread = read(0, buf, sizeof(buf) < end - pos ? sizeof(buf) : end - pos); if (bread < 0) { perror("read"); return 1; } if (bread == 0) return 0; while (bwritten < bread) { long this; this = write(fd, buf + bwritten, bread - bwritten); if (this < 0) { perror("write"); return 1; } bwritten += this; pos += bwritten; } } } } repro.sh:: #!/bin/bash set -e set -x sysctl -w vm.dirty_expire_centisecs=300000 sysctl -w vm.dirty_writeback_centisecs=300000 sysctl -w vm.dirtytime_expire_seconds=300000 echo 3 > /proc/sys/vm/drop_caches TEST=/sys/fs/cgroup/test A=$TEST/A B=$TEST/B mkdir -p $A $B echo "+memory +io" > $TEST/cgroup.subtree_control echo $((1<<30)) > $A/memory.high echo $((32<<30)) > $B/memory.high rm -f testfile touch testfile fallocate -l 4G testfile echo "Starting B" (echo $BASHPID > $B/cgroup.procs pv -q --rate-limit 70M < /dev/urandom | ./write-range testfile $((2<<30)) $((2<<30))) & echo "Waiting 10s to ensure B claims the testfile inode" sleep 5 sync sleep 5 sync echo "Starting A" (echo $BASHPID > $A/cgroup.procs pv < /dev/urandom | ./write-range testfile 0 $((2<<30))) v2: Added comments explaining why the specific intervals are being used. v3: Use 0 @nr when calling cgroup_writeback_by_id() to use best-effort flushing while avoding possible livelocks. v4: Use get_jiffies_64() and time_before/after64() instead of raw jiffies_64 and arthimetic comparisons as suggested by Jan. Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-08-26 16:06:56 +00:00
for (i = 0; i < MEMCG_CGWB_FRN_CNT; i++)
memcg->cgwb_frn[i].done =
__WB_COMPLETION_INIT(&memcg_cgwb_frn_waitq);
mm: thp: make deferred split shrinker memcg aware Currently THP deferred split shrinker is not memcg aware, this may cause premature OOM with some configuration. For example the below test would run into premature OOM easily: $ cgcreate -g memory:thp $ echo 4G > /sys/fs/cgroup/memory/thp/memory/limit_in_bytes $ cgexec -g memory:thp transhuge-stress 4000 transhuge-stress comes from kernel selftest. It is easy to hit OOM, but there are still a lot THP on the deferred split queue, memcg direct reclaim can't touch them since the deferred split shrinker is not memcg aware. Convert deferred split shrinker memcg aware by introducing per memcg deferred split queue. The THP should be on either per node or per memcg deferred split queue if it belongs to a memcg. When the page is immigrated to the other memcg, it will be immigrated to the target memcg's deferred split queue too. Reuse the second tail page's deferred_list for per memcg list since the same THP can't be on multiple deferred split queues. [yang.shi@linux.alibaba.com: simplify deferred split queue dereference per Kirill Tkhai] Link: http://lkml.kernel.org/r/1566496227-84952-5-git-send-email-yang.shi@linux.alibaba.com Link: http://lkml.kernel.org/r/1565144277-36240-5-git-send-email-yang.shi@linux.alibaba.com Signed-off-by: Yang Shi <yang.shi@linux.alibaba.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Reviewed-by: Kirill Tkhai <ktkhai@virtuozzo.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: "Kirill A . Shutemov" <kirill.shutemov@linux.intel.com> Cc: Hugh Dickins <hughd@google.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: David Rientjes <rientjes@google.com> Cc: Qian Cai <cai@lca.pw> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-09-23 22:38:15 +00:00
#endif
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
spin_lock_init(&memcg->deferred_split_queue.split_queue_lock);
INIT_LIST_HEAD(&memcg->deferred_split_queue.split_queue);
memcg->deferred_split_queue.split_queue_len = 0;
writeback: make backing_dev_info host cgroup-specific bdi_writebacks For the planned cgroup writeback support, on each bdi (backing_dev_info), each memcg will be served by a separate wb (bdi_writeback). This patch updates bdi so that a bdi can host multiple wbs (bdi_writebacks). On the default hierarchy, blkcg implicitly enables memcg. This allows using memcg's page ownership for attributing writeback IOs, and every memcg - blkcg combination can be served by its own wb by assigning a dedicated wb to each memcg. This means that there may be multiple wb's of a bdi mapped to the same blkcg. As congested state is per blkcg - bdi combination, those wb's should share the same congested state. This is achieved by tracking congested state via bdi_writeback_congested structs which are keyed by blkcg. bdi->wb remains unchanged and will keep serving the root cgroup. cgwb's (cgroup wb's) for non-root cgroups are created on-demand or looked up while dirtying an inode according to the memcg of the page being dirtied or current task. Each cgwb is indexed on bdi->cgwb_tree by its memcg id. Once an inode is associated with its wb, it can be retrieved using inode_to_wb(). Currently, none of the filesystems has FS_CGROUP_WRITEBACK and all pages will keep being associated with bdi->wb. v3: inode_attach_wb() in account_page_dirtied() moved inside mapping_cap_account_dirty() block where it's known to be !NULL. Also, an unnecessary NULL check before kfree() removed. Both detected by the kbuild bot. v2: Updated so that wb association is per inode and wb is per memcg rather than blkcg. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: kbuild test robot <fengguang.wu@intel.com> Cc: Dan Carpenter <dan.carpenter@oracle.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-22 21:13:37 +00:00
#endif
mm: memcontrol: fix cgroup creation failure after many small jobs The memory controller has quite a bit of state that usually outlives the cgroup and pins its CSS until said state disappears. At the same time it imposes a 16-bit limit on the CSS ID space to economically store IDs in the wild. Consequently, when we use cgroups to contain frequent but small and short-lived jobs that leave behind some page cache, we quickly run into the 64k limitations of outstanding CSSs. Creating a new cgroup fails with -ENOSPC while there are only a few, or even no user-visible cgroups in existence. Although pinning CSSs past cgroup removal is common, there are only two instances that actually need an ID after a cgroup is deleted: cache shadow entries and swapout records. Cache shadow entries reference the ID weakly and can deal with the CSS having disappeared when it's looked up later. They pose no hurdle. Swap-out records do need to pin the css to hierarchically attribute swapins after the cgroup has been deleted; though the only pages that remain swapped out after offlining are tmpfs/shmem pages. And those references are under the user's control, so they are manageable. This patch introduces a private 16-bit memcg ID and switches swap and cache shadow entries over to using that. This ID can then be recycled after offlining when the CSS remains pinned only by objects that don't specifically need it. This script demonstrates the problem by faulting one cache page in a new cgroup and deleting it again: set -e mkdir -p pages for x in `seq 128000`; do [ $((x % 1000)) -eq 0 ] && echo $x mkdir /cgroup/foo echo $$ >/cgroup/foo/cgroup.procs echo trex >pages/$x echo $$ >/cgroup/cgroup.procs rmdir /cgroup/foo done When run on an unpatched kernel, we eventually run out of possible IDs even though there are no visible cgroups: [root@ham ~]# ./cssidstress.sh [...] 65000 mkdir: cannot create directory '/cgroup/foo': No space left on device After this patch, the IDs get released upon cgroup destruction and the cache and css objects get released once memory reclaim kicks in. [hannes@cmpxchg.org: init the IDR] Link: http://lkml.kernel.org/r/20160621154601.GA22431@cmpxchg.org Fixes: b2052564e66d ("mm: memcontrol: continue cache reclaim from offlined groups") Link: http://lkml.kernel.org/r/20160617162516.GD19084@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: John Garcia <john.garcia@mesosphere.io> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: Nikolay Borisov <kernel@kyup.com> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-20 22:44:57 +00:00
idr_replace(&mem_cgroup_idr, memcg, memcg->id.id);
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
return memcg;
fail:
mem_cgroup_id_remove(memcg);
mm: do not call mem_cgroup_free() from within mem_cgroup_alloc() mem_cgroup_free() indirectly calls wb_domain_exit() which is not prepared to deal with a struct wb_domain object that hasn't executed wb_domain_init(). For instance, the following warning message is printed by lockdep if alloc_percpu() fails in mem_cgroup_alloc(): INFO: trying to register non-static key. the code is fine but needs lockdep annotation. turning off the locking correctness validator. CPU: 1 PID: 1950 Comm: mkdir Not tainted 4.10.0+ #151 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS Bochs 01/01/2011 Call Trace: dump_stack+0x67/0x99 register_lock_class+0x36d/0x540 __lock_acquire+0x7f/0x1a30 lock_acquire+0xcc/0x200 del_timer_sync+0x3c/0xc0 wb_domain_exit+0x14/0x20 mem_cgroup_free+0x14/0x40 mem_cgroup_css_alloc+0x3f9/0x620 cgroup_apply_control_enable+0x190/0x390 cgroup_mkdir+0x290/0x3d0 kernfs_iop_mkdir+0x58/0x80 vfs_mkdir+0x10e/0x1a0 SyS_mkdirat+0xa8/0xd0 SyS_mkdir+0x14/0x20 entry_SYSCALL_64_fastpath+0x18/0xad Add __mem_cgroup_free() which skips wb_domain_exit(). This is used by both mem_cgroup_free() and mem_cgroup_alloc() clean up. Fixes: 0b8f73e104285 ("mm: memcontrol: clean up alloc, online, offline, free functions") Link: http://lkml.kernel.org/r/20170306192122.24262-1-tahsin@google.com Signed-off-by: Tahsin Erdogan <tahsin@google.com> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-03-10 00:17:26 +00:00
__mem_cgroup_free(memcg);
return ERR_PTR(error);
}
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
static struct cgroup_subsys_state * __ref
mem_cgroup_css_alloc(struct cgroup_subsys_state *parent_css)
{
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
struct mem_cgroup *parent = mem_cgroup_from_css(parent_css);
struct mem_cgroup *memcg;
long error = -ENOMEM;
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
memcg = mem_cgroup_alloc();
if (IS_ERR(memcg))
return ERR_CAST(memcg);
page_counter_set_high(&memcg->memory, PAGE_COUNTER_MAX);
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
memcg->soft_limit = PAGE_COUNTER_MAX;
mm/memcg: automatically penalize tasks with high swap use Add a memory.swap.high knob, which can be used to protect the system from SWAP exhaustion. The mechanism used for penalizing is similar to memory.high penalty (sleep on return to user space). That is not to say that the knob itself is equivalent to memory.high. The objective is more to protect the system from potentially buggy tasks consuming a lot of swap and impacting other tasks, or even bringing the whole system to stand still with complete SWAP exhaustion. Hopefully without the need to find per-task hard limits. Slowing misbehaving tasks down gradually allows user space oom killers or other protection mechanisms to react. oomd and earlyoom already do killing based on swap exhaustion, and memory.swap.high protection will help implement such userspace oom policies more reliably. We can use one counter for number of pages allocated under pressure to save struct task space and avoid two separate hierarchy walks on the hot path. The exact overage is calculated on return to user space, anyway. Take the new high limit into account when determining if swap is "full". Borrowing the explanation from Johannes: The idea behind "swap full" is that as long as the workload has plenty of swap space available and it's not changing its memory contents, it makes sense to generously hold on to copies of data in the swap device, even after the swapin. A later reclaim cycle can drop the page without any IO. Trading disk space for IO. But the only two ways to reclaim a swap slot is when they're faulted in and the references go away, or by scanning the virtual address space like swapoff does - which is very expensive (one could argue it's too expensive even for swapoff, it's often more practical to just reboot). So at some point in the fill level, we have to start freeing up swap slots on fault/swapin. Otherwise we could eventually run out of swap slots while they're filled with copies of data that is also in RAM. We don't want to OOM a workload because its available swap space is filled with redundant cache. Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Chris Down <chris@chrisdown.name> Cc: Shakeel Butt <shakeelb@google.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Hugh Dickins <hughd@google.com> Link: http://lkml.kernel.org/r/20200527195846.102707-5-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:52 +00:00
page_counter_set_high(&memcg->swap, PAGE_COUNTER_MAX);
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
if (parent) {
memcg->swappiness = mem_cgroup_swappiness(parent);
memcg->oom_kill_disable = parent->oom_kill_disable;
}
if (parent && parent->use_hierarchy) {
memcg->use_hierarchy = true;
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
page_counter_init(&memcg->memory, &parent->memory);
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
page_counter_init(&memcg->swap, &parent->swap);
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
page_counter_init(&memcg->memsw, &parent->memsw);
page_counter_init(&memcg->kmem, &parent->kmem);
page_counter_init(&memcg->tcpmem, &parent->tcpmem);
} else {
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
page_counter_init(&memcg->memory, NULL);
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
page_counter_init(&memcg->swap, NULL);
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
page_counter_init(&memcg->memsw, NULL);
page_counter_init(&memcg->kmem, NULL);
page_counter_init(&memcg->tcpmem, NULL);
cgroup: mark subsystems with broken hierarchy support and whine if cgroups are nested for them Currently, cgroup hierarchy support is a mess. cpu related subsystems behave correctly - configuration, accounting and control on a parent properly cover its children. blkio and freezer completely ignore hierarchy and treat all cgroups as if they're directly under the root cgroup. Others show yet different behaviors. These differing interpretations of cgroup hierarchy make using cgroup confusing and it impossible to co-mount controllers into the same hierarchy and obtain sane behavior. Eventually, we want full hierarchy support from all subsystems and probably a unified hierarchy. Users using separate hierarchies expecting completely different behaviors depending on the mounted subsystem is deterimental to making any progress on this front. This patch adds cgroup_subsys.broken_hierarchy and sets it to %true for controllers which are lacking in hierarchy support. The goal of this patch is two-fold. * Move users away from using hierarchy on currently non-hierarchical subsystems, so that implementing proper hierarchy support on those doesn't surprise them. * Keep track of which controllers are broken how and nudge the subsystems to implement proper hierarchy support. For now, start with a single warning message. We can whine louder later on. v2: Fixed a typo spotted by Michal. Warning message updated. v3: Updated memcg part so that it doesn't generate warning in the cases where .use_hierarchy=false doesn't make the behavior different from root.use_hierarchy=true. Fixed a typo spotted by Glauber. v4: Check ->broken_hierarchy after cgroup creation is complete so that ->create() can affect the result per Michal. Dropped unnecessary memcg root handling per Michal. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Serge E. Hallyn <serue@us.ibm.com> Cc: Glauber Costa <glommer@parallels.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Paul Turner <pjt@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Thomas Graf <tgraf@suug.ch> Cc: Vivek Goyal <vgoyal@redhat.com> Cc: Paul Mackerras <paulus@samba.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Arnaldo Carvalho de Melo <acme@ghostprotocols.net> Cc: Neil Horman <nhorman@tuxdriver.com> Cc: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com>
2012-09-13 19:20:58 +00:00
/*
* Deeper hierachy with use_hierarchy == false doesn't make
* much sense so let cgroup subsystem know about this
* unfortunate state in our controller.
*/
if (parent != root_mem_cgroup)
cgroup: clean up cgroup_subsys names and initialization cgroup_subsys is a bit messier than it needs to be. * The name of a subsys can be different from its internal identifier defined in cgroup_subsys.h. Most subsystems use the matching name but three - cpu, memory and perf_event - use different ones. * cgroup_subsys_id enums are postfixed with _subsys_id and each cgroup_subsys is postfixed with _subsys. cgroup.h is widely included throughout various subsystems, it doesn't and shouldn't have claim on such generic names which don't have any qualifier indicating that they belong to cgroup. * cgroup_subsys->subsys_id should always equal the matching cgroup_subsys_id enum; however, we require each controller to initialize it and then BUG if they don't match, which is a bit silly. This patch cleans up cgroup_subsys names and initialization by doing the followings. * cgroup_subsys_id enums are now postfixed with _cgrp_id, and each cgroup_subsys with _cgrp_subsys. * With the above, renaming subsys identifiers to match the userland visible names doesn't cause any naming conflicts. All non-matching identifiers are renamed to match the official names. cpu_cgroup -> cpu mem_cgroup -> memory perf -> perf_event * controllers no longer need to initialize ->subsys_id and ->name. They're generated in cgroup core and set automatically during boot. * Redundant cgroup_subsys declarations removed. * While updating BUG_ON()s in cgroup_init_early(), convert them to WARN()s. BUGging that early during boot is stupid - the kernel can't print anything, even through serial console and the trap handler doesn't even link stack frame properly for back-tracing. This patch doesn't introduce any behavior changes. v2: Rebased on top of fe1217c4f3f7 ("net: net_cls: move cgroupfs classid handling into core"). Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Neil Horman <nhorman@tuxdriver.com> Acked-by: "David S. Miller" <davem@davemloft.net> Acked-by: "Rafael J. Wysocki" <rjw@rjwysocki.net> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Peter Zijlstra <peterz@infradead.org> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Ingo Molnar <mingo@redhat.com> Acked-by: Li Zefan <lizefan@huawei.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Serge E. Hallyn <serue@us.ibm.com> Cc: Vivek Goyal <vgoyal@redhat.com> Cc: Thomas Graf <tgraf@suug.ch>
2014-02-08 15:36:58 +00:00
memory_cgrp_subsys.broken_hierarchy = true;
}
memcg: rework memcg_update_kmem_limit synchronization Currently we take both the memcg_create_mutex and the set_limit_mutex when we enable kmem accounting for a memory cgroup, which makes kmem activation events serialize with both memcg creations and other memcg limit updates (memory.limit, memory.memsw.limit). However, there is no point in such strict synchronization rules there. First, the set_limit_mutex was introduced to keep the memory.limit and memory.memsw.limit values in sync. Since memory.kmem.limit can be set independently of them, it is better to introduce a separate mutex to synchronize against concurrent kmem limit updates. Second, we take the memcg_create_mutex in order to make sure all children of this memcg will be kmem-active as well. For achieving that, it is enough to hold this mutex only while checking if memcg_has_children() though. This guarantees that if a child is added after we checked that the memcg has no children, the newly added cgroup will see its parent kmem-active (of course if the latter succeeded), and call kmem activation for itself. This patch simplifies the locking rules of memcg_update_kmem_limit() according to these considerations. [vdavydov@parallels.com: fix unintialized var warning] Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Michal Hocko <mhocko@suse.cz> Cc: Glauber Costa <glommer@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-01-23 23:53:09 +00:00
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
/* The following stuff does not apply to the root */
if (!parent) {
mm: memcg/slab: reparent memcg kmem_caches on cgroup removal Let's reparent non-root kmem_caches on memcg offlining. This allows us to release the memory cgroup without waiting for the last outstanding kernel object (e.g. dentry used by another application). Since the parent cgroup is already charged, everything we need to do is to splice the list of kmem_caches to the parent's kmem_caches list, swap the memcg pointer, drop the css refcounter for each kmem_cache and adjust the parent's css refcounter. Please, note that kmem_cache->memcg_params.memcg isn't a stable pointer anymore. It's safe to read it under rcu_read_lock(), cgroup_mutex held, or any other way that protects the memory cgroup from being released. We can race with the slab allocation and deallocation paths. It's not a big problem: parent's charge and slab global stats are always correct, and we don't care anymore about the child usage and global stats. The child cgroup is already offline, so we don't use or show it anywhere. Local slab stats (NR_SLAB_RECLAIMABLE and NR_SLAB_UNRECLAIMABLE) aren't used anywhere except count_shadow_nodes(). But even there it won't break anything: after reparenting "nodes" will be 0 on child level (because we're already reparenting shrinker lists), and on parent level page stats always were 0, and this patch won't change anything. [guro@fb.com: properly handle kmem_caches reparented to root_mem_cgroup] Link: http://lkml.kernel.org/r/20190620213427.1691847-1-guro@fb.com Link: http://lkml.kernel.org/r/20190611231813.3148843-11-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Reviewed-by: Shakeel Butt <shakeelb@google.com> Acked-by: David Rientjes <rientjes@google.com> Cc: Christoph Lameter <cl@linux.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Waiman Long <longman@redhat.com> Cc: David Rientjes <rientjes@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Pekka Enberg <penberg@kernel.org> Cc: Andrei Vagin <avagin@gmail.com> Cc: Qian Cai <cai@lca.pw> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:56:34 +00:00
#ifdef CONFIG_MEMCG_KMEM
INIT_LIST_HEAD(&memcg->kmem_caches);
#endif
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
root_mem_cgroup = memcg;
return &memcg->css;
}
mm: memcontrol: enable kmem accounting for all cgroups in the legacy hierarchy Workingset code was recently made memcg aware, but shadow node shrinker is still global. As a result, one small cgroup can consume all memory available for shadow nodes, possibly hurting other cgroups by reclaiming their shadow nodes, even though reclaim distances stored in its shadow nodes have no effect. To avoid this, we need to make shadow node shrinker memcg aware. The actual work is done in patch 6 of the series. Patches 1 and 2 prepare memcg/shrinker infrastructure for the change. Patch 3 is just a collateral cleanup. Patch 4 makes radix_tree_node accounted, which is necessary for making shadow node shrinker memcg aware. Patch 5 reduces shadow nodes overhead in case workload mostly uses anonymous pages. This patch: Currently, in the legacy hierarchy kmem accounting is off for all cgroups by default and must be enabled explicitly by writing something to memory.kmem.limit_in_bytes. Since we don't support reclaim on hitting kmem limit, nor do we have any plans to implement it, this is likely to be -1, just to enable kmem accounting and limit kernel memory consumption by the memory.limit_in_bytes along with user memory. This user API was introduced when the implementation of kmem accounting lacked slab shrinker support and hence was useless in practice. Things have changed since then - slab shrinkers were made memcg aware, the accounting overhead seems to be negligible, and a failure to charge a kmem allocation should not have critical consequences, because we only account those kernel objects that should be safe to fail. That's why kmem accounting is enabled by default for all cgroups in the default hierarchy, which will eventually replace the legacy one. The ability to enable kmem accounting for some cgroups while keeping it disabled for others is getting difficult to maintain. E.g. to make shadow node shrinker memcg aware (see mm/workingset.c), we need to know the relationship between the number of shadow nodes allocated for a cgroup and the size of its lru list. If kmem accounting is enabled for all cgroups there is no problem, but what should we do if kmem accounting is enabled only for half of cgroups? We've no other choice but use global lru stats while scanning root cgroup's shadow nodes, but that would be wrong if kmem accounting was enabled for all cgroups (which is the case if the unified hierarchy is used), in which case we should use lru stats of the root cgroup's lruvec. That being said, let's enable kmem accounting for all memory cgroups by default. If one finds it unstable or too costly, it can always be disabled system-wide by passing cgroup.memory=nokmem to the kernel at boot time. Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-03-17 21:18:27 +00:00
error = memcg_online_kmem(memcg);
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
if (error)
goto fail;
if (cgroup_subsys_on_dfl(memory_cgrp_subsys) && !cgroup_memory_nosocket)
static_branch_inc(&memcg_sockets_enabled_key);
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
return &memcg->css;
fail:
mem_cgroup_id_remove(memcg);
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
mem_cgroup_free(memcg);
return ERR_PTR(error);
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
}
mm: memcontrol: fix cgroup creation failure after many small jobs The memory controller has quite a bit of state that usually outlives the cgroup and pins its CSS until said state disappears. At the same time it imposes a 16-bit limit on the CSS ID space to economically store IDs in the wild. Consequently, when we use cgroups to contain frequent but small and short-lived jobs that leave behind some page cache, we quickly run into the 64k limitations of outstanding CSSs. Creating a new cgroup fails with -ENOSPC while there are only a few, or even no user-visible cgroups in existence. Although pinning CSSs past cgroup removal is common, there are only two instances that actually need an ID after a cgroup is deleted: cache shadow entries and swapout records. Cache shadow entries reference the ID weakly and can deal with the CSS having disappeared when it's looked up later. They pose no hurdle. Swap-out records do need to pin the css to hierarchically attribute swapins after the cgroup has been deleted; though the only pages that remain swapped out after offlining are tmpfs/shmem pages. And those references are under the user's control, so they are manageable. This patch introduces a private 16-bit memcg ID and switches swap and cache shadow entries over to using that. This ID can then be recycled after offlining when the CSS remains pinned only by objects that don't specifically need it. This script demonstrates the problem by faulting one cache page in a new cgroup and deleting it again: set -e mkdir -p pages for x in `seq 128000`; do [ $((x % 1000)) -eq 0 ] && echo $x mkdir /cgroup/foo echo $$ >/cgroup/foo/cgroup.procs echo trex >pages/$x echo $$ >/cgroup/cgroup.procs rmdir /cgroup/foo done When run on an unpatched kernel, we eventually run out of possible IDs even though there are no visible cgroups: [root@ham ~]# ./cssidstress.sh [...] 65000 mkdir: cannot create directory '/cgroup/foo': No space left on device After this patch, the IDs get released upon cgroup destruction and the cache and css objects get released once memory reclaim kicks in. [hannes@cmpxchg.org: init the IDR] Link: http://lkml.kernel.org/r/20160621154601.GA22431@cmpxchg.org Fixes: b2052564e66d ("mm: memcontrol: continue cache reclaim from offlined groups") Link: http://lkml.kernel.org/r/20160617162516.GD19084@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: John Garcia <john.garcia@mesosphere.io> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: Nikolay Borisov <kernel@kyup.com> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-20 22:44:57 +00:00
static int mem_cgroup_css_online(struct cgroup_subsys_state *css)
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
{
struct mem_cgroup *memcg = mem_cgroup_from_css(css);
mm, memcg: assign memcg-aware shrinkers bitmap to memcg Imagine a big node with many cpus, memory cgroups and containers. Let we have 200 containers, every container has 10 mounts, and 10 cgroups. All container tasks don't touch foreign containers mounts. If there is intensive pages write, and global reclaim happens, a writing task has to iterate over all memcgs to shrink slab, before it's able to go to shrink_page_list(). Iteration over all the memcg slabs is very expensive: the task has to visit 200 * 10 = 2000 shrinkers for every memcg, and since there are 2000 memcgs, the total calls are 2000 * 2000 = 4000000. So, the shrinker makes 4 million do_shrink_slab() calls just to try to isolate SWAP_CLUSTER_MAX pages in one of the actively writing memcg via shrink_page_list(). I've observed a node spending almost 100% in kernel, making useless iteration over already shrinked slab. This patch adds bitmap of memcg-aware shrinkers to memcg. The size of the bitmap depends on bitmap_nr_ids, and during memcg life it's maintained to be enough to fit bitmap_nr_ids shrinkers. Every bit in the map is related to corresponding shrinker id. Next patches will maintain set bit only for really charged memcg. This will allow shrink_slab() to increase its performance in significant way. See the last patch for the numbers. [ktkhai@virtuozzo.com: v9] Link: http://lkml.kernel.org/r/153112549031.4097.3576147070498769979.stgit@localhost.localdomain [ktkhai@virtuozzo.com: add comment to mem_cgroup_css_online()] Link: http://lkml.kernel.org/r/521f9e5f-c436-b388-fe83-4dc870bfb489@virtuozzo.com Link: http://lkml.kernel.org/r/153063056619.1818.12550500883688681076.stgit@localhost.localdomain Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Tested-by: Shakeel Butt <shakeelb@google.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Guenter Roeck <linux@roeck-us.net> Cc: "Huang, Ying" <ying.huang@intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Josef Bacik <jbacik@fb.com> Cc: Li RongQing <lirongqing@baidu.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Matthias Kaehlcke <mka@chromium.org> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Michal Hocko <mhocko@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Philippe Ombredanne <pombredanne@nexb.com> Cc: Roman Gushchin <guro@fb.com> Cc: Sahitya Tummala <stummala@codeaurora.org> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Waiman Long <longman@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:37 +00:00
/*
* A memcg must be visible for memcg_expand_shrinker_maps()
* by the time the maps are allocated. So, we allocate maps
* here, when for_each_mem_cgroup() can't skip it.
*/
if (memcg_alloc_shrinker_maps(memcg)) {
mem_cgroup_id_remove(memcg);
return -ENOMEM;
}
mm: memcontrol: fix cgroup creation failure after many small jobs The memory controller has quite a bit of state that usually outlives the cgroup and pins its CSS until said state disappears. At the same time it imposes a 16-bit limit on the CSS ID space to economically store IDs in the wild. Consequently, when we use cgroups to contain frequent but small and short-lived jobs that leave behind some page cache, we quickly run into the 64k limitations of outstanding CSSs. Creating a new cgroup fails with -ENOSPC while there are only a few, or even no user-visible cgroups in existence. Although pinning CSSs past cgroup removal is common, there are only two instances that actually need an ID after a cgroup is deleted: cache shadow entries and swapout records. Cache shadow entries reference the ID weakly and can deal with the CSS having disappeared when it's looked up later. They pose no hurdle. Swap-out records do need to pin the css to hierarchically attribute swapins after the cgroup has been deleted; though the only pages that remain swapped out after offlining are tmpfs/shmem pages. And those references are under the user's control, so they are manageable. This patch introduces a private 16-bit memcg ID and switches swap and cache shadow entries over to using that. This ID can then be recycled after offlining when the CSS remains pinned only by objects that don't specifically need it. This script demonstrates the problem by faulting one cache page in a new cgroup and deleting it again: set -e mkdir -p pages for x in `seq 128000`; do [ $((x % 1000)) -eq 0 ] && echo $x mkdir /cgroup/foo echo $$ >/cgroup/foo/cgroup.procs echo trex >pages/$x echo $$ >/cgroup/cgroup.procs rmdir /cgroup/foo done When run on an unpatched kernel, we eventually run out of possible IDs even though there are no visible cgroups: [root@ham ~]# ./cssidstress.sh [...] 65000 mkdir: cannot create directory '/cgroup/foo': No space left on device After this patch, the IDs get released upon cgroup destruction and the cache and css objects get released once memory reclaim kicks in. [hannes@cmpxchg.org: init the IDR] Link: http://lkml.kernel.org/r/20160621154601.GA22431@cmpxchg.org Fixes: b2052564e66d ("mm: memcontrol: continue cache reclaim from offlined groups") Link: http://lkml.kernel.org/r/20160617162516.GD19084@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: John Garcia <john.garcia@mesosphere.io> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: Nikolay Borisov <kernel@kyup.com> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-20 22:44:57 +00:00
/* Online state pins memcg ID, memcg ID pins CSS */
refcount_set(&memcg->id.ref, 1);
mm: memcontrol: fix cgroup creation failure after many small jobs The memory controller has quite a bit of state that usually outlives the cgroup and pins its CSS until said state disappears. At the same time it imposes a 16-bit limit on the CSS ID space to economically store IDs in the wild. Consequently, when we use cgroups to contain frequent but small and short-lived jobs that leave behind some page cache, we quickly run into the 64k limitations of outstanding CSSs. Creating a new cgroup fails with -ENOSPC while there are only a few, or even no user-visible cgroups in existence. Although pinning CSSs past cgroup removal is common, there are only two instances that actually need an ID after a cgroup is deleted: cache shadow entries and swapout records. Cache shadow entries reference the ID weakly and can deal with the CSS having disappeared when it's looked up later. They pose no hurdle. Swap-out records do need to pin the css to hierarchically attribute swapins after the cgroup has been deleted; though the only pages that remain swapped out after offlining are tmpfs/shmem pages. And those references are under the user's control, so they are manageable. This patch introduces a private 16-bit memcg ID and switches swap and cache shadow entries over to using that. This ID can then be recycled after offlining when the CSS remains pinned only by objects that don't specifically need it. This script demonstrates the problem by faulting one cache page in a new cgroup and deleting it again: set -e mkdir -p pages for x in `seq 128000`; do [ $((x % 1000)) -eq 0 ] && echo $x mkdir /cgroup/foo echo $$ >/cgroup/foo/cgroup.procs echo trex >pages/$x echo $$ >/cgroup/cgroup.procs rmdir /cgroup/foo done When run on an unpatched kernel, we eventually run out of possible IDs even though there are no visible cgroups: [root@ham ~]# ./cssidstress.sh [...] 65000 mkdir: cannot create directory '/cgroup/foo': No space left on device After this patch, the IDs get released upon cgroup destruction and the cache and css objects get released once memory reclaim kicks in. [hannes@cmpxchg.org: init the IDR] Link: http://lkml.kernel.org/r/20160621154601.GA22431@cmpxchg.org Fixes: b2052564e66d ("mm: memcontrol: continue cache reclaim from offlined groups") Link: http://lkml.kernel.org/r/20160617162516.GD19084@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: John Garcia <john.garcia@mesosphere.io> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: Nikolay Borisov <kernel@kyup.com> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-20 22:44:57 +00:00
css_get(css);
return 0;
}
cgroup: pass around cgroup_subsys_state instead of cgroup in subsystem methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup * in subsystem implementations for the following reasons. * With unified hierarchy, subsystems will be dynamically bound and unbound from cgroups and thus css's (cgroup_subsys_state) may be created and destroyed dynamically over the lifetime of a cgroup, which is different from the current state where all css's are allocated and destroyed together with the associated cgroup. This in turn means that cgroup_css() should be synchronized and may return NULL, making it more cumbersome to use. * Differing levels of per-subsystem granularity in the unified hierarchy means that the task and descendant iterators should behave differently depending on the specific subsystem the iteration is being performed for. * In majority of the cases, subsystems only care about its part in the cgroup hierarchy - ie. the hierarchy of css's. Subsystem methods often obtain the matching css pointer from the cgroup and don't bother with the cgroup pointer itself. Passing around css fits much better. This patch converts all cgroup_subsys methods to take @css instead of @cgroup. The conversions are mostly straight-forward. A few noteworthy changes are * ->css_alloc() now takes css of the parent cgroup rather than the pointer to the new cgroup as the css for the new cgroup doesn't exist yet. Knowing the parent css is enough for all the existing subsystems. * In kernel/cgroup.c::offline_css(), unnecessary open coded css dereference is replaced with local variable access. This patch shouldn't cause any behavior differences. v2: Unnecessary explicit cgrp->subsys[] deref in css_online() replaced with local variable @css as suggested by Li Zefan. Rebased on top of new for-3.12 which includes for-3.11-fixes so that ->css_free() invocation added by da0a12caff ("cgroup: fix a leak when percpu_ref_init() fails") is converted too. Suggested by Li Zefan. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:23 +00:00
static void mem_cgroup_css_offline(struct cgroup_subsys_state *css)
{
cgroup: pass around cgroup_subsys_state instead of cgroup in subsystem methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup * in subsystem implementations for the following reasons. * With unified hierarchy, subsystems will be dynamically bound and unbound from cgroups and thus css's (cgroup_subsys_state) may be created and destroyed dynamically over the lifetime of a cgroup, which is different from the current state where all css's are allocated and destroyed together with the associated cgroup. This in turn means that cgroup_css() should be synchronized and may return NULL, making it more cumbersome to use. * Differing levels of per-subsystem granularity in the unified hierarchy means that the task and descendant iterators should behave differently depending on the specific subsystem the iteration is being performed for. * In majority of the cases, subsystems only care about its part in the cgroup hierarchy - ie. the hierarchy of css's. Subsystem methods often obtain the matching css pointer from the cgroup and don't bother with the cgroup pointer itself. Passing around css fits much better. This patch converts all cgroup_subsys methods to take @css instead of @cgroup. The conversions are mostly straight-forward. A few noteworthy changes are * ->css_alloc() now takes css of the parent cgroup rather than the pointer to the new cgroup as the css for the new cgroup doesn't exist yet. Knowing the parent css is enough for all the existing subsystems. * In kernel/cgroup.c::offline_css(), unnecessary open coded css dereference is replaced with local variable access. This patch shouldn't cause any behavior differences. v2: Unnecessary explicit cgrp->subsys[] deref in css_online() replaced with local variable @css as suggested by Li Zefan. Rebased on top of new for-3.12 which includes for-3.11-fixes so that ->css_free() invocation added by da0a12caff ("cgroup: fix a leak when percpu_ref_init() fails") is converted too. Suggested by Li Zefan. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:23 +00:00
struct mem_cgroup *memcg = mem_cgroup_from_css(css);
struct mem_cgroup_event *event, *tmp;
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
/*
* Unregister events and notify userspace.
* Notify userspace about cgroup removing only after rmdir of cgroup
* directory to avoid race between userspace and kernelspace.
*/
spin_lock(&memcg->event_list_lock);
list_for_each_entry_safe(event, tmp, &memcg->event_list, list) {
cgroup, memcg: move cgroup_event implementation to memcg cgroup_event is way over-designed and tries to build a generic flexible event mechanism into cgroup - fully customizable event specification for each user of the interface. This is utterly unnecessary and overboard especially in the light of the planned unified hierarchy as there's gonna be single agent. Simply generating events at fixed points, or if that's too restrictive, configureable cadence or single set of configureable points should be enough. Thankfully, memcg is the only user and gets to keep it. Replacing it with something simpler on sane_behavior is strongly recommended. This patch moves cgroup_event and "cgroup.event_control" implementation to mm/memcontrol.c. Clearing of events on cgroup destruction is moved from cgroup_destroy_locked() to mem_cgroup_css_offline(), which shouldn't make any noticeable difference. cgroup_css() and __file_cft() are exported to enable the move; however, this will soon be reverted once the event code is updated to be memcg specific. Note that "cgroup.event_control" will now exist only on the hierarchy with memcg attached to it. While this change is visible to userland, it is unlikely to be noticeable as the file has never been meaningful outside memcg. Aside from the above change, this is pure code relocation. v2: Per Li Zefan's comments, init/Kconfig updated accordingly and poll.h inclusion moved from cgroup.c to memcontrol.c. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com>
2013-11-22 23:20:42 +00:00
list_del_init(&event->list);
schedule_work(&event->remove);
}
spin_unlock(&memcg->event_list_lock);
memcg: introduce memory.min Memory controller implements the memory.low best-effort memory protection mechanism, which works perfectly in many cases and allows protecting working sets of important workloads from sudden reclaim. But its semantics has a significant limitation: it works only as long as there is a supply of reclaimable memory. This makes it pretty useless against any sort of slow memory leaks or memory usage increases. This is especially true for swapless systems. If swap is enabled, memory soft protection effectively postpones problems, allowing a leaking application to fill all swap area, which makes no sense. The only effective way to guarantee the memory protection in this case is to invoke the OOM killer. It's possible to handle this case in userspace by reacting on MEMCG_LOW events; but there is still a place for a fail-safe in-kernel mechanism to provide stronger guarantees. This patch introduces the memory.min interface for cgroup v2 memory controller. It works very similarly to memory.low (sharing the same hierarchical behavior), except that it's not disabled if there is no more reclaimable memory in the system. If cgroup is not populated, its memory.min is ignored, because otherwise even the OOM killer wouldn't be able to reclaim the protected memory, and the system can stall. [guro@fb.com: s/low/min/ in docs] Link: http://lkml.kernel.org/r/20180510130758.GA9129@castle.DHCP.thefacebook.com Link: http://lkml.kernel.org/r/20180509180734.GA4856@castle.DHCP.thefacebook.com Signed-off-by: Roman Gushchin <guro@fb.com> Reviewed-by: Randy Dunlap <rdunlap@infradead.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:07:46 +00:00
page_counter_set_min(&memcg->memory, 0);
mm: memory.low hierarchical behavior This patch aims to address an issue in current memory.low semantics, which makes it hard to use it in a hierarchy, where some leaf memory cgroups are more valuable than others. For example, there are memcgs A, A/B, A/C, A/D and A/E: A A/memory.low = 2G, A/memory.current = 6G //\\ BC DE B/memory.low = 3G B/memory.current = 2G C/memory.low = 1G C/memory.current = 2G D/memory.low = 0 D/memory.current = 2G E/memory.low = 10G E/memory.current = 0 If we apply memory pressure, B, C and D are reclaimed at the same pace while A's usage exceeds 2G. This is obviously wrong, as B's usage is fully below B's memory.low, and C has 1G of protection as well. Also, A is pushed to the size, which is less than A's 2G memory.low, which is also wrong. A simple bash script (provided below) can be used to reproduce the problem. Current results are: A: 1430097920 A/B: 711929856 A/C: 717426688 A/D: 741376 A/E: 0 To address the issue a concept of effective memory.low is introduced. Effective memory.low is always equal or less than original memory.low. In a case, when there is no memory.low overcommittment (and also for top-level cgroups), these two values are equal. Otherwise it's a part of parent's effective memory.low, calculated as a cgroup's memory.low usage divided by sum of sibling's memory.low usages (under memory.low usage I mean the size of actually protected memory: memory.current if memory.current < memory.low, 0 otherwise). It's necessary to track the actual usage, because otherwise an empty cgroup with memory.low set (A/E in my example) will affect actual memory distribution, which makes no sense. To avoid traversing the cgroup tree twice, page_counters code is reused. Calculating effective memory.low can be done in the reclaim path, as we conveniently traversing the cgroup tree from top to bottom and check memory.low on each level. So, it's a perfect place to calculate effective memory low and save it to use it for children cgroups. This also eliminates a need to traverse the cgroup tree from bottom to top each time to check if parent's guarantee is not exceeded. Setting/resetting effective memory.low is intentionally racy, but it's fine and shouldn't lead to any significant differences in actual memory distribution. With this patch applied results are matching the expectations: A: 2147930112 A/B: 1428721664 A/C: 718393344 A/D: 815104 A/E: 0 Test script: #!/bin/bash CGPATH="/sys/fs/cgroup" truncate /file1 --size 2G truncate /file2 --size 2G truncate /file3 --size 2G truncate /file4 --size 50G mkdir "${CGPATH}/A" echo "+memory" > "${CGPATH}/A/cgroup.subtree_control" mkdir "${CGPATH}/A/B" "${CGPATH}/A/C" "${CGPATH}/A/D" "${CGPATH}/A/E" echo 2G > "${CGPATH}/A/memory.low" echo 3G > "${CGPATH}/A/B/memory.low" echo 1G > "${CGPATH}/A/C/memory.low" echo 0 > "${CGPATH}/A/D/memory.low" echo 10G > "${CGPATH}/A/E/memory.low" echo $$ > "${CGPATH}/A/B/cgroup.procs" && vmtouch -qt /file1 echo $$ > "${CGPATH}/A/C/cgroup.procs" && vmtouch -qt /file2 echo $$ > "${CGPATH}/A/D/cgroup.procs" && vmtouch -qt /file3 echo $$ > "${CGPATH}/cgroup.procs" && vmtouch -qt /file4 echo "A: " `cat "${CGPATH}/A/memory.current"` echo "A/B: " `cat "${CGPATH}/A/B/memory.current"` echo "A/C: " `cat "${CGPATH}/A/C/memory.current"` echo "A/D: " `cat "${CGPATH}/A/D/memory.current"` echo "A/E: " `cat "${CGPATH}/A/E/memory.current"` rmdir "${CGPATH}/A/B" "${CGPATH}/A/C" "${CGPATH}/A/D" "${CGPATH}/A/E" rmdir "${CGPATH}/A" rm /file1 /file2 /file3 /file4 Link: http://lkml.kernel.org/r/20180405185921.4942-2-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:06:22 +00:00
page_counter_set_low(&memcg->memory, 0);
memcg_offline_kmem(memcg);
writeback: make backing_dev_info host cgroup-specific bdi_writebacks For the planned cgroup writeback support, on each bdi (backing_dev_info), each memcg will be served by a separate wb (bdi_writeback). This patch updates bdi so that a bdi can host multiple wbs (bdi_writebacks). On the default hierarchy, blkcg implicitly enables memcg. This allows using memcg's page ownership for attributing writeback IOs, and every memcg - blkcg combination can be served by its own wb by assigning a dedicated wb to each memcg. This means that there may be multiple wb's of a bdi mapped to the same blkcg. As congested state is per blkcg - bdi combination, those wb's should share the same congested state. This is achieved by tracking congested state via bdi_writeback_congested structs which are keyed by blkcg. bdi->wb remains unchanged and will keep serving the root cgroup. cgwb's (cgroup wb's) for non-root cgroups are created on-demand or looked up while dirtying an inode according to the memcg of the page being dirtied or current task. Each cgwb is indexed on bdi->cgwb_tree by its memcg id. Once an inode is associated with its wb, it can be retrieved using inode_to_wb(). Currently, none of the filesystems has FS_CGROUP_WRITEBACK and all pages will keep being associated with bdi->wb. v3: inode_attach_wb() in account_page_dirtied() moved inside mapping_cap_account_dirty() block where it's known to be !NULL. Also, an unnecessary NULL check before kfree() removed. Both detected by the kbuild bot. v2: Updated so that wb association is per inode and wb is per memcg rather than blkcg. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: kbuild test robot <fengguang.wu@intel.com> Cc: Dan Carpenter <dan.carpenter@oracle.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Jan Kara <jack@suse.cz> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-22 21:13:37 +00:00
wb_memcg_offline(memcg);
mm: memcontrol: fix cgroup creation failure after many small jobs The memory controller has quite a bit of state that usually outlives the cgroup and pins its CSS until said state disappears. At the same time it imposes a 16-bit limit on the CSS ID space to economically store IDs in the wild. Consequently, when we use cgroups to contain frequent but small and short-lived jobs that leave behind some page cache, we quickly run into the 64k limitations of outstanding CSSs. Creating a new cgroup fails with -ENOSPC while there are only a few, or even no user-visible cgroups in existence. Although pinning CSSs past cgroup removal is common, there are only two instances that actually need an ID after a cgroup is deleted: cache shadow entries and swapout records. Cache shadow entries reference the ID weakly and can deal with the CSS having disappeared when it's looked up later. They pose no hurdle. Swap-out records do need to pin the css to hierarchically attribute swapins after the cgroup has been deleted; though the only pages that remain swapped out after offlining are tmpfs/shmem pages. And those references are under the user's control, so they are manageable. This patch introduces a private 16-bit memcg ID and switches swap and cache shadow entries over to using that. This ID can then be recycled after offlining when the CSS remains pinned only by objects that don't specifically need it. This script demonstrates the problem by faulting one cache page in a new cgroup and deleting it again: set -e mkdir -p pages for x in `seq 128000`; do [ $((x % 1000)) -eq 0 ] && echo $x mkdir /cgroup/foo echo $$ >/cgroup/foo/cgroup.procs echo trex >pages/$x echo $$ >/cgroup/cgroup.procs rmdir /cgroup/foo done When run on an unpatched kernel, we eventually run out of possible IDs even though there are no visible cgroups: [root@ham ~]# ./cssidstress.sh [...] 65000 mkdir: cannot create directory '/cgroup/foo': No space left on device After this patch, the IDs get released upon cgroup destruction and the cache and css objects get released once memory reclaim kicks in. [hannes@cmpxchg.org: init the IDR] Link: http://lkml.kernel.org/r/20160621154601.GA22431@cmpxchg.org Fixes: b2052564e66d ("mm: memcontrol: continue cache reclaim from offlined groups") Link: http://lkml.kernel.org/r/20160617162516.GD19084@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: John Garcia <john.garcia@mesosphere.io> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: Nikolay Borisov <kernel@kyup.com> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-20 22:44:57 +00:00
drain_all_stock(memcg);
mm: memcontrol: fix cgroup creation failure after many small jobs The memory controller has quite a bit of state that usually outlives the cgroup and pins its CSS until said state disappears. At the same time it imposes a 16-bit limit on the CSS ID space to economically store IDs in the wild. Consequently, when we use cgroups to contain frequent but small and short-lived jobs that leave behind some page cache, we quickly run into the 64k limitations of outstanding CSSs. Creating a new cgroup fails with -ENOSPC while there are only a few, or even no user-visible cgroups in existence. Although pinning CSSs past cgroup removal is common, there are only two instances that actually need an ID after a cgroup is deleted: cache shadow entries and swapout records. Cache shadow entries reference the ID weakly and can deal with the CSS having disappeared when it's looked up later. They pose no hurdle. Swap-out records do need to pin the css to hierarchically attribute swapins after the cgroup has been deleted; though the only pages that remain swapped out after offlining are tmpfs/shmem pages. And those references are under the user's control, so they are manageable. This patch introduces a private 16-bit memcg ID and switches swap and cache shadow entries over to using that. This ID can then be recycled after offlining when the CSS remains pinned only by objects that don't specifically need it. This script demonstrates the problem by faulting one cache page in a new cgroup and deleting it again: set -e mkdir -p pages for x in `seq 128000`; do [ $((x % 1000)) -eq 0 ] && echo $x mkdir /cgroup/foo echo $$ >/cgroup/foo/cgroup.procs echo trex >pages/$x echo $$ >/cgroup/cgroup.procs rmdir /cgroup/foo done When run on an unpatched kernel, we eventually run out of possible IDs even though there are no visible cgroups: [root@ham ~]# ./cssidstress.sh [...] 65000 mkdir: cannot create directory '/cgroup/foo': No space left on device After this patch, the IDs get released upon cgroup destruction and the cache and css objects get released once memory reclaim kicks in. [hannes@cmpxchg.org: init the IDR] Link: http://lkml.kernel.org/r/20160621154601.GA22431@cmpxchg.org Fixes: b2052564e66d ("mm: memcontrol: continue cache reclaim from offlined groups") Link: http://lkml.kernel.org/r/20160617162516.GD19084@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: John Garcia <john.garcia@mesosphere.io> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: Nikolay Borisov <kernel@kyup.com> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-20 22:44:57 +00:00
mem_cgroup_id_put(memcg);
}
mm: memcontrol: fix possible memcg leak due to interrupted reclaim Memory cgroup reclaim can be interrupted with mem_cgroup_iter_break() once enough pages have been reclaimed, in which case, in contrast to a full round-trip over a cgroup sub-tree, the current position stored in mem_cgroup_reclaim_iter of the target cgroup does not get invalidated and so is left holding the reference to the last scanned cgroup. If the target cgroup does not get scanned again (we might have just reclaimed the last page or all processes might exit and free their memory voluntary), we will leak it, because there is nobody to put the reference held by the iterator. The problem is easy to reproduce by running the following command sequence in a loop: mkdir /sys/fs/cgroup/memory/test echo 100M > /sys/fs/cgroup/memory/test/memory.limit_in_bytes echo $$ > /sys/fs/cgroup/memory/test/cgroup.procs memhog 150M echo $$ > /sys/fs/cgroup/memory/cgroup.procs rmdir test The cgroups generated by it will never get freed. This patch fixes this issue by making mem_cgroup_iter avoid taking reference to the current position. In order not to hit use-after-free bug while running reclaim in parallel with cgroup deletion, we make use of ->css_released cgroup callback to clear references to the dying cgroup in all reclaim iterators that might refer to it. This callback is called right before scheduling rcu work which will free css, so if we access iter->position from rcu read section, we might be sure it won't go away under us. [hannes@cmpxchg.org: clean up css ref handling] Fixes: 5ac8fb31ad2e ("mm: memcontrol: convert reclaim iterator to simple css refcounting") Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-12-29 22:54:10 +00:00
static void mem_cgroup_css_released(struct cgroup_subsys_state *css)
{
struct mem_cgroup *memcg = mem_cgroup_from_css(css);
invalidate_reclaim_iterators(memcg);
}
cgroup: pass around cgroup_subsys_state instead of cgroup in subsystem methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup * in subsystem implementations for the following reasons. * With unified hierarchy, subsystems will be dynamically bound and unbound from cgroups and thus css's (cgroup_subsys_state) may be created and destroyed dynamically over the lifetime of a cgroup, which is different from the current state where all css's are allocated and destroyed together with the associated cgroup. This in turn means that cgroup_css() should be synchronized and may return NULL, making it more cumbersome to use. * Differing levels of per-subsystem granularity in the unified hierarchy means that the task and descendant iterators should behave differently depending on the specific subsystem the iteration is being performed for. * In majority of the cases, subsystems only care about its part in the cgroup hierarchy - ie. the hierarchy of css's. Subsystem methods often obtain the matching css pointer from the cgroup and don't bother with the cgroup pointer itself. Passing around css fits much better. This patch converts all cgroup_subsys methods to take @css instead of @cgroup. The conversions are mostly straight-forward. A few noteworthy changes are * ->css_alloc() now takes css of the parent cgroup rather than the pointer to the new cgroup as the css for the new cgroup doesn't exist yet. Knowing the parent css is enough for all the existing subsystems. * In kernel/cgroup.c::offline_css(), unnecessary open coded css dereference is replaced with local variable access. This patch shouldn't cause any behavior differences. v2: Unnecessary explicit cgrp->subsys[] deref in css_online() replaced with local variable @css as suggested by Li Zefan. Rebased on top of new for-3.12 which includes for-3.11-fixes so that ->css_free() invocation added by da0a12caff ("cgroup: fix a leak when percpu_ref_init() fails") is converted too. Suggested by Li Zefan. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:23 +00:00
static void mem_cgroup_css_free(struct cgroup_subsys_state *css)
{
cgroup: pass around cgroup_subsys_state instead of cgroup in subsystem methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup * in subsystem implementations for the following reasons. * With unified hierarchy, subsystems will be dynamically bound and unbound from cgroups and thus css's (cgroup_subsys_state) may be created and destroyed dynamically over the lifetime of a cgroup, which is different from the current state where all css's are allocated and destroyed together with the associated cgroup. This in turn means that cgroup_css() should be synchronized and may return NULL, making it more cumbersome to use. * Differing levels of per-subsystem granularity in the unified hierarchy means that the task and descendant iterators should behave differently depending on the specific subsystem the iteration is being performed for. * In majority of the cases, subsystems only care about its part in the cgroup hierarchy - ie. the hierarchy of css's. Subsystem methods often obtain the matching css pointer from the cgroup and don't bother with the cgroup pointer itself. Passing around css fits much better. This patch converts all cgroup_subsys methods to take @css instead of @cgroup. The conversions are mostly straight-forward. A few noteworthy changes are * ->css_alloc() now takes css of the parent cgroup rather than the pointer to the new cgroup as the css for the new cgroup doesn't exist yet. Knowing the parent css is enough for all the existing subsystems. * In kernel/cgroup.c::offline_css(), unnecessary open coded css dereference is replaced with local variable access. This patch shouldn't cause any behavior differences. v2: Unnecessary explicit cgrp->subsys[] deref in css_online() replaced with local variable @css as suggested by Li Zefan. Rebased on top of new for-3.12 which includes for-3.11-fixes so that ->css_free() invocation added by da0a12caff ("cgroup: fix a leak when percpu_ref_init() fails") is converted too. Suggested by Li Zefan. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:23 +00:00
struct mem_cgroup *memcg = mem_cgroup_from_css(css);
writeback, memcg: Implement foreign dirty flushing There's an inherent mismatch between memcg and writeback. The former trackes ownership per-page while the latter per-inode. This was a deliberate design decision because honoring per-page ownership in the writeback path is complicated, may lead to higher CPU and IO overheads and deemed unnecessary given that write-sharing an inode across different cgroups isn't a common use-case. Combined with inode majority-writer ownership switching, this works well enough in most cases but there are some pathological cases. For example, let's say there are two cgroups A and B which keep writing to different but confined parts of the same inode. B owns the inode and A's memory is limited far below B's. A's dirty ratio can rise enough to trigger balance_dirty_pages() sleeps but B's can be low enough to avoid triggering background writeback. A will be slowed down without a way to make writeback of the dirty pages happen. This patch implements foreign dirty recording and foreign mechanism so that when a memcg encounters a condition as above it can trigger flushes on bdi_writebacks which can clean its pages. Please see the comment on top of mem_cgroup_track_foreign_dirty_slowpath() for details. A reproducer follows. write-range.c:: #include <stdio.h> #include <stdlib.h> #include <unistd.h> #include <fcntl.h> #include <sys/types.h> static const char *usage = "write-range FILE START SIZE\n"; int main(int argc, char **argv) { int fd; unsigned long start, size, end, pos; char *endp; char buf[4096]; if (argc < 4) { fprintf(stderr, usage); return 1; } fd = open(argv[1], O_WRONLY); if (fd < 0) { perror("open"); return 1; } start = strtoul(argv[2], &endp, 0); if (*endp != '\0') { fprintf(stderr, usage); return 1; } size = strtoul(argv[3], &endp, 0); if (*endp != '\0') { fprintf(stderr, usage); return 1; } end = start + size; while (1) { for (pos = start; pos < end; ) { long bread, bwritten = 0; if (lseek(fd, pos, SEEK_SET) < 0) { perror("lseek"); return 1; } bread = read(0, buf, sizeof(buf) < end - pos ? sizeof(buf) : end - pos); if (bread < 0) { perror("read"); return 1; } if (bread == 0) return 0; while (bwritten < bread) { long this; this = write(fd, buf + bwritten, bread - bwritten); if (this < 0) { perror("write"); return 1; } bwritten += this; pos += bwritten; } } } } repro.sh:: #!/bin/bash set -e set -x sysctl -w vm.dirty_expire_centisecs=300000 sysctl -w vm.dirty_writeback_centisecs=300000 sysctl -w vm.dirtytime_expire_seconds=300000 echo 3 > /proc/sys/vm/drop_caches TEST=/sys/fs/cgroup/test A=$TEST/A B=$TEST/B mkdir -p $A $B echo "+memory +io" > $TEST/cgroup.subtree_control echo $((1<<30)) > $A/memory.high echo $((32<<30)) > $B/memory.high rm -f testfile touch testfile fallocate -l 4G testfile echo "Starting B" (echo $BASHPID > $B/cgroup.procs pv -q --rate-limit 70M < /dev/urandom | ./write-range testfile $((2<<30)) $((2<<30))) & echo "Waiting 10s to ensure B claims the testfile inode" sleep 5 sync sleep 5 sync echo "Starting A" (echo $BASHPID > $A/cgroup.procs pv < /dev/urandom | ./write-range testfile 0 $((2<<30))) v2: Added comments explaining why the specific intervals are being used. v3: Use 0 @nr when calling cgroup_writeback_by_id() to use best-effort flushing while avoding possible livelocks. v4: Use get_jiffies_64() and time_before/after64() instead of raw jiffies_64 and arthimetic comparisons as suggested by Jan. Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-08-26 16:06:56 +00:00
int __maybe_unused i;
writeback, memcg: Implement foreign dirty flushing There's an inherent mismatch between memcg and writeback. The former trackes ownership per-page while the latter per-inode. This was a deliberate design decision because honoring per-page ownership in the writeback path is complicated, may lead to higher CPU and IO overheads and deemed unnecessary given that write-sharing an inode across different cgroups isn't a common use-case. Combined with inode majority-writer ownership switching, this works well enough in most cases but there are some pathological cases. For example, let's say there are two cgroups A and B which keep writing to different but confined parts of the same inode. B owns the inode and A's memory is limited far below B's. A's dirty ratio can rise enough to trigger balance_dirty_pages() sleeps but B's can be low enough to avoid triggering background writeback. A will be slowed down without a way to make writeback of the dirty pages happen. This patch implements foreign dirty recording and foreign mechanism so that when a memcg encounters a condition as above it can trigger flushes on bdi_writebacks which can clean its pages. Please see the comment on top of mem_cgroup_track_foreign_dirty_slowpath() for details. A reproducer follows. write-range.c:: #include <stdio.h> #include <stdlib.h> #include <unistd.h> #include <fcntl.h> #include <sys/types.h> static const char *usage = "write-range FILE START SIZE\n"; int main(int argc, char **argv) { int fd; unsigned long start, size, end, pos; char *endp; char buf[4096]; if (argc < 4) { fprintf(stderr, usage); return 1; } fd = open(argv[1], O_WRONLY); if (fd < 0) { perror("open"); return 1; } start = strtoul(argv[2], &endp, 0); if (*endp != '\0') { fprintf(stderr, usage); return 1; } size = strtoul(argv[3], &endp, 0); if (*endp != '\0') { fprintf(stderr, usage); return 1; } end = start + size; while (1) { for (pos = start; pos < end; ) { long bread, bwritten = 0; if (lseek(fd, pos, SEEK_SET) < 0) { perror("lseek"); return 1; } bread = read(0, buf, sizeof(buf) < end - pos ? sizeof(buf) : end - pos); if (bread < 0) { perror("read"); return 1; } if (bread == 0) return 0; while (bwritten < bread) { long this; this = write(fd, buf + bwritten, bread - bwritten); if (this < 0) { perror("write"); return 1; } bwritten += this; pos += bwritten; } } } } repro.sh:: #!/bin/bash set -e set -x sysctl -w vm.dirty_expire_centisecs=300000 sysctl -w vm.dirty_writeback_centisecs=300000 sysctl -w vm.dirtytime_expire_seconds=300000 echo 3 > /proc/sys/vm/drop_caches TEST=/sys/fs/cgroup/test A=$TEST/A B=$TEST/B mkdir -p $A $B echo "+memory +io" > $TEST/cgroup.subtree_control echo $((1<<30)) > $A/memory.high echo $((32<<30)) > $B/memory.high rm -f testfile touch testfile fallocate -l 4G testfile echo "Starting B" (echo $BASHPID > $B/cgroup.procs pv -q --rate-limit 70M < /dev/urandom | ./write-range testfile $((2<<30)) $((2<<30))) & echo "Waiting 10s to ensure B claims the testfile inode" sleep 5 sync sleep 5 sync echo "Starting A" (echo $BASHPID > $A/cgroup.procs pv < /dev/urandom | ./write-range testfile 0 $((2<<30))) v2: Added comments explaining why the specific intervals are being used. v3: Use 0 @nr when calling cgroup_writeback_by_id() to use best-effort flushing while avoding possible livelocks. v4: Use get_jiffies_64() and time_before/after64() instead of raw jiffies_64 and arthimetic comparisons as suggested by Jan. Reviewed-by: Jan Kara <jack@suse.cz> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@kernel.dk>
2019-08-26 16:06:56 +00:00
#ifdef CONFIG_CGROUP_WRITEBACK
for (i = 0; i < MEMCG_CGWB_FRN_CNT; i++)
wb_wait_for_completion(&memcg->cgwb_frn[i].done);
#endif
if (cgroup_subsys_on_dfl(memory_cgrp_subsys) && !cgroup_memory_nosocket)
static_branch_dec(&memcg_sockets_enabled_key);
if (!cgroup_subsys_on_dfl(memory_cgrp_subsys) && memcg->tcpmem_active)
static_branch_dec(&memcg_sockets_enabled_key);
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
vmpressure_cleanup(&memcg->vmpressure);
cancel_work_sync(&memcg->high_work);
mem_cgroup_remove_from_trees(memcg);
mm, memcg: assign memcg-aware shrinkers bitmap to memcg Imagine a big node with many cpus, memory cgroups and containers. Let we have 200 containers, every container has 10 mounts, and 10 cgroups. All container tasks don't touch foreign containers mounts. If there is intensive pages write, and global reclaim happens, a writing task has to iterate over all memcgs to shrink slab, before it's able to go to shrink_page_list(). Iteration over all the memcg slabs is very expensive: the task has to visit 200 * 10 = 2000 shrinkers for every memcg, and since there are 2000 memcgs, the total calls are 2000 * 2000 = 4000000. So, the shrinker makes 4 million do_shrink_slab() calls just to try to isolate SWAP_CLUSTER_MAX pages in one of the actively writing memcg via shrink_page_list(). I've observed a node spending almost 100% in kernel, making useless iteration over already shrinked slab. This patch adds bitmap of memcg-aware shrinkers to memcg. The size of the bitmap depends on bitmap_nr_ids, and during memcg life it's maintained to be enough to fit bitmap_nr_ids shrinkers. Every bit in the map is related to corresponding shrinker id. Next patches will maintain set bit only for really charged memcg. This will allow shrink_slab() to increase its performance in significant way. See the last patch for the numbers. [ktkhai@virtuozzo.com: v9] Link: http://lkml.kernel.org/r/153112549031.4097.3576147070498769979.stgit@localhost.localdomain [ktkhai@virtuozzo.com: add comment to mem_cgroup_css_online()] Link: http://lkml.kernel.org/r/521f9e5f-c436-b388-fe83-4dc870bfb489@virtuozzo.com Link: http://lkml.kernel.org/r/153063056619.1818.12550500883688681076.stgit@localhost.localdomain Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Tested-by: Shakeel Butt <shakeelb@google.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Guenter Roeck <linux@roeck-us.net> Cc: "Huang, Ying" <ying.huang@intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Josef Bacik <jbacik@fb.com> Cc: Li RongQing <lirongqing@baidu.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Matthias Kaehlcke <mka@chromium.org> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Michal Hocko <mhocko@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Philippe Ombredanne <pombredanne@nexb.com> Cc: Roman Gushchin <guro@fb.com> Cc: Sahitya Tummala <stummala@codeaurora.org> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Waiman Long <longman@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:37 +00:00
memcg_free_shrinker_maps(memcg);
memcg_free_kmem(memcg);
mm: memcontrol: clean up alloc, online, offline, free functions The creation and teardown of struct mem_cgroup is fairly messy and that has attracted mistakes and subtle bugs before. The main cause for this is that there is no clear model about what needs to happen when, and that attracts more chaos. So create one: 1. mem_cgroup_alloc() should allocate struct mem_cgroup and its auxiliary members and initialize work items, locks etc. so that the object it returns is fully initialized and in a neutral state. 2. mem_cgroup_css_alloc() will use mem_cgroup_alloc() to obtain a new memcg object and configure it and the system according to the role of the new memory-controlled cgroup in the hierarchy. 3. mem_cgroup_css_online() is no longer needed to synchronize with iterators, but it verifies css->id which isn't available earlier. 4. mem_cgroup_css_offline() implements stuff that needs to happen upon the user-visible destruction of a cgroup, which includes stopping all user interfacing as well as releasing certain structures when continued memory consumption would be unexpected at that point. 5. mem_cgroup_css_free() prepares the system and the memcg object for the object's disappearance, neutralizes its state, and then gives it back to mem_cgroup_free(). 6. mem_cgroup_free() releases struct mem_cgroup and auxiliary memory. [arnd@arndb.de: fix SLOB build regression] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: Michal Hocko <mhocko@suse.cz> Signed-off-by: Arnd Bergmann <arnd@arndb.de> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:53 +00:00
mem_cgroup_free(memcg);
}
/**
* mem_cgroup_css_reset - reset the states of a mem_cgroup
* @css: the target css
*
* Reset the states of the mem_cgroup associated with @css. This is
* invoked when the userland requests disabling on the default hierarchy
* but the memcg is pinned through dependency. The memcg should stop
* applying policies and should revert to the vanilla state as it may be
* made visible again.
*
* The current implementation only resets the essential configurations.
* This needs to be expanded to cover all the visible parts.
*/
static void mem_cgroup_css_reset(struct cgroup_subsys_state *css)
{
struct mem_cgroup *memcg = mem_cgroup_from_css(css);
page_counter_set_max(&memcg->memory, PAGE_COUNTER_MAX);
page_counter_set_max(&memcg->swap, PAGE_COUNTER_MAX);
page_counter_set_max(&memcg->memsw, PAGE_COUNTER_MAX);
page_counter_set_max(&memcg->kmem, PAGE_COUNTER_MAX);
page_counter_set_max(&memcg->tcpmem, PAGE_COUNTER_MAX);
memcg: introduce memory.min Memory controller implements the memory.low best-effort memory protection mechanism, which works perfectly in many cases and allows protecting working sets of important workloads from sudden reclaim. But its semantics has a significant limitation: it works only as long as there is a supply of reclaimable memory. This makes it pretty useless against any sort of slow memory leaks or memory usage increases. This is especially true for swapless systems. If swap is enabled, memory soft protection effectively postpones problems, allowing a leaking application to fill all swap area, which makes no sense. The only effective way to guarantee the memory protection in this case is to invoke the OOM killer. It's possible to handle this case in userspace by reacting on MEMCG_LOW events; but there is still a place for a fail-safe in-kernel mechanism to provide stronger guarantees. This patch introduces the memory.min interface for cgroup v2 memory controller. It works very similarly to memory.low (sharing the same hierarchical behavior), except that it's not disabled if there is no more reclaimable memory in the system. If cgroup is not populated, its memory.min is ignored, because otherwise even the OOM killer wouldn't be able to reclaim the protected memory, and the system can stall. [guro@fb.com: s/low/min/ in docs] Link: http://lkml.kernel.org/r/20180510130758.GA9129@castle.DHCP.thefacebook.com Link: http://lkml.kernel.org/r/20180509180734.GA4856@castle.DHCP.thefacebook.com Signed-off-by: Roman Gushchin <guro@fb.com> Reviewed-by: Randy Dunlap <rdunlap@infradead.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:07:46 +00:00
page_counter_set_min(&memcg->memory, 0);
mm: memory.low hierarchical behavior This patch aims to address an issue in current memory.low semantics, which makes it hard to use it in a hierarchy, where some leaf memory cgroups are more valuable than others. For example, there are memcgs A, A/B, A/C, A/D and A/E: A A/memory.low = 2G, A/memory.current = 6G //\\ BC DE B/memory.low = 3G B/memory.current = 2G C/memory.low = 1G C/memory.current = 2G D/memory.low = 0 D/memory.current = 2G E/memory.low = 10G E/memory.current = 0 If we apply memory pressure, B, C and D are reclaimed at the same pace while A's usage exceeds 2G. This is obviously wrong, as B's usage is fully below B's memory.low, and C has 1G of protection as well. Also, A is pushed to the size, which is less than A's 2G memory.low, which is also wrong. A simple bash script (provided below) can be used to reproduce the problem. Current results are: A: 1430097920 A/B: 711929856 A/C: 717426688 A/D: 741376 A/E: 0 To address the issue a concept of effective memory.low is introduced. Effective memory.low is always equal or less than original memory.low. In a case, when there is no memory.low overcommittment (and also for top-level cgroups), these two values are equal. Otherwise it's a part of parent's effective memory.low, calculated as a cgroup's memory.low usage divided by sum of sibling's memory.low usages (under memory.low usage I mean the size of actually protected memory: memory.current if memory.current < memory.low, 0 otherwise). It's necessary to track the actual usage, because otherwise an empty cgroup with memory.low set (A/E in my example) will affect actual memory distribution, which makes no sense. To avoid traversing the cgroup tree twice, page_counters code is reused. Calculating effective memory.low can be done in the reclaim path, as we conveniently traversing the cgroup tree from top to bottom and check memory.low on each level. So, it's a perfect place to calculate effective memory low and save it to use it for children cgroups. This also eliminates a need to traverse the cgroup tree from bottom to top each time to check if parent's guarantee is not exceeded. Setting/resetting effective memory.low is intentionally racy, but it's fine and shouldn't lead to any significant differences in actual memory distribution. With this patch applied results are matching the expectations: A: 2147930112 A/B: 1428721664 A/C: 718393344 A/D: 815104 A/E: 0 Test script: #!/bin/bash CGPATH="/sys/fs/cgroup" truncate /file1 --size 2G truncate /file2 --size 2G truncate /file3 --size 2G truncate /file4 --size 50G mkdir "${CGPATH}/A" echo "+memory" > "${CGPATH}/A/cgroup.subtree_control" mkdir "${CGPATH}/A/B" "${CGPATH}/A/C" "${CGPATH}/A/D" "${CGPATH}/A/E" echo 2G > "${CGPATH}/A/memory.low" echo 3G > "${CGPATH}/A/B/memory.low" echo 1G > "${CGPATH}/A/C/memory.low" echo 0 > "${CGPATH}/A/D/memory.low" echo 10G > "${CGPATH}/A/E/memory.low" echo $$ > "${CGPATH}/A/B/cgroup.procs" && vmtouch -qt /file1 echo $$ > "${CGPATH}/A/C/cgroup.procs" && vmtouch -qt /file2 echo $$ > "${CGPATH}/A/D/cgroup.procs" && vmtouch -qt /file3 echo $$ > "${CGPATH}/cgroup.procs" && vmtouch -qt /file4 echo "A: " `cat "${CGPATH}/A/memory.current"` echo "A/B: " `cat "${CGPATH}/A/B/memory.current"` echo "A/C: " `cat "${CGPATH}/A/C/memory.current"` echo "A/D: " `cat "${CGPATH}/A/D/memory.current"` echo "A/E: " `cat "${CGPATH}/A/E/memory.current"` rmdir "${CGPATH}/A/B" "${CGPATH}/A/C" "${CGPATH}/A/D" "${CGPATH}/A/E" rmdir "${CGPATH}/A" rm /file1 /file2 /file3 /file4 Link: http://lkml.kernel.org/r/20180405185921.4942-2-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:06:22 +00:00
page_counter_set_low(&memcg->memory, 0);
page_counter_set_high(&memcg->memory, PAGE_COUNTER_MAX);
memcg->soft_limit = PAGE_COUNTER_MAX;
mm/memcg: automatically penalize tasks with high swap use Add a memory.swap.high knob, which can be used to protect the system from SWAP exhaustion. The mechanism used for penalizing is similar to memory.high penalty (sleep on return to user space). That is not to say that the knob itself is equivalent to memory.high. The objective is more to protect the system from potentially buggy tasks consuming a lot of swap and impacting other tasks, or even bringing the whole system to stand still with complete SWAP exhaustion. Hopefully without the need to find per-task hard limits. Slowing misbehaving tasks down gradually allows user space oom killers or other protection mechanisms to react. oomd and earlyoom already do killing based on swap exhaustion, and memory.swap.high protection will help implement such userspace oom policies more reliably. We can use one counter for number of pages allocated under pressure to save struct task space and avoid two separate hierarchy walks on the hot path. The exact overage is calculated on return to user space, anyway. Take the new high limit into account when determining if swap is "full". Borrowing the explanation from Johannes: The idea behind "swap full" is that as long as the workload has plenty of swap space available and it's not changing its memory contents, it makes sense to generously hold on to copies of data in the swap device, even after the swapin. A later reclaim cycle can drop the page without any IO. Trading disk space for IO. But the only two ways to reclaim a swap slot is when they're faulted in and the references go away, or by scanning the virtual address space like swapoff does - which is very expensive (one could argue it's too expensive even for swapoff, it's often more practical to just reboot). So at some point in the fill level, we have to start freeing up swap slots on fault/swapin. Otherwise we could eventually run out of swap slots while they're filled with copies of data that is also in RAM. We don't want to OOM a workload because its available swap space is filled with redundant cache. Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Chris Down <chris@chrisdown.name> Cc: Shakeel Butt <shakeelb@google.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Hugh Dickins <hughd@google.com> Link: http://lkml.kernel.org/r/20200527195846.102707-5-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:52 +00:00
page_counter_set_high(&memcg->swap, PAGE_COUNTER_MAX);
memcg_wb_domain_size_changed(memcg);
}
#ifdef CONFIG_MMU
/* Handlers for move charge at task migration. */
static int mem_cgroup_do_precharge(unsigned long count)
{
int ret;
mm, page_alloc: distinguish between being unable to sleep, unwilling to sleep and avoiding waking kswapd __GFP_WAIT has been used to identify atomic context in callers that hold spinlocks or are in interrupts. They are expected to be high priority and have access one of two watermarks lower than "min" which can be referred to as the "atomic reserve". __GFP_HIGH users get access to the first lower watermark and can be called the "high priority reserve". Over time, callers had a requirement to not block when fallback options were available. Some have abused __GFP_WAIT leading to a situation where an optimisitic allocation with a fallback option can access atomic reserves. This patch uses __GFP_ATOMIC to identify callers that are truely atomic, cannot sleep and have no alternative. High priority users continue to use __GFP_HIGH. __GFP_DIRECT_RECLAIM identifies callers that can sleep and are willing to enter direct reclaim. __GFP_KSWAPD_RECLAIM to identify callers that want to wake kswapd for background reclaim. __GFP_WAIT is redefined as a caller that is willing to enter direct reclaim and wake kswapd for background reclaim. This patch then converts a number of sites o __GFP_ATOMIC is used by callers that are high priority and have memory pools for those requests. GFP_ATOMIC uses this flag. o Callers that have a limited mempool to guarantee forward progress clear __GFP_DIRECT_RECLAIM but keep __GFP_KSWAPD_RECLAIM. bio allocations fall into this category where kswapd will still be woken but atomic reserves are not used as there is a one-entry mempool to guarantee progress. o Callers that are checking if they are non-blocking should use the helper gfpflags_allow_blocking() where possible. This is because checking for __GFP_WAIT as was done historically now can trigger false positives. Some exceptions like dm-crypt.c exist where the code intent is clearer if __GFP_DIRECT_RECLAIM is used instead of the helper due to flag manipulations. o Callers that built their own GFP flags instead of starting with GFP_KERNEL and friends now also need to specify __GFP_KSWAPD_RECLAIM. The first key hazard to watch out for is callers that removed __GFP_WAIT and was depending on access to atomic reserves for inconspicuous reasons. In some cases it may be appropriate for them to use __GFP_HIGH. The second key hazard is callers that assembled their own combination of GFP flags instead of starting with something like GFP_KERNEL. They may now wish to specify __GFP_KSWAPD_RECLAIM. It's almost certainly harmless if it's missed in most cases as other activity will wake kswapd. Signed-off-by: Mel Gorman <mgorman@techsingularity.net> Acked-by: Vlastimil Babka <vbabka@suse.cz> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Vitaly Wool <vitalywool@gmail.com> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-11-07 00:28:21 +00:00
/* Try a single bulk charge without reclaim first, kswapd may wake */
ret = try_charge(mc.to, GFP_KERNEL & ~__GFP_DIRECT_RECLAIM, count);
if (!ret) {
mc.precharge += count;
return ret;
}
/* Try charges one by one with reclaim, but do not retry */
while (count--) {
ret = try_charge(mc.to, GFP_KERNEL | __GFP_NORETRY, 1);
if (ret)
return ret;
mc.precharge++;
cond_resched();
}
return 0;
}
union mc_target {
struct page *page;
swp_entry_t ent;
};
enum mc_target_type {
MC_TARGET_NONE = 0,
MC_TARGET_PAGE,
MC_TARGET_SWAP,
MC_TARGET_DEVICE,
};
static struct page *mc_handle_present_pte(struct vm_area_struct *vma,
unsigned long addr, pte_t ptent)
{
struct page *page = vm_normal_page(vma, addr, ptent);
if (!page || !page_mapped(page))
return NULL;
if (PageAnon(page)) {
if (!(mc.flags & MOVE_ANON))
return NULL;
} else {
if (!(mc.flags & MOVE_FILE))
return NULL;
}
if (!get_page_unless_zero(page))
return NULL;
return page;
}
#if defined(CONFIG_SWAP) || defined(CONFIG_DEVICE_PRIVATE)
static struct page *mc_handle_swap_pte(struct vm_area_struct *vma,
pte_t ptent, swp_entry_t *entry)
{
struct page *page = NULL;
swp_entry_t ent = pte_to_swp_entry(ptent);
if (!(mc.flags & MOVE_ANON) || non_swap_entry(ent))
return NULL;
/*
* Handle MEMORY_DEVICE_PRIVATE which are ZONE_DEVICE page belonging to
* a device and because they are not accessible by CPU they are store
* as special swap entry in the CPU page table.
*/
if (is_device_private_entry(ent)) {
page = device_private_entry_to_page(ent);
/*
* MEMORY_DEVICE_PRIVATE means ZONE_DEVICE page and which have
* a refcount of 1 when free (unlike normal page)
*/
if (!page_ref_add_unless(page, 1, 1))
return NULL;
return page;
}
memcg: fix/change behavior of shared anon at moving task This patch changes memcg's behavior at task_move(). At task_move(), the kernel scans a task's page table and move the changes for mapped pages from source cgroup to target cgroup. There has been a bug at handling shared anonymous pages for a long time. Before patch: - The spec says 'shared anonymous pages are not moved.' - The implementation was 'shared anonymoys pages may be moved'. If page_mapcount <=2, shared anonymous pages's charge were moved. After patch: - The spec says 'all anonymous pages are moved'. - The implementation is 'all anonymous pages are moved'. Considering usage of memcg, this will not affect user's experience. 'shared anonymous' pages only exists between a tree of processes which don't do exec(). Moving one of process without exec() seems not sane. For example, libcgroup will not be affected by this change. (Anyway, no one noticed the implementation for a long time...) Below is a discussion log: - current spec/implementation are complex - Now, shared file caches are moved - It adds unclear check as page_mapcount(). To do correct check, we should check swap users, etc. - No one notice this implementation behavior. So, no one get benefit from the design. - In general, once task is moved to a cgroup for running, it will not be moved.... - Finally, we have control knob as memory.move_charge_at_immigrate. Here is a patch to allow moving shared pages, completely. This makes memcg simpler and fix current broken code. Suggested-by: Hugh Dickins <hughd@google.com> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Cc: Glauber Costa <glommer@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-05-29 22:06:51 +00:00
/*
* Because lookup_swap_cache() updates some statistics counter,
* we call find_get_page() with swapper_space directly.
*/
mm, swap: use offset of swap entry as key of swap cache This patch is to improve the performance of swap cache operations when the type of the swap device is not 0. Originally, the whole swap entry value is used as the key of the swap cache, even though there is one radix tree for each swap device. If the type of the swap device is not 0, the height of the radix tree of the swap cache will be increased unnecessary, especially on 64bit architecture. For example, for a 1GB swap device on the x86_64 architecture, the height of the radix tree of the swap cache is 11. But if the offset of the swap entry is used as the key of the swap cache, the height of the radix tree of the swap cache is 4. The increased height causes unnecessary radix tree descending and increased cache footprint. This patch reduces the height of the radix tree of the swap cache via using the offset of the swap entry instead of the whole swap entry value as the key of the swap cache. In 32 processes sequential swap out test case on a Xeon E5 v3 system with RAM disk as swap, the lock contention for the spinlock of the swap cache is reduced from 20.15% to 12.19%, when the type of the swap device is 1. Use the whole swap entry as key, perf-profile.calltrace.cycles-pp._raw_spin_lock_irq.__add_to_swap_cache.add_to_swap_cache.add_to_swap.shrink_page_list: 10.37, perf-profile.calltrace.cycles-pp._raw_spin_lock_irqsave.__remove_mapping.shrink_page_list.shrink_inactive_list.shrink_node_memcg: 9.78, Use the swap offset as key, perf-profile.calltrace.cycles-pp._raw_spin_lock_irq.__add_to_swap_cache.add_to_swap_cache.add_to_swap.shrink_page_list: 6.25, perf-profile.calltrace.cycles-pp._raw_spin_lock_irqsave.__remove_mapping.shrink_page_list.shrink_inactive_list.shrink_node_memcg: 5.94, Link: http://lkml.kernel.org/r/1473270649-27229-1-git-send-email-ying.huang@intel.com Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> Cc: Dave Hansen <dave.hansen@linux.intel.com> Cc: Dan Williams <dan.j.williams@intel.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Hugh Dickins <hughd@google.com> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Minchan Kim <minchan@kernel.org> Cc: Aaron Lu <aaron.lu@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-10-08 00:00:21 +00:00
page = find_get_page(swap_address_space(ent), swp_offset(ent));
mm: memcontrol: make swap tracking an integral part of memory control Without swap page tracking, users that are otherwise memory controlled can easily escape their containment and allocate significant amounts of memory that they're not being charged for. That's because swap does readahead, but without the cgroup records of who owned the page at swapout, readahead pages don't get charged until somebody actually faults them into their page table and we can identify an owner task. This can be maliciously exploited with MADV_WILLNEED, which triggers arbitrary readahead allocations without charging the pages. Make swap swap page tracking an integral part of memcg and remove the Kconfig options. In the first place, it was only made configurable to allow users to save some memory. But the overhead of tracking cgroup ownership per swap page is minimal - 2 byte per page, or 512k per 1G of swap, or 0.04%. Saving that at the expense of broken containment semantics is not something we should present as a coequal option. The swapaccount=0 boot option will continue to exist, and it will eliminate the page_counter overhead and hide the swap control files, but it won't disable swap slot ownership tracking. This patch makes sure we always have the cgroup records at swapin time; the next patch will fix the actual bug by charging readahead swap pages at swapin time rather than at fault time. v2: fix double swap charge bug in cgroup1/cgroup2 code gating [hannes@cmpxchg.org: fix crash with cgroup_disable=memory] Link: http://lkml.kernel.org/r/20200521215855.GB815153@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Alex Shi <alex.shi@linux.alibaba.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Roman Gushchin <guro@fb.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Naresh Kamboju <naresh.kamboju@linaro.org> Link: http://lkml.kernel.org/r/20200508183105.225460-16-hannes@cmpxchg.org Debugged-by: Hugh Dickins <hughd@google.com> Debugged-by: Michal Hocko <mhocko@kernel.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:02:14 +00:00
entry->val = ent.val;
return page;
}
memcg: fix/change behavior of shared anon at moving task This patch changes memcg's behavior at task_move(). At task_move(), the kernel scans a task's page table and move the changes for mapped pages from source cgroup to target cgroup. There has been a bug at handling shared anonymous pages for a long time. Before patch: - The spec says 'shared anonymous pages are not moved.' - The implementation was 'shared anonymoys pages may be moved'. If page_mapcount <=2, shared anonymous pages's charge were moved. After patch: - The spec says 'all anonymous pages are moved'. - The implementation is 'all anonymous pages are moved'. Considering usage of memcg, this will not affect user's experience. 'shared anonymous' pages only exists between a tree of processes which don't do exec(). Moving one of process without exec() seems not sane. For example, libcgroup will not be affected by this change. (Anyway, no one noticed the implementation for a long time...) Below is a discussion log: - current spec/implementation are complex - Now, shared file caches are moved - It adds unclear check as page_mapcount(). To do correct check, we should check swap users, etc. - No one notice this implementation behavior. So, no one get benefit from the design. - In general, once task is moved to a cgroup for running, it will not be moved.... - Finally, we have control knob as memory.move_charge_at_immigrate. Here is a patch to allow moving shared pages, completely. This makes memcg simpler and fix current broken code. Suggested-by: Hugh Dickins <hughd@google.com> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Cc: Glauber Costa <glommer@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-05-29 22:06:51 +00:00
#else
static struct page *mc_handle_swap_pte(struct vm_area_struct *vma,
pte_t ptent, swp_entry_t *entry)
memcg: fix/change behavior of shared anon at moving task This patch changes memcg's behavior at task_move(). At task_move(), the kernel scans a task's page table and move the changes for mapped pages from source cgroup to target cgroup. There has been a bug at handling shared anonymous pages for a long time. Before patch: - The spec says 'shared anonymous pages are not moved.' - The implementation was 'shared anonymoys pages may be moved'. If page_mapcount <=2, shared anonymous pages's charge were moved. After patch: - The spec says 'all anonymous pages are moved'. - The implementation is 'all anonymous pages are moved'. Considering usage of memcg, this will not affect user's experience. 'shared anonymous' pages only exists between a tree of processes which don't do exec(). Moving one of process without exec() seems not sane. For example, libcgroup will not be affected by this change. (Anyway, no one noticed the implementation for a long time...) Below is a discussion log: - current spec/implementation are complex - Now, shared file caches are moved - It adds unclear check as page_mapcount(). To do correct check, we should check swap users, etc. - No one notice this implementation behavior. So, no one get benefit from the design. - In general, once task is moved to a cgroup for running, it will not be moved.... - Finally, we have control knob as memory.move_charge_at_immigrate. Here is a patch to allow moving shared pages, completely. This makes memcg simpler and fix current broken code. Suggested-by: Hugh Dickins <hughd@google.com> Signed-off-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Cc: Glauber Costa <glommer@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-05-29 22:06:51 +00:00
{
return NULL;
}
#endif
static struct page *mc_handle_file_pte(struct vm_area_struct *vma,
unsigned long addr, pte_t ptent, swp_entry_t *entry)
{
struct page *page = NULL;
struct address_space *mapping;
pgoff_t pgoff;
if (!vma->vm_file) /* anonymous vma */
return NULL;
if (!(mc.flags & MOVE_FILE))
return NULL;
mapping = vma->vm_file->f_mapping;
pgoff = linear_page_index(vma, addr);
/* page is moved even if it's not RSS of this task(page-faulted). */
#ifdef CONFIG_SWAP
/* shmem/tmpfs may report page out on swap: account for that too. */
mm: filemap: update find_get_pages_tag() to deal with shadow entries Dave Jones reports the following crash when find_get_pages_tag() runs into an exceptional entry: kernel BUG at mm/filemap.c:1347! RIP: find_get_pages_tag+0x1cb/0x220 Call Trace: find_get_pages_tag+0x36/0x220 pagevec_lookup_tag+0x21/0x30 filemap_fdatawait_range+0xbe/0x1e0 filemap_fdatawait+0x27/0x30 sync_inodes_sb+0x204/0x2a0 sync_inodes_one_sb+0x19/0x20 iterate_supers+0xb2/0x110 sys_sync+0x44/0xb0 ia32_do_call+0x13/0x13 1343 /* 1344 * This function is never used on a shmem/tmpfs 1345 * mapping, so a swap entry won't be found here. 1346 */ 1347 BUG(); After commit 0cd6144aadd2 ("mm + fs: prepare for non-page entries in page cache radix trees") this comment and BUG() are out of date because exceptional entries can now appear in all mappings - as shadows of recently evicted pages. However, as Hugh Dickins notes, "it is truly surprising for a PAGECACHE_TAG_WRITEBACK (and probably any other PAGECACHE_TAG_*) to appear on an exceptional entry. I expect it comes down to an occasional race in RCU lookup of the radix_tree: lacking absolute synchronization, we might sometimes catch an exceptional entry, with the tag which really belongs with the unexceptional entry which was there an instant before." And indeed, not only is the tree walk lockless, the tags are also read in chunks, one radix tree node at a time. There is plenty of time for page reclaim to swoop in and replace a page that was already looked up as tagged with a shadow entry. Remove the BUG() and update the comment. While reviewing all other lookup sites for whether they properly deal with shadow entries of evicted pages, update all the comments and fix memcg file charge moving to not miss shmem/tmpfs swapcache pages. Fixes: 0cd6144aadd2 ("mm + fs: prepare for non-page entries in page cache radix trees") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Dave Jones <davej@redhat.com> Acked-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-05-06 19:50:05 +00:00
if (shmem_mapping(mapping)) {
page = find_get_entry(mapping, pgoff);
if (xa_is_value(page)) {
mm: filemap: update find_get_pages_tag() to deal with shadow entries Dave Jones reports the following crash when find_get_pages_tag() runs into an exceptional entry: kernel BUG at mm/filemap.c:1347! RIP: find_get_pages_tag+0x1cb/0x220 Call Trace: find_get_pages_tag+0x36/0x220 pagevec_lookup_tag+0x21/0x30 filemap_fdatawait_range+0xbe/0x1e0 filemap_fdatawait+0x27/0x30 sync_inodes_sb+0x204/0x2a0 sync_inodes_one_sb+0x19/0x20 iterate_supers+0xb2/0x110 sys_sync+0x44/0xb0 ia32_do_call+0x13/0x13 1343 /* 1344 * This function is never used on a shmem/tmpfs 1345 * mapping, so a swap entry won't be found here. 1346 */ 1347 BUG(); After commit 0cd6144aadd2 ("mm + fs: prepare for non-page entries in page cache radix trees") this comment and BUG() are out of date because exceptional entries can now appear in all mappings - as shadows of recently evicted pages. However, as Hugh Dickins notes, "it is truly surprising for a PAGECACHE_TAG_WRITEBACK (and probably any other PAGECACHE_TAG_*) to appear on an exceptional entry. I expect it comes down to an occasional race in RCU lookup of the radix_tree: lacking absolute synchronization, we might sometimes catch an exceptional entry, with the tag which really belongs with the unexceptional entry which was there an instant before." And indeed, not only is the tree walk lockless, the tags are also read in chunks, one radix tree node at a time. There is plenty of time for page reclaim to swoop in and replace a page that was already looked up as tagged with a shadow entry. Remove the BUG() and update the comment. While reviewing all other lookup sites for whether they properly deal with shadow entries of evicted pages, update all the comments and fix memcg file charge moving to not miss shmem/tmpfs swapcache pages. Fixes: 0cd6144aadd2 ("mm + fs: prepare for non-page entries in page cache radix trees") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Dave Jones <davej@redhat.com> Acked-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-05-06 19:50:05 +00:00
swp_entry_t swp = radix_to_swp_entry(page);
mm: memcontrol: make swap tracking an integral part of memory control Without swap page tracking, users that are otherwise memory controlled can easily escape their containment and allocate significant amounts of memory that they're not being charged for. That's because swap does readahead, but without the cgroup records of who owned the page at swapout, readahead pages don't get charged until somebody actually faults them into their page table and we can identify an owner task. This can be maliciously exploited with MADV_WILLNEED, which triggers arbitrary readahead allocations without charging the pages. Make swap swap page tracking an integral part of memcg and remove the Kconfig options. In the first place, it was only made configurable to allow users to save some memory. But the overhead of tracking cgroup ownership per swap page is minimal - 2 byte per page, or 512k per 1G of swap, or 0.04%. Saving that at the expense of broken containment semantics is not something we should present as a coequal option. The swapaccount=0 boot option will continue to exist, and it will eliminate the page_counter overhead and hide the swap control files, but it won't disable swap slot ownership tracking. This patch makes sure we always have the cgroup records at swapin time; the next patch will fix the actual bug by charging readahead swap pages at swapin time rather than at fault time. v2: fix double swap charge bug in cgroup1/cgroup2 code gating [hannes@cmpxchg.org: fix crash with cgroup_disable=memory] Link: http://lkml.kernel.org/r/20200521215855.GB815153@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Alex Shi <alex.shi@linux.alibaba.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Roman Gushchin <guro@fb.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Naresh Kamboju <naresh.kamboju@linaro.org> Link: http://lkml.kernel.org/r/20200508183105.225460-16-hannes@cmpxchg.org Debugged-by: Hugh Dickins <hughd@google.com> Debugged-by: Michal Hocko <mhocko@kernel.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:02:14 +00:00
*entry = swp;
mm, swap: use offset of swap entry as key of swap cache This patch is to improve the performance of swap cache operations when the type of the swap device is not 0. Originally, the whole swap entry value is used as the key of the swap cache, even though there is one radix tree for each swap device. If the type of the swap device is not 0, the height of the radix tree of the swap cache will be increased unnecessary, especially on 64bit architecture. For example, for a 1GB swap device on the x86_64 architecture, the height of the radix tree of the swap cache is 11. But if the offset of the swap entry is used as the key of the swap cache, the height of the radix tree of the swap cache is 4. The increased height causes unnecessary radix tree descending and increased cache footprint. This patch reduces the height of the radix tree of the swap cache via using the offset of the swap entry instead of the whole swap entry value as the key of the swap cache. In 32 processes sequential swap out test case on a Xeon E5 v3 system with RAM disk as swap, the lock contention for the spinlock of the swap cache is reduced from 20.15% to 12.19%, when the type of the swap device is 1. Use the whole swap entry as key, perf-profile.calltrace.cycles-pp._raw_spin_lock_irq.__add_to_swap_cache.add_to_swap_cache.add_to_swap.shrink_page_list: 10.37, perf-profile.calltrace.cycles-pp._raw_spin_lock_irqsave.__remove_mapping.shrink_page_list.shrink_inactive_list.shrink_node_memcg: 9.78, Use the swap offset as key, perf-profile.calltrace.cycles-pp._raw_spin_lock_irq.__add_to_swap_cache.add_to_swap_cache.add_to_swap.shrink_page_list: 6.25, perf-profile.calltrace.cycles-pp._raw_spin_lock_irqsave.__remove_mapping.shrink_page_list.shrink_inactive_list.shrink_node_memcg: 5.94, Link: http://lkml.kernel.org/r/1473270649-27229-1-git-send-email-ying.huang@intel.com Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: "Kirill A. Shutemov" <kirill.shutemov@linux.intel.com> Cc: Dave Hansen <dave.hansen@linux.intel.com> Cc: Dan Williams <dan.j.williams@intel.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Hugh Dickins <hughd@google.com> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Minchan Kim <minchan@kernel.org> Cc: Aaron Lu <aaron.lu@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-10-08 00:00:21 +00:00
page = find_get_page(swap_address_space(swp),
swp_offset(swp));
mm: filemap: update find_get_pages_tag() to deal with shadow entries Dave Jones reports the following crash when find_get_pages_tag() runs into an exceptional entry: kernel BUG at mm/filemap.c:1347! RIP: find_get_pages_tag+0x1cb/0x220 Call Trace: find_get_pages_tag+0x36/0x220 pagevec_lookup_tag+0x21/0x30 filemap_fdatawait_range+0xbe/0x1e0 filemap_fdatawait+0x27/0x30 sync_inodes_sb+0x204/0x2a0 sync_inodes_one_sb+0x19/0x20 iterate_supers+0xb2/0x110 sys_sync+0x44/0xb0 ia32_do_call+0x13/0x13 1343 /* 1344 * This function is never used on a shmem/tmpfs 1345 * mapping, so a swap entry won't be found here. 1346 */ 1347 BUG(); After commit 0cd6144aadd2 ("mm + fs: prepare for non-page entries in page cache radix trees") this comment and BUG() are out of date because exceptional entries can now appear in all mappings - as shadows of recently evicted pages. However, as Hugh Dickins notes, "it is truly surprising for a PAGECACHE_TAG_WRITEBACK (and probably any other PAGECACHE_TAG_*) to appear on an exceptional entry. I expect it comes down to an occasional race in RCU lookup of the radix_tree: lacking absolute synchronization, we might sometimes catch an exceptional entry, with the tag which really belongs with the unexceptional entry which was there an instant before." And indeed, not only is the tree walk lockless, the tags are also read in chunks, one radix tree node at a time. There is plenty of time for page reclaim to swoop in and replace a page that was already looked up as tagged with a shadow entry. Remove the BUG() and update the comment. While reviewing all other lookup sites for whether they properly deal with shadow entries of evicted pages, update all the comments and fix memcg file charge moving to not miss shmem/tmpfs swapcache pages. Fixes: 0cd6144aadd2 ("mm + fs: prepare for non-page entries in page cache radix trees") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Dave Jones <davej@redhat.com> Acked-by: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-05-06 19:50:05 +00:00
}
} else
page = find_get_page(mapping, pgoff);
#else
page = find_get_page(mapping, pgoff);
#endif
return page;
}
/**
* mem_cgroup_move_account - move account of the page
* @page: the page
* @compound: charge the page as compound or small page
* @from: mem_cgroup which the page is moved from.
* @to: mem_cgroup which the page is moved to. @from != @to.
*
* The caller must make sure the page is not on LRU (isolate_page() is useful.)
*
* This function doesn't do "charge" to new cgroup and doesn't do "uncharge"
* from old cgroup.
*/
static int mem_cgroup_move_account(struct page *page,
memcg: adjust to support new THP refcounting As with rmap, with new refcounting we cannot rely on PageTransHuge() to check if we need to charge size of huge page form the cgroup. We need to get information from caller to know whether it was mapped with PMD or PTE. We do uncharge when last reference on the page gone. At that point if we see PageTransHuge() it means we need to unchange whole huge page. The tricky part is partial unmap -- when we try to unmap part of huge page. We don't do a special handing of this situation, meaning we don't uncharge the part of huge page unless last user is gone or split_huge_page() is triggered. In case of cgroup memory pressure happens the partial unmapped page will be split through shrinker. This should be good enough. Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Tested-by: Sasha Levin <sasha.levin@oracle.com> Tested-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Acked-by: Vlastimil Babka <vbabka@suse.cz> Acked-by: Jerome Marchand <jmarchan@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Hugh Dickins <hughd@google.com> Cc: Dave Hansen <dave.hansen@intel.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Rik van Riel <riel@redhat.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Cc: Steve Capper <steve.capper@linaro.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-16 00:52:20 +00:00
bool compound,
struct mem_cgroup *from,
struct mem_cgroup *to)
{
struct lruvec *from_vec, *to_vec;
struct pglist_data *pgdat;
memcg: adjust to support new THP refcounting As with rmap, with new refcounting we cannot rely on PageTransHuge() to check if we need to charge size of huge page form the cgroup. We need to get information from caller to know whether it was mapped with PMD or PTE. We do uncharge when last reference on the page gone. At that point if we see PageTransHuge() it means we need to unchange whole huge page. The tricky part is partial unmap -- when we try to unmap part of huge page. We don't do a special handing of this situation, meaning we don't uncharge the part of huge page unless last user is gone or split_huge_page() is triggered. In case of cgroup memory pressure happens the partial unmapped page will be split through shrinker. This should be good enough. Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Tested-by: Sasha Levin <sasha.levin@oracle.com> Tested-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Acked-by: Vlastimil Babka <vbabka@suse.cz> Acked-by: Jerome Marchand <jmarchan@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Hugh Dickins <hughd@google.com> Cc: Dave Hansen <dave.hansen@intel.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Rik van Riel <riel@redhat.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Cc: Steve Capper <steve.capper@linaro.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-16 00:52:20 +00:00
unsigned int nr_pages = compound ? hpage_nr_pages(page) : 1;
int ret;
VM_BUG_ON(from == to);
VM_BUG_ON_PAGE(PageLRU(page), page);
memcg: adjust to support new THP refcounting As with rmap, with new refcounting we cannot rely on PageTransHuge() to check if we need to charge size of huge page form the cgroup. We need to get information from caller to know whether it was mapped with PMD or PTE. We do uncharge when last reference on the page gone. At that point if we see PageTransHuge() it means we need to unchange whole huge page. The tricky part is partial unmap -- when we try to unmap part of huge page. We don't do a special handing of this situation, meaning we don't uncharge the part of huge page unless last user is gone or split_huge_page() is triggered. In case of cgroup memory pressure happens the partial unmapped page will be split through shrinker. This should be good enough. Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Tested-by: Sasha Levin <sasha.levin@oracle.com> Tested-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Acked-by: Vlastimil Babka <vbabka@suse.cz> Acked-by: Jerome Marchand <jmarchan@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Hugh Dickins <hughd@google.com> Cc: Dave Hansen <dave.hansen@intel.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Rik van Riel <riel@redhat.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Cc: Steve Capper <steve.capper@linaro.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-16 00:52:20 +00:00
VM_BUG_ON(compound && !PageTransHuge(page));
/*
* Prevent mem_cgroup_migrate() from looking at
* page->mem_cgroup of its source page while we change it.
*/
memcg: adjust to support new THP refcounting As with rmap, with new refcounting we cannot rely on PageTransHuge() to check if we need to charge size of huge page form the cgroup. We need to get information from caller to know whether it was mapped with PMD or PTE. We do uncharge when last reference on the page gone. At that point if we see PageTransHuge() it means we need to unchange whole huge page. The tricky part is partial unmap -- when we try to unmap part of huge page. We don't do a special handing of this situation, meaning we don't uncharge the part of huge page unless last user is gone or split_huge_page() is triggered. In case of cgroup memory pressure happens the partial unmapped page will be split through shrinker. This should be good enough. Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Tested-by: Sasha Levin <sasha.levin@oracle.com> Tested-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Acked-by: Vlastimil Babka <vbabka@suse.cz> Acked-by: Jerome Marchand <jmarchan@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Hugh Dickins <hughd@google.com> Cc: Dave Hansen <dave.hansen@intel.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Rik van Riel <riel@redhat.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Cc: Steve Capper <steve.capper@linaro.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-16 00:52:20 +00:00
ret = -EBUSY;
if (!trylock_page(page))
goto out;
ret = -EINVAL;
if (page->mem_cgroup != from)
goto out_unlock;
pgdat = page_pgdat(page);
from_vec = mem_cgroup_lruvec(from, pgdat);
to_vec = mem_cgroup_lruvec(to, pgdat);
lock_page_memcg(page);
mm: memcontrol: switch to native NR_ANON_MAPPED counter Memcg maintains a private MEMCG_RSS counter. This divergence from the generic VM accounting means unnecessary code overhead, and creates a dependency for memcg that page->mapping is set up at the time of charging, so that page types can be told apart. Convert the generic accounting sites to mod_lruvec_page_state and friends to maintain the per-cgroup vmstat counter of NR_ANON_MAPPED. We use lock_page_memcg() to stabilize page->mem_cgroup during rmap changes, the same way we do for NR_FILE_MAPPED. With the previous patch removing MEMCG_CACHE and the private NR_SHMEM counter, this patch finally eliminates the need to have page->mapping set up at charge time. However, we need to have page->mem_cgroup set up by the time rmap runs and does the accounting, so switch the commit and the rmap callbacks around. v2: fix temporary accounting bug by switching rmap<->commit (Joonsoo) Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Cc: Alex Shi <alex.shi@linux.alibaba.com> Cc: Hugh Dickins <hughd@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.com> Cc: Roman Gushchin <guro@fb.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Balbir Singh <bsingharora@gmail.com> Link: http://lkml.kernel.org/r/20200508183105.225460-11-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:01:57 +00:00
if (PageAnon(page)) {
if (page_mapped(page)) {
__mod_lruvec_state(from_vec, NR_ANON_MAPPED, -nr_pages);
__mod_lruvec_state(to_vec, NR_ANON_MAPPED, nr_pages);
if (PageTransHuge(page)) {
__mod_lruvec_state(from_vec, NR_ANON_THPS,
-nr_pages);
__mod_lruvec_state(to_vec, NR_ANON_THPS,
nr_pages);
}
mm: memcontrol: switch to native NR_ANON_MAPPED counter Memcg maintains a private MEMCG_RSS counter. This divergence from the generic VM accounting means unnecessary code overhead, and creates a dependency for memcg that page->mapping is set up at the time of charging, so that page types can be told apart. Convert the generic accounting sites to mod_lruvec_page_state and friends to maintain the per-cgroup vmstat counter of NR_ANON_MAPPED. We use lock_page_memcg() to stabilize page->mem_cgroup during rmap changes, the same way we do for NR_FILE_MAPPED. With the previous patch removing MEMCG_CACHE and the private NR_SHMEM counter, this patch finally eliminates the need to have page->mapping set up at charge time. However, we need to have page->mem_cgroup set up by the time rmap runs and does the accounting, so switch the commit and the rmap callbacks around. v2: fix temporary accounting bug by switching rmap<->commit (Joonsoo) Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Cc: Alex Shi <alex.shi@linux.alibaba.com> Cc: Hugh Dickins <hughd@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.com> Cc: Roman Gushchin <guro@fb.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Balbir Singh <bsingharora@gmail.com> Link: http://lkml.kernel.org/r/20200508183105.225460-11-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:01:57 +00:00
}
} else {
__mod_lruvec_state(from_vec, NR_FILE_PAGES, -nr_pages);
__mod_lruvec_state(to_vec, NR_FILE_PAGES, nr_pages);
if (PageSwapBacked(page)) {
__mod_lruvec_state(from_vec, NR_SHMEM, -nr_pages);
__mod_lruvec_state(to_vec, NR_SHMEM, nr_pages);
}
if (page_mapped(page)) {
__mod_lruvec_state(from_vec, NR_FILE_MAPPED, -nr_pages);
__mod_lruvec_state(to_vec, NR_FILE_MAPPED, nr_pages);
}
if (PageDirty(page)) {
struct address_space *mapping = page_mapping(page);
memcg: add per cgroup dirty page accounting When modifying PG_Dirty on cached file pages, update the new MEM_CGROUP_STAT_DIRTY counter. This is done in the same places where global NR_FILE_DIRTY is managed. The new memcg stat is visible in the per memcg memory.stat cgroupfs file. The most recent past attempt at this was http://thread.gmane.org/gmane.linux.kernel.cgroups/8632 The new accounting supports future efforts to add per cgroup dirty page throttling and writeback. It also helps an administrator break down a container's memory usage and provides evidence to understand memcg oom kills (the new dirty count is included in memcg oom kill messages). The ability to move page accounting between memcg (memory.move_charge_at_immigrate) makes this accounting more complicated than the global counter. The existing mem_cgroup_{begin,end}_page_stat() lock is used to serialize move accounting with stat updates. Typical update operation: memcg = mem_cgroup_begin_page_stat(page) if (TestSetPageDirty()) { [...] mem_cgroup_update_page_stat(memcg) } mem_cgroup_end_page_stat(memcg) Summary of mem_cgroup_end_page_stat() overhead: - Without CONFIG_MEMCG it's a no-op - With CONFIG_MEMCG and no inter memcg task movement, it's just rcu_read_lock() - With CONFIG_MEMCG and inter memcg task movement, it's rcu_read_lock() + spin_lock_irqsave() A memcg parameter is added to several routines because their callers now grab mem_cgroup_begin_page_stat() which returns the memcg later needed by for mem_cgroup_update_page_stat(). Because mem_cgroup_begin_page_stat() may disable interrupts, some adjustments are needed: - move __mark_inode_dirty() from __set_page_dirty() to its caller. __mark_inode_dirty() locking does not want interrupts disabled. - use spin_lock_irqsave(tree_lock) rather than spin_lock_irq() in __delete_from_page_cache(), replace_page_cache_page(), invalidate_complete_page2(), and __remove_mapping(). text data bss dec hex filename 8925147 1774832 1785856 12485835 be84cb vmlinux-!CONFIG_MEMCG-before 8925339 1774832 1785856 12486027 be858b vmlinux-!CONFIG_MEMCG-after +192 text bytes 8965977 1784992 1785856 12536825 bf4bf9 vmlinux-CONFIG_MEMCG-before 8966750 1784992 1785856 12537598 bf4efe vmlinux-CONFIG_MEMCG-after +773 text bytes Performance tests run on v4.0-rc1-36-g4f671fe2f952. Lower is better for all metrics, they're all wall clock or cycle counts. The read and write fault benchmarks just measure fault time, they do not include I/O time. * CONFIG_MEMCG not set: baseline patched kbuild 1m25.030000(+-0.088% 3 samples) 1m25.426667(+-0.120% 3 samples) dd write 100 MiB 0.859211561 +-15.10% 0.874162885 +-15.03% dd write 200 MiB 1.670653105 +-17.87% 1.669384764 +-11.99% dd write 1000 MiB 8.434691190 +-14.15% 8.474733215 +-14.77% read fault cycles 254.0(+-0.000% 10 samples) 253.0(+-0.000% 10 samples) write fault cycles 2021.2(+-3.070% 10 samples) 1984.5(+-1.036% 10 samples) * CONFIG_MEMCG=y root_memcg: baseline patched kbuild 1m25.716667(+-0.105% 3 samples) 1m25.686667(+-0.153% 3 samples) dd write 100 MiB 0.855650830 +-14.90% 0.887557919 +-14.90% dd write 200 MiB 1.688322953 +-12.72% 1.667682724 +-13.33% dd write 1000 MiB 8.418601605 +-14.30% 8.673532299 +-15.00% read fault cycles 266.0(+-0.000% 10 samples) 266.0(+-0.000% 10 samples) write fault cycles 2051.7(+-1.349% 10 samples) 2049.6(+-1.686% 10 samples) * CONFIG_MEMCG=y non-root_memcg: baseline patched kbuild 1m26.120000(+-0.273% 3 samples) 1m25.763333(+-0.127% 3 samples) dd write 100 MiB 0.861723964 +-15.25% 0.818129350 +-14.82% dd write 200 MiB 1.669887569 +-13.30% 1.698645885 +-13.27% dd write 1000 MiB 8.383191730 +-14.65% 8.351742280 +-14.52% read fault cycles 265.7(+-0.172% 10 samples) 267.0(+-0.000% 10 samples) write fault cycles 2070.6(+-1.512% 10 samples) 2084.4(+-2.148% 10 samples) As expected anon page faults are not affected by this patch. tj: Updated to apply on top of the recent cancel_dirty_page() changes. Signed-off-by: Sha Zhengju <handai.szj@gmail.com> Signed-off-by: Greg Thelen <gthelen@google.com> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-22 21:13:16 +00:00
if (mapping_cap_account_dirty(mapping)) {
__mod_lruvec_state(from_vec, NR_FILE_DIRTY,
-nr_pages);
__mod_lruvec_state(to_vec, NR_FILE_DIRTY,
nr_pages);
}
memcg: add per cgroup dirty page accounting When modifying PG_Dirty on cached file pages, update the new MEM_CGROUP_STAT_DIRTY counter. This is done in the same places where global NR_FILE_DIRTY is managed. The new memcg stat is visible in the per memcg memory.stat cgroupfs file. The most recent past attempt at this was http://thread.gmane.org/gmane.linux.kernel.cgroups/8632 The new accounting supports future efforts to add per cgroup dirty page throttling and writeback. It also helps an administrator break down a container's memory usage and provides evidence to understand memcg oom kills (the new dirty count is included in memcg oom kill messages). The ability to move page accounting between memcg (memory.move_charge_at_immigrate) makes this accounting more complicated than the global counter. The existing mem_cgroup_{begin,end}_page_stat() lock is used to serialize move accounting with stat updates. Typical update operation: memcg = mem_cgroup_begin_page_stat(page) if (TestSetPageDirty()) { [...] mem_cgroup_update_page_stat(memcg) } mem_cgroup_end_page_stat(memcg) Summary of mem_cgroup_end_page_stat() overhead: - Without CONFIG_MEMCG it's a no-op - With CONFIG_MEMCG and no inter memcg task movement, it's just rcu_read_lock() - With CONFIG_MEMCG and inter memcg task movement, it's rcu_read_lock() + spin_lock_irqsave() A memcg parameter is added to several routines because their callers now grab mem_cgroup_begin_page_stat() which returns the memcg later needed by for mem_cgroup_update_page_stat(). Because mem_cgroup_begin_page_stat() may disable interrupts, some adjustments are needed: - move __mark_inode_dirty() from __set_page_dirty() to its caller. __mark_inode_dirty() locking does not want interrupts disabled. - use spin_lock_irqsave(tree_lock) rather than spin_lock_irq() in __delete_from_page_cache(), replace_page_cache_page(), invalidate_complete_page2(), and __remove_mapping(). text data bss dec hex filename 8925147 1774832 1785856 12485835 be84cb vmlinux-!CONFIG_MEMCG-before 8925339 1774832 1785856 12486027 be858b vmlinux-!CONFIG_MEMCG-after +192 text bytes 8965977 1784992 1785856 12536825 bf4bf9 vmlinux-CONFIG_MEMCG-before 8966750 1784992 1785856 12537598 bf4efe vmlinux-CONFIG_MEMCG-after +773 text bytes Performance tests run on v4.0-rc1-36-g4f671fe2f952. Lower is better for all metrics, they're all wall clock or cycle counts. The read and write fault benchmarks just measure fault time, they do not include I/O time. * CONFIG_MEMCG not set: baseline patched kbuild 1m25.030000(+-0.088% 3 samples) 1m25.426667(+-0.120% 3 samples) dd write 100 MiB 0.859211561 +-15.10% 0.874162885 +-15.03% dd write 200 MiB 1.670653105 +-17.87% 1.669384764 +-11.99% dd write 1000 MiB 8.434691190 +-14.15% 8.474733215 +-14.77% read fault cycles 254.0(+-0.000% 10 samples) 253.0(+-0.000% 10 samples) write fault cycles 2021.2(+-3.070% 10 samples) 1984.5(+-1.036% 10 samples) * CONFIG_MEMCG=y root_memcg: baseline patched kbuild 1m25.716667(+-0.105% 3 samples) 1m25.686667(+-0.153% 3 samples) dd write 100 MiB 0.855650830 +-14.90% 0.887557919 +-14.90% dd write 200 MiB 1.688322953 +-12.72% 1.667682724 +-13.33% dd write 1000 MiB 8.418601605 +-14.30% 8.673532299 +-15.00% read fault cycles 266.0(+-0.000% 10 samples) 266.0(+-0.000% 10 samples) write fault cycles 2051.7(+-1.349% 10 samples) 2049.6(+-1.686% 10 samples) * CONFIG_MEMCG=y non-root_memcg: baseline patched kbuild 1m26.120000(+-0.273% 3 samples) 1m25.763333(+-0.127% 3 samples) dd write 100 MiB 0.861723964 +-15.25% 0.818129350 +-14.82% dd write 200 MiB 1.669887569 +-13.30% 1.698645885 +-13.27% dd write 1000 MiB 8.383191730 +-14.65% 8.351742280 +-14.52% read fault cycles 265.7(+-0.172% 10 samples) 267.0(+-0.000% 10 samples) write fault cycles 2070.6(+-1.512% 10 samples) 2084.4(+-2.148% 10 samples) As expected anon page faults are not affected by this patch. tj: Updated to apply on top of the recent cancel_dirty_page() changes. Signed-off-by: Sha Zhengju <handai.szj@gmail.com> Signed-off-by: Greg Thelen <gthelen@google.com> Signed-off-by: Tejun Heo <tj@kernel.org> Signed-off-by: Jens Axboe <axboe@fb.com>
2015-05-22 21:13:16 +00:00
}
}
if (PageWriteback(page)) {
__mod_lruvec_state(from_vec, NR_WRITEBACK, -nr_pages);
__mod_lruvec_state(to_vec, NR_WRITEBACK, nr_pages);
}
/*
* All state has been migrated, let's switch to the new memcg.
*
* It is safe to change page->mem_cgroup here because the page
* is referenced, charged, isolated, and locked: we can't race
* with (un)charging, migration, LRU putback, or anything else
* that would rely on a stable page->mem_cgroup.
*
* Note that lock_page_memcg is a memcg lock, not a page lock,
* to save space. As soon as we switch page->mem_cgroup to a
* new memcg that isn't locked, the above state can change
* concurrently again. Make sure we're truly done with it.
*/
smp_mb();
page->mem_cgroup = to; /* caller should have done css_get */
mm: thp: make deferred split shrinker memcg aware Currently THP deferred split shrinker is not memcg aware, this may cause premature OOM with some configuration. For example the below test would run into premature OOM easily: $ cgcreate -g memory:thp $ echo 4G > /sys/fs/cgroup/memory/thp/memory/limit_in_bytes $ cgexec -g memory:thp transhuge-stress 4000 transhuge-stress comes from kernel selftest. It is easy to hit OOM, but there are still a lot THP on the deferred split queue, memcg direct reclaim can't touch them since the deferred split shrinker is not memcg aware. Convert deferred split shrinker memcg aware by introducing per memcg deferred split queue. The THP should be on either per node or per memcg deferred split queue if it belongs to a memcg. When the page is immigrated to the other memcg, it will be immigrated to the target memcg's deferred split queue too. Reuse the second tail page's deferred_list for per memcg list since the same THP can't be on multiple deferred split queues. [yang.shi@linux.alibaba.com: simplify deferred split queue dereference per Kirill Tkhai] Link: http://lkml.kernel.org/r/1566496227-84952-5-git-send-email-yang.shi@linux.alibaba.com Link: http://lkml.kernel.org/r/1565144277-36240-5-git-send-email-yang.shi@linux.alibaba.com Signed-off-by: Yang Shi <yang.shi@linux.alibaba.com> Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Reviewed-by: Kirill Tkhai <ktkhai@virtuozzo.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: "Kirill A . Shutemov" <kirill.shutemov@linux.intel.com> Cc: Hugh Dickins <hughd@google.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: David Rientjes <rientjes@google.com> Cc: Qian Cai <cai@lca.pw> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-09-23 22:38:15 +00:00
__unlock_page_memcg(from);
ret = 0;
local_irq_disable();
mm: memcontrol: drop @compound parameter from memcg charging API The memcg charging API carries a boolean @compound parameter that tells whether the page we're dealing with is a hugepage. mem_cgroup_commit_charge() has another boolean @lrucare that indicates whether the page needs LRU locking or not while charging. The majority of callsites know those parameters at compile time, which results in a lot of naked "false, false" argument lists. This makes for cryptic code and is a breeding ground for subtle mistakes. Thankfully, the huge page state can be inferred from the page itself and doesn't need to be passed along. This is safe because charging completes before the page is published and somebody may split it. Simplify the callsites by removing @compound, and let memcg infer the state by using hpage_nr_pages() unconditionally. That function does PageTransHuge() to identify huge pages, which also helpfully asserts that nobody passes in tail pages by accident. The following patches will introduce a new charging API, best not to carry over unnecessary weight. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Alex Shi <alex.shi@linux.alibaba.com> Reviewed-by: Joonsoo Kim <iamjoonsoo.kim@lge.com> Reviewed-by: Shakeel Butt <shakeelb@google.com> Cc: Hugh Dickins <hughd@google.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.com> Cc: Roman Gushchin <guro@fb.com> Cc: Balbir Singh <bsingharora@gmail.com> Link: http://lkml.kernel.org/r/20200508183105.225460-4-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:01:31 +00:00
mem_cgroup_charge_statistics(to, page, nr_pages);
memcg_check_events(to, page);
mm: memcontrol: drop @compound parameter from memcg charging API The memcg charging API carries a boolean @compound parameter that tells whether the page we're dealing with is a hugepage. mem_cgroup_commit_charge() has another boolean @lrucare that indicates whether the page needs LRU locking or not while charging. The majority of callsites know those parameters at compile time, which results in a lot of naked "false, false" argument lists. This makes for cryptic code and is a breeding ground for subtle mistakes. Thankfully, the huge page state can be inferred from the page itself and doesn't need to be passed along. This is safe because charging completes before the page is published and somebody may split it. Simplify the callsites by removing @compound, and let memcg infer the state by using hpage_nr_pages() unconditionally. That function does PageTransHuge() to identify huge pages, which also helpfully asserts that nobody passes in tail pages by accident. The following patches will introduce a new charging API, best not to carry over unnecessary weight. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Alex Shi <alex.shi@linux.alibaba.com> Reviewed-by: Joonsoo Kim <iamjoonsoo.kim@lge.com> Reviewed-by: Shakeel Butt <shakeelb@google.com> Cc: Hugh Dickins <hughd@google.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.com> Cc: Roman Gushchin <guro@fb.com> Cc: Balbir Singh <bsingharora@gmail.com> Link: http://lkml.kernel.org/r/20200508183105.225460-4-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:01:31 +00:00
mem_cgroup_charge_statistics(from, page, -nr_pages);
memcg_check_events(from, page);
local_irq_enable();
out_unlock:
unlock_page(page);
out:
return ret;
}
/**
* get_mctgt_type - get target type of moving charge
* @vma: the vma the pte to be checked belongs
* @addr: the address corresponding to the pte to be checked
* @ptent: the pte to be checked
* @target: the pointer the target page or swap ent will be stored(can be NULL)
*
* Returns
* 0(MC_TARGET_NONE): if the pte is not a target for move charge.
* 1(MC_TARGET_PAGE): if the page corresponding to this pte is a target for
* move charge. if @target is not NULL, the page is stored in target->page
* with extra refcnt got(Callers should handle it).
* 2(MC_TARGET_SWAP): if the swap entry corresponding to this pte is a
* target for charge migration. if @target is not NULL, the entry is stored
* in target->ent.
* 3(MC_TARGET_DEVICE): like MC_TARGET_PAGE but page is MEMORY_DEVICE_PRIVATE
* (so ZONE_DEVICE page and thus not on the lru).
* For now we such page is charge like a regular page would be as for all
* intent and purposes it is just special memory taking the place of a
* regular page.
*
* See Documentations/vm/hmm.txt and include/linux/hmm.h
*
* Called with pte lock held.
*/
static enum mc_target_type get_mctgt_type(struct vm_area_struct *vma,
unsigned long addr, pte_t ptent, union mc_target *target)
{
struct page *page = NULL;
enum mc_target_type ret = MC_TARGET_NONE;
swp_entry_t ent = { .val = 0 };
if (pte_present(ptent))
page = mc_handle_present_pte(vma, addr, ptent);
else if (is_swap_pte(ptent))
page = mc_handle_swap_pte(vma, ptent, &ent);
else if (pte_none(ptent))
page = mc_handle_file_pte(vma, addr, ptent, &ent);
if (!page && !ent.val)
return ret;
if (page) {
/*
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
* Do only loose check w/o serialization.
mm: embed the memcg pointer directly into struct page Memory cgroups used to have 5 per-page pointers. To allow users to disable that amount of overhead during runtime, those pointers were allocated in a separate array, with a translation layer between them and struct page. There is now only one page pointer remaining: the memcg pointer, that indicates which cgroup the page is associated with when charged. The complexity of runtime allocation and the runtime translation overhead is no longer justified to save that *potential* 0.19% of memory. With CONFIG_SLUB, page->mem_cgroup actually sits in the doubleword padding after the page->private member and doesn't even increase struct page, and then this patch actually saves space. Remaining users that care can still compile their kernels without CONFIG_MEMCG. text data bss dec hex filename 8828345 1725264 983040 11536649 b00909 vmlinux.old 8827425 1725264 966656 11519345 afc571 vmlinux.new [mhocko@suse.cz: update Documentation/cgroups/memory.txt] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: David S. Miller <davem@davemloft.net> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Acked-by: Konstantin Khlebnikov <koct9i@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:44:52 +00:00
* mem_cgroup_move_account() checks the page is valid or
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
* not under LRU exclusion.
*/
mm: embed the memcg pointer directly into struct page Memory cgroups used to have 5 per-page pointers. To allow users to disable that amount of overhead during runtime, those pointers were allocated in a separate array, with a translation layer between them and struct page. There is now only one page pointer remaining: the memcg pointer, that indicates which cgroup the page is associated with when charged. The complexity of runtime allocation and the runtime translation overhead is no longer justified to save that *potential* 0.19% of memory. With CONFIG_SLUB, page->mem_cgroup actually sits in the doubleword padding after the page->private member and doesn't even increase struct page, and then this patch actually saves space. Remaining users that care can still compile their kernels without CONFIG_MEMCG. text data bss dec hex filename 8828345 1725264 983040 11536649 b00909 vmlinux.old 8827425 1725264 966656 11519345 afc571 vmlinux.new [mhocko@suse.cz: update Documentation/cgroups/memory.txt] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: David S. Miller <davem@davemloft.net> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Acked-by: Konstantin Khlebnikov <koct9i@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:44:52 +00:00
if (page->mem_cgroup == mc.from) {
ret = MC_TARGET_PAGE;
if (is_device_private_page(page))
ret = MC_TARGET_DEVICE;
if (target)
target->page = page;
}
if (!ret || !target)
put_page(page);
}
/*
* There is a swap entry and a page doesn't exist or isn't charged.
* But we cannot move a tail-page in a THP.
*/
if (ent.val && !ret && (!page || !PageTransCompound(page)) &&
mem_cgroup_id(mc.from) == lookup_swap_cgroup_id(ent)) {
ret = MC_TARGET_SWAP;
if (target)
target->ent = ent;
}
return ret;
}
#ifdef CONFIG_TRANSPARENT_HUGEPAGE
/*
* We don't consider PMD mapped swapping or file mapped pages because THP does
* not support them for now.
* Caller should make sure that pmd_trans_huge(pmd) is true.
*/
static enum mc_target_type get_mctgt_type_thp(struct vm_area_struct *vma,
unsigned long addr, pmd_t pmd, union mc_target *target)
{
struct page *page = NULL;
enum mc_target_type ret = MC_TARGET_NONE;
mm: thp: check pmd migration entry in common path When THP migration is being used, memory management code needs to handle pmd migration entries properly. This patch uses !pmd_present() or is_swap_pmd() (depending on whether pmd_none() needs separate code or not) to check pmd migration entries at the places where a pmd entry is present. Since pmd-related code uses split_huge_page(), split_huge_pmd(), pmd_trans_huge(), pmd_trans_unstable(), or pmd_none_or_trans_huge_or_clear_bad(), this patch: 1. adds pmd migration entry split code in split_huge_pmd(), 2. takes care of pmd migration entries whenever pmd_trans_huge() is present, 3. makes pmd_none_or_trans_huge_or_clear_bad() pmd migration entry aware. Since split_huge_page() uses split_huge_pmd() and pmd_trans_unstable() is equivalent to pmd_none_or_trans_huge_or_clear_bad(), we do not change them. Until this commit, a pmd entry should be: 1. pointing to a pte page, 2. is_swap_pmd(), 3. pmd_trans_huge(), 4. pmd_devmap(), or 5. pmd_none(). Signed-off-by: Zi Yan <zi.yan@cs.rutgers.edu> Cc: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Cc: "H. Peter Anvin" <hpa@zytor.com> Cc: Anshuman Khandual <khandual@linux.vnet.ibm.com> Cc: Dave Hansen <dave.hansen@intel.com> Cc: David Nellans <dnellans@nvidia.com> Cc: Ingo Molnar <mingo@elte.hu> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Minchan Kim <minchan@kernel.org> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Vlastimil Babka <vbabka@suse.cz> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-09-08 23:11:01 +00:00
if (unlikely(is_swap_pmd(pmd))) {
VM_BUG_ON(thp_migration_supported() &&
!is_pmd_migration_entry(pmd));
return ret;
}
page = pmd_page(pmd);
VM_BUG_ON_PAGE(!page || !PageHead(page), page);
if (!(mc.flags & MOVE_ANON))
return ret;
mm: embed the memcg pointer directly into struct page Memory cgroups used to have 5 per-page pointers. To allow users to disable that amount of overhead during runtime, those pointers were allocated in a separate array, with a translation layer between them and struct page. There is now only one page pointer remaining: the memcg pointer, that indicates which cgroup the page is associated with when charged. The complexity of runtime allocation and the runtime translation overhead is no longer justified to save that *potential* 0.19% of memory. With CONFIG_SLUB, page->mem_cgroup actually sits in the doubleword padding after the page->private member and doesn't even increase struct page, and then this patch actually saves space. Remaining users that care can still compile their kernels without CONFIG_MEMCG. text data bss dec hex filename 8828345 1725264 983040 11536649 b00909 vmlinux.old 8827425 1725264 966656 11519345 afc571 vmlinux.new [mhocko@suse.cz: update Documentation/cgroups/memory.txt] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: David S. Miller <davem@davemloft.net> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Acked-by: Konstantin Khlebnikov <koct9i@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:44:52 +00:00
if (page->mem_cgroup == mc.from) {
ret = MC_TARGET_PAGE;
if (target) {
get_page(page);
target->page = page;
}
}
return ret;
}
#else
static inline enum mc_target_type get_mctgt_type_thp(struct vm_area_struct *vma,
unsigned long addr, pmd_t pmd, union mc_target *target)
{
return MC_TARGET_NONE;
}
#endif
static int mem_cgroup_count_precharge_pte_range(pmd_t *pmd,
unsigned long addr, unsigned long end,
struct mm_walk *walk)
{
struct vm_area_struct *vma = walk->vma;
pte_t *pte;
spinlock_t *ptl;
ptl = pmd_trans_huge_lock(pmd, vma);
if (ptl) {
/*
* Note their can not be MC_TARGET_DEVICE for now as we do not
* support transparent huge page with MEMORY_DEVICE_PRIVATE but
* this might change.
*/
if (get_mctgt_type_thp(vma, addr, *pmd, NULL) == MC_TARGET_PAGE)
mc.precharge += HPAGE_PMD_NR;
mm, thp: change pmd_trans_huge_lock() to return taken lock With split ptlock it's important to know which lock pmd_trans_huge_lock() took. This patch adds one more parameter to the function to return the lock. In most places migration to new api is trivial. Exception is move_huge_pmd(): we need to take two locks if pmd tables are different. Signed-off-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Tested-by: Alex Thorlton <athorlton@sgi.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: "Eric W . Biederman" <ebiederm@xmission.com> Cc: "Paul E . McKenney" <paulmck@linux.vnet.ibm.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andi Kleen <ak@linux.intel.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Dave Hansen <dave.hansen@intel.com> Cc: Dave Jones <davej@redhat.com> Cc: David Howells <dhowells@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Kees Cook <keescook@chromium.org> Cc: Mel Gorman <mgorman@suse.de> Cc: Michael Kerrisk <mtk.manpages@gmail.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Rik van Riel <riel@redhat.com> Cc: Robin Holt <robinmholt@gmail.com> Cc: Sedat Dilek <sedat.dilek@gmail.com> Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-11-14 22:30:54 +00:00
spin_unlock(ptl);
mm: thp: fix pmd_bad() triggering in code paths holding mmap_sem read mode In some cases it may happen that pmd_none_or_clear_bad() is called with the mmap_sem hold in read mode. In those cases the huge page faults can allocate hugepmds under pmd_none_or_clear_bad() and that can trigger a false positive from pmd_bad() that will not like to see a pmd materializing as trans huge. It's not khugepaged causing the problem, khugepaged holds the mmap_sem in write mode (and all those sites must hold the mmap_sem in read mode to prevent pagetables to go away from under them, during code review it seems vm86 mode on 32bit kernels requires that too unless it's restricted to 1 thread per process or UP builds). The race is only with the huge pagefaults that can convert a pmd_none() into a pmd_trans_huge(). Effectively all these pmd_none_or_clear_bad() sites running with mmap_sem in read mode are somewhat speculative with the page faults, and the result is always undefined when they run simultaneously. This is probably why it wasn't common to run into this. For example if the madvise(MADV_DONTNEED) runs zap_page_range() shortly before the page fault, the hugepage will not be zapped, if the page fault runs first it will be zapped. Altering pmd_bad() not to error out if it finds hugepmds won't be enough to fix this, because zap_pmd_range would then proceed to call zap_pte_range (which would be incorrect if the pmd become a pmd_trans_huge()). The simplest way to fix this is to read the pmd in the local stack (regardless of what we read, no need of actual CPU barriers, only compiler barrier needed), and be sure it is not changing under the code that computes its value. Even if the real pmd is changing under the value we hold on the stack, we don't care. If we actually end up in zap_pte_range it means the pmd was not none already and it was not huge, and it can't become huge from under us (khugepaged locking explained above). All we need is to enforce that there is no way anymore that in a code path like below, pmd_trans_huge can be false, but pmd_none_or_clear_bad can run into a hugepmd. The overhead of a barrier() is just a compiler tweak and should not be measurable (I only added it for THP builds). I don't exclude different compiler versions may have prevented the race too by caching the value of *pmd on the stack (that hasn't been verified, but it wouldn't be impossible considering pmd_none_or_clear_bad, pmd_bad, pmd_trans_huge, pmd_none are all inlines and there's no external function called in between pmd_trans_huge and pmd_none_or_clear_bad). if (pmd_trans_huge(*pmd)) { if (next-addr != HPAGE_PMD_SIZE) { VM_BUG_ON(!rwsem_is_locked(&tlb->mm->mmap_sem)); split_huge_page_pmd(vma->vm_mm, pmd); } else if (zap_huge_pmd(tlb, vma, pmd, addr)) continue; /* fall through */ } if (pmd_none_or_clear_bad(pmd)) Because this race condition could be exercised without special privileges this was reported in CVE-2012-1179. The race was identified and fully explained by Ulrich who debugged it. I'm quoting his accurate explanation below, for reference. ====== start quote ======= mapcount 0 page_mapcount 1 kernel BUG at mm/huge_memory.c:1384! At some point prior to the panic, a "bad pmd ..." message similar to the following is logged on the console: mm/memory.c:145: bad pmd ffff8800376e1f98(80000000314000e7). The "bad pmd ..." message is logged by pmd_clear_bad() before it clears the page's PMD table entry. 143 void pmd_clear_bad(pmd_t *pmd) 144 { -> 145 pmd_ERROR(*pmd); 146 pmd_clear(pmd); 147 } After the PMD table entry has been cleared, there is an inconsistency between the actual number of PMD table entries that are mapping the page and the page's map count (_mapcount field in struct page). When the page is subsequently reclaimed, __split_huge_page() detects this inconsistency. 1381 if (mapcount != page_mapcount(page)) 1382 printk(KERN_ERR "mapcount %d page_mapcount %d\n", 1383 mapcount, page_mapcount(page)); -> 1384 BUG_ON(mapcount != page_mapcount(page)); The root cause of the problem is a race of two threads in a multithreaded process. Thread B incurs a page fault on a virtual address that has never been accessed (PMD entry is zero) while Thread A is executing an madvise() system call on a virtual address within the same 2 MB (huge page) range. virtual address space .---------------------. | | | | .-|---------------------| | | | | | |<-- B(fault) | | | 2 MB | |/////////////////////|-. huge < |/////////////////////| > A(range) page | |/////////////////////|-' | | | | | | '-|---------------------| | | | | '---------------------' - Thread A is executing an madvise(..., MADV_DONTNEED) system call on the virtual address range "A(range)" shown in the picture. sys_madvise // Acquire the semaphore in shared mode. down_read(&current->mm->mmap_sem) ... madvise_vma switch (behavior) case MADV_DONTNEED: madvise_dontneed zap_page_range unmap_vmas unmap_page_range zap_pud_range zap_pmd_range // // Assume that this huge page has never been accessed. // I.e. content of the PMD entry is zero (not mapped). // if (pmd_trans_huge(*pmd)) { // We don't get here due to the above assumption. } // // Assume that Thread B incurred a page fault and .---------> // sneaks in here as shown below. | // | if (pmd_none_or_clear_bad(pmd)) | { | if (unlikely(pmd_bad(*pmd))) | pmd_clear_bad | { | pmd_ERROR | // Log "bad pmd ..." message here. | pmd_clear | // Clear the page's PMD entry. | // Thread B incremented the map count | // in page_add_new_anon_rmap(), but | // now the page is no longer mapped | // by a PMD entry (-> inconsistency). | } | } | v - Thread B is handling a page fault on virtual address "B(fault)" shown in the picture. ... do_page_fault __do_page_fault // Acquire the semaphore in shared mode. down_read_trylock(&mm->mmap_sem) ... handle_mm_fault if (pmd_none(*pmd) && transparent_hugepage_enabled(vma)) // We get here due to the above assumption (PMD entry is zero). do_huge_pmd_anonymous_page alloc_hugepage_vma // Allocate a new transparent huge page here. ... __do_huge_pmd_anonymous_page ... spin_lock(&mm->page_table_lock) ... page_add_new_anon_rmap // Here we increment the page's map count (starts at -1). atomic_set(&page->_mapcount, 0) set_pmd_at // Here we set the page's PMD entry which will be cleared // when Thread A calls pmd_clear_bad(). ... spin_unlock(&mm->page_table_lock) The mmap_sem does not prevent the race because both threads are acquiring it in shared mode (down_read). Thread B holds the page_table_lock while the page's map count and PMD table entry are updated. However, Thread A does not synchronize on that lock. ====== end quote ======= [akpm@linux-foundation.org: checkpatch fixes] Reported-by: Ulrich Obergfell <uobergfe@redhat.com> Signed-off-by: Andrea Arcangeli <aarcange@redhat.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Mel Gorman <mgorman@suse.de> Cc: Hugh Dickins <hughd@google.com> Cc: Dave Jones <davej@redhat.com> Acked-by: Larry Woodman <lwoodman@redhat.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: <stable@vger.kernel.org> [2.6.38+] Cc: Mark Salter <msalter@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-21 23:33:42 +00:00
return 0;
}
if (pmd_trans_unstable(pmd))
return 0;
pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
for (; addr != end; pte++, addr += PAGE_SIZE)
if (get_mctgt_type(vma, addr, *pte, NULL))
mc.precharge++; /* increment precharge temporarily */
pte_unmap_unlock(pte - 1, ptl);
cond_resched();
return 0;
}
static const struct mm_walk_ops precharge_walk_ops = {
.pmd_entry = mem_cgroup_count_precharge_pte_range,
};
static unsigned long mem_cgroup_count_precharge(struct mm_struct *mm)
{
unsigned long precharge;
memcg: fix deadlock between cpuset and memcg Commit b1dd693e ("memcg: avoid deadlock between move charge and try_charge()") can cause another deadlock about mmap_sem on task migration if cpuset and memcg are mounted onto the same mount point. After the commit, cgroup_attach_task() has sequence like: cgroup_attach_task() ss->can_attach() cpuset_can_attach() mem_cgroup_can_attach() down_read(&mmap_sem) (1) ss->attach() cpuset_attach() mpol_rebind_mm() down_write(&mmap_sem) (2) up_write(&mmap_sem) cpuset_migrate_mm() do_migrate_pages() down_read(&mmap_sem) up_read(&mmap_sem) mem_cgroup_move_task() mem_cgroup_clear_mc() up_read(&mmap_sem) We can cause deadlock at (2) because we've already aquire the mmap_sem at (1). But the commit itself is necessary to fix deadlocks which have existed before the commit like: Ex.1) move charge | try charge --------------------------------------+------------------------------ mem_cgroup_can_attach() | down_write(&mmap_sem) mc.moving_task = current | .. mem_cgroup_precharge_mc() | __mem_cgroup_try_charge() mem_cgroup_count_precharge() | prepare_to_wait() down_read(&mmap_sem) | if (mc.moving_task) -> cannot aquire the lock | -> true | schedule() | -> move charge should wake it up Ex.2) move charge | try charge --------------------------------------+------------------------------ mem_cgroup_can_attach() | mc.moving_task = current | mem_cgroup_precharge_mc() | mem_cgroup_count_precharge() | down_read(&mmap_sem) | .. | up_read(&mmap_sem) | | down_write(&mmap_sem) mem_cgroup_move_task() | .. mem_cgroup_move_charge() | __mem_cgroup_try_charge() down_read(&mmap_sem) | prepare_to_wait() -> cannot aquire the lock | if (mc.moving_task) | -> true | schedule() | -> move charge should wake it up This patch fixes all of these problems by: 1. revert the commit. 2. To fix the Ex.1, we set mc.moving_task after mem_cgroup_count_precharge() has released the mmap_sem. 3. To fix the Ex.2, we use down_read_trylock() instead of down_read() in mem_cgroup_move_charge() and, if it has failed to aquire the lock, cancel all extra charges, wake up all waiters, and retry trylock. Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Reported-by: Ben Blum <bblum@andrew.cmu.edu> Cc: Miao Xie <miaox@cn.fujitsu.com> Cc: David Rientjes <rientjes@google.com> Cc: Paul Menage <menage@google.com> Cc: Hiroyuki Kamezawa <kamezawa.hiroyuki@gmail.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-01-13 23:47:41 +00:00
down_read(&mm->mmap_sem);
walk_page_range(mm, 0, mm->highest_vm_end, &precharge_walk_ops, NULL);
memcg: fix deadlock between cpuset and memcg Commit b1dd693e ("memcg: avoid deadlock between move charge and try_charge()") can cause another deadlock about mmap_sem on task migration if cpuset and memcg are mounted onto the same mount point. After the commit, cgroup_attach_task() has sequence like: cgroup_attach_task() ss->can_attach() cpuset_can_attach() mem_cgroup_can_attach() down_read(&mmap_sem) (1) ss->attach() cpuset_attach() mpol_rebind_mm() down_write(&mmap_sem) (2) up_write(&mmap_sem) cpuset_migrate_mm() do_migrate_pages() down_read(&mmap_sem) up_read(&mmap_sem) mem_cgroup_move_task() mem_cgroup_clear_mc() up_read(&mmap_sem) We can cause deadlock at (2) because we've already aquire the mmap_sem at (1). But the commit itself is necessary to fix deadlocks which have existed before the commit like: Ex.1) move charge | try charge --------------------------------------+------------------------------ mem_cgroup_can_attach() | down_write(&mmap_sem) mc.moving_task = current | .. mem_cgroup_precharge_mc() | __mem_cgroup_try_charge() mem_cgroup_count_precharge() | prepare_to_wait() down_read(&mmap_sem) | if (mc.moving_task) -> cannot aquire the lock | -> true | schedule() | -> move charge should wake it up Ex.2) move charge | try charge --------------------------------------+------------------------------ mem_cgroup_can_attach() | mc.moving_task = current | mem_cgroup_precharge_mc() | mem_cgroup_count_precharge() | down_read(&mmap_sem) | .. | up_read(&mmap_sem) | | down_write(&mmap_sem) mem_cgroup_move_task() | .. mem_cgroup_move_charge() | __mem_cgroup_try_charge() down_read(&mmap_sem) | prepare_to_wait() -> cannot aquire the lock | if (mc.moving_task) | -> true | schedule() | -> move charge should wake it up This patch fixes all of these problems by: 1. revert the commit. 2. To fix the Ex.1, we set mc.moving_task after mem_cgroup_count_precharge() has released the mmap_sem. 3. To fix the Ex.2, we use down_read_trylock() instead of down_read() in mem_cgroup_move_charge() and, if it has failed to aquire the lock, cancel all extra charges, wake up all waiters, and retry trylock. Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Reported-by: Ben Blum <bblum@andrew.cmu.edu> Cc: Miao Xie <miaox@cn.fujitsu.com> Cc: David Rientjes <rientjes@google.com> Cc: Paul Menage <menage@google.com> Cc: Hiroyuki Kamezawa <kamezawa.hiroyuki@gmail.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-01-13 23:47:41 +00:00
up_read(&mm->mmap_sem);
precharge = mc.precharge;
mc.precharge = 0;
return precharge;
}
static int mem_cgroup_precharge_mc(struct mm_struct *mm)
{
memcg: fix deadlock between cpuset and memcg Commit b1dd693e ("memcg: avoid deadlock between move charge and try_charge()") can cause another deadlock about mmap_sem on task migration if cpuset and memcg are mounted onto the same mount point. After the commit, cgroup_attach_task() has sequence like: cgroup_attach_task() ss->can_attach() cpuset_can_attach() mem_cgroup_can_attach() down_read(&mmap_sem) (1) ss->attach() cpuset_attach() mpol_rebind_mm() down_write(&mmap_sem) (2) up_write(&mmap_sem) cpuset_migrate_mm() do_migrate_pages() down_read(&mmap_sem) up_read(&mmap_sem) mem_cgroup_move_task() mem_cgroup_clear_mc() up_read(&mmap_sem) We can cause deadlock at (2) because we've already aquire the mmap_sem at (1). But the commit itself is necessary to fix deadlocks which have existed before the commit like: Ex.1) move charge | try charge --------------------------------------+------------------------------ mem_cgroup_can_attach() | down_write(&mmap_sem) mc.moving_task = current | .. mem_cgroup_precharge_mc() | __mem_cgroup_try_charge() mem_cgroup_count_precharge() | prepare_to_wait() down_read(&mmap_sem) | if (mc.moving_task) -> cannot aquire the lock | -> true | schedule() | -> move charge should wake it up Ex.2) move charge | try charge --------------------------------------+------------------------------ mem_cgroup_can_attach() | mc.moving_task = current | mem_cgroup_precharge_mc() | mem_cgroup_count_precharge() | down_read(&mmap_sem) | .. | up_read(&mmap_sem) | | down_write(&mmap_sem) mem_cgroup_move_task() | .. mem_cgroup_move_charge() | __mem_cgroup_try_charge() down_read(&mmap_sem) | prepare_to_wait() -> cannot aquire the lock | if (mc.moving_task) | -> true | schedule() | -> move charge should wake it up This patch fixes all of these problems by: 1. revert the commit. 2. To fix the Ex.1, we set mc.moving_task after mem_cgroup_count_precharge() has released the mmap_sem. 3. To fix the Ex.2, we use down_read_trylock() instead of down_read() in mem_cgroup_move_charge() and, if it has failed to aquire the lock, cancel all extra charges, wake up all waiters, and retry trylock. Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Reported-by: Ben Blum <bblum@andrew.cmu.edu> Cc: Miao Xie <miaox@cn.fujitsu.com> Cc: David Rientjes <rientjes@google.com> Cc: Paul Menage <menage@google.com> Cc: Hiroyuki Kamezawa <kamezawa.hiroyuki@gmail.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-01-13 23:47:41 +00:00
unsigned long precharge = mem_cgroup_count_precharge(mm);
VM_BUG_ON(mc.moving_task);
mc.moving_task = current;
return mem_cgroup_do_precharge(precharge);
}
memcg: fix deadlock between cpuset and memcg Commit b1dd693e ("memcg: avoid deadlock between move charge and try_charge()") can cause another deadlock about mmap_sem on task migration if cpuset and memcg are mounted onto the same mount point. After the commit, cgroup_attach_task() has sequence like: cgroup_attach_task() ss->can_attach() cpuset_can_attach() mem_cgroup_can_attach() down_read(&mmap_sem) (1) ss->attach() cpuset_attach() mpol_rebind_mm() down_write(&mmap_sem) (2) up_write(&mmap_sem) cpuset_migrate_mm() do_migrate_pages() down_read(&mmap_sem) up_read(&mmap_sem) mem_cgroup_move_task() mem_cgroup_clear_mc() up_read(&mmap_sem) We can cause deadlock at (2) because we've already aquire the mmap_sem at (1). But the commit itself is necessary to fix deadlocks which have existed before the commit like: Ex.1) move charge | try charge --------------------------------------+------------------------------ mem_cgroup_can_attach() | down_write(&mmap_sem) mc.moving_task = current | .. mem_cgroup_precharge_mc() | __mem_cgroup_try_charge() mem_cgroup_count_precharge() | prepare_to_wait() down_read(&mmap_sem) | if (mc.moving_task) -> cannot aquire the lock | -> true | schedule() | -> move charge should wake it up Ex.2) move charge | try charge --------------------------------------+------------------------------ mem_cgroup_can_attach() | mc.moving_task = current | mem_cgroup_precharge_mc() | mem_cgroup_count_precharge() | down_read(&mmap_sem) | .. | up_read(&mmap_sem) | | down_write(&mmap_sem) mem_cgroup_move_task() | .. mem_cgroup_move_charge() | __mem_cgroup_try_charge() down_read(&mmap_sem) | prepare_to_wait() -> cannot aquire the lock | if (mc.moving_task) | -> true | schedule() | -> move charge should wake it up This patch fixes all of these problems by: 1. revert the commit. 2. To fix the Ex.1, we set mc.moving_task after mem_cgroup_count_precharge() has released the mmap_sem. 3. To fix the Ex.2, we use down_read_trylock() instead of down_read() in mem_cgroup_move_charge() and, if it has failed to aquire the lock, cancel all extra charges, wake up all waiters, and retry trylock. Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Reported-by: Ben Blum <bblum@andrew.cmu.edu> Cc: Miao Xie <miaox@cn.fujitsu.com> Cc: David Rientjes <rientjes@google.com> Cc: Paul Menage <menage@google.com> Cc: Hiroyuki Kamezawa <kamezawa.hiroyuki@gmail.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-01-13 23:47:41 +00:00
/* cancels all extra charges on mc.from and mc.to, and wakes up all waiters. */
static void __mem_cgroup_clear_mc(void)
{
struct mem_cgroup *from = mc.from;
struct mem_cgroup *to = mc.to;
/* we must uncharge all the leftover precharges from mc.to */
if (mc.precharge) {
mm: memcontrol: rewrite charge API These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 4): The memcg charge API charges pages before they are rmapped - i.e. have an actual "type" - and so every callsite needs its own set of charge and uncharge functions to know what type is being operated on. Worse, uncharge has to happen from a context that is still type-specific, rather than at the end of the page's lifetime with exclusive access, and so requires a lot of synchronization. Rewrite the charge API to provide a generic set of try_charge(), commit_charge() and cancel_charge() transaction operations, much like what's currently done for swap-in: mem_cgroup_try_charge() attempts to reserve a charge, reclaiming pages from the memcg if necessary. mem_cgroup_commit_charge() commits the page to the charge once it has a valid page->mapping and PageAnon() reliably tells the type. mem_cgroup_cancel_charge() aborts the transaction. This reduces the charge API and enables subsequent patches to drastically simplify uncharging. As pages need to be committed after rmap is established but before they are added to the LRU, page_add_new_anon_rmap() must stop doing LRU additions again. Revive lru_cache_add_active_or_unevictable(). [hughd@google.com: fix shmem_unuse] [hughd@google.com: Add comments on the private use of -EAGAIN] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Hugh Dickins <hughd@google.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:20 +00:00
cancel_charge(mc.to, mc.precharge);
mc.precharge = 0;
}
/*
* we didn't uncharge from mc.from at mem_cgroup_move_account(), so
* we must uncharge here.
*/
if (mc.moved_charge) {
mm: memcontrol: rewrite charge API These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 4): The memcg charge API charges pages before they are rmapped - i.e. have an actual "type" - and so every callsite needs its own set of charge and uncharge functions to know what type is being operated on. Worse, uncharge has to happen from a context that is still type-specific, rather than at the end of the page's lifetime with exclusive access, and so requires a lot of synchronization. Rewrite the charge API to provide a generic set of try_charge(), commit_charge() and cancel_charge() transaction operations, much like what's currently done for swap-in: mem_cgroup_try_charge() attempts to reserve a charge, reclaiming pages from the memcg if necessary. mem_cgroup_commit_charge() commits the page to the charge once it has a valid page->mapping and PageAnon() reliably tells the type. mem_cgroup_cancel_charge() aborts the transaction. This reduces the charge API and enables subsequent patches to drastically simplify uncharging. As pages need to be committed after rmap is established but before they are added to the LRU, page_add_new_anon_rmap() must stop doing LRU additions again. Revive lru_cache_add_active_or_unevictable(). [hughd@google.com: fix shmem_unuse] [hughd@google.com: Add comments on the private use of -EAGAIN] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Hugh Dickins <hughd@google.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:20 +00:00
cancel_charge(mc.from, mc.moved_charge);
mc.moved_charge = 0;
}
/* we must fixup refcnts and charges */
if (mc.moved_swap) {
/* uncharge swap account from the old cgroup */
if (!mem_cgroup_is_root(mc.from))
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
page_counter_uncharge(&mc.from->memsw, mc.moved_swap);
mem_cgroup_id_put_many(mc.from, mc.moved_swap);
/*
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
* we charged both to->memory and to->memsw, so we
* should uncharge to->memory.
*/
if (!mem_cgroup_is_root(mc.to))
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
page_counter_uncharge(&mc.to->memory, mc.moved_swap);
mem_cgroup_id_get_many(mc.to, mc.moved_swap);
css_put_many(&mc.to->css, mc.moved_swap);
mm: memcontrol: lockless page counters Memory is internally accounted in bytes, using spinlock-protected 64-bit counters, even though the smallest accounting delta is a page. The counter interface is also convoluted and does too many things. Introduce a new lockless word-sized page counter API, then change all memory accounting over to it. The translation from and to bytes then only happens when interfacing with userspace. The removed locking overhead is noticable when scaling beyond the per-cpu charge caches - on a 4-socket machine with 144-threads, the following test shows the performance differences of 288 memcgs concurrently running a page fault benchmark: vanilla: 18631648.500498 task-clock (msec) # 140.643 CPUs utilized ( +- 0.33% ) 1,380,638 context-switches # 0.074 K/sec ( +- 0.75% ) 24,390 cpu-migrations # 0.001 K/sec ( +- 8.44% ) 1,843,305,768 page-faults # 0.099 M/sec ( +- 0.00% ) 50,134,994,088,218 cycles # 2.691 GHz ( +- 0.33% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 8,049,712,224,651 instructions # 0.16 insns per cycle ( +- 0.04% ) 1,586,970,584,979 branches # 85.176 M/sec ( +- 0.05% ) 1,724,989,949 branch-misses # 0.11% of all branches ( +- 0.48% ) 132.474343877 seconds time elapsed ( +- 0.21% ) lockless: 12195979.037525 task-clock (msec) # 133.480 CPUs utilized ( +- 0.18% ) 832,850 context-switches # 0.068 K/sec ( +- 0.54% ) 15,624 cpu-migrations # 0.001 K/sec ( +- 10.17% ) 1,843,304,774 page-faults # 0.151 M/sec ( +- 0.00% ) 32,811,216,801,141 cycles # 2.690 GHz ( +- 0.18% ) <not supported> stalled-cycles-frontend <not supported> stalled-cycles-backend 9,999,265,091,727 instructions # 0.30 insns per cycle ( +- 0.10% ) 2,076,759,325,203 branches # 170.282 M/sec ( +- 0.12% ) 1,656,917,214 branch-misses # 0.08% of all branches ( +- 0.55% ) 91.369330729 seconds time elapsed ( +- 0.45% ) On top of improved scalability, this also gets rid of the icky long long types in the very heart of memcg, which is great for 32 bit and also makes the code a lot more readable. Notable differences between the old and new API: - res_counter_charge() and res_counter_charge_nofail() become page_counter_try_charge() and page_counter_charge() resp. to match the more common kernel naming scheme of try_do()/do() - res_counter_uncharge_until() is only ever used to cancel a local counter and never to uncharge bigger segments of a hierarchy, so it's replaced by the simpler page_counter_cancel() - res_counter_set_limit() is replaced by page_counter_limit(), which expects its callers to serialize against themselves - res_counter_memparse_write_strategy() is replaced by page_counter_limit(), which rounds down to the nearest page size - rather than up. This is more reasonable for explicitely requested hard upper limits. - to keep charging light-weight, page_counter_try_charge() charges speculatively, only to roll back if the result exceeds the limit. Because of this, a failing bigger charge can temporarily lock out smaller charges that would otherwise succeed. The error is bounded to the difference between the smallest and the biggest possible charge size, so for memcg, this means that a failing THP charge can send base page charges into reclaim upto 2MB (4MB) before the limit would have been reached. This should be acceptable. [akpm@linux-foundation.org: add includes for WARN_ON_ONCE and memparse] [akpm@linux-foundation.org: add includes for WARN_ON_ONCE, memparse, strncmp, and PAGE_SIZE] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: David Rientjes <rientjes@google.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:42:31 +00:00
mc.moved_swap = 0;
}
memcg: fix deadlock between cpuset and memcg Commit b1dd693e ("memcg: avoid deadlock between move charge and try_charge()") can cause another deadlock about mmap_sem on task migration if cpuset and memcg are mounted onto the same mount point. After the commit, cgroup_attach_task() has sequence like: cgroup_attach_task() ss->can_attach() cpuset_can_attach() mem_cgroup_can_attach() down_read(&mmap_sem) (1) ss->attach() cpuset_attach() mpol_rebind_mm() down_write(&mmap_sem) (2) up_write(&mmap_sem) cpuset_migrate_mm() do_migrate_pages() down_read(&mmap_sem) up_read(&mmap_sem) mem_cgroup_move_task() mem_cgroup_clear_mc() up_read(&mmap_sem) We can cause deadlock at (2) because we've already aquire the mmap_sem at (1). But the commit itself is necessary to fix deadlocks which have existed before the commit like: Ex.1) move charge | try charge --------------------------------------+------------------------------ mem_cgroup_can_attach() | down_write(&mmap_sem) mc.moving_task = current | .. mem_cgroup_precharge_mc() | __mem_cgroup_try_charge() mem_cgroup_count_precharge() | prepare_to_wait() down_read(&mmap_sem) | if (mc.moving_task) -> cannot aquire the lock | -> true | schedule() | -> move charge should wake it up Ex.2) move charge | try charge --------------------------------------+------------------------------ mem_cgroup_can_attach() | mc.moving_task = current | mem_cgroup_precharge_mc() | mem_cgroup_count_precharge() | down_read(&mmap_sem) | .. | up_read(&mmap_sem) | | down_write(&mmap_sem) mem_cgroup_move_task() | .. mem_cgroup_move_charge() | __mem_cgroup_try_charge() down_read(&mmap_sem) | prepare_to_wait() -> cannot aquire the lock | if (mc.moving_task) | -> true | schedule() | -> move charge should wake it up This patch fixes all of these problems by: 1. revert the commit. 2. To fix the Ex.1, we set mc.moving_task after mem_cgroup_count_precharge() has released the mmap_sem. 3. To fix the Ex.2, we use down_read_trylock() instead of down_read() in mem_cgroup_move_charge() and, if it has failed to aquire the lock, cancel all extra charges, wake up all waiters, and retry trylock. Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Reported-by: Ben Blum <bblum@andrew.cmu.edu> Cc: Miao Xie <miaox@cn.fujitsu.com> Cc: David Rientjes <rientjes@google.com> Cc: Paul Menage <menage@google.com> Cc: Hiroyuki Kamezawa <kamezawa.hiroyuki@gmail.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-01-13 23:47:41 +00:00
memcg_oom_recover(from);
memcg_oom_recover(to);
wake_up_all(&mc.waitq);
}
static void mem_cgroup_clear_mc(void)
{
memcg: relocate charge moving from ->attach to ->post_attach Hello, So, this ended up a lot simpler than I originally expected. I tested it lightly and it seems to work fine. Petr, can you please test these two patches w/o the lru drain drop patch and see whether the problem is gone? Thanks. ------ 8< ------ If charge moving is used, memcg performs relabeling of the affected pages from its ->attach callback which is called under both cgroup_threadgroup_rwsem and thus can't create new kthreads. This is fragile as various operations may depend on workqueues making forward progress which relies on the ability to create new kthreads. There's no reason to perform charge moving from ->attach which is deep in the task migration path. Move it to ->post_attach which is called after the actual migration is finished and cgroup_threadgroup_rwsem is dropped. * move_charge_struct->mm is added and ->can_attach is now responsible for pinning and recording the target mm. mem_cgroup_clear_mc() is updated accordingly. This also simplifies mem_cgroup_move_task(). * mem_cgroup_move_task() is now called from ->post_attach instead of ->attach. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Debugged-and-tested-by: Petr Mladek <pmladek@suse.com> Reported-by: Cyril Hrubis <chrubis@suse.cz> Reported-by: Johannes Weiner <hannes@cmpxchg.org> Fixes: 1ed1328792ff ("sched, cgroup: replace signal_struct->group_rwsem with a global percpu_rwsem") Cc: <stable@vger.kernel.org> # 4.4+
2016-04-21 23:09:02 +00:00
struct mm_struct *mm = mc.mm;
memcg: fix deadlock between cpuset and memcg Commit b1dd693e ("memcg: avoid deadlock between move charge and try_charge()") can cause another deadlock about mmap_sem on task migration if cpuset and memcg are mounted onto the same mount point. After the commit, cgroup_attach_task() has sequence like: cgroup_attach_task() ss->can_attach() cpuset_can_attach() mem_cgroup_can_attach() down_read(&mmap_sem) (1) ss->attach() cpuset_attach() mpol_rebind_mm() down_write(&mmap_sem) (2) up_write(&mmap_sem) cpuset_migrate_mm() do_migrate_pages() down_read(&mmap_sem) up_read(&mmap_sem) mem_cgroup_move_task() mem_cgroup_clear_mc() up_read(&mmap_sem) We can cause deadlock at (2) because we've already aquire the mmap_sem at (1). But the commit itself is necessary to fix deadlocks which have existed before the commit like: Ex.1) move charge | try charge --------------------------------------+------------------------------ mem_cgroup_can_attach() | down_write(&mmap_sem) mc.moving_task = current | .. mem_cgroup_precharge_mc() | __mem_cgroup_try_charge() mem_cgroup_count_precharge() | prepare_to_wait() down_read(&mmap_sem) | if (mc.moving_task) -> cannot aquire the lock | -> true | schedule() | -> move charge should wake it up Ex.2) move charge | try charge --------------------------------------+------------------------------ mem_cgroup_can_attach() | mc.moving_task = current | mem_cgroup_precharge_mc() | mem_cgroup_count_precharge() | down_read(&mmap_sem) | .. | up_read(&mmap_sem) | | down_write(&mmap_sem) mem_cgroup_move_task() | .. mem_cgroup_move_charge() | __mem_cgroup_try_charge() down_read(&mmap_sem) | prepare_to_wait() -> cannot aquire the lock | if (mc.moving_task) | -> true | schedule() | -> move charge should wake it up This patch fixes all of these problems by: 1. revert the commit. 2. To fix the Ex.1, we set mc.moving_task after mem_cgroup_count_precharge() has released the mmap_sem. 3. To fix the Ex.2, we use down_read_trylock() instead of down_read() in mem_cgroup_move_charge() and, if it has failed to aquire the lock, cancel all extra charges, wake up all waiters, and retry trylock. Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Reported-by: Ben Blum <bblum@andrew.cmu.edu> Cc: Miao Xie <miaox@cn.fujitsu.com> Cc: David Rientjes <rientjes@google.com> Cc: Paul Menage <menage@google.com> Cc: Hiroyuki Kamezawa <kamezawa.hiroyuki@gmail.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-01-13 23:47:41 +00:00
/*
* we must clear moving_task before waking up waiters at the end of
* task migration.
*/
mc.moving_task = NULL;
__mem_cgroup_clear_mc();
spin_lock(&mc.lock);
mc.from = NULL;
mc.to = NULL;
memcg: relocate charge moving from ->attach to ->post_attach Hello, So, this ended up a lot simpler than I originally expected. I tested it lightly and it seems to work fine. Petr, can you please test these two patches w/o the lru drain drop patch and see whether the problem is gone? Thanks. ------ 8< ------ If charge moving is used, memcg performs relabeling of the affected pages from its ->attach callback which is called under both cgroup_threadgroup_rwsem and thus can't create new kthreads. This is fragile as various operations may depend on workqueues making forward progress which relies on the ability to create new kthreads. There's no reason to perform charge moving from ->attach which is deep in the task migration path. Move it to ->post_attach which is called after the actual migration is finished and cgroup_threadgroup_rwsem is dropped. * move_charge_struct->mm is added and ->can_attach is now responsible for pinning and recording the target mm. mem_cgroup_clear_mc() is updated accordingly. This also simplifies mem_cgroup_move_task(). * mem_cgroup_move_task() is now called from ->post_attach instead of ->attach. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Debugged-and-tested-by: Petr Mladek <pmladek@suse.com> Reported-by: Cyril Hrubis <chrubis@suse.cz> Reported-by: Johannes Weiner <hannes@cmpxchg.org> Fixes: 1ed1328792ff ("sched, cgroup: replace signal_struct->group_rwsem with a global percpu_rwsem") Cc: <stable@vger.kernel.org> # 4.4+
2016-04-21 23:09:02 +00:00
mc.mm = NULL;
spin_unlock(&mc.lock);
memcg: relocate charge moving from ->attach to ->post_attach Hello, So, this ended up a lot simpler than I originally expected. I tested it lightly and it seems to work fine. Petr, can you please test these two patches w/o the lru drain drop patch and see whether the problem is gone? Thanks. ------ 8< ------ If charge moving is used, memcg performs relabeling of the affected pages from its ->attach callback which is called under both cgroup_threadgroup_rwsem and thus can't create new kthreads. This is fragile as various operations may depend on workqueues making forward progress which relies on the ability to create new kthreads. There's no reason to perform charge moving from ->attach which is deep in the task migration path. Move it to ->post_attach which is called after the actual migration is finished and cgroup_threadgroup_rwsem is dropped. * move_charge_struct->mm is added and ->can_attach is now responsible for pinning and recording the target mm. mem_cgroup_clear_mc() is updated accordingly. This also simplifies mem_cgroup_move_task(). * mem_cgroup_move_task() is now called from ->post_attach instead of ->attach. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Debugged-and-tested-by: Petr Mladek <pmladek@suse.com> Reported-by: Cyril Hrubis <chrubis@suse.cz> Reported-by: Johannes Weiner <hannes@cmpxchg.org> Fixes: 1ed1328792ff ("sched, cgroup: replace signal_struct->group_rwsem with a global percpu_rwsem") Cc: <stable@vger.kernel.org> # 4.4+
2016-04-21 23:09:02 +00:00
mmput(mm);
}
cgroup: fix handling of multi-destination migration from subtree_control enabling Consider the following v2 hierarchy. P0 (+memory) --- P1 (-memory) --- A \- B P0 has memory enabled in its subtree_control while P1 doesn't. If both A and B contain processes, they would belong to the memory css of P1. Now if memory is enabled on P1's subtree_control, memory csses should be created on both A and B and A's processes should be moved to the former and B's processes the latter. IOW, enabling controllers can cause atomic migrations into different csses. The core cgroup migration logic has been updated accordingly but the controller migration methods haven't and still assume that all tasks migrate to a single target css; furthermore, the methods were fed the css in which subtree_control was updated which is the parent of the target csses. pids controller depends on the migration methods to move charges and this made the controller attribute charges to the wrong csses often triggering the following warning by driving a counter negative. WARNING: CPU: 1 PID: 1 at kernel/cgroup_pids.c:97 pids_cancel.constprop.6+0x31/0x40() Modules linked in: CPU: 1 PID: 1 Comm: systemd Not tainted 4.4.0-rc1+ #29 ... ffffffff81f65382 ffff88007c043b90 ffffffff81551ffc 0000000000000000 ffff88007c043bc8 ffffffff810de202 ffff88007a752000 ffff88007a29ab00 ffff88007c043c80 ffff88007a1d8400 0000000000000001 ffff88007c043bd8 Call Trace: [<ffffffff81551ffc>] dump_stack+0x4e/0x82 [<ffffffff810de202>] warn_slowpath_common+0x82/0xc0 [<ffffffff810de2fa>] warn_slowpath_null+0x1a/0x20 [<ffffffff8118e031>] pids_cancel.constprop.6+0x31/0x40 [<ffffffff8118e0fd>] pids_can_attach+0x6d/0xf0 [<ffffffff81188a4c>] cgroup_taskset_migrate+0x6c/0x330 [<ffffffff81188e05>] cgroup_migrate+0xf5/0x190 [<ffffffff81189016>] cgroup_attach_task+0x176/0x200 [<ffffffff8118949d>] __cgroup_procs_write+0x2ad/0x460 [<ffffffff81189684>] cgroup_procs_write+0x14/0x20 [<ffffffff811854e5>] cgroup_file_write+0x35/0x1c0 [<ffffffff812e26f1>] kernfs_fop_write+0x141/0x190 [<ffffffff81265f88>] __vfs_write+0x28/0xe0 [<ffffffff812666fc>] vfs_write+0xac/0x1a0 [<ffffffff81267019>] SyS_write+0x49/0xb0 [<ffffffff81bcef32>] entry_SYSCALL_64_fastpath+0x12/0x76 This patch fixes the bug by removing @css parameter from the three migration methods, ->can_attach, ->cancel_attach() and ->attach() and updating cgroup_taskset iteration helpers also return the destination css in addition to the task being migrated. All controllers are updated accordingly. * Controllers which don't care whether there are one or multiple target csses can be converted trivially. cpu, io, freezer, perf, netclassid and netprio fall in this category. * cpuset's current implementation assumes that there's single source and destination and thus doesn't support v2 hierarchy already. The only change made by this patchset is how that single destination css is obtained. * memory migration path already doesn't do anything on v2. How the single destination css is obtained is updated and the prep stage of mem_cgroup_can_attach() is reordered to accomodate the change. * pids is the only controller which was affected by this bug. It now correctly handles multi-destination migrations and no longer causes counter underflow from incorrect accounting. Signed-off-by: Tejun Heo <tj@kernel.org> Reported-and-tested-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Aleksa Sarai <cyphar@cyphar.com>
2015-12-03 15:18:21 +00:00
static int mem_cgroup_can_attach(struct cgroup_taskset *tset)
{
cgroup: fix handling of multi-destination migration from subtree_control enabling Consider the following v2 hierarchy. P0 (+memory) --- P1 (-memory) --- A \- B P0 has memory enabled in its subtree_control while P1 doesn't. If both A and B contain processes, they would belong to the memory css of P1. Now if memory is enabled on P1's subtree_control, memory csses should be created on both A and B and A's processes should be moved to the former and B's processes the latter. IOW, enabling controllers can cause atomic migrations into different csses. The core cgroup migration logic has been updated accordingly but the controller migration methods haven't and still assume that all tasks migrate to a single target css; furthermore, the methods were fed the css in which subtree_control was updated which is the parent of the target csses. pids controller depends on the migration methods to move charges and this made the controller attribute charges to the wrong csses often triggering the following warning by driving a counter negative. WARNING: CPU: 1 PID: 1 at kernel/cgroup_pids.c:97 pids_cancel.constprop.6+0x31/0x40() Modules linked in: CPU: 1 PID: 1 Comm: systemd Not tainted 4.4.0-rc1+ #29 ... ffffffff81f65382 ffff88007c043b90 ffffffff81551ffc 0000000000000000 ffff88007c043bc8 ffffffff810de202 ffff88007a752000 ffff88007a29ab00 ffff88007c043c80 ffff88007a1d8400 0000000000000001 ffff88007c043bd8 Call Trace: [<ffffffff81551ffc>] dump_stack+0x4e/0x82 [<ffffffff810de202>] warn_slowpath_common+0x82/0xc0 [<ffffffff810de2fa>] warn_slowpath_null+0x1a/0x20 [<ffffffff8118e031>] pids_cancel.constprop.6+0x31/0x40 [<ffffffff8118e0fd>] pids_can_attach+0x6d/0xf0 [<ffffffff81188a4c>] cgroup_taskset_migrate+0x6c/0x330 [<ffffffff81188e05>] cgroup_migrate+0xf5/0x190 [<ffffffff81189016>] cgroup_attach_task+0x176/0x200 [<ffffffff8118949d>] __cgroup_procs_write+0x2ad/0x460 [<ffffffff81189684>] cgroup_procs_write+0x14/0x20 [<ffffffff811854e5>] cgroup_file_write+0x35/0x1c0 [<ffffffff812e26f1>] kernfs_fop_write+0x141/0x190 [<ffffffff81265f88>] __vfs_write+0x28/0xe0 [<ffffffff812666fc>] vfs_write+0xac/0x1a0 [<ffffffff81267019>] SyS_write+0x49/0xb0 [<ffffffff81bcef32>] entry_SYSCALL_64_fastpath+0x12/0x76 This patch fixes the bug by removing @css parameter from the three migration methods, ->can_attach, ->cancel_attach() and ->attach() and updating cgroup_taskset iteration helpers also return the destination css in addition to the task being migrated. All controllers are updated accordingly. * Controllers which don't care whether there are one or multiple target csses can be converted trivially. cpu, io, freezer, perf, netclassid and netprio fall in this category. * cpuset's current implementation assumes that there's single source and destination and thus doesn't support v2 hierarchy already. The only change made by this patchset is how that single destination css is obtained. * memory migration path already doesn't do anything on v2. How the single destination css is obtained is updated and the prep stage of mem_cgroup_can_attach() is reordered to accomodate the change. * pids is the only controller which was affected by this bug. It now correctly handles multi-destination migrations and no longer causes counter underflow from incorrect accounting. Signed-off-by: Tejun Heo <tj@kernel.org> Reported-and-tested-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Aleksa Sarai <cyphar@cyphar.com>
2015-12-03 15:18:21 +00:00
struct cgroup_subsys_state *css;
struct mem_cgroup *memcg = NULL; /* unneeded init to make gcc happy */
struct mem_cgroup *from;
struct task_struct *leader, *p;
struct mm_struct *mm;
unsigned long move_flags;
int ret = 0;
cgroup: fix handling of multi-destination migration from subtree_control enabling Consider the following v2 hierarchy. P0 (+memory) --- P1 (-memory) --- A \- B P0 has memory enabled in its subtree_control while P1 doesn't. If both A and B contain processes, they would belong to the memory css of P1. Now if memory is enabled on P1's subtree_control, memory csses should be created on both A and B and A's processes should be moved to the former and B's processes the latter. IOW, enabling controllers can cause atomic migrations into different csses. The core cgroup migration logic has been updated accordingly but the controller migration methods haven't and still assume that all tasks migrate to a single target css; furthermore, the methods were fed the css in which subtree_control was updated which is the parent of the target csses. pids controller depends on the migration methods to move charges and this made the controller attribute charges to the wrong csses often triggering the following warning by driving a counter negative. WARNING: CPU: 1 PID: 1 at kernel/cgroup_pids.c:97 pids_cancel.constprop.6+0x31/0x40() Modules linked in: CPU: 1 PID: 1 Comm: systemd Not tainted 4.4.0-rc1+ #29 ... ffffffff81f65382 ffff88007c043b90 ffffffff81551ffc 0000000000000000 ffff88007c043bc8 ffffffff810de202 ffff88007a752000 ffff88007a29ab00 ffff88007c043c80 ffff88007a1d8400 0000000000000001 ffff88007c043bd8 Call Trace: [<ffffffff81551ffc>] dump_stack+0x4e/0x82 [<ffffffff810de202>] warn_slowpath_common+0x82/0xc0 [<ffffffff810de2fa>] warn_slowpath_null+0x1a/0x20 [<ffffffff8118e031>] pids_cancel.constprop.6+0x31/0x40 [<ffffffff8118e0fd>] pids_can_attach+0x6d/0xf0 [<ffffffff81188a4c>] cgroup_taskset_migrate+0x6c/0x330 [<ffffffff81188e05>] cgroup_migrate+0xf5/0x190 [<ffffffff81189016>] cgroup_attach_task+0x176/0x200 [<ffffffff8118949d>] __cgroup_procs_write+0x2ad/0x460 [<ffffffff81189684>] cgroup_procs_write+0x14/0x20 [<ffffffff811854e5>] cgroup_file_write+0x35/0x1c0 [<ffffffff812e26f1>] kernfs_fop_write+0x141/0x190 [<ffffffff81265f88>] __vfs_write+0x28/0xe0 [<ffffffff812666fc>] vfs_write+0xac/0x1a0 [<ffffffff81267019>] SyS_write+0x49/0xb0 [<ffffffff81bcef32>] entry_SYSCALL_64_fastpath+0x12/0x76 This patch fixes the bug by removing @css parameter from the three migration methods, ->can_attach, ->cancel_attach() and ->attach() and updating cgroup_taskset iteration helpers also return the destination css in addition to the task being migrated. All controllers are updated accordingly. * Controllers which don't care whether there are one or multiple target csses can be converted trivially. cpu, io, freezer, perf, netclassid and netprio fall in this category. * cpuset's current implementation assumes that there's single source and destination and thus doesn't support v2 hierarchy already. The only change made by this patchset is how that single destination css is obtained. * memory migration path already doesn't do anything on v2. How the single destination css is obtained is updated and the prep stage of mem_cgroup_can_attach() is reordered to accomodate the change. * pids is the only controller which was affected by this bug. It now correctly handles multi-destination migrations and no longer causes counter underflow from incorrect accounting. Signed-off-by: Tejun Heo <tj@kernel.org> Reported-and-tested-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Aleksa Sarai <cyphar@cyphar.com>
2015-12-03 15:18:21 +00:00
/* charge immigration isn't supported on the default hierarchy */
if (cgroup_subsys_on_dfl(memory_cgrp_subsys))
return 0;
/*
* Multi-process migrations only happen on the default hierarchy
* where charge immigration is not used. Perform charge
* immigration if @tset contains a leader and whine if there are
* multiple.
*/
p = NULL;
cgroup: fix handling of multi-destination migration from subtree_control enabling Consider the following v2 hierarchy. P0 (+memory) --- P1 (-memory) --- A \- B P0 has memory enabled in its subtree_control while P1 doesn't. If both A and B contain processes, they would belong to the memory css of P1. Now if memory is enabled on P1's subtree_control, memory csses should be created on both A and B and A's processes should be moved to the former and B's processes the latter. IOW, enabling controllers can cause atomic migrations into different csses. The core cgroup migration logic has been updated accordingly but the controller migration methods haven't and still assume that all tasks migrate to a single target css; furthermore, the methods were fed the css in which subtree_control was updated which is the parent of the target csses. pids controller depends on the migration methods to move charges and this made the controller attribute charges to the wrong csses often triggering the following warning by driving a counter negative. WARNING: CPU: 1 PID: 1 at kernel/cgroup_pids.c:97 pids_cancel.constprop.6+0x31/0x40() Modules linked in: CPU: 1 PID: 1 Comm: systemd Not tainted 4.4.0-rc1+ #29 ... ffffffff81f65382 ffff88007c043b90 ffffffff81551ffc 0000000000000000 ffff88007c043bc8 ffffffff810de202 ffff88007a752000 ffff88007a29ab00 ffff88007c043c80 ffff88007a1d8400 0000000000000001 ffff88007c043bd8 Call Trace: [<ffffffff81551ffc>] dump_stack+0x4e/0x82 [<ffffffff810de202>] warn_slowpath_common+0x82/0xc0 [<ffffffff810de2fa>] warn_slowpath_null+0x1a/0x20 [<ffffffff8118e031>] pids_cancel.constprop.6+0x31/0x40 [<ffffffff8118e0fd>] pids_can_attach+0x6d/0xf0 [<ffffffff81188a4c>] cgroup_taskset_migrate+0x6c/0x330 [<ffffffff81188e05>] cgroup_migrate+0xf5/0x190 [<ffffffff81189016>] cgroup_attach_task+0x176/0x200 [<ffffffff8118949d>] __cgroup_procs_write+0x2ad/0x460 [<ffffffff81189684>] cgroup_procs_write+0x14/0x20 [<ffffffff811854e5>] cgroup_file_write+0x35/0x1c0 [<ffffffff812e26f1>] kernfs_fop_write+0x141/0x190 [<ffffffff81265f88>] __vfs_write+0x28/0xe0 [<ffffffff812666fc>] vfs_write+0xac/0x1a0 [<ffffffff81267019>] SyS_write+0x49/0xb0 [<ffffffff81bcef32>] entry_SYSCALL_64_fastpath+0x12/0x76 This patch fixes the bug by removing @css parameter from the three migration methods, ->can_attach, ->cancel_attach() and ->attach() and updating cgroup_taskset iteration helpers also return the destination css in addition to the task being migrated. All controllers are updated accordingly. * Controllers which don't care whether there are one or multiple target csses can be converted trivially. cpu, io, freezer, perf, netclassid and netprio fall in this category. * cpuset's current implementation assumes that there's single source and destination and thus doesn't support v2 hierarchy already. The only change made by this patchset is how that single destination css is obtained. * memory migration path already doesn't do anything on v2. How the single destination css is obtained is updated and the prep stage of mem_cgroup_can_attach() is reordered to accomodate the change. * pids is the only controller which was affected by this bug. It now correctly handles multi-destination migrations and no longer causes counter underflow from incorrect accounting. Signed-off-by: Tejun Heo <tj@kernel.org> Reported-and-tested-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Aleksa Sarai <cyphar@cyphar.com>
2015-12-03 15:18:21 +00:00
cgroup_taskset_for_each_leader(leader, css, tset) {
WARN_ON_ONCE(p);
p = leader;
cgroup: fix handling of multi-destination migration from subtree_control enabling Consider the following v2 hierarchy. P0 (+memory) --- P1 (-memory) --- A \- B P0 has memory enabled in its subtree_control while P1 doesn't. If both A and B contain processes, they would belong to the memory css of P1. Now if memory is enabled on P1's subtree_control, memory csses should be created on both A and B and A's processes should be moved to the former and B's processes the latter. IOW, enabling controllers can cause atomic migrations into different csses. The core cgroup migration logic has been updated accordingly but the controller migration methods haven't and still assume that all tasks migrate to a single target css; furthermore, the methods were fed the css in which subtree_control was updated which is the parent of the target csses. pids controller depends on the migration methods to move charges and this made the controller attribute charges to the wrong csses often triggering the following warning by driving a counter negative. WARNING: CPU: 1 PID: 1 at kernel/cgroup_pids.c:97 pids_cancel.constprop.6+0x31/0x40() Modules linked in: CPU: 1 PID: 1 Comm: systemd Not tainted 4.4.0-rc1+ #29 ... ffffffff81f65382 ffff88007c043b90 ffffffff81551ffc 0000000000000000 ffff88007c043bc8 ffffffff810de202 ffff88007a752000 ffff88007a29ab00 ffff88007c043c80 ffff88007a1d8400 0000000000000001 ffff88007c043bd8 Call Trace: [<ffffffff81551ffc>] dump_stack+0x4e/0x82 [<ffffffff810de202>] warn_slowpath_common+0x82/0xc0 [<ffffffff810de2fa>] warn_slowpath_null+0x1a/0x20 [<ffffffff8118e031>] pids_cancel.constprop.6+0x31/0x40 [<ffffffff8118e0fd>] pids_can_attach+0x6d/0xf0 [<ffffffff81188a4c>] cgroup_taskset_migrate+0x6c/0x330 [<ffffffff81188e05>] cgroup_migrate+0xf5/0x190 [<ffffffff81189016>] cgroup_attach_task+0x176/0x200 [<ffffffff8118949d>] __cgroup_procs_write+0x2ad/0x460 [<ffffffff81189684>] cgroup_procs_write+0x14/0x20 [<ffffffff811854e5>] cgroup_file_write+0x35/0x1c0 [<ffffffff812e26f1>] kernfs_fop_write+0x141/0x190 [<ffffffff81265f88>] __vfs_write+0x28/0xe0 [<ffffffff812666fc>] vfs_write+0xac/0x1a0 [<ffffffff81267019>] SyS_write+0x49/0xb0 [<ffffffff81bcef32>] entry_SYSCALL_64_fastpath+0x12/0x76 This patch fixes the bug by removing @css parameter from the three migration methods, ->can_attach, ->cancel_attach() and ->attach() and updating cgroup_taskset iteration helpers also return the destination css in addition to the task being migrated. All controllers are updated accordingly. * Controllers which don't care whether there are one or multiple target csses can be converted trivially. cpu, io, freezer, perf, netclassid and netprio fall in this category. * cpuset's current implementation assumes that there's single source and destination and thus doesn't support v2 hierarchy already. The only change made by this patchset is how that single destination css is obtained. * memory migration path already doesn't do anything on v2. How the single destination css is obtained is updated and the prep stage of mem_cgroup_can_attach() is reordered to accomodate the change. * pids is the only controller which was affected by this bug. It now correctly handles multi-destination migrations and no longer causes counter underflow from incorrect accounting. Signed-off-by: Tejun Heo <tj@kernel.org> Reported-and-tested-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Aleksa Sarai <cyphar@cyphar.com>
2015-12-03 15:18:21 +00:00
memcg = mem_cgroup_from_css(css);
}
if (!p)
return 0;
cgroup: fix handling of multi-destination migration from subtree_control enabling Consider the following v2 hierarchy. P0 (+memory) --- P1 (-memory) --- A \- B P0 has memory enabled in its subtree_control while P1 doesn't. If both A and B contain processes, they would belong to the memory css of P1. Now if memory is enabled on P1's subtree_control, memory csses should be created on both A and B and A's processes should be moved to the former and B's processes the latter. IOW, enabling controllers can cause atomic migrations into different csses. The core cgroup migration logic has been updated accordingly but the controller migration methods haven't and still assume that all tasks migrate to a single target css; furthermore, the methods were fed the css in which subtree_control was updated which is the parent of the target csses. pids controller depends on the migration methods to move charges and this made the controller attribute charges to the wrong csses often triggering the following warning by driving a counter negative. WARNING: CPU: 1 PID: 1 at kernel/cgroup_pids.c:97 pids_cancel.constprop.6+0x31/0x40() Modules linked in: CPU: 1 PID: 1 Comm: systemd Not tainted 4.4.0-rc1+ #29 ... ffffffff81f65382 ffff88007c043b90 ffffffff81551ffc 0000000000000000 ffff88007c043bc8 ffffffff810de202 ffff88007a752000 ffff88007a29ab00 ffff88007c043c80 ffff88007a1d8400 0000000000000001 ffff88007c043bd8 Call Trace: [<ffffffff81551ffc>] dump_stack+0x4e/0x82 [<ffffffff810de202>] warn_slowpath_common+0x82/0xc0 [<ffffffff810de2fa>] warn_slowpath_null+0x1a/0x20 [<ffffffff8118e031>] pids_cancel.constprop.6+0x31/0x40 [<ffffffff8118e0fd>] pids_can_attach+0x6d/0xf0 [<ffffffff81188a4c>] cgroup_taskset_migrate+0x6c/0x330 [<ffffffff81188e05>] cgroup_migrate+0xf5/0x190 [<ffffffff81189016>] cgroup_attach_task+0x176/0x200 [<ffffffff8118949d>] __cgroup_procs_write+0x2ad/0x460 [<ffffffff81189684>] cgroup_procs_write+0x14/0x20 [<ffffffff811854e5>] cgroup_file_write+0x35/0x1c0 [<ffffffff812e26f1>] kernfs_fop_write+0x141/0x190 [<ffffffff81265f88>] __vfs_write+0x28/0xe0 [<ffffffff812666fc>] vfs_write+0xac/0x1a0 [<ffffffff81267019>] SyS_write+0x49/0xb0 [<ffffffff81bcef32>] entry_SYSCALL_64_fastpath+0x12/0x76 This patch fixes the bug by removing @css parameter from the three migration methods, ->can_attach, ->cancel_attach() and ->attach() and updating cgroup_taskset iteration helpers also return the destination css in addition to the task being migrated. All controllers are updated accordingly. * Controllers which don't care whether there are one or multiple target csses can be converted trivially. cpu, io, freezer, perf, netclassid and netprio fall in this category. * cpuset's current implementation assumes that there's single source and destination and thus doesn't support v2 hierarchy already. The only change made by this patchset is how that single destination css is obtained. * memory migration path already doesn't do anything on v2. How the single destination css is obtained is updated and the prep stage of mem_cgroup_can_attach() is reordered to accomodate the change. * pids is the only controller which was affected by this bug. It now correctly handles multi-destination migrations and no longer causes counter underflow from incorrect accounting. Signed-off-by: Tejun Heo <tj@kernel.org> Reported-and-tested-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Aleksa Sarai <cyphar@cyphar.com>
2015-12-03 15:18:21 +00:00
/*
* We are now commited to this value whatever it is. Changes in this
* tunable will only affect upcoming migrations, not the current one.
* So we need to save it, and keep it going.
*/
move_flags = READ_ONCE(memcg->move_charge_at_immigrate);
if (!move_flags)
return 0;
from = mem_cgroup_from_task(p);
VM_BUG_ON(from == memcg);
mm = get_task_mm(p);
if (!mm)
return 0;
/* We move charges only when we move a owner of the mm */
if (mm->owner == p) {
VM_BUG_ON(mc.from);
VM_BUG_ON(mc.to);
VM_BUG_ON(mc.precharge);
VM_BUG_ON(mc.moved_charge);
VM_BUG_ON(mc.moved_swap);
spin_lock(&mc.lock);
memcg: relocate charge moving from ->attach to ->post_attach Hello, So, this ended up a lot simpler than I originally expected. I tested it lightly and it seems to work fine. Petr, can you please test these two patches w/o the lru drain drop patch and see whether the problem is gone? Thanks. ------ 8< ------ If charge moving is used, memcg performs relabeling of the affected pages from its ->attach callback which is called under both cgroup_threadgroup_rwsem and thus can't create new kthreads. This is fragile as various operations may depend on workqueues making forward progress which relies on the ability to create new kthreads. There's no reason to perform charge moving from ->attach which is deep in the task migration path. Move it to ->post_attach which is called after the actual migration is finished and cgroup_threadgroup_rwsem is dropped. * move_charge_struct->mm is added and ->can_attach is now responsible for pinning and recording the target mm. mem_cgroup_clear_mc() is updated accordingly. This also simplifies mem_cgroup_move_task(). * mem_cgroup_move_task() is now called from ->post_attach instead of ->attach. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Debugged-and-tested-by: Petr Mladek <pmladek@suse.com> Reported-by: Cyril Hrubis <chrubis@suse.cz> Reported-by: Johannes Weiner <hannes@cmpxchg.org> Fixes: 1ed1328792ff ("sched, cgroup: replace signal_struct->group_rwsem with a global percpu_rwsem") Cc: <stable@vger.kernel.org> # 4.4+
2016-04-21 23:09:02 +00:00
mc.mm = mm;
mc.from = from;
mc.to = memcg;
mc.flags = move_flags;
spin_unlock(&mc.lock);
/* We set mc.moving_task later */
ret = mem_cgroup_precharge_mc(mm);
if (ret)
mem_cgroup_clear_mc();
memcg: relocate charge moving from ->attach to ->post_attach Hello, So, this ended up a lot simpler than I originally expected. I tested it lightly and it seems to work fine. Petr, can you please test these two patches w/o the lru drain drop patch and see whether the problem is gone? Thanks. ------ 8< ------ If charge moving is used, memcg performs relabeling of the affected pages from its ->attach callback which is called under both cgroup_threadgroup_rwsem and thus can't create new kthreads. This is fragile as various operations may depend on workqueues making forward progress which relies on the ability to create new kthreads. There's no reason to perform charge moving from ->attach which is deep in the task migration path. Move it to ->post_attach which is called after the actual migration is finished and cgroup_threadgroup_rwsem is dropped. * move_charge_struct->mm is added and ->can_attach is now responsible for pinning and recording the target mm. mem_cgroup_clear_mc() is updated accordingly. This also simplifies mem_cgroup_move_task(). * mem_cgroup_move_task() is now called from ->post_attach instead of ->attach. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Debugged-and-tested-by: Petr Mladek <pmladek@suse.com> Reported-by: Cyril Hrubis <chrubis@suse.cz> Reported-by: Johannes Weiner <hannes@cmpxchg.org> Fixes: 1ed1328792ff ("sched, cgroup: replace signal_struct->group_rwsem with a global percpu_rwsem") Cc: <stable@vger.kernel.org> # 4.4+
2016-04-21 23:09:02 +00:00
} else {
mmput(mm);
}
return ret;
}
cgroup: fix handling of multi-destination migration from subtree_control enabling Consider the following v2 hierarchy. P0 (+memory) --- P1 (-memory) --- A \- B P0 has memory enabled in its subtree_control while P1 doesn't. If both A and B contain processes, they would belong to the memory css of P1. Now if memory is enabled on P1's subtree_control, memory csses should be created on both A and B and A's processes should be moved to the former and B's processes the latter. IOW, enabling controllers can cause atomic migrations into different csses. The core cgroup migration logic has been updated accordingly but the controller migration methods haven't and still assume that all tasks migrate to a single target css; furthermore, the methods were fed the css in which subtree_control was updated which is the parent of the target csses. pids controller depends on the migration methods to move charges and this made the controller attribute charges to the wrong csses often triggering the following warning by driving a counter negative. WARNING: CPU: 1 PID: 1 at kernel/cgroup_pids.c:97 pids_cancel.constprop.6+0x31/0x40() Modules linked in: CPU: 1 PID: 1 Comm: systemd Not tainted 4.4.0-rc1+ #29 ... ffffffff81f65382 ffff88007c043b90 ffffffff81551ffc 0000000000000000 ffff88007c043bc8 ffffffff810de202 ffff88007a752000 ffff88007a29ab00 ffff88007c043c80 ffff88007a1d8400 0000000000000001 ffff88007c043bd8 Call Trace: [<ffffffff81551ffc>] dump_stack+0x4e/0x82 [<ffffffff810de202>] warn_slowpath_common+0x82/0xc0 [<ffffffff810de2fa>] warn_slowpath_null+0x1a/0x20 [<ffffffff8118e031>] pids_cancel.constprop.6+0x31/0x40 [<ffffffff8118e0fd>] pids_can_attach+0x6d/0xf0 [<ffffffff81188a4c>] cgroup_taskset_migrate+0x6c/0x330 [<ffffffff81188e05>] cgroup_migrate+0xf5/0x190 [<ffffffff81189016>] cgroup_attach_task+0x176/0x200 [<ffffffff8118949d>] __cgroup_procs_write+0x2ad/0x460 [<ffffffff81189684>] cgroup_procs_write+0x14/0x20 [<ffffffff811854e5>] cgroup_file_write+0x35/0x1c0 [<ffffffff812e26f1>] kernfs_fop_write+0x141/0x190 [<ffffffff81265f88>] __vfs_write+0x28/0xe0 [<ffffffff812666fc>] vfs_write+0xac/0x1a0 [<ffffffff81267019>] SyS_write+0x49/0xb0 [<ffffffff81bcef32>] entry_SYSCALL_64_fastpath+0x12/0x76 This patch fixes the bug by removing @css parameter from the three migration methods, ->can_attach, ->cancel_attach() and ->attach() and updating cgroup_taskset iteration helpers also return the destination css in addition to the task being migrated. All controllers are updated accordingly. * Controllers which don't care whether there are one or multiple target csses can be converted trivially. cpu, io, freezer, perf, netclassid and netprio fall in this category. * cpuset's current implementation assumes that there's single source and destination and thus doesn't support v2 hierarchy already. The only change made by this patchset is how that single destination css is obtained. * memory migration path already doesn't do anything on v2. How the single destination css is obtained is updated and the prep stage of mem_cgroup_can_attach() is reordered to accomodate the change. * pids is the only controller which was affected by this bug. It now correctly handles multi-destination migrations and no longer causes counter underflow from incorrect accounting. Signed-off-by: Tejun Heo <tj@kernel.org> Reported-and-tested-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Aleksa Sarai <cyphar@cyphar.com>
2015-12-03 15:18:21 +00:00
static void mem_cgroup_cancel_attach(struct cgroup_taskset *tset)
{
if (mc.to)
mem_cgroup_clear_mc();
}
static int mem_cgroup_move_charge_pte_range(pmd_t *pmd,
unsigned long addr, unsigned long end,
struct mm_walk *walk)
{
int ret = 0;
struct vm_area_struct *vma = walk->vma;
pte_t *pte;
spinlock_t *ptl;
enum mc_target_type target_type;
union mc_target target;
struct page *page;
ptl = pmd_trans_huge_lock(pmd, vma);
if (ptl) {
if (mc.precharge < HPAGE_PMD_NR) {
mm, thp: change pmd_trans_huge_lock() to return taken lock With split ptlock it's important to know which lock pmd_trans_huge_lock() took. This patch adds one more parameter to the function to return the lock. In most places migration to new api is trivial. Exception is move_huge_pmd(): we need to take two locks if pmd tables are different. Signed-off-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Tested-by: Alex Thorlton <athorlton@sgi.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: "Eric W . Biederman" <ebiederm@xmission.com> Cc: "Paul E . McKenney" <paulmck@linux.vnet.ibm.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andi Kleen <ak@linux.intel.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Dave Hansen <dave.hansen@intel.com> Cc: Dave Jones <davej@redhat.com> Cc: David Howells <dhowells@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Kees Cook <keescook@chromium.org> Cc: Mel Gorman <mgorman@suse.de> Cc: Michael Kerrisk <mtk.manpages@gmail.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Rik van Riel <riel@redhat.com> Cc: Robin Holt <robinmholt@gmail.com> Cc: Sedat Dilek <sedat.dilek@gmail.com> Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-11-14 22:30:54 +00:00
spin_unlock(ptl);
return 0;
}
target_type = get_mctgt_type_thp(vma, addr, *pmd, &target);
if (target_type == MC_TARGET_PAGE) {
page = target.page;
if (!isolate_lru_page(page)) {
memcg: adjust to support new THP refcounting As with rmap, with new refcounting we cannot rely on PageTransHuge() to check if we need to charge size of huge page form the cgroup. We need to get information from caller to know whether it was mapped with PMD or PTE. We do uncharge when last reference on the page gone. At that point if we see PageTransHuge() it means we need to unchange whole huge page. The tricky part is partial unmap -- when we try to unmap part of huge page. We don't do a special handing of this situation, meaning we don't uncharge the part of huge page unless last user is gone or split_huge_page() is triggered. In case of cgroup memory pressure happens the partial unmapped page will be split through shrinker. This should be good enough. Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Tested-by: Sasha Levin <sasha.levin@oracle.com> Tested-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Acked-by: Vlastimil Babka <vbabka@suse.cz> Acked-by: Jerome Marchand <jmarchan@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Hugh Dickins <hughd@google.com> Cc: Dave Hansen <dave.hansen@intel.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Rik van Riel <riel@redhat.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Cc: Steve Capper <steve.capper@linaro.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-16 00:52:20 +00:00
if (!mem_cgroup_move_account(page, true,
mm: embed the memcg pointer directly into struct page Memory cgroups used to have 5 per-page pointers. To allow users to disable that amount of overhead during runtime, those pointers were allocated in a separate array, with a translation layer between them and struct page. There is now only one page pointer remaining: the memcg pointer, that indicates which cgroup the page is associated with when charged. The complexity of runtime allocation and the runtime translation overhead is no longer justified to save that *potential* 0.19% of memory. With CONFIG_SLUB, page->mem_cgroup actually sits in the doubleword padding after the page->private member and doesn't even increase struct page, and then this patch actually saves space. Remaining users that care can still compile their kernels without CONFIG_MEMCG. text data bss dec hex filename 8828345 1725264 983040 11536649 b00909 vmlinux.old 8827425 1725264 966656 11519345 afc571 vmlinux.new [mhocko@suse.cz: update Documentation/cgroups/memory.txt] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: David S. Miller <davem@davemloft.net> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Acked-by: Konstantin Khlebnikov <koct9i@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:44:52 +00:00
mc.from, mc.to)) {
mc.precharge -= HPAGE_PMD_NR;
mc.moved_charge += HPAGE_PMD_NR;
}
putback_lru_page(page);
}
put_page(page);
} else if (target_type == MC_TARGET_DEVICE) {
page = target.page;
if (!mem_cgroup_move_account(page, true,
mc.from, mc.to)) {
mc.precharge -= HPAGE_PMD_NR;
mc.moved_charge += HPAGE_PMD_NR;
}
put_page(page);
}
mm, thp: change pmd_trans_huge_lock() to return taken lock With split ptlock it's important to know which lock pmd_trans_huge_lock() took. This patch adds one more parameter to the function to return the lock. In most places migration to new api is trivial. Exception is move_huge_pmd(): we need to take two locks if pmd tables are different. Signed-off-by: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Tested-by: Alex Thorlton <athorlton@sgi.com> Cc: Ingo Molnar <mingo@redhat.com> Cc: "Eric W . Biederman" <ebiederm@xmission.com> Cc: "Paul E . McKenney" <paulmck@linux.vnet.ibm.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andi Kleen <ak@linux.intel.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Dave Hansen <dave.hansen@intel.com> Cc: Dave Jones <davej@redhat.com> Cc: David Howells <dhowells@redhat.com> Cc: Frederic Weisbecker <fweisbec@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Kees Cook <keescook@chromium.org> Cc: Mel Gorman <mgorman@suse.de> Cc: Michael Kerrisk <mtk.manpages@gmail.com> Cc: Oleg Nesterov <oleg@redhat.com> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Rik van Riel <riel@redhat.com> Cc: Robin Holt <robinmholt@gmail.com> Cc: Sedat Dilek <sedat.dilek@gmail.com> Cc: Srikar Dronamraju <srikar@linux.vnet.ibm.com> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Hugh Dickins <hughd@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2013-11-14 22:30:54 +00:00
spin_unlock(ptl);
mm: thp: fix pmd_bad() triggering in code paths holding mmap_sem read mode In some cases it may happen that pmd_none_or_clear_bad() is called with the mmap_sem hold in read mode. In those cases the huge page faults can allocate hugepmds under pmd_none_or_clear_bad() and that can trigger a false positive from pmd_bad() that will not like to see a pmd materializing as trans huge. It's not khugepaged causing the problem, khugepaged holds the mmap_sem in write mode (and all those sites must hold the mmap_sem in read mode to prevent pagetables to go away from under them, during code review it seems vm86 mode on 32bit kernels requires that too unless it's restricted to 1 thread per process or UP builds). The race is only with the huge pagefaults that can convert a pmd_none() into a pmd_trans_huge(). Effectively all these pmd_none_or_clear_bad() sites running with mmap_sem in read mode are somewhat speculative with the page faults, and the result is always undefined when they run simultaneously. This is probably why it wasn't common to run into this. For example if the madvise(MADV_DONTNEED) runs zap_page_range() shortly before the page fault, the hugepage will not be zapped, if the page fault runs first it will be zapped. Altering pmd_bad() not to error out if it finds hugepmds won't be enough to fix this, because zap_pmd_range would then proceed to call zap_pte_range (which would be incorrect if the pmd become a pmd_trans_huge()). The simplest way to fix this is to read the pmd in the local stack (regardless of what we read, no need of actual CPU barriers, only compiler barrier needed), and be sure it is not changing under the code that computes its value. Even if the real pmd is changing under the value we hold on the stack, we don't care. If we actually end up in zap_pte_range it means the pmd was not none already and it was not huge, and it can't become huge from under us (khugepaged locking explained above). All we need is to enforce that there is no way anymore that in a code path like below, pmd_trans_huge can be false, but pmd_none_or_clear_bad can run into a hugepmd. The overhead of a barrier() is just a compiler tweak and should not be measurable (I only added it for THP builds). I don't exclude different compiler versions may have prevented the race too by caching the value of *pmd on the stack (that hasn't been verified, but it wouldn't be impossible considering pmd_none_or_clear_bad, pmd_bad, pmd_trans_huge, pmd_none are all inlines and there's no external function called in between pmd_trans_huge and pmd_none_or_clear_bad). if (pmd_trans_huge(*pmd)) { if (next-addr != HPAGE_PMD_SIZE) { VM_BUG_ON(!rwsem_is_locked(&tlb->mm->mmap_sem)); split_huge_page_pmd(vma->vm_mm, pmd); } else if (zap_huge_pmd(tlb, vma, pmd, addr)) continue; /* fall through */ } if (pmd_none_or_clear_bad(pmd)) Because this race condition could be exercised without special privileges this was reported in CVE-2012-1179. The race was identified and fully explained by Ulrich who debugged it. I'm quoting his accurate explanation below, for reference. ====== start quote ======= mapcount 0 page_mapcount 1 kernel BUG at mm/huge_memory.c:1384! At some point prior to the panic, a "bad pmd ..." message similar to the following is logged on the console: mm/memory.c:145: bad pmd ffff8800376e1f98(80000000314000e7). The "bad pmd ..." message is logged by pmd_clear_bad() before it clears the page's PMD table entry. 143 void pmd_clear_bad(pmd_t *pmd) 144 { -> 145 pmd_ERROR(*pmd); 146 pmd_clear(pmd); 147 } After the PMD table entry has been cleared, there is an inconsistency between the actual number of PMD table entries that are mapping the page and the page's map count (_mapcount field in struct page). When the page is subsequently reclaimed, __split_huge_page() detects this inconsistency. 1381 if (mapcount != page_mapcount(page)) 1382 printk(KERN_ERR "mapcount %d page_mapcount %d\n", 1383 mapcount, page_mapcount(page)); -> 1384 BUG_ON(mapcount != page_mapcount(page)); The root cause of the problem is a race of two threads in a multithreaded process. Thread B incurs a page fault on a virtual address that has never been accessed (PMD entry is zero) while Thread A is executing an madvise() system call on a virtual address within the same 2 MB (huge page) range. virtual address space .---------------------. | | | | .-|---------------------| | | | | | |<-- B(fault) | | | 2 MB | |/////////////////////|-. huge < |/////////////////////| > A(range) page | |/////////////////////|-' | | | | | | '-|---------------------| | | | | '---------------------' - Thread A is executing an madvise(..., MADV_DONTNEED) system call on the virtual address range "A(range)" shown in the picture. sys_madvise // Acquire the semaphore in shared mode. down_read(&current->mm->mmap_sem) ... madvise_vma switch (behavior) case MADV_DONTNEED: madvise_dontneed zap_page_range unmap_vmas unmap_page_range zap_pud_range zap_pmd_range // // Assume that this huge page has never been accessed. // I.e. content of the PMD entry is zero (not mapped). // if (pmd_trans_huge(*pmd)) { // We don't get here due to the above assumption. } // // Assume that Thread B incurred a page fault and .---------> // sneaks in here as shown below. | // | if (pmd_none_or_clear_bad(pmd)) | { | if (unlikely(pmd_bad(*pmd))) | pmd_clear_bad | { | pmd_ERROR | // Log "bad pmd ..." message here. | pmd_clear | // Clear the page's PMD entry. | // Thread B incremented the map count | // in page_add_new_anon_rmap(), but | // now the page is no longer mapped | // by a PMD entry (-> inconsistency). | } | } | v - Thread B is handling a page fault on virtual address "B(fault)" shown in the picture. ... do_page_fault __do_page_fault // Acquire the semaphore in shared mode. down_read_trylock(&mm->mmap_sem) ... handle_mm_fault if (pmd_none(*pmd) && transparent_hugepage_enabled(vma)) // We get here due to the above assumption (PMD entry is zero). do_huge_pmd_anonymous_page alloc_hugepage_vma // Allocate a new transparent huge page here. ... __do_huge_pmd_anonymous_page ... spin_lock(&mm->page_table_lock) ... page_add_new_anon_rmap // Here we increment the page's map count (starts at -1). atomic_set(&page->_mapcount, 0) set_pmd_at // Here we set the page's PMD entry which will be cleared // when Thread A calls pmd_clear_bad(). ... spin_unlock(&mm->page_table_lock) The mmap_sem does not prevent the race because both threads are acquiring it in shared mode (down_read). Thread B holds the page_table_lock while the page's map count and PMD table entry are updated. However, Thread A does not synchronize on that lock. ====== end quote ======= [akpm@linux-foundation.org: checkpatch fixes] Reported-by: Ulrich Obergfell <uobergfe@redhat.com> Signed-off-by: Andrea Arcangeli <aarcange@redhat.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Mel Gorman <mgorman@suse.de> Cc: Hugh Dickins <hughd@google.com> Cc: Dave Jones <davej@redhat.com> Acked-by: Larry Woodman <lwoodman@redhat.com> Acked-by: Rik van Riel <riel@redhat.com> Cc: <stable@vger.kernel.org> [2.6.38+] Cc: Mark Salter <msalter@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2012-03-21 23:33:42 +00:00
return 0;
}
if (pmd_trans_unstable(pmd))
return 0;
retry:
pte = pte_offset_map_lock(vma->vm_mm, pmd, addr, &ptl);
for (; addr != end; addr += PAGE_SIZE) {
pte_t ptent = *(pte++);
bool device = false;
swp_entry_t ent;
if (!mc.precharge)
break;
switch (get_mctgt_type(vma, addr, ptent, &target)) {
case MC_TARGET_DEVICE:
device = true;
fallthrough;
case MC_TARGET_PAGE:
page = target.page;
mm: rework mapcount accounting to enable 4k mapping of THPs We're going to allow mapping of individual 4k pages of THP compound. It means we need to track mapcount on per small page basis. Straight-forward approach is to use ->_mapcount in all subpages to track how many time this subpage is mapped with PMDs or PTEs combined. But this is rather expensive: mapping or unmapping of a THP page with PMD would require HPAGE_PMD_NR atomic operations instead of single we have now. The idea is to store separately how many times the page was mapped as whole -- compound_mapcount. This frees up ->_mapcount in subpages to track PTE mapcount. We use the same approach as with compound page destructor and compound order to store compound_mapcount: use space in first tail page, ->mapping this time. Any time we map/unmap whole compound page (THP or hugetlb) -- we increment/decrement compound_mapcount. When we map part of compound page with PTE we operate on ->_mapcount of the subpage. page_mapcount() counts both: PTE and PMD mappings of the page. Basically, we have mapcount for a subpage spread over two counters. It makes tricky to detect when last mapcount for a page goes away. We introduced PageDoubleMap() for this. When we split THP PMD for the first time and there's other PMD mapping left we offset up ->_mapcount in all subpages by one and set PG_double_map on the compound page. These additional references go away with last compound_mapcount. This approach provides a way to detect when last mapcount goes away on per small page basis without introducing new overhead for most common cases. [akpm@linux-foundation.org: fix typo in comment] [mhocko@suse.com: ignore partial THP when moving task] Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Tested-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Acked-by: Jerome Marchand <jmarchan@redhat.com> Cc: Sasha Levin <sasha.levin@oracle.com> Cc: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Cc: Jerome Marchand <jmarchan@redhat.com> Cc: Vlastimil Babka <vbabka@suse.cz> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Hugh Dickins <hughd@google.com> Cc: Dave Hansen <dave.hansen@intel.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Rik van Riel <riel@redhat.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Cc: Steve Capper <steve.capper@linaro.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Michal Hocko <mhocko@suse.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-16 00:53:42 +00:00
/*
* We can have a part of the split pmd here. Moving it
* can be done but it would be too convoluted so simply
* ignore such a partial THP and keep it in original
* memcg. There should be somebody mapping the head.
*/
if (PageTransCompound(page))
goto put;
if (!device && isolate_lru_page(page))
goto put;
memcg: adjust to support new THP refcounting As with rmap, with new refcounting we cannot rely on PageTransHuge() to check if we need to charge size of huge page form the cgroup. We need to get information from caller to know whether it was mapped with PMD or PTE. We do uncharge when last reference on the page gone. At that point if we see PageTransHuge() it means we need to unchange whole huge page. The tricky part is partial unmap -- when we try to unmap part of huge page. We don't do a special handing of this situation, meaning we don't uncharge the part of huge page unless last user is gone or split_huge_page() is triggered. In case of cgroup memory pressure happens the partial unmapped page will be split through shrinker. This should be good enough. Signed-off-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> Tested-by: Sasha Levin <sasha.levin@oracle.com> Tested-by: Aneesh Kumar K.V <aneesh.kumar@linux.vnet.ibm.com> Acked-by: Vlastimil Babka <vbabka@suse.cz> Acked-by: Jerome Marchand <jmarchan@redhat.com> Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Hugh Dickins <hughd@google.com> Cc: Dave Hansen <dave.hansen@intel.com> Cc: Mel Gorman <mgorman@suse.de> Cc: Rik van Riel <riel@redhat.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Cc: Steve Capper <steve.capper@linaro.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.cz> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-16 00:52:20 +00:00
if (!mem_cgroup_move_account(page, false,
mc.from, mc.to)) {
mc.precharge--;
/* we uncharge from mc.from later. */
mc.moved_charge++;
}
if (!device)
putback_lru_page(page);
put: /* get_mctgt_type() gets the page */
put_page(page);
break;
case MC_TARGET_SWAP:
ent = target.ent;
if (!mem_cgroup_move_swap_account(ent, mc.from, mc.to)) {
mc.precharge--;
/* we fixup refcnts and charges later. */
mc.moved_swap++;
}
break;
default:
break;
}
}
pte_unmap_unlock(pte - 1, ptl);
cond_resched();
if (addr != end) {
/*
* We have consumed all precharges we got in can_attach().
* We try charge one by one, but don't do any additional
* charges to mc.to if we have failed in charge once in attach()
* phase.
*/
ret = mem_cgroup_do_precharge(1);
if (!ret)
goto retry;
}
return ret;
}
static const struct mm_walk_ops charge_walk_ops = {
.pmd_entry = mem_cgroup_move_charge_pte_range,
};
memcg: relocate charge moving from ->attach to ->post_attach Hello, So, this ended up a lot simpler than I originally expected. I tested it lightly and it seems to work fine. Petr, can you please test these two patches w/o the lru drain drop patch and see whether the problem is gone? Thanks. ------ 8< ------ If charge moving is used, memcg performs relabeling of the affected pages from its ->attach callback which is called under both cgroup_threadgroup_rwsem and thus can't create new kthreads. This is fragile as various operations may depend on workqueues making forward progress which relies on the ability to create new kthreads. There's no reason to perform charge moving from ->attach which is deep in the task migration path. Move it to ->post_attach which is called after the actual migration is finished and cgroup_threadgroup_rwsem is dropped. * move_charge_struct->mm is added and ->can_attach is now responsible for pinning and recording the target mm. mem_cgroup_clear_mc() is updated accordingly. This also simplifies mem_cgroup_move_task(). * mem_cgroup_move_task() is now called from ->post_attach instead of ->attach. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Debugged-and-tested-by: Petr Mladek <pmladek@suse.com> Reported-by: Cyril Hrubis <chrubis@suse.cz> Reported-by: Johannes Weiner <hannes@cmpxchg.org> Fixes: 1ed1328792ff ("sched, cgroup: replace signal_struct->group_rwsem with a global percpu_rwsem") Cc: <stable@vger.kernel.org> # 4.4+
2016-04-21 23:09:02 +00:00
static void mem_cgroup_move_charge(void)
{
lru_add_drain_all();
/*
* Signal lock_page_memcg() to take the memcg's move_lock
* while we're moving its pages to another memcg. Then wait
* for already started RCU-only updates to finish.
*/
atomic_inc(&mc.from->moving_account);
synchronize_rcu();
memcg: fix deadlock between cpuset and memcg Commit b1dd693e ("memcg: avoid deadlock between move charge and try_charge()") can cause another deadlock about mmap_sem on task migration if cpuset and memcg are mounted onto the same mount point. After the commit, cgroup_attach_task() has sequence like: cgroup_attach_task() ss->can_attach() cpuset_can_attach() mem_cgroup_can_attach() down_read(&mmap_sem) (1) ss->attach() cpuset_attach() mpol_rebind_mm() down_write(&mmap_sem) (2) up_write(&mmap_sem) cpuset_migrate_mm() do_migrate_pages() down_read(&mmap_sem) up_read(&mmap_sem) mem_cgroup_move_task() mem_cgroup_clear_mc() up_read(&mmap_sem) We can cause deadlock at (2) because we've already aquire the mmap_sem at (1). But the commit itself is necessary to fix deadlocks which have existed before the commit like: Ex.1) move charge | try charge --------------------------------------+------------------------------ mem_cgroup_can_attach() | down_write(&mmap_sem) mc.moving_task = current | .. mem_cgroup_precharge_mc() | __mem_cgroup_try_charge() mem_cgroup_count_precharge() | prepare_to_wait() down_read(&mmap_sem) | if (mc.moving_task) -> cannot aquire the lock | -> true | schedule() | -> move charge should wake it up Ex.2) move charge | try charge --------------------------------------+------------------------------ mem_cgroup_can_attach() | mc.moving_task = current | mem_cgroup_precharge_mc() | mem_cgroup_count_precharge() | down_read(&mmap_sem) | .. | up_read(&mmap_sem) | | down_write(&mmap_sem) mem_cgroup_move_task() | .. mem_cgroup_move_charge() | __mem_cgroup_try_charge() down_read(&mmap_sem) | prepare_to_wait() -> cannot aquire the lock | if (mc.moving_task) | -> true | schedule() | -> move charge should wake it up This patch fixes all of these problems by: 1. revert the commit. 2. To fix the Ex.1, we set mc.moving_task after mem_cgroup_count_precharge() has released the mmap_sem. 3. To fix the Ex.2, we use down_read_trylock() instead of down_read() in mem_cgroup_move_charge() and, if it has failed to aquire the lock, cancel all extra charges, wake up all waiters, and retry trylock. Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Reported-by: Ben Blum <bblum@andrew.cmu.edu> Cc: Miao Xie <miaox@cn.fujitsu.com> Cc: David Rientjes <rientjes@google.com> Cc: Paul Menage <menage@google.com> Cc: Hiroyuki Kamezawa <kamezawa.hiroyuki@gmail.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-01-13 23:47:41 +00:00
retry:
memcg: relocate charge moving from ->attach to ->post_attach Hello, So, this ended up a lot simpler than I originally expected. I tested it lightly and it seems to work fine. Petr, can you please test these two patches w/o the lru drain drop patch and see whether the problem is gone? Thanks. ------ 8< ------ If charge moving is used, memcg performs relabeling of the affected pages from its ->attach callback which is called under both cgroup_threadgroup_rwsem and thus can't create new kthreads. This is fragile as various operations may depend on workqueues making forward progress which relies on the ability to create new kthreads. There's no reason to perform charge moving from ->attach which is deep in the task migration path. Move it to ->post_attach which is called after the actual migration is finished and cgroup_threadgroup_rwsem is dropped. * move_charge_struct->mm is added and ->can_attach is now responsible for pinning and recording the target mm. mem_cgroup_clear_mc() is updated accordingly. This also simplifies mem_cgroup_move_task(). * mem_cgroup_move_task() is now called from ->post_attach instead of ->attach. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Debugged-and-tested-by: Petr Mladek <pmladek@suse.com> Reported-by: Cyril Hrubis <chrubis@suse.cz> Reported-by: Johannes Weiner <hannes@cmpxchg.org> Fixes: 1ed1328792ff ("sched, cgroup: replace signal_struct->group_rwsem with a global percpu_rwsem") Cc: <stable@vger.kernel.org> # 4.4+
2016-04-21 23:09:02 +00:00
if (unlikely(!down_read_trylock(&mc.mm->mmap_sem))) {
memcg: fix deadlock between cpuset and memcg Commit b1dd693e ("memcg: avoid deadlock between move charge and try_charge()") can cause another deadlock about mmap_sem on task migration if cpuset and memcg are mounted onto the same mount point. After the commit, cgroup_attach_task() has sequence like: cgroup_attach_task() ss->can_attach() cpuset_can_attach() mem_cgroup_can_attach() down_read(&mmap_sem) (1) ss->attach() cpuset_attach() mpol_rebind_mm() down_write(&mmap_sem) (2) up_write(&mmap_sem) cpuset_migrate_mm() do_migrate_pages() down_read(&mmap_sem) up_read(&mmap_sem) mem_cgroup_move_task() mem_cgroup_clear_mc() up_read(&mmap_sem) We can cause deadlock at (2) because we've already aquire the mmap_sem at (1). But the commit itself is necessary to fix deadlocks which have existed before the commit like: Ex.1) move charge | try charge --------------------------------------+------------------------------ mem_cgroup_can_attach() | down_write(&mmap_sem) mc.moving_task = current | .. mem_cgroup_precharge_mc() | __mem_cgroup_try_charge() mem_cgroup_count_precharge() | prepare_to_wait() down_read(&mmap_sem) | if (mc.moving_task) -> cannot aquire the lock | -> true | schedule() | -> move charge should wake it up Ex.2) move charge | try charge --------------------------------------+------------------------------ mem_cgroup_can_attach() | mc.moving_task = current | mem_cgroup_precharge_mc() | mem_cgroup_count_precharge() | down_read(&mmap_sem) | .. | up_read(&mmap_sem) | | down_write(&mmap_sem) mem_cgroup_move_task() | .. mem_cgroup_move_charge() | __mem_cgroup_try_charge() down_read(&mmap_sem) | prepare_to_wait() -> cannot aquire the lock | if (mc.moving_task) | -> true | schedule() | -> move charge should wake it up This patch fixes all of these problems by: 1. revert the commit. 2. To fix the Ex.1, we set mc.moving_task after mem_cgroup_count_precharge() has released the mmap_sem. 3. To fix the Ex.2, we use down_read_trylock() instead of down_read() in mem_cgroup_move_charge() and, if it has failed to aquire the lock, cancel all extra charges, wake up all waiters, and retry trylock. Signed-off-by: Daisuke Nishimura <nishimura@mxp.nes.nec.co.jp> Reported-by: Ben Blum <bblum@andrew.cmu.edu> Cc: Miao Xie <miaox@cn.fujitsu.com> Cc: David Rientjes <rientjes@google.com> Cc: Paul Menage <menage@google.com> Cc: Hiroyuki Kamezawa <kamezawa.hiroyuki@gmail.com> Cc: Balbir Singh <balbir@in.ibm.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2011-01-13 23:47:41 +00:00
/*
* Someone who are holding the mmap_sem might be waiting in
* waitq. So we cancel all extra charges, wake up all waiters,
* and retry. Because we cancel precharges, we might not be able
* to move enough charges, but moving charge is a best-effort
* feature anyway, so it wouldn't be a big problem.
*/
__mem_cgroup_clear_mc();
cond_resched();
goto retry;
}
/*
* When we have consumed all precharges and failed in doing
* additional charge, the page walk just aborts.
*/
walk_page_range(mc.mm, 0, mc.mm->highest_vm_end, &charge_walk_ops,
NULL);
memcg: relocate charge moving from ->attach to ->post_attach Hello, So, this ended up a lot simpler than I originally expected. I tested it lightly and it seems to work fine. Petr, can you please test these two patches w/o the lru drain drop patch and see whether the problem is gone? Thanks. ------ 8< ------ If charge moving is used, memcg performs relabeling of the affected pages from its ->attach callback which is called under both cgroup_threadgroup_rwsem and thus can't create new kthreads. This is fragile as various operations may depend on workqueues making forward progress which relies on the ability to create new kthreads. There's no reason to perform charge moving from ->attach which is deep in the task migration path. Move it to ->post_attach which is called after the actual migration is finished and cgroup_threadgroup_rwsem is dropped. * move_charge_struct->mm is added and ->can_attach is now responsible for pinning and recording the target mm. mem_cgroup_clear_mc() is updated accordingly. This also simplifies mem_cgroup_move_task(). * mem_cgroup_move_task() is now called from ->post_attach instead of ->attach. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Debugged-and-tested-by: Petr Mladek <pmladek@suse.com> Reported-by: Cyril Hrubis <chrubis@suse.cz> Reported-by: Johannes Weiner <hannes@cmpxchg.org> Fixes: 1ed1328792ff ("sched, cgroup: replace signal_struct->group_rwsem with a global percpu_rwsem") Cc: <stable@vger.kernel.org> # 4.4+
2016-04-21 23:09:02 +00:00
up_read(&mc.mm->mmap_sem);
atomic_dec(&mc.from->moving_account);
}
memcg: relocate charge moving from ->attach to ->post_attach Hello, So, this ended up a lot simpler than I originally expected. I tested it lightly and it seems to work fine. Petr, can you please test these two patches w/o the lru drain drop patch and see whether the problem is gone? Thanks. ------ 8< ------ If charge moving is used, memcg performs relabeling of the affected pages from its ->attach callback which is called under both cgroup_threadgroup_rwsem and thus can't create new kthreads. This is fragile as various operations may depend on workqueues making forward progress which relies on the ability to create new kthreads. There's no reason to perform charge moving from ->attach which is deep in the task migration path. Move it to ->post_attach which is called after the actual migration is finished and cgroup_threadgroup_rwsem is dropped. * move_charge_struct->mm is added and ->can_attach is now responsible for pinning and recording the target mm. mem_cgroup_clear_mc() is updated accordingly. This also simplifies mem_cgroup_move_task(). * mem_cgroup_move_task() is now called from ->post_attach instead of ->attach. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Debugged-and-tested-by: Petr Mladek <pmladek@suse.com> Reported-by: Cyril Hrubis <chrubis@suse.cz> Reported-by: Johannes Weiner <hannes@cmpxchg.org> Fixes: 1ed1328792ff ("sched, cgroup: replace signal_struct->group_rwsem with a global percpu_rwsem") Cc: <stable@vger.kernel.org> # 4.4+
2016-04-21 23:09:02 +00:00
static void mem_cgroup_move_task(void)
{
memcg: relocate charge moving from ->attach to ->post_attach Hello, So, this ended up a lot simpler than I originally expected. I tested it lightly and it seems to work fine. Petr, can you please test these two patches w/o the lru drain drop patch and see whether the problem is gone? Thanks. ------ 8< ------ If charge moving is used, memcg performs relabeling of the affected pages from its ->attach callback which is called under both cgroup_threadgroup_rwsem and thus can't create new kthreads. This is fragile as various operations may depend on workqueues making forward progress which relies on the ability to create new kthreads. There's no reason to perform charge moving from ->attach which is deep in the task migration path. Move it to ->post_attach which is called after the actual migration is finished and cgroup_threadgroup_rwsem is dropped. * move_charge_struct->mm is added and ->can_attach is now responsible for pinning and recording the target mm. mem_cgroup_clear_mc() is updated accordingly. This also simplifies mem_cgroup_move_task(). * mem_cgroup_move_task() is now called from ->post_attach instead of ->attach. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Debugged-and-tested-by: Petr Mladek <pmladek@suse.com> Reported-by: Cyril Hrubis <chrubis@suse.cz> Reported-by: Johannes Weiner <hannes@cmpxchg.org> Fixes: 1ed1328792ff ("sched, cgroup: replace signal_struct->group_rwsem with a global percpu_rwsem") Cc: <stable@vger.kernel.org> # 4.4+
2016-04-21 23:09:02 +00:00
if (mc.to) {
mem_cgroup_move_charge();
mem_cgroup_clear_mc();
memcg: relocate charge moving from ->attach to ->post_attach Hello, So, this ended up a lot simpler than I originally expected. I tested it lightly and it seems to work fine. Petr, can you please test these two patches w/o the lru drain drop patch and see whether the problem is gone? Thanks. ------ 8< ------ If charge moving is used, memcg performs relabeling of the affected pages from its ->attach callback which is called under both cgroup_threadgroup_rwsem and thus can't create new kthreads. This is fragile as various operations may depend on workqueues making forward progress which relies on the ability to create new kthreads. There's no reason to perform charge moving from ->attach which is deep in the task migration path. Move it to ->post_attach which is called after the actual migration is finished and cgroup_threadgroup_rwsem is dropped. * move_charge_struct->mm is added and ->can_attach is now responsible for pinning and recording the target mm. mem_cgroup_clear_mc() is updated accordingly. This also simplifies mem_cgroup_move_task(). * mem_cgroup_move_task() is now called from ->post_attach instead of ->attach. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Debugged-and-tested-by: Petr Mladek <pmladek@suse.com> Reported-by: Cyril Hrubis <chrubis@suse.cz> Reported-by: Johannes Weiner <hannes@cmpxchg.org> Fixes: 1ed1328792ff ("sched, cgroup: replace signal_struct->group_rwsem with a global percpu_rwsem") Cc: <stable@vger.kernel.org> # 4.4+
2016-04-21 23:09:02 +00:00
}
}
#else /* !CONFIG_MMU */
cgroup: fix handling of multi-destination migration from subtree_control enabling Consider the following v2 hierarchy. P0 (+memory) --- P1 (-memory) --- A \- B P0 has memory enabled in its subtree_control while P1 doesn't. If both A and B contain processes, they would belong to the memory css of P1. Now if memory is enabled on P1's subtree_control, memory csses should be created on both A and B and A's processes should be moved to the former and B's processes the latter. IOW, enabling controllers can cause atomic migrations into different csses. The core cgroup migration logic has been updated accordingly but the controller migration methods haven't and still assume that all tasks migrate to a single target css; furthermore, the methods were fed the css in which subtree_control was updated which is the parent of the target csses. pids controller depends on the migration methods to move charges and this made the controller attribute charges to the wrong csses often triggering the following warning by driving a counter negative. WARNING: CPU: 1 PID: 1 at kernel/cgroup_pids.c:97 pids_cancel.constprop.6+0x31/0x40() Modules linked in: CPU: 1 PID: 1 Comm: systemd Not tainted 4.4.0-rc1+ #29 ... ffffffff81f65382 ffff88007c043b90 ffffffff81551ffc 0000000000000000 ffff88007c043bc8 ffffffff810de202 ffff88007a752000 ffff88007a29ab00 ffff88007c043c80 ffff88007a1d8400 0000000000000001 ffff88007c043bd8 Call Trace: [<ffffffff81551ffc>] dump_stack+0x4e/0x82 [<ffffffff810de202>] warn_slowpath_common+0x82/0xc0 [<ffffffff810de2fa>] warn_slowpath_null+0x1a/0x20 [<ffffffff8118e031>] pids_cancel.constprop.6+0x31/0x40 [<ffffffff8118e0fd>] pids_can_attach+0x6d/0xf0 [<ffffffff81188a4c>] cgroup_taskset_migrate+0x6c/0x330 [<ffffffff81188e05>] cgroup_migrate+0xf5/0x190 [<ffffffff81189016>] cgroup_attach_task+0x176/0x200 [<ffffffff8118949d>] __cgroup_procs_write+0x2ad/0x460 [<ffffffff81189684>] cgroup_procs_write+0x14/0x20 [<ffffffff811854e5>] cgroup_file_write+0x35/0x1c0 [<ffffffff812e26f1>] kernfs_fop_write+0x141/0x190 [<ffffffff81265f88>] __vfs_write+0x28/0xe0 [<ffffffff812666fc>] vfs_write+0xac/0x1a0 [<ffffffff81267019>] SyS_write+0x49/0xb0 [<ffffffff81bcef32>] entry_SYSCALL_64_fastpath+0x12/0x76 This patch fixes the bug by removing @css parameter from the three migration methods, ->can_attach, ->cancel_attach() and ->attach() and updating cgroup_taskset iteration helpers also return the destination css in addition to the task being migrated. All controllers are updated accordingly. * Controllers which don't care whether there are one or multiple target csses can be converted trivially. cpu, io, freezer, perf, netclassid and netprio fall in this category. * cpuset's current implementation assumes that there's single source and destination and thus doesn't support v2 hierarchy already. The only change made by this patchset is how that single destination css is obtained. * memory migration path already doesn't do anything on v2. How the single destination css is obtained is updated and the prep stage of mem_cgroup_can_attach() is reordered to accomodate the change. * pids is the only controller which was affected by this bug. It now correctly handles multi-destination migrations and no longer causes counter underflow from incorrect accounting. Signed-off-by: Tejun Heo <tj@kernel.org> Reported-and-tested-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Aleksa Sarai <cyphar@cyphar.com>
2015-12-03 15:18:21 +00:00
static int mem_cgroup_can_attach(struct cgroup_taskset *tset)
{
return 0;
}
cgroup: fix handling of multi-destination migration from subtree_control enabling Consider the following v2 hierarchy. P0 (+memory) --- P1 (-memory) --- A \- B P0 has memory enabled in its subtree_control while P1 doesn't. If both A and B contain processes, they would belong to the memory css of P1. Now if memory is enabled on P1's subtree_control, memory csses should be created on both A and B and A's processes should be moved to the former and B's processes the latter. IOW, enabling controllers can cause atomic migrations into different csses. The core cgroup migration logic has been updated accordingly but the controller migration methods haven't and still assume that all tasks migrate to a single target css; furthermore, the methods were fed the css in which subtree_control was updated which is the parent of the target csses. pids controller depends on the migration methods to move charges and this made the controller attribute charges to the wrong csses often triggering the following warning by driving a counter negative. WARNING: CPU: 1 PID: 1 at kernel/cgroup_pids.c:97 pids_cancel.constprop.6+0x31/0x40() Modules linked in: CPU: 1 PID: 1 Comm: systemd Not tainted 4.4.0-rc1+ #29 ... ffffffff81f65382 ffff88007c043b90 ffffffff81551ffc 0000000000000000 ffff88007c043bc8 ffffffff810de202 ffff88007a752000 ffff88007a29ab00 ffff88007c043c80 ffff88007a1d8400 0000000000000001 ffff88007c043bd8 Call Trace: [<ffffffff81551ffc>] dump_stack+0x4e/0x82 [<ffffffff810de202>] warn_slowpath_common+0x82/0xc0 [<ffffffff810de2fa>] warn_slowpath_null+0x1a/0x20 [<ffffffff8118e031>] pids_cancel.constprop.6+0x31/0x40 [<ffffffff8118e0fd>] pids_can_attach+0x6d/0xf0 [<ffffffff81188a4c>] cgroup_taskset_migrate+0x6c/0x330 [<ffffffff81188e05>] cgroup_migrate+0xf5/0x190 [<ffffffff81189016>] cgroup_attach_task+0x176/0x200 [<ffffffff8118949d>] __cgroup_procs_write+0x2ad/0x460 [<ffffffff81189684>] cgroup_procs_write+0x14/0x20 [<ffffffff811854e5>] cgroup_file_write+0x35/0x1c0 [<ffffffff812e26f1>] kernfs_fop_write+0x141/0x190 [<ffffffff81265f88>] __vfs_write+0x28/0xe0 [<ffffffff812666fc>] vfs_write+0xac/0x1a0 [<ffffffff81267019>] SyS_write+0x49/0xb0 [<ffffffff81bcef32>] entry_SYSCALL_64_fastpath+0x12/0x76 This patch fixes the bug by removing @css parameter from the three migration methods, ->can_attach, ->cancel_attach() and ->attach() and updating cgroup_taskset iteration helpers also return the destination css in addition to the task being migrated. All controllers are updated accordingly. * Controllers which don't care whether there are one or multiple target csses can be converted trivially. cpu, io, freezer, perf, netclassid and netprio fall in this category. * cpuset's current implementation assumes that there's single source and destination and thus doesn't support v2 hierarchy already. The only change made by this patchset is how that single destination css is obtained. * memory migration path already doesn't do anything on v2. How the single destination css is obtained is updated and the prep stage of mem_cgroup_can_attach() is reordered to accomodate the change. * pids is the only controller which was affected by this bug. It now correctly handles multi-destination migrations and no longer causes counter underflow from incorrect accounting. Signed-off-by: Tejun Heo <tj@kernel.org> Reported-and-tested-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Aleksa Sarai <cyphar@cyphar.com>
2015-12-03 15:18:21 +00:00
static void mem_cgroup_cancel_attach(struct cgroup_taskset *tset)
{
}
memcg: relocate charge moving from ->attach to ->post_attach Hello, So, this ended up a lot simpler than I originally expected. I tested it lightly and it seems to work fine. Petr, can you please test these two patches w/o the lru drain drop patch and see whether the problem is gone? Thanks. ------ 8< ------ If charge moving is used, memcg performs relabeling of the affected pages from its ->attach callback which is called under both cgroup_threadgroup_rwsem and thus can't create new kthreads. This is fragile as various operations may depend on workqueues making forward progress which relies on the ability to create new kthreads. There's no reason to perform charge moving from ->attach which is deep in the task migration path. Move it to ->post_attach which is called after the actual migration is finished and cgroup_threadgroup_rwsem is dropped. * move_charge_struct->mm is added and ->can_attach is now responsible for pinning and recording the target mm. mem_cgroup_clear_mc() is updated accordingly. This also simplifies mem_cgroup_move_task(). * mem_cgroup_move_task() is now called from ->post_attach instead of ->attach. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Debugged-and-tested-by: Petr Mladek <pmladek@suse.com> Reported-by: Cyril Hrubis <chrubis@suse.cz> Reported-by: Johannes Weiner <hannes@cmpxchg.org> Fixes: 1ed1328792ff ("sched, cgroup: replace signal_struct->group_rwsem with a global percpu_rwsem") Cc: <stable@vger.kernel.org> # 4.4+
2016-04-21 23:09:02 +00:00
static void mem_cgroup_move_task(void)
{
}
#endif
/*
* Cgroup retains root cgroups across [un]mount cycles making it necessary
* to verify whether we're attached to the default hierarchy on each mount
* attempt.
*/
cgroup: pass around cgroup_subsys_state instead of cgroup in subsystem methods cgroup is currently in the process of transitioning to using struct cgroup_subsys_state * as the primary handle instead of struct cgroup * in subsystem implementations for the following reasons. * With unified hierarchy, subsystems will be dynamically bound and unbound from cgroups and thus css's (cgroup_subsys_state) may be created and destroyed dynamically over the lifetime of a cgroup, which is different from the current state where all css's are allocated and destroyed together with the associated cgroup. This in turn means that cgroup_css() should be synchronized and may return NULL, making it more cumbersome to use. * Differing levels of per-subsystem granularity in the unified hierarchy means that the task and descendant iterators should behave differently depending on the specific subsystem the iteration is being performed for. * In majority of the cases, subsystems only care about its part in the cgroup hierarchy - ie. the hierarchy of css's. Subsystem methods often obtain the matching css pointer from the cgroup and don't bother with the cgroup pointer itself. Passing around css fits much better. This patch converts all cgroup_subsys methods to take @css instead of @cgroup. The conversions are mostly straight-forward. A few noteworthy changes are * ->css_alloc() now takes css of the parent cgroup rather than the pointer to the new cgroup as the css for the new cgroup doesn't exist yet. Knowing the parent css is enough for all the existing subsystems. * In kernel/cgroup.c::offline_css(), unnecessary open coded css dereference is replaced with local variable access. This patch shouldn't cause any behavior differences. v2: Unnecessary explicit cgrp->subsys[] deref in css_online() replaced with local variable @css as suggested by Li Zefan. Rebased on top of new for-3.12 which includes for-3.11-fixes so that ->css_free() invocation added by da0a12caff ("cgroup: fix a leak when percpu_ref_init() fails") is converted too. Suggested by Li Zefan. Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Li Zefan <lizefan@huawei.com> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vivek Goyal <vgoyal@redhat.com> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Daniel Wagner <daniel.wagner@bmw-carit.de> Cc: Peter Zijlstra <peterz@infradead.org> Cc: Ingo Molnar <mingo@redhat.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Matt Helsley <matthltc@us.ibm.com> Cc: Jens Axboe <axboe@kernel.dk> Cc: Steven Rostedt <rostedt@goodmis.org>
2013-08-09 00:11:23 +00:00
static void mem_cgroup_bind(struct cgroup_subsys_state *root_css)
{
/*
* use_hierarchy is forced on the default hierarchy. cgroup core
* guarantees that @root doesn't have any children, so turning it
* on for the root memcg is enough.
*/
if (cgroup_subsys_on_dfl(memory_cgrp_subsys))
root_mem_cgroup->use_hierarchy = true;
else
root_mem_cgroup->use_hierarchy = false;
}
static int seq_puts_memcg_tunable(struct seq_file *m, unsigned long value)
{
if (value == PAGE_COUNTER_MAX)
seq_puts(m, "max\n");
else
seq_printf(m, "%llu\n", (u64)value * PAGE_SIZE);
return 0;
}
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
static u64 memory_current_read(struct cgroup_subsys_state *css,
struct cftype *cft)
{
struct mem_cgroup *memcg = mem_cgroup_from_css(css);
return (u64)page_counter_read(&memcg->memory) * PAGE_SIZE;
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
}
memcg: introduce memory.min Memory controller implements the memory.low best-effort memory protection mechanism, which works perfectly in many cases and allows protecting working sets of important workloads from sudden reclaim. But its semantics has a significant limitation: it works only as long as there is a supply of reclaimable memory. This makes it pretty useless against any sort of slow memory leaks or memory usage increases. This is especially true for swapless systems. If swap is enabled, memory soft protection effectively postpones problems, allowing a leaking application to fill all swap area, which makes no sense. The only effective way to guarantee the memory protection in this case is to invoke the OOM killer. It's possible to handle this case in userspace by reacting on MEMCG_LOW events; but there is still a place for a fail-safe in-kernel mechanism to provide stronger guarantees. This patch introduces the memory.min interface for cgroup v2 memory controller. It works very similarly to memory.low (sharing the same hierarchical behavior), except that it's not disabled if there is no more reclaimable memory in the system. If cgroup is not populated, its memory.min is ignored, because otherwise even the OOM killer wouldn't be able to reclaim the protected memory, and the system can stall. [guro@fb.com: s/low/min/ in docs] Link: http://lkml.kernel.org/r/20180510130758.GA9129@castle.DHCP.thefacebook.com Link: http://lkml.kernel.org/r/20180509180734.GA4856@castle.DHCP.thefacebook.com Signed-off-by: Roman Gushchin <guro@fb.com> Reviewed-by: Randy Dunlap <rdunlap@infradead.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:07:46 +00:00
static int memory_min_show(struct seq_file *m, void *v)
{
return seq_puts_memcg_tunable(m,
READ_ONCE(mem_cgroup_from_seq(m)->memory.min));
memcg: introduce memory.min Memory controller implements the memory.low best-effort memory protection mechanism, which works perfectly in many cases and allows protecting working sets of important workloads from sudden reclaim. But its semantics has a significant limitation: it works only as long as there is a supply of reclaimable memory. This makes it pretty useless against any sort of slow memory leaks or memory usage increases. This is especially true for swapless systems. If swap is enabled, memory soft protection effectively postpones problems, allowing a leaking application to fill all swap area, which makes no sense. The only effective way to guarantee the memory protection in this case is to invoke the OOM killer. It's possible to handle this case in userspace by reacting on MEMCG_LOW events; but there is still a place for a fail-safe in-kernel mechanism to provide stronger guarantees. This patch introduces the memory.min interface for cgroup v2 memory controller. It works very similarly to memory.low (sharing the same hierarchical behavior), except that it's not disabled if there is no more reclaimable memory in the system. If cgroup is not populated, its memory.min is ignored, because otherwise even the OOM killer wouldn't be able to reclaim the protected memory, and the system can stall. [guro@fb.com: s/low/min/ in docs] Link: http://lkml.kernel.org/r/20180510130758.GA9129@castle.DHCP.thefacebook.com Link: http://lkml.kernel.org/r/20180509180734.GA4856@castle.DHCP.thefacebook.com Signed-off-by: Roman Gushchin <guro@fb.com> Reviewed-by: Randy Dunlap <rdunlap@infradead.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:07:46 +00:00
}
static ssize_t memory_min_write(struct kernfs_open_file *of,
char *buf, size_t nbytes, loff_t off)
{
struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of));
unsigned long min;
int err;
buf = strstrip(buf);
err = page_counter_memparse(buf, "max", &min);
if (err)
return err;
page_counter_set_min(&memcg->memory, min);
return nbytes;
}
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
static int memory_low_show(struct seq_file *m, void *v)
{
return seq_puts_memcg_tunable(m,
READ_ONCE(mem_cgroup_from_seq(m)->memory.low));
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
}
static ssize_t memory_low_write(struct kernfs_open_file *of,
char *buf, size_t nbytes, loff_t off)
{
struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of));
unsigned long low;
int err;
buf = strstrip(buf);
err = page_counter_memparse(buf, "max", &low);
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
if (err)
return err;
mm: memory.low hierarchical behavior This patch aims to address an issue in current memory.low semantics, which makes it hard to use it in a hierarchy, where some leaf memory cgroups are more valuable than others. For example, there are memcgs A, A/B, A/C, A/D and A/E: A A/memory.low = 2G, A/memory.current = 6G //\\ BC DE B/memory.low = 3G B/memory.current = 2G C/memory.low = 1G C/memory.current = 2G D/memory.low = 0 D/memory.current = 2G E/memory.low = 10G E/memory.current = 0 If we apply memory pressure, B, C and D are reclaimed at the same pace while A's usage exceeds 2G. This is obviously wrong, as B's usage is fully below B's memory.low, and C has 1G of protection as well. Also, A is pushed to the size, which is less than A's 2G memory.low, which is also wrong. A simple bash script (provided below) can be used to reproduce the problem. Current results are: A: 1430097920 A/B: 711929856 A/C: 717426688 A/D: 741376 A/E: 0 To address the issue a concept of effective memory.low is introduced. Effective memory.low is always equal or less than original memory.low. In a case, when there is no memory.low overcommittment (and also for top-level cgroups), these two values are equal. Otherwise it's a part of parent's effective memory.low, calculated as a cgroup's memory.low usage divided by sum of sibling's memory.low usages (under memory.low usage I mean the size of actually protected memory: memory.current if memory.current < memory.low, 0 otherwise). It's necessary to track the actual usage, because otherwise an empty cgroup with memory.low set (A/E in my example) will affect actual memory distribution, which makes no sense. To avoid traversing the cgroup tree twice, page_counters code is reused. Calculating effective memory.low can be done in the reclaim path, as we conveniently traversing the cgroup tree from top to bottom and check memory.low on each level. So, it's a perfect place to calculate effective memory low and save it to use it for children cgroups. This also eliminates a need to traverse the cgroup tree from bottom to top each time to check if parent's guarantee is not exceeded. Setting/resetting effective memory.low is intentionally racy, but it's fine and shouldn't lead to any significant differences in actual memory distribution. With this patch applied results are matching the expectations: A: 2147930112 A/B: 1428721664 A/C: 718393344 A/D: 815104 A/E: 0 Test script: #!/bin/bash CGPATH="/sys/fs/cgroup" truncate /file1 --size 2G truncate /file2 --size 2G truncate /file3 --size 2G truncate /file4 --size 50G mkdir "${CGPATH}/A" echo "+memory" > "${CGPATH}/A/cgroup.subtree_control" mkdir "${CGPATH}/A/B" "${CGPATH}/A/C" "${CGPATH}/A/D" "${CGPATH}/A/E" echo 2G > "${CGPATH}/A/memory.low" echo 3G > "${CGPATH}/A/B/memory.low" echo 1G > "${CGPATH}/A/C/memory.low" echo 0 > "${CGPATH}/A/D/memory.low" echo 10G > "${CGPATH}/A/E/memory.low" echo $$ > "${CGPATH}/A/B/cgroup.procs" && vmtouch -qt /file1 echo $$ > "${CGPATH}/A/C/cgroup.procs" && vmtouch -qt /file2 echo $$ > "${CGPATH}/A/D/cgroup.procs" && vmtouch -qt /file3 echo $$ > "${CGPATH}/cgroup.procs" && vmtouch -qt /file4 echo "A: " `cat "${CGPATH}/A/memory.current"` echo "A/B: " `cat "${CGPATH}/A/B/memory.current"` echo "A/C: " `cat "${CGPATH}/A/C/memory.current"` echo "A/D: " `cat "${CGPATH}/A/D/memory.current"` echo "A/E: " `cat "${CGPATH}/A/E/memory.current"` rmdir "${CGPATH}/A/B" "${CGPATH}/A/C" "${CGPATH}/A/D" "${CGPATH}/A/E" rmdir "${CGPATH}/A" rm /file1 /file2 /file3 /file4 Link: http://lkml.kernel.org/r/20180405185921.4942-2-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:06:22 +00:00
page_counter_set_low(&memcg->memory, low);
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
return nbytes;
}
static int memory_high_show(struct seq_file *m, void *v)
{
return seq_puts_memcg_tunable(m,
READ_ONCE(mem_cgroup_from_seq(m)->memory.high));
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
}
static ssize_t memory_high_write(struct kernfs_open_file *of,
char *buf, size_t nbytes, loff_t off)
{
struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of));
mm: memcontrol: try harder to set a new memory.high Setting a memory.high limit below the usage makes almost no effort to shrink the cgroup to the new target size. While memory.high is a "soft" limit that isn't supposed to cause OOM situations, we should still try harder to meet a user request through persistent reclaim. For example, after setting a 10M memory.high on an 800M cgroup full of file cache, the usage shrinks to about 350M: + cat /cgroup/workingset/memory.current 841568256 + echo 10M + cat /cgroup/workingset/memory.current 355729408 This isn't exactly what the user would expect to happen. Setting the value a few more times eventually whittles the usage down to what we are asking for: + echo 10M + cat /cgroup/workingset/memory.current 104181760 + echo 10M + cat /cgroup/workingset/memory.current 31801344 + echo 10M + cat /cgroup/workingset/memory.current 10440704 To improve this, add reclaim retry loops to the memory.high write() callback, similar to what we do for memory.max, to make a reasonable effort that the usage meets the requested size after the call returns. Afterwards, a single write() to memory.high is enough in all but extreme cases: + cat /cgroup/workingset/memory.current 841609216 + echo 10M + cat /cgroup/workingset/memory.current 10182656 790M is not a reasonable reclaim target to ask of a single reclaim invocation. And it wouldn't be reasonable to optimize the reclaim code for it. So asking for the full size but retrying is not a bad choice here: we express our intent, and benefit if reclaim becomes better at handling larger requests, but we also acknowledge that some of the deltas we can encounter in memory_high_write() are just too ridiculously big for a single reclaim invocation to manage. Link: http://lkml.kernel.org/r/20191022201518.341216-2-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-12-01 01:50:09 +00:00
unsigned int nr_retries = MEM_CGROUP_RECLAIM_RETRIES;
bool drained = false;
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
unsigned long high;
int err;
buf = strstrip(buf);
err = page_counter_memparse(buf, "max", &high);
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
if (err)
return err;
page_counter_set_high(&memcg->memory, high);
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
mm: memcontrol: try harder to set a new memory.high Setting a memory.high limit below the usage makes almost no effort to shrink the cgroup to the new target size. While memory.high is a "soft" limit that isn't supposed to cause OOM situations, we should still try harder to meet a user request through persistent reclaim. For example, after setting a 10M memory.high on an 800M cgroup full of file cache, the usage shrinks to about 350M: + cat /cgroup/workingset/memory.current 841568256 + echo 10M + cat /cgroup/workingset/memory.current 355729408 This isn't exactly what the user would expect to happen. Setting the value a few more times eventually whittles the usage down to what we are asking for: + echo 10M + cat /cgroup/workingset/memory.current 104181760 + echo 10M + cat /cgroup/workingset/memory.current 31801344 + echo 10M + cat /cgroup/workingset/memory.current 10440704 To improve this, add reclaim retry loops to the memory.high write() callback, similar to what we do for memory.max, to make a reasonable effort that the usage meets the requested size after the call returns. Afterwards, a single write() to memory.high is enough in all but extreme cases: + cat /cgroup/workingset/memory.current 841609216 + echo 10M + cat /cgroup/workingset/memory.current 10182656 790M is not a reasonable reclaim target to ask of a single reclaim invocation. And it wouldn't be reasonable to optimize the reclaim code for it. So asking for the full size but retrying is not a bad choice here: we express our intent, and benefit if reclaim becomes better at handling larger requests, but we also acknowledge that some of the deltas we can encounter in memory_high_write() are just too ridiculously big for a single reclaim invocation to manage. Link: http://lkml.kernel.org/r/20191022201518.341216-2-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-12-01 01:50:09 +00:00
for (;;) {
unsigned long nr_pages = page_counter_read(&memcg->memory);
unsigned long reclaimed;
if (nr_pages <= high)
break;
if (signal_pending(current))
break;
if (!drained) {
drain_all_stock(memcg);
drained = true;
continue;
}
reclaimed = try_to_free_mem_cgroup_pages(memcg, nr_pages - high,
GFP_KERNEL, true);
if (!reclaimed && !nr_retries--)
break;
}
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
return nbytes;
}
static int memory_max_show(struct seq_file *m, void *v)
{
return seq_puts_memcg_tunable(m,
READ_ONCE(mem_cgroup_from_seq(m)->memory.max));
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
}
static ssize_t memory_max_write(struct kernfs_open_file *of,
char *buf, size_t nbytes, loff_t off)
{
struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of));
unsigned int nr_reclaims = MEM_CGROUP_RECLAIM_RETRIES;
bool drained = false;
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
unsigned long max;
int err;
buf = strstrip(buf);
err = page_counter_memparse(buf, "max", &max);
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
if (err)
return err;
xchg(&memcg->memory.max, max);
for (;;) {
unsigned long nr_pages = page_counter_read(&memcg->memory);
if (nr_pages <= max)
break;
if (signal_pending(current))
break;
if (!drained) {
drain_all_stock(memcg);
drained = true;
continue;
}
if (nr_reclaims) {
if (!try_to_free_mem_cgroup_pages(memcg, nr_pages - max,
GFP_KERNEL, true))
nr_reclaims--;
continue;
}
mm: memcg: make sure memory.events is uptodate when waking pollers Commit a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting") added per-cpu drift to all memory cgroup stats and events shown in memory.stat and memory.events. For memory.stat this is acceptable. But memory.events issues file notifications, and somebody polling the file for changes will be confused when the counters in it are unchanged after a wakeup. Luckily, the events in memory.events - MEMCG_LOW, MEMCG_HIGH, MEMCG_MAX, MEMCG_OOM - are sufficiently rare and high-level that we don't need per-cpu buffering for them: MEMCG_HIGH and MEMCG_MAX would be the most frequent, but they're counting invocations of reclaim, which is a complex operation that touches many shared cachelines. This splits memory.events from the generic VM events and tracks them in their own, unbuffered atomic counters. That's also cleaner, as it eliminates the ugly enum nesting of VM and cgroup events. [hannes@cmpxchg.org: "array subscript is above array bounds"] Link: http://lkml.kernel.org/r/20180406155441.GA20806@cmpxchg.org Link: http://lkml.kernel.org/r/20180405175507.GA24817@cmpxchg.org Fixes: a983b5ebee57 ("mm: memcontrol: fix excessive complexity in memory.stat reporting") Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: Tejun Heo <tj@kernel.org> Acked-by: Tejun Heo <tj@kernel.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Cc: Rik van Riel <riel@surriel.com> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-04-10 23:29:45 +00:00
memcg_memory_event(memcg, MEMCG_OOM);
if (!mem_cgroup_out_of_memory(memcg, GFP_KERNEL, 0))
break;
}
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
memcg_wb_domain_size_changed(memcg);
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
return nbytes;
}
mm, memcg: introduce memory.events.local The memory controller in cgroup v2 exposes memory.events file for each memcg which shows the number of times events like low, high, max, oom and oom_kill have happened for the whole tree rooted at that memcg. Users can also poll or register notification to monitor the changes in that file. Any event at any level of the tree rooted at memcg will notify all the listeners along the path till root_mem_cgroup. There are existing users which depend on this behavior. However there are users which are only interested in the events happening at a specific level of the memcg tree and not in the events in the underlying tree rooted at that memcg. One such use-case is a centralized resource monitor which can dynamically adjust the limits of the jobs running on a system. The jobs can create their sub-hierarchy for their own sub-tasks. The centralized monitor is only interested in the events at the top level memcgs of the jobs as it can then act and adjust the limits of the jobs. Using the current memory.events for such centralized monitor is very inconvenient. The monitor will keep receiving events which it is not interested and to find if the received event is interesting, it has to read memory.event files of the next level and compare it with the top level one. So, let's introduce memory.events.local to the memcg which shows and notify for the events at the memcg level. Now, does memory.stat and memory.pressure need their local versions. IMHO no due to the no internal process contraint of the cgroup v2. The memory.stat file of the top level memcg of a job shows the stats and vmevents of the whole tree. The local stats or vmevents of the top level memcg will only change if there is a process running in that memcg but v2 does not allow that. Similarly for memory.pressure there will not be any process in the internal nodes and thus no chance of local pressure. Link: http://lkml.kernel.org/r/20190527174643.209172-1-shakeelb@google.com Signed-off-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Chris Down <chris@chrisdown.name> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:55 +00:00
static void __memory_events_show(struct seq_file *m, atomic_long_t *events)
{
seq_printf(m, "low %lu\n", atomic_long_read(&events[MEMCG_LOW]));
seq_printf(m, "high %lu\n", atomic_long_read(&events[MEMCG_HIGH]));
seq_printf(m, "max %lu\n", atomic_long_read(&events[MEMCG_MAX]));
seq_printf(m, "oom %lu\n", atomic_long_read(&events[MEMCG_OOM]));
seq_printf(m, "oom_kill %lu\n",
atomic_long_read(&events[MEMCG_OOM_KILL]));
}
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
static int memory_events_show(struct seq_file *m, void *v)
{
struct mem_cgroup *memcg = mem_cgroup_from_seq(m);
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
mm, memcg: introduce memory.events.local The memory controller in cgroup v2 exposes memory.events file for each memcg which shows the number of times events like low, high, max, oom and oom_kill have happened for the whole tree rooted at that memcg. Users can also poll or register notification to monitor the changes in that file. Any event at any level of the tree rooted at memcg will notify all the listeners along the path till root_mem_cgroup. There are existing users which depend on this behavior. However there are users which are only interested in the events happening at a specific level of the memcg tree and not in the events in the underlying tree rooted at that memcg. One such use-case is a centralized resource monitor which can dynamically adjust the limits of the jobs running on a system. The jobs can create their sub-hierarchy for their own sub-tasks. The centralized monitor is only interested in the events at the top level memcgs of the jobs as it can then act and adjust the limits of the jobs. Using the current memory.events for such centralized monitor is very inconvenient. The monitor will keep receiving events which it is not interested and to find if the received event is interesting, it has to read memory.event files of the next level and compare it with the top level one. So, let's introduce memory.events.local to the memcg which shows and notify for the events at the memcg level. Now, does memory.stat and memory.pressure need their local versions. IMHO no due to the no internal process contraint of the cgroup v2. The memory.stat file of the top level memcg of a job shows the stats and vmevents of the whole tree. The local stats or vmevents of the top level memcg will only change if there is a process running in that memcg but v2 does not allow that. Similarly for memory.pressure there will not be any process in the internal nodes and thus no chance of local pressure. Link: http://lkml.kernel.org/r/20190527174643.209172-1-shakeelb@google.com Signed-off-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Chris Down <chris@chrisdown.name> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:55 +00:00
__memory_events_show(m, memcg->memory_events);
return 0;
}
static int memory_events_local_show(struct seq_file *m, void *v)
{
struct mem_cgroup *memcg = mem_cgroup_from_seq(m);
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
mm, memcg: introduce memory.events.local The memory controller in cgroup v2 exposes memory.events file for each memcg which shows the number of times events like low, high, max, oom and oom_kill have happened for the whole tree rooted at that memcg. Users can also poll or register notification to monitor the changes in that file. Any event at any level of the tree rooted at memcg will notify all the listeners along the path till root_mem_cgroup. There are existing users which depend on this behavior. However there are users which are only interested in the events happening at a specific level of the memcg tree and not in the events in the underlying tree rooted at that memcg. One such use-case is a centralized resource monitor which can dynamically adjust the limits of the jobs running on a system. The jobs can create their sub-hierarchy for their own sub-tasks. The centralized monitor is only interested in the events at the top level memcgs of the jobs as it can then act and adjust the limits of the jobs. Using the current memory.events for such centralized monitor is very inconvenient. The monitor will keep receiving events which it is not interested and to find if the received event is interesting, it has to read memory.event files of the next level and compare it with the top level one. So, let's introduce memory.events.local to the memcg which shows and notify for the events at the memcg level. Now, does memory.stat and memory.pressure need their local versions. IMHO no due to the no internal process contraint of the cgroup v2. The memory.stat file of the top level memcg of a job shows the stats and vmevents of the whole tree. The local stats or vmevents of the top level memcg will only change if there is a process running in that memcg but v2 does not allow that. Similarly for memory.pressure there will not be any process in the internal nodes and thus no chance of local pressure. Link: http://lkml.kernel.org/r/20190527174643.209172-1-shakeelb@google.com Signed-off-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Chris Down <chris@chrisdown.name> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:55 +00:00
__memory_events_show(m, memcg->memory_events_local);
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
return 0;
}
static int memory_stat_show(struct seq_file *m, void *v)
{
struct mem_cgroup *memcg = mem_cgroup_from_seq(m);
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
char *buf;
mm: memcontrol: dump memory.stat during cgroup OOM The current cgroup OOM memory info dump doesn't include all the memory we are tracking, nor does it give insight into what the VM tried to do leading up to the OOM. All that useful info is in memory.stat. Furthermore, the recursive printing for every child cgroup can generate absurd amounts of data on the console for larger cgroup trees, and it's not like we provide a per-cgroup breakdown during global OOM kills. When an OOM kill is triggered, print one set of recursive memory.stat items at the level whose limit triggered the OOM condition. Example output: stress invoked oom-killer: gfp_mask=0x100cca(GFP_HIGHUSER_MOVABLE), order=0, oom_score_adj=0 CPU: 2 PID: 210 Comm: stress Not tainted 5.2.0-rc2-mm1-00247-g47d49835983c #135 Hardware name: QEMU Standard PC (i440FX + PIIX, 1996), BIOS 1.12.0-20181126_142135-anatol 04/01/2014 Call Trace: dump_stack+0x46/0x60 dump_header+0x4c/0x2d0 oom_kill_process.cold.10+0xb/0x10 out_of_memory+0x200/0x270 ? try_to_free_mem_cgroup_pages+0xdf/0x130 mem_cgroup_out_of_memory+0xb7/0xc0 try_charge+0x680/0x6f0 mem_cgroup_try_charge+0xb5/0x160 __add_to_page_cache_locked+0xc6/0x300 ? list_lru_destroy+0x80/0x80 add_to_page_cache_lru+0x45/0xc0 pagecache_get_page+0x11b/0x290 filemap_fault+0x458/0x6d0 ext4_filemap_fault+0x27/0x36 __do_fault+0x2f/0xb0 __handle_mm_fault+0x9c5/0x1140 ? apic_timer_interrupt+0xa/0x20 handle_mm_fault+0xc5/0x180 __do_page_fault+0x1ab/0x440 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x55c32167fc10 Code: Bad RIP value. RSP: 002b:00007fff1d031c50 EFLAGS: 00010206 RAX: 000000000dc00000 RBX: 00007fd2db000010 RCX: 00007fd2db000010 RDX: 0000000000000000 RSI: 0000000010001000 RDI: 0000000000000000 RBP: 000055c321680a54 R08: 00000000ffffffff R09: 0000000000000000 R10: 0000000000000022 R11: 0000000000000246 R12: ffffffffffffffff R13: 0000000000000002 R14: 0000000000001000 R15: 0000000010000000 memory: usage 1024kB, limit 1024kB, failcnt 75131 swap: usage 0kB, limit 9007199254740988kB, failcnt 0 Memory cgroup stats for /foo: anon 0 file 0 kernel_stack 36864 slab 274432 sock 0 shmem 0 file_mapped 0 file_dirty 0 file_writeback 0 anon_thp 0 inactive_anon 126976 active_anon 0 inactive_file 0 active_file 0 unevictable 0 slab_reclaimable 0 slab_unreclaimable 274432 pgfault 59466 pgmajfault 1617 workingset_refault 2145 workingset_activate 0 workingset_nodereclaim 0 pgrefill 98952 pgscan 200060 pgsteal 59340 pgactivate 40095 pgdeactivate 96787 pglazyfree 0 pglazyfreed 0 thp_fault_alloc 0 thp_collapse_alloc 0 Tasks state (memory values in pages): [ pid ] uid tgid total_vm rss pgtables_bytes swapents oom_score_adj name [ 200] 0 200 1121 884 53248 29 0 bash [ 209] 0 209 905 246 45056 19 0 stress [ 210] 0 210 66442 56 499712 56349 0 stress oom-kill:constraint=CONSTRAINT_NONE,nodemask=(null),oom_memcg=/foo,task_memcg=/foo,task=stress,pid=210,uid=0 Memory cgroup out of memory: Killed process 210 (stress) total-vm:265768kB, anon-rss:0kB, file-rss:224kB, shmem-rss:0kB oom_reaper: reaped process 210 (stress), now anon-rss:0kB, file-rss:0kB, shmem-rss:0kB [hannes@cmpxchg.org: s/kvmalloc/kmalloc/ per Michal] Link: http://lkml.kernel.org/r/20190605161133.GA12453@cmpxchg.org Link: http://lkml.kernel.org/r/20190604210509.9744-1-hannes@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:59 +00:00
buf = memory_stat_format(memcg);
if (!buf)
return -ENOMEM;
seq_puts(m, buf);
kfree(buf);
return 0;
}
mm, oom: introduce memory.oom.group For some workloads an intervention from the OOM killer can be painful. Killing a random task can bring the workload into an inconsistent state. Historically, there are two common solutions for this problem: 1) enabling panic_on_oom, 2) using a userspace daemon to monitor OOMs and kill all outstanding processes. Both approaches have their downsides: rebooting on each OOM is an obvious waste of capacity, and handling all in userspace is tricky and requires a userspace agent, which will monitor all cgroups for OOMs. In most cases an in-kernel after-OOM cleaning-up mechanism can eliminate the necessity of enabling panic_on_oom. Also, it can simplify the cgroup management for userspace applications. This commit introduces a new knob for cgroup v2 memory controller: memory.oom.group. The knob determines whether the cgroup should be treated as an indivisible workload by the OOM killer. If set, all tasks belonging to the cgroup or to its descendants (if the memory cgroup is not a leaf cgroup) are killed together or not at all. To determine which cgroup has to be killed, we do traverse the cgroup hierarchy from the victim task's cgroup up to the OOMing cgroup (or root) and looking for the highest-level cgroup with memory.oom.group set. Tasks with the OOM protection (oom_score_adj set to -1000) are treated as an exception and are never killed. This patch doesn't change the OOM victim selection algorithm. Link: http://lkml.kernel.org/r/20180802003201.817-4-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: David Rientjes <rientjes@google.com> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-22 04:53:54 +00:00
static int memory_oom_group_show(struct seq_file *m, void *v)
{
struct mem_cgroup *memcg = mem_cgroup_from_seq(m);
mm, oom: introduce memory.oom.group For some workloads an intervention from the OOM killer can be painful. Killing a random task can bring the workload into an inconsistent state. Historically, there are two common solutions for this problem: 1) enabling panic_on_oom, 2) using a userspace daemon to monitor OOMs and kill all outstanding processes. Both approaches have their downsides: rebooting on each OOM is an obvious waste of capacity, and handling all in userspace is tricky and requires a userspace agent, which will monitor all cgroups for OOMs. In most cases an in-kernel after-OOM cleaning-up mechanism can eliminate the necessity of enabling panic_on_oom. Also, it can simplify the cgroup management for userspace applications. This commit introduces a new knob for cgroup v2 memory controller: memory.oom.group. The knob determines whether the cgroup should be treated as an indivisible workload by the OOM killer. If set, all tasks belonging to the cgroup or to its descendants (if the memory cgroup is not a leaf cgroup) are killed together or not at all. To determine which cgroup has to be killed, we do traverse the cgroup hierarchy from the victim task's cgroup up to the OOMing cgroup (or root) and looking for the highest-level cgroup with memory.oom.group set. Tasks with the OOM protection (oom_score_adj set to -1000) are treated as an exception and are never killed. This patch doesn't change the OOM victim selection algorithm. Link: http://lkml.kernel.org/r/20180802003201.817-4-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: David Rientjes <rientjes@google.com> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-22 04:53:54 +00:00
seq_printf(m, "%d\n", memcg->oom_group);
return 0;
}
static ssize_t memory_oom_group_write(struct kernfs_open_file *of,
char *buf, size_t nbytes, loff_t off)
{
struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of));
int ret, oom_group;
buf = strstrip(buf);
if (!buf)
return -EINVAL;
ret = kstrtoint(buf, 0, &oom_group);
if (ret)
return ret;
if (oom_group != 0 && oom_group != 1)
return -EINVAL;
memcg->oom_group = oom_group;
return nbytes;
}
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
static struct cftype memory_files[] = {
{
.name = "current",
.flags = CFTYPE_NOT_ON_ROOT,
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
.read_u64 = memory_current_read,
},
memcg: introduce memory.min Memory controller implements the memory.low best-effort memory protection mechanism, which works perfectly in many cases and allows protecting working sets of important workloads from sudden reclaim. But its semantics has a significant limitation: it works only as long as there is a supply of reclaimable memory. This makes it pretty useless against any sort of slow memory leaks or memory usage increases. This is especially true for swapless systems. If swap is enabled, memory soft protection effectively postpones problems, allowing a leaking application to fill all swap area, which makes no sense. The only effective way to guarantee the memory protection in this case is to invoke the OOM killer. It's possible to handle this case in userspace by reacting on MEMCG_LOW events; but there is still a place for a fail-safe in-kernel mechanism to provide stronger guarantees. This patch introduces the memory.min interface for cgroup v2 memory controller. It works very similarly to memory.low (sharing the same hierarchical behavior), except that it's not disabled if there is no more reclaimable memory in the system. If cgroup is not populated, its memory.min is ignored, because otherwise even the OOM killer wouldn't be able to reclaim the protected memory, and the system can stall. [guro@fb.com: s/low/min/ in docs] Link: http://lkml.kernel.org/r/20180510130758.GA9129@castle.DHCP.thefacebook.com Link: http://lkml.kernel.org/r/20180509180734.GA4856@castle.DHCP.thefacebook.com Signed-off-by: Roman Gushchin <guro@fb.com> Reviewed-by: Randy Dunlap <rdunlap@infradead.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:07:46 +00:00
{
.name = "min",
.flags = CFTYPE_NOT_ON_ROOT,
.seq_show = memory_min_show,
.write = memory_min_write,
},
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
{
.name = "low",
.flags = CFTYPE_NOT_ON_ROOT,
.seq_show = memory_low_show,
.write = memory_low_write,
},
{
.name = "high",
.flags = CFTYPE_NOT_ON_ROOT,
.seq_show = memory_high_show,
.write = memory_high_write,
},
{
.name = "max",
.flags = CFTYPE_NOT_ON_ROOT,
.seq_show = memory_max_show,
.write = memory_max_write,
},
{
.name = "events",
.flags = CFTYPE_NOT_ON_ROOT,
.file_offset = offsetof(struct mem_cgroup, events_file),
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
.seq_show = memory_events_show,
},
mm, memcg: introduce memory.events.local The memory controller in cgroup v2 exposes memory.events file for each memcg which shows the number of times events like low, high, max, oom and oom_kill have happened for the whole tree rooted at that memcg. Users can also poll or register notification to monitor the changes in that file. Any event at any level of the tree rooted at memcg will notify all the listeners along the path till root_mem_cgroup. There are existing users which depend on this behavior. However there are users which are only interested in the events happening at a specific level of the memcg tree and not in the events in the underlying tree rooted at that memcg. One such use-case is a centralized resource monitor which can dynamically adjust the limits of the jobs running on a system. The jobs can create their sub-hierarchy for their own sub-tasks. The centralized monitor is only interested in the events at the top level memcgs of the jobs as it can then act and adjust the limits of the jobs. Using the current memory.events for such centralized monitor is very inconvenient. The monitor will keep receiving events which it is not interested and to find if the received event is interesting, it has to read memory.event files of the next level and compare it with the top level one. So, let's introduce memory.events.local to the memcg which shows and notify for the events at the memcg level. Now, does memory.stat and memory.pressure need their local versions. IMHO no due to the no internal process contraint of the cgroup v2. The memory.stat file of the top level memcg of a job shows the stats and vmevents of the whole tree. The local stats or vmevents of the top level memcg will only change if there is a process running in that memcg but v2 does not allow that. Similarly for memory.pressure there will not be any process in the internal nodes and thus no chance of local pressure. Link: http://lkml.kernel.org/r/20190527174643.209172-1-shakeelb@google.com Signed-off-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Chris Down <chris@chrisdown.name> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2019-07-12 03:55:55 +00:00
{
.name = "events.local",
.flags = CFTYPE_NOT_ON_ROOT,
.file_offset = offsetof(struct mem_cgroup, events_local_file),
.seq_show = memory_events_local_show,
},
{
.name = "stat",
.seq_show = memory_stat_show,
},
mm, oom: introduce memory.oom.group For some workloads an intervention from the OOM killer can be painful. Killing a random task can bring the workload into an inconsistent state. Historically, there are two common solutions for this problem: 1) enabling panic_on_oom, 2) using a userspace daemon to monitor OOMs and kill all outstanding processes. Both approaches have their downsides: rebooting on each OOM is an obvious waste of capacity, and handling all in userspace is tricky and requires a userspace agent, which will monitor all cgroups for OOMs. In most cases an in-kernel after-OOM cleaning-up mechanism can eliminate the necessity of enabling panic_on_oom. Also, it can simplify the cgroup management for userspace applications. This commit introduces a new knob for cgroup v2 memory controller: memory.oom.group. The knob determines whether the cgroup should be treated as an indivisible workload by the OOM killer. If set, all tasks belonging to the cgroup or to its descendants (if the memory cgroup is not a leaf cgroup) are killed together or not at all. To determine which cgroup has to be killed, we do traverse the cgroup hierarchy from the victim task's cgroup up to the OOMing cgroup (or root) and looking for the highest-level cgroup with memory.oom.group set. Tasks with the OOM protection (oom_score_adj set to -1000) are treated as an exception and are never killed. This patch doesn't change the OOM victim selection algorithm. Link: http://lkml.kernel.org/r/20180802003201.817-4-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Michal Hocko <mhocko@suse.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: David Rientjes <rientjes@google.com> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-22 04:53:54 +00:00
{
.name = "oom.group",
.flags = CFTYPE_NOT_ON_ROOT | CFTYPE_NS_DELEGATABLE,
.seq_show = memory_oom_group_show,
.write = memory_oom_group_write,
},
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
{ } /* terminate */
};
cgroup: clean up cgroup_subsys names and initialization cgroup_subsys is a bit messier than it needs to be. * The name of a subsys can be different from its internal identifier defined in cgroup_subsys.h. Most subsystems use the matching name but three - cpu, memory and perf_event - use different ones. * cgroup_subsys_id enums are postfixed with _subsys_id and each cgroup_subsys is postfixed with _subsys. cgroup.h is widely included throughout various subsystems, it doesn't and shouldn't have claim on such generic names which don't have any qualifier indicating that they belong to cgroup. * cgroup_subsys->subsys_id should always equal the matching cgroup_subsys_id enum; however, we require each controller to initialize it and then BUG if they don't match, which is a bit silly. This patch cleans up cgroup_subsys names and initialization by doing the followings. * cgroup_subsys_id enums are now postfixed with _cgrp_id, and each cgroup_subsys with _cgrp_subsys. * With the above, renaming subsys identifiers to match the userland visible names doesn't cause any naming conflicts. All non-matching identifiers are renamed to match the official names. cpu_cgroup -> cpu mem_cgroup -> memory perf -> perf_event * controllers no longer need to initialize ->subsys_id and ->name. They're generated in cgroup core and set automatically during boot. * Redundant cgroup_subsys declarations removed. * While updating BUG_ON()s in cgroup_init_early(), convert them to WARN()s. BUGging that early during boot is stupid - the kernel can't print anything, even through serial console and the trap handler doesn't even link stack frame properly for back-tracing. This patch doesn't introduce any behavior changes. v2: Rebased on top of fe1217c4f3f7 ("net: net_cls: move cgroupfs classid handling into core"). Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Neil Horman <nhorman@tuxdriver.com> Acked-by: "David S. Miller" <davem@davemloft.net> Acked-by: "Rafael J. Wysocki" <rjw@rjwysocki.net> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Peter Zijlstra <peterz@infradead.org> Acked-by: Aristeu Rozanski <aris@redhat.com> Acked-by: Ingo Molnar <mingo@redhat.com> Acked-by: Li Zefan <lizefan@huawei.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Balbir Singh <bsingharora@gmail.com> Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: Serge E. Hallyn <serue@us.ibm.com> Cc: Vivek Goyal <vgoyal@redhat.com> Cc: Thomas Graf <tgraf@suug.ch>
2014-02-08 15:36:58 +00:00
struct cgroup_subsys memory_cgrp_subsys = {
.css_alloc = mem_cgroup_css_alloc,
.css_online = mem_cgroup_css_online,
.css_offline = mem_cgroup_css_offline,
mm: memcontrol: fix possible memcg leak due to interrupted reclaim Memory cgroup reclaim can be interrupted with mem_cgroup_iter_break() once enough pages have been reclaimed, in which case, in contrast to a full round-trip over a cgroup sub-tree, the current position stored in mem_cgroup_reclaim_iter of the target cgroup does not get invalidated and so is left holding the reference to the last scanned cgroup. If the target cgroup does not get scanned again (we might have just reclaimed the last page or all processes might exit and free their memory voluntary), we will leak it, because there is nobody to put the reference held by the iterator. The problem is easy to reproduce by running the following command sequence in a loop: mkdir /sys/fs/cgroup/memory/test echo 100M > /sys/fs/cgroup/memory/test/memory.limit_in_bytes echo $$ > /sys/fs/cgroup/memory/test/cgroup.procs memhog 150M echo $$ > /sys/fs/cgroup/memory/cgroup.procs rmdir test The cgroups generated by it will never get freed. This patch fixes this issue by making mem_cgroup_iter avoid taking reference to the current position. In order not to hit use-after-free bug while running reclaim in parallel with cgroup deletion, we make use of ->css_released cgroup callback to clear references to the dying cgroup in all reclaim iterators that might refer to it. This callback is called right before scheduling rcu work which will free css, so if we access iter->position from rcu read section, we might be sure it won't go away under us. [hannes@cmpxchg.org: clean up css ref handling] Fixes: 5ac8fb31ad2e ("mm: memcontrol: convert reclaim iterator to simple css refcounting") Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-12-29 22:54:10 +00:00
.css_released = mem_cgroup_css_released,
.css_free = mem_cgroup_css_free,
.css_reset = mem_cgroup_css_reset,
.can_attach = mem_cgroup_can_attach,
.cancel_attach = mem_cgroup_cancel_attach,
memcg: relocate charge moving from ->attach to ->post_attach Hello, So, this ended up a lot simpler than I originally expected. I tested it lightly and it seems to work fine. Petr, can you please test these two patches w/o the lru drain drop patch and see whether the problem is gone? Thanks. ------ 8< ------ If charge moving is used, memcg performs relabeling of the affected pages from its ->attach callback which is called under both cgroup_threadgroup_rwsem and thus can't create new kthreads. This is fragile as various operations may depend on workqueues making forward progress which relies on the ability to create new kthreads. There's no reason to perform charge moving from ->attach which is deep in the task migration path. Move it to ->post_attach which is called after the actual migration is finished and cgroup_threadgroup_rwsem is dropped. * move_charge_struct->mm is added and ->can_attach is now responsible for pinning and recording the target mm. mem_cgroup_clear_mc() is updated accordingly. This also simplifies mem_cgroup_move_task(). * mem_cgroup_move_task() is now called from ->post_attach instead of ->attach. Signed-off-by: Tejun Heo <tj@kernel.org> Cc: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Debugged-and-tested-by: Petr Mladek <pmladek@suse.com> Reported-by: Cyril Hrubis <chrubis@suse.cz> Reported-by: Johannes Weiner <hannes@cmpxchg.org> Fixes: 1ed1328792ff ("sched, cgroup: replace signal_struct->group_rwsem with a global percpu_rwsem") Cc: <stable@vger.kernel.org> # 4.4+
2016-04-21 23:09:02 +00:00
.post_attach = mem_cgroup_move_task,
.bind = mem_cgroup_bind,
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
.dfl_cftypes = memory_files,
.legacy_cftypes = mem_cgroup_legacy_files,
.early_init = 0,
};
2020-04-02 04:07:03 +00:00
/*
* This function calculates an individual cgroup's effective
* protection which is derived from its own memory.min/low, its
* parent's and siblings' settings, as well as the actual memory
* distribution in the tree.
*
* The following rules apply to the effective protection values:
*
* 1. At the first level of reclaim, effective protection is equal to
* the declared protection in memory.min and memory.low.
*
* 2. To enable safe delegation of the protection configuration, at
* subsequent levels the effective protection is capped to the
* parent's effective protection.
*
* 3. To make complex and dynamic subtrees easier to configure, the
* user is allowed to overcommit the declared protection at a given
* level. If that is the case, the parent's effective protection is
* distributed to the children in proportion to how much protection
* they have declared and how much of it they are utilizing.
*
* This makes distribution proportional, but also work-conserving:
* if one cgroup claims much more protection than it uses memory,
* the unused remainder is available to its siblings.
*
* 4. Conversely, when the declared protection is undercommitted at a
* given level, the distribution of the larger parental protection
* budget is NOT proportional. A cgroup's protection from a sibling
* is capped to its own memory.min/low setting.
*
mm: memcontrol: recursive memory.low protection Right now, the effective protection of any given cgroup is capped by its own explicit memory.low setting, regardless of what the parent says. The reasons for this are mostly historical and ease of implementation: to make delegation of memory.low safe, effective protection is the min() of all memory.low up the tree. Unfortunately, this limitation makes it impossible to protect an entire subtree from another without forcing the user to make explicit protection allocations all the way to the leaf cgroups - something that is highly undesirable in real life scenarios. Consider memory in a data center host. At the cgroup top level, we have a distinction between system management software and the actual workload the system is executing. Both branches are further subdivided into individual services, job components etc. We want to protect the workload as a whole from the system management software, but that doesn't mean we want to protect and prioritize individual workload wrt each other. Their memory demand can vary over time, and we'd want the VM to simply cache the hottest data within the workload subtree. Yet, the current memory.low limitations force us to allocate a fixed amount of protection to each workload component in order to get protection from system management software in general. This results in very inefficient resource distribution. Another concern with mandating downward allocation is that, as the complexity of the cgroup tree grows, it gets harder for the lower levels to be informed about decisions made at the host-level. Consider a container inside a namespace that in turn creates its own nested tree of cgroups to run multiple workloads. It'd be extremely difficult to configure memory.low parameters in those leaf cgroups that on one hand balance pressure among siblings as the container desires, while also reflecting the host-level protection from e.g. rpm upgrades, that lie beyond one or more delegation and namespacing points in the tree. It's highly unusual from a cgroup interface POV that nested levels have to be aware of and reflect decisions made at higher levels for them to be effective. To enable such use cases and scale configurability for complex trees, this patch implements a resource inheritance model for memory that is similar to how the CPU and the IO controller implement work-conserving resource allocations: a share of a resource allocated to a subree always applies to the entire subtree recursively, while allowing, but not mandating, children to further specify distribution rules. That means that if protection is explicitly allocated among siblings, those configured shares are being followed during page reclaim just like they are now. However, if the memory.low set at a higher level is not fully claimed by the children in that subtree, the "floating" remainder is applied to each cgroup in the tree in proportion to its size. Since reclaim pressure is applied in proportion to size as well, each child in that tree gets the same boost, and the effect is neutral among siblings - with respect to each other, they behave as if no memory control was enabled at all, and the VM simply balances the memory demands optimally within the subtree. But collectively those cgroups enjoy a boost over the cgroups in neighboring trees. E.g. a leaf cgroup with a memory.low setting of 0 no longer means that it's not getting a share of the hierarchically assigned resource, just that it doesn't claim a fixed amount of it to protect from its siblings. This allows us to recursively protect one subtree (workload) from another (system management), while letting subgroups compete freely among each other - without having to assign fixed shares to each leaf, and without nested groups having to echo higher-level settings. The floating protection composes naturally with fixed protection. Consider the following example tree: A A: low = 2G / \ A1: low = 1G A1 A2 A2: low = 0G As outside pressure is applied to this tree, A1 will enjoy a fixed protection from A2 of 1G, but the remaining, unclaimed 1G from A is split evenly among A1 and A2, coming out to 1.5G and 0.5G. There is a slight risk of regressing theoretical setups where the top-level cgroups don't know about the true budgeting and set bogusly high "bypass" values that are meaningfully allocated down the tree. Such setups would rely on unclaimed protection to be discarded, and distributing it would change the intended behavior. Be safe and hide the new behavior behind a mount option, 'memory_recursiveprot'. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Tejun Heo <tj@kernel.org> Acked-by: Roman Gushchin <guro@fb.com> Acked-by: Chris Down <chris@chrisdown.name> Cc: Michal Hocko <mhocko@suse.com> Cc: Michal Koutný <mkoutny@suse.com> Link: http://lkml.kernel.org/r/20200227195606.46212-4-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-04-02 04:07:07 +00:00
* 5. However, to allow protecting recursive subtrees from each other
* without having to declare each individual cgroup's fixed share
* of the ancestor's claim to protection, any unutilized -
* "floating" - protection from up the tree is distributed in
* proportion to each cgroup's *usage*. This makes the protection
* neutral wrt sibling cgroups and lets them compete freely over
* the shared parental protection budget, but it protects the
* subtree as a whole from neighboring subtrees.
*
* Note that 4. and 5. are not in conflict: 4. is about protecting
* against immediate siblings whereas 5. is about protecting against
* neighboring subtrees.
2020-04-02 04:07:03 +00:00
*/
static unsigned long effective_protection(unsigned long usage,
mm: memcontrol: recursive memory.low protection Right now, the effective protection of any given cgroup is capped by its own explicit memory.low setting, regardless of what the parent says. The reasons for this are mostly historical and ease of implementation: to make delegation of memory.low safe, effective protection is the min() of all memory.low up the tree. Unfortunately, this limitation makes it impossible to protect an entire subtree from another without forcing the user to make explicit protection allocations all the way to the leaf cgroups - something that is highly undesirable in real life scenarios. Consider memory in a data center host. At the cgroup top level, we have a distinction between system management software and the actual workload the system is executing. Both branches are further subdivided into individual services, job components etc. We want to protect the workload as a whole from the system management software, but that doesn't mean we want to protect and prioritize individual workload wrt each other. Their memory demand can vary over time, and we'd want the VM to simply cache the hottest data within the workload subtree. Yet, the current memory.low limitations force us to allocate a fixed amount of protection to each workload component in order to get protection from system management software in general. This results in very inefficient resource distribution. Another concern with mandating downward allocation is that, as the complexity of the cgroup tree grows, it gets harder for the lower levels to be informed about decisions made at the host-level. Consider a container inside a namespace that in turn creates its own nested tree of cgroups to run multiple workloads. It'd be extremely difficult to configure memory.low parameters in those leaf cgroups that on one hand balance pressure among siblings as the container desires, while also reflecting the host-level protection from e.g. rpm upgrades, that lie beyond one or more delegation and namespacing points in the tree. It's highly unusual from a cgroup interface POV that nested levels have to be aware of and reflect decisions made at higher levels for them to be effective. To enable such use cases and scale configurability for complex trees, this patch implements a resource inheritance model for memory that is similar to how the CPU and the IO controller implement work-conserving resource allocations: a share of a resource allocated to a subree always applies to the entire subtree recursively, while allowing, but not mandating, children to further specify distribution rules. That means that if protection is explicitly allocated among siblings, those configured shares are being followed during page reclaim just like they are now. However, if the memory.low set at a higher level is not fully claimed by the children in that subtree, the "floating" remainder is applied to each cgroup in the tree in proportion to its size. Since reclaim pressure is applied in proportion to size as well, each child in that tree gets the same boost, and the effect is neutral among siblings - with respect to each other, they behave as if no memory control was enabled at all, and the VM simply balances the memory demands optimally within the subtree. But collectively those cgroups enjoy a boost over the cgroups in neighboring trees. E.g. a leaf cgroup with a memory.low setting of 0 no longer means that it's not getting a share of the hierarchically assigned resource, just that it doesn't claim a fixed amount of it to protect from its siblings. This allows us to recursively protect one subtree (workload) from another (system management), while letting subgroups compete freely among each other - without having to assign fixed shares to each leaf, and without nested groups having to echo higher-level settings. The floating protection composes naturally with fixed protection. Consider the following example tree: A A: low = 2G / \ A1: low = 1G A1 A2 A2: low = 0G As outside pressure is applied to this tree, A1 will enjoy a fixed protection from A2 of 1G, but the remaining, unclaimed 1G from A is split evenly among A1 and A2, coming out to 1.5G and 0.5G. There is a slight risk of regressing theoretical setups where the top-level cgroups don't know about the true budgeting and set bogusly high "bypass" values that are meaningfully allocated down the tree. Such setups would rely on unclaimed protection to be discarded, and distributing it would change the intended behavior. Be safe and hide the new behavior behind a mount option, 'memory_recursiveprot'. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Tejun Heo <tj@kernel.org> Acked-by: Roman Gushchin <guro@fb.com> Acked-by: Chris Down <chris@chrisdown.name> Cc: Michal Hocko <mhocko@suse.com> Cc: Michal Koutný <mkoutny@suse.com> Link: http://lkml.kernel.org/r/20200227195606.46212-4-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-04-02 04:07:07 +00:00
unsigned long parent_usage,
2020-04-02 04:07:03 +00:00
unsigned long setting,
unsigned long parent_effective,
unsigned long siblings_protected)
{
unsigned long protected;
mm: memcontrol: recursive memory.low protection Right now, the effective protection of any given cgroup is capped by its own explicit memory.low setting, regardless of what the parent says. The reasons for this are mostly historical and ease of implementation: to make delegation of memory.low safe, effective protection is the min() of all memory.low up the tree. Unfortunately, this limitation makes it impossible to protect an entire subtree from another without forcing the user to make explicit protection allocations all the way to the leaf cgroups - something that is highly undesirable in real life scenarios. Consider memory in a data center host. At the cgroup top level, we have a distinction between system management software and the actual workload the system is executing. Both branches are further subdivided into individual services, job components etc. We want to protect the workload as a whole from the system management software, but that doesn't mean we want to protect and prioritize individual workload wrt each other. Their memory demand can vary over time, and we'd want the VM to simply cache the hottest data within the workload subtree. Yet, the current memory.low limitations force us to allocate a fixed amount of protection to each workload component in order to get protection from system management software in general. This results in very inefficient resource distribution. Another concern with mandating downward allocation is that, as the complexity of the cgroup tree grows, it gets harder for the lower levels to be informed about decisions made at the host-level. Consider a container inside a namespace that in turn creates its own nested tree of cgroups to run multiple workloads. It'd be extremely difficult to configure memory.low parameters in those leaf cgroups that on one hand balance pressure among siblings as the container desires, while also reflecting the host-level protection from e.g. rpm upgrades, that lie beyond one or more delegation and namespacing points in the tree. It's highly unusual from a cgroup interface POV that nested levels have to be aware of and reflect decisions made at higher levels for them to be effective. To enable such use cases and scale configurability for complex trees, this patch implements a resource inheritance model for memory that is similar to how the CPU and the IO controller implement work-conserving resource allocations: a share of a resource allocated to a subree always applies to the entire subtree recursively, while allowing, but not mandating, children to further specify distribution rules. That means that if protection is explicitly allocated among siblings, those configured shares are being followed during page reclaim just like they are now. However, if the memory.low set at a higher level is not fully claimed by the children in that subtree, the "floating" remainder is applied to each cgroup in the tree in proportion to its size. Since reclaim pressure is applied in proportion to size as well, each child in that tree gets the same boost, and the effect is neutral among siblings - with respect to each other, they behave as if no memory control was enabled at all, and the VM simply balances the memory demands optimally within the subtree. But collectively those cgroups enjoy a boost over the cgroups in neighboring trees. E.g. a leaf cgroup with a memory.low setting of 0 no longer means that it's not getting a share of the hierarchically assigned resource, just that it doesn't claim a fixed amount of it to protect from its siblings. This allows us to recursively protect one subtree (workload) from another (system management), while letting subgroups compete freely among each other - without having to assign fixed shares to each leaf, and without nested groups having to echo higher-level settings. The floating protection composes naturally with fixed protection. Consider the following example tree: A A: low = 2G / \ A1: low = 1G A1 A2 A2: low = 0G As outside pressure is applied to this tree, A1 will enjoy a fixed protection from A2 of 1G, but the remaining, unclaimed 1G from A is split evenly among A1 and A2, coming out to 1.5G and 0.5G. There is a slight risk of regressing theoretical setups where the top-level cgroups don't know about the true budgeting and set bogusly high "bypass" values that are meaningfully allocated down the tree. Such setups would rely on unclaimed protection to be discarded, and distributing it would change the intended behavior. Be safe and hide the new behavior behind a mount option, 'memory_recursiveprot'. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Tejun Heo <tj@kernel.org> Acked-by: Roman Gushchin <guro@fb.com> Acked-by: Chris Down <chris@chrisdown.name> Cc: Michal Hocko <mhocko@suse.com> Cc: Michal Koutný <mkoutny@suse.com> Link: http://lkml.kernel.org/r/20200227195606.46212-4-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-04-02 04:07:07 +00:00
unsigned long ep;
2020-04-02 04:07:03 +00:00
protected = min(usage, setting);
/*
* If all cgroups at this level combined claim and use more
* protection then what the parent affords them, distribute
* shares in proportion to utilization.
*
* We are using actual utilization rather than the statically
* claimed protection in order to be work-conserving: claimed
* but unused protection is available to siblings that would
* otherwise get a smaller chunk than what they claimed.
*/
if (siblings_protected > parent_effective)
return protected * parent_effective / siblings_protected;
/*
* Ok, utilized protection of all children is within what the
* parent affords them, so we know whatever this child claims
* and utilizes is effectively protected.
*
* If there is unprotected usage beyond this value, reclaim
* will apply pressure in proportion to that amount.
*
* If there is unutilized protection, the cgroup will be fully
* shielded from reclaim, but we do return a smaller value for
* protection than what the group could enjoy in theory. This
* is okay. With the overcommit distribution above, effective
* protection is always dependent on how memory is actually
* consumed among the siblings anyway.
*/
mm: memcontrol: recursive memory.low protection Right now, the effective protection of any given cgroup is capped by its own explicit memory.low setting, regardless of what the parent says. The reasons for this are mostly historical and ease of implementation: to make delegation of memory.low safe, effective protection is the min() of all memory.low up the tree. Unfortunately, this limitation makes it impossible to protect an entire subtree from another without forcing the user to make explicit protection allocations all the way to the leaf cgroups - something that is highly undesirable in real life scenarios. Consider memory in a data center host. At the cgroup top level, we have a distinction between system management software and the actual workload the system is executing. Both branches are further subdivided into individual services, job components etc. We want to protect the workload as a whole from the system management software, but that doesn't mean we want to protect and prioritize individual workload wrt each other. Their memory demand can vary over time, and we'd want the VM to simply cache the hottest data within the workload subtree. Yet, the current memory.low limitations force us to allocate a fixed amount of protection to each workload component in order to get protection from system management software in general. This results in very inefficient resource distribution. Another concern with mandating downward allocation is that, as the complexity of the cgroup tree grows, it gets harder for the lower levels to be informed about decisions made at the host-level. Consider a container inside a namespace that in turn creates its own nested tree of cgroups to run multiple workloads. It'd be extremely difficult to configure memory.low parameters in those leaf cgroups that on one hand balance pressure among siblings as the container desires, while also reflecting the host-level protection from e.g. rpm upgrades, that lie beyond one or more delegation and namespacing points in the tree. It's highly unusual from a cgroup interface POV that nested levels have to be aware of and reflect decisions made at higher levels for them to be effective. To enable such use cases and scale configurability for complex trees, this patch implements a resource inheritance model for memory that is similar to how the CPU and the IO controller implement work-conserving resource allocations: a share of a resource allocated to a subree always applies to the entire subtree recursively, while allowing, but not mandating, children to further specify distribution rules. That means that if protection is explicitly allocated among siblings, those configured shares are being followed during page reclaim just like they are now. However, if the memory.low set at a higher level is not fully claimed by the children in that subtree, the "floating" remainder is applied to each cgroup in the tree in proportion to its size. Since reclaim pressure is applied in proportion to size as well, each child in that tree gets the same boost, and the effect is neutral among siblings - with respect to each other, they behave as if no memory control was enabled at all, and the VM simply balances the memory demands optimally within the subtree. But collectively those cgroups enjoy a boost over the cgroups in neighboring trees. E.g. a leaf cgroup with a memory.low setting of 0 no longer means that it's not getting a share of the hierarchically assigned resource, just that it doesn't claim a fixed amount of it to protect from its siblings. This allows us to recursively protect one subtree (workload) from another (system management), while letting subgroups compete freely among each other - without having to assign fixed shares to each leaf, and without nested groups having to echo higher-level settings. The floating protection composes naturally with fixed protection. Consider the following example tree: A A: low = 2G / \ A1: low = 1G A1 A2 A2: low = 0G As outside pressure is applied to this tree, A1 will enjoy a fixed protection from A2 of 1G, but the remaining, unclaimed 1G from A is split evenly among A1 and A2, coming out to 1.5G and 0.5G. There is a slight risk of regressing theoretical setups where the top-level cgroups don't know about the true budgeting and set bogusly high "bypass" values that are meaningfully allocated down the tree. Such setups would rely on unclaimed protection to be discarded, and distributing it would change the intended behavior. Be safe and hide the new behavior behind a mount option, 'memory_recursiveprot'. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Tejun Heo <tj@kernel.org> Acked-by: Roman Gushchin <guro@fb.com> Acked-by: Chris Down <chris@chrisdown.name> Cc: Michal Hocko <mhocko@suse.com> Cc: Michal Koutný <mkoutny@suse.com> Link: http://lkml.kernel.org/r/20200227195606.46212-4-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-04-02 04:07:07 +00:00
ep = protected;
/*
* If the children aren't claiming (all of) the protection
* afforded to them by the parent, distribute the remainder in
* proportion to the (unprotected) memory of each cgroup. That
* way, cgroups that aren't explicitly prioritized wrt each
* other compete freely over the allowance, but they are
* collectively protected from neighboring trees.
*
* We're using unprotected memory for the weight so that if
* some cgroups DO claim explicit protection, we don't protect
* the same bytes twice.
*/
if (!(cgrp_dfl_root.flags & CGRP_ROOT_MEMORY_RECURSIVE_PROT))
return ep;
if (parent_effective > siblings_protected && usage > protected) {
unsigned long unclaimed;
unclaimed = parent_effective - siblings_protected;
unclaimed *= usage - protected;
unclaimed /= parent_usage - siblings_protected;
ep += unclaimed;
}
return ep;
2020-04-02 04:07:03 +00:00
}
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
/**
memcg: introduce memory.min Memory controller implements the memory.low best-effort memory protection mechanism, which works perfectly in many cases and allows protecting working sets of important workloads from sudden reclaim. But its semantics has a significant limitation: it works only as long as there is a supply of reclaimable memory. This makes it pretty useless against any sort of slow memory leaks or memory usage increases. This is especially true for swapless systems. If swap is enabled, memory soft protection effectively postpones problems, allowing a leaking application to fill all swap area, which makes no sense. The only effective way to guarantee the memory protection in this case is to invoke the OOM killer. It's possible to handle this case in userspace by reacting on MEMCG_LOW events; but there is still a place for a fail-safe in-kernel mechanism to provide stronger guarantees. This patch introduces the memory.min interface for cgroup v2 memory controller. It works very similarly to memory.low (sharing the same hierarchical behavior), except that it's not disabled if there is no more reclaimable memory in the system. If cgroup is not populated, its memory.min is ignored, because otherwise even the OOM killer wouldn't be able to reclaim the protected memory, and the system can stall. [guro@fb.com: s/low/min/ in docs] Link: http://lkml.kernel.org/r/20180510130758.GA9129@castle.DHCP.thefacebook.com Link: http://lkml.kernel.org/r/20180509180734.GA4856@castle.DHCP.thefacebook.com Signed-off-by: Roman Gushchin <guro@fb.com> Reviewed-by: Randy Dunlap <rdunlap@infradead.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:07:46 +00:00
* mem_cgroup_protected - check if memory consumption is in the normal range
mm/memcontrol: exclude @root from checks in mem_cgroup_low Make @root exclusive in mem_cgroup_low; it is never considered low when looked at directly and is not checked when traversing the tree. In effect, @root is handled identically to how root_mem_cgroup was previously handled by mem_cgroup_low. If @root is not excluded from the checks, a cgroup underneath @root will never be considered low during targeted reclaim of @root, e.g. due to memory.current > memory.high, unless @root is misconfigured to have memory.low > memory.high. Excluding @root enables using memory.low to prioritize memory usage between cgroups within a subtree of the hierarchy that is limited by memory.high or memory.max, e.g. when ROOT owns @root's controls but delegates the @root directory to a USER so that USER can create and administer children of @root. For example, given cgroup A with children B and C: A / \ B C and 1. A/memory.current > A/memory.high 2. A/B/memory.current < A/B/memory.low 3. A/C/memory.current >= A/C/memory.low As 'A' is high, i.e. triggers reclaim from 'A', and 'B' is low, we should reclaim from 'C' until 'A' is no longer high or until we can no longer reclaim from 'C'. If 'A', i.e. @root, isn't excluded by mem_cgroup_low when reclaming from 'A', then 'B' won't be considered low and we will reclaim indiscriminately from both 'B' and 'C'. Here is the test I used to confirm the bug and the patch. 20:00:55@sjchrist-vm ? ~ $ cat ~/.bin/memcg_low_test #!/bin/bash x62mb=$((62<<20)) x66mb=$((66<<20)) x94mb=$((94<<20)) x98mb=$((98<<20)) setup() { set -e if [[ -n $DEBUG ]]; then set -x fi trap teardown EXIT HUP INT TERM if [[ ! -e /mnt/1gb.swap ]]; then sudo fallocate -l 1G /mnt/1gb.swap > /dev/null sudo mkswap /mnt/1gb.swap > /dev/null fi if ! swapon --show=NAME | grep -q "/mnt/1gb.swap"; then sudo swapon /mnt/1gb.swap fi if [[ ! -e /cgroup/cgroup.controllers ]]; then sudo mount -t cgroup2 none /cgroup fi grep -q memory /cgroup/cgroup.controllers sudo sh -c "echo '+memory' > /cgroup/cgroup.subtree_control" sudo mkdir /cgroup/A && sudo chown $USER:$USER /cgroup/A sudo sh -c "echo '+memory' > /cgroup/A/cgroup.subtree_control" sudo sh -c "echo '96m' > /cgroup/A/memory.high" mkdir /cgroup/A/0 mkdir /cgroup/A/1 echo 64m > /cgroup/A/0/memory.low } teardown() { set +e trap - EXIT HUP INT TERM if [[ -z $1 ]]; then printf "\n" printf "%0.s*" {1..35} printf "\nFAILED!\n\n" tail /cgroup/A/**/memory.current printf "%0.s*" {1..35} printf "\n\n" fi ps | grep stress | tr -s ' ' | cut -f 2 -d ' ' | xargs -I % kill % sleep 2 if [[ -e /cgroup/A/0 ]]; then rmdir /cgroup/A/0 fi if [[ -e /cgroup/A/1 ]]; then rmdir /cgroup/A/1 fi if [[ -e /cgroup/A ]]; then sudo rmdir /cgroup/A fi } stress_test() { sudo sh -c "echo $$ > /cgroup/A/$1/cgroup.procs" stress --vm 1 --vm-bytes 64M --vm-keep > /dev/null & sudo sh -c "echo $$ > /cgroup/A/$2/cgroup.procs" stress --vm 1 --vm-bytes 64M --vm-keep > /dev/null & sudo sh -c "echo $$ > /cgroup/cgroup.procs" sleep 1 # A/0 should be consuming more memory than A/1 [[ $(cat /cgroup/A/0/memory.current) -ge $(cat /cgroup/A/1/memory.current) ]] # A/0 should be consuming ~64mb [[ $(cat /cgroup/A/0/memory.current) -ge $x62mb ]] && [[ $(cat /cgroup/A/0/memory.current) -le $x66mb ]] # A should cumulatively be consuming ~96mb [[ $(cat /cgroup/A/memory.current) -ge $x94mb ]] && [[ $(cat /cgroup/A/memory.current) -le $x98mb ]] # Stop the stressors ps | grep stress | tr -s ' ' | cut -f 2 -d ' ' | xargs -I % kill % } teardown 1 setup for ((i=1;i<=$1;i++)); do printf "ITERATION $i of $1 - stress_test 0 1" stress_test 0 1 printf "\x1b[2K\r" printf "ITERATION $i of $1 - stress_test 1 0" stress_test 1 0 printf "\x1b[2K\r" printf "ITERATION $i of $1 - PASSED\n" done teardown 1 echo PASSED! 20:11:26@sjchrist-vm ? ~ $ memcg_low_test 10 Link: http://lkml.kernel.org/r/1496434412-21005-1-git-send-email-sean.j.christopherson@intel.com Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Acked-by: Balbir Singh <bsingharora@gmail.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-07-10 22:48:05 +00:00
* @root: the top ancestor of the sub-tree being checked
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
* @memcg: the memory cgroup to check
*
mm: memory.low hierarchical behavior This patch aims to address an issue in current memory.low semantics, which makes it hard to use it in a hierarchy, where some leaf memory cgroups are more valuable than others. For example, there are memcgs A, A/B, A/C, A/D and A/E: A A/memory.low = 2G, A/memory.current = 6G //\\ BC DE B/memory.low = 3G B/memory.current = 2G C/memory.low = 1G C/memory.current = 2G D/memory.low = 0 D/memory.current = 2G E/memory.low = 10G E/memory.current = 0 If we apply memory pressure, B, C and D are reclaimed at the same pace while A's usage exceeds 2G. This is obviously wrong, as B's usage is fully below B's memory.low, and C has 1G of protection as well. Also, A is pushed to the size, which is less than A's 2G memory.low, which is also wrong. A simple bash script (provided below) can be used to reproduce the problem. Current results are: A: 1430097920 A/B: 711929856 A/C: 717426688 A/D: 741376 A/E: 0 To address the issue a concept of effective memory.low is introduced. Effective memory.low is always equal or less than original memory.low. In a case, when there is no memory.low overcommittment (and also for top-level cgroups), these two values are equal. Otherwise it's a part of parent's effective memory.low, calculated as a cgroup's memory.low usage divided by sum of sibling's memory.low usages (under memory.low usage I mean the size of actually protected memory: memory.current if memory.current < memory.low, 0 otherwise). It's necessary to track the actual usage, because otherwise an empty cgroup with memory.low set (A/E in my example) will affect actual memory distribution, which makes no sense. To avoid traversing the cgroup tree twice, page_counters code is reused. Calculating effective memory.low can be done in the reclaim path, as we conveniently traversing the cgroup tree from top to bottom and check memory.low on each level. So, it's a perfect place to calculate effective memory low and save it to use it for children cgroups. This also eliminates a need to traverse the cgroup tree from bottom to top each time to check if parent's guarantee is not exceeded. Setting/resetting effective memory.low is intentionally racy, but it's fine and shouldn't lead to any significant differences in actual memory distribution. With this patch applied results are matching the expectations: A: 2147930112 A/B: 1428721664 A/C: 718393344 A/D: 815104 A/E: 0 Test script: #!/bin/bash CGPATH="/sys/fs/cgroup" truncate /file1 --size 2G truncate /file2 --size 2G truncate /file3 --size 2G truncate /file4 --size 50G mkdir "${CGPATH}/A" echo "+memory" > "${CGPATH}/A/cgroup.subtree_control" mkdir "${CGPATH}/A/B" "${CGPATH}/A/C" "${CGPATH}/A/D" "${CGPATH}/A/E" echo 2G > "${CGPATH}/A/memory.low" echo 3G > "${CGPATH}/A/B/memory.low" echo 1G > "${CGPATH}/A/C/memory.low" echo 0 > "${CGPATH}/A/D/memory.low" echo 10G > "${CGPATH}/A/E/memory.low" echo $$ > "${CGPATH}/A/B/cgroup.procs" && vmtouch -qt /file1 echo $$ > "${CGPATH}/A/C/cgroup.procs" && vmtouch -qt /file2 echo $$ > "${CGPATH}/A/D/cgroup.procs" && vmtouch -qt /file3 echo $$ > "${CGPATH}/cgroup.procs" && vmtouch -qt /file4 echo "A: " `cat "${CGPATH}/A/memory.current"` echo "A/B: " `cat "${CGPATH}/A/B/memory.current"` echo "A/C: " `cat "${CGPATH}/A/C/memory.current"` echo "A/D: " `cat "${CGPATH}/A/D/memory.current"` echo "A/E: " `cat "${CGPATH}/A/E/memory.current"` rmdir "${CGPATH}/A/B" "${CGPATH}/A/C" "${CGPATH}/A/D" "${CGPATH}/A/E" rmdir "${CGPATH}/A" rm /file1 /file2 /file3 /file4 Link: http://lkml.kernel.org/r/20180405185921.4942-2-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:06:22 +00:00
* WARNING: This function is not stateless! It can only be used as part
* of a top-down tree iteration, not for isolated queries.
mm/memcontrol: exclude @root from checks in mem_cgroup_low Make @root exclusive in mem_cgroup_low; it is never considered low when looked at directly and is not checked when traversing the tree. In effect, @root is handled identically to how root_mem_cgroup was previously handled by mem_cgroup_low. If @root is not excluded from the checks, a cgroup underneath @root will never be considered low during targeted reclaim of @root, e.g. due to memory.current > memory.high, unless @root is misconfigured to have memory.low > memory.high. Excluding @root enables using memory.low to prioritize memory usage between cgroups within a subtree of the hierarchy that is limited by memory.high or memory.max, e.g. when ROOT owns @root's controls but delegates the @root directory to a USER so that USER can create and administer children of @root. For example, given cgroup A with children B and C: A / \ B C and 1. A/memory.current > A/memory.high 2. A/B/memory.current < A/B/memory.low 3. A/C/memory.current >= A/C/memory.low As 'A' is high, i.e. triggers reclaim from 'A', and 'B' is low, we should reclaim from 'C' until 'A' is no longer high or until we can no longer reclaim from 'C'. If 'A', i.e. @root, isn't excluded by mem_cgroup_low when reclaming from 'A', then 'B' won't be considered low and we will reclaim indiscriminately from both 'B' and 'C'. Here is the test I used to confirm the bug and the patch. 20:00:55@sjchrist-vm ? ~ $ cat ~/.bin/memcg_low_test #!/bin/bash x62mb=$((62<<20)) x66mb=$((66<<20)) x94mb=$((94<<20)) x98mb=$((98<<20)) setup() { set -e if [[ -n $DEBUG ]]; then set -x fi trap teardown EXIT HUP INT TERM if [[ ! -e /mnt/1gb.swap ]]; then sudo fallocate -l 1G /mnt/1gb.swap > /dev/null sudo mkswap /mnt/1gb.swap > /dev/null fi if ! swapon --show=NAME | grep -q "/mnt/1gb.swap"; then sudo swapon /mnt/1gb.swap fi if [[ ! -e /cgroup/cgroup.controllers ]]; then sudo mount -t cgroup2 none /cgroup fi grep -q memory /cgroup/cgroup.controllers sudo sh -c "echo '+memory' > /cgroup/cgroup.subtree_control" sudo mkdir /cgroup/A && sudo chown $USER:$USER /cgroup/A sudo sh -c "echo '+memory' > /cgroup/A/cgroup.subtree_control" sudo sh -c "echo '96m' > /cgroup/A/memory.high" mkdir /cgroup/A/0 mkdir /cgroup/A/1 echo 64m > /cgroup/A/0/memory.low } teardown() { set +e trap - EXIT HUP INT TERM if [[ -z $1 ]]; then printf "\n" printf "%0.s*" {1..35} printf "\nFAILED!\n\n" tail /cgroup/A/**/memory.current printf "%0.s*" {1..35} printf "\n\n" fi ps | grep stress | tr -s ' ' | cut -f 2 -d ' ' | xargs -I % kill % sleep 2 if [[ -e /cgroup/A/0 ]]; then rmdir /cgroup/A/0 fi if [[ -e /cgroup/A/1 ]]; then rmdir /cgroup/A/1 fi if [[ -e /cgroup/A ]]; then sudo rmdir /cgroup/A fi } stress_test() { sudo sh -c "echo $$ > /cgroup/A/$1/cgroup.procs" stress --vm 1 --vm-bytes 64M --vm-keep > /dev/null & sudo sh -c "echo $$ > /cgroup/A/$2/cgroup.procs" stress --vm 1 --vm-bytes 64M --vm-keep > /dev/null & sudo sh -c "echo $$ > /cgroup/cgroup.procs" sleep 1 # A/0 should be consuming more memory than A/1 [[ $(cat /cgroup/A/0/memory.current) -ge $(cat /cgroup/A/1/memory.current) ]] # A/0 should be consuming ~64mb [[ $(cat /cgroup/A/0/memory.current) -ge $x62mb ]] && [[ $(cat /cgroup/A/0/memory.current) -le $x66mb ]] # A should cumulatively be consuming ~96mb [[ $(cat /cgroup/A/memory.current) -ge $x94mb ]] && [[ $(cat /cgroup/A/memory.current) -le $x98mb ]] # Stop the stressors ps | grep stress | tr -s ' ' | cut -f 2 -d ' ' | xargs -I % kill % } teardown 1 setup for ((i=1;i<=$1;i++)); do printf "ITERATION $i of $1 - stress_test 0 1" stress_test 0 1 printf "\x1b[2K\r" printf "ITERATION $i of $1 - stress_test 1 0" stress_test 1 0 printf "\x1b[2K\r" printf "ITERATION $i of $1 - PASSED\n" done teardown 1 echo PASSED! 20:11:26@sjchrist-vm ? ~ $ memcg_low_test 10 Link: http://lkml.kernel.org/r/1496434412-21005-1-git-send-email-sean.j.christopherson@intel.com Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Acked-by: Balbir Singh <bsingharora@gmail.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-07-10 22:48:05 +00:00
*
memcg: introduce memory.min Memory controller implements the memory.low best-effort memory protection mechanism, which works perfectly in many cases and allows protecting working sets of important workloads from sudden reclaim. But its semantics has a significant limitation: it works only as long as there is a supply of reclaimable memory. This makes it pretty useless against any sort of slow memory leaks or memory usage increases. This is especially true for swapless systems. If swap is enabled, memory soft protection effectively postpones problems, allowing a leaking application to fill all swap area, which makes no sense. The only effective way to guarantee the memory protection in this case is to invoke the OOM killer. It's possible to handle this case in userspace by reacting on MEMCG_LOW events; but there is still a place for a fail-safe in-kernel mechanism to provide stronger guarantees. This patch introduces the memory.min interface for cgroup v2 memory controller. It works very similarly to memory.low (sharing the same hierarchical behavior), except that it's not disabled if there is no more reclaimable memory in the system. If cgroup is not populated, its memory.min is ignored, because otherwise even the OOM killer wouldn't be able to reclaim the protected memory, and the system can stall. [guro@fb.com: s/low/min/ in docs] Link: http://lkml.kernel.org/r/20180510130758.GA9129@castle.DHCP.thefacebook.com Link: http://lkml.kernel.org/r/20180509180734.GA4856@castle.DHCP.thefacebook.com Signed-off-by: Roman Gushchin <guro@fb.com> Reviewed-by: Randy Dunlap <rdunlap@infradead.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:07:46 +00:00
* Returns one of the following:
* MEMCG_PROT_NONE: cgroup memory is not protected
* MEMCG_PROT_LOW: cgroup memory is protected as long there is
* an unprotected supply of reclaimable memory from other cgroups.
* MEMCG_PROT_MIN: cgroup memory is protected
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
*/
memcg: introduce memory.min Memory controller implements the memory.low best-effort memory protection mechanism, which works perfectly in many cases and allows protecting working sets of important workloads from sudden reclaim. But its semantics has a significant limitation: it works only as long as there is a supply of reclaimable memory. This makes it pretty useless against any sort of slow memory leaks or memory usage increases. This is especially true for swapless systems. If swap is enabled, memory soft protection effectively postpones problems, allowing a leaking application to fill all swap area, which makes no sense. The only effective way to guarantee the memory protection in this case is to invoke the OOM killer. It's possible to handle this case in userspace by reacting on MEMCG_LOW events; but there is still a place for a fail-safe in-kernel mechanism to provide stronger guarantees. This patch introduces the memory.min interface for cgroup v2 memory controller. It works very similarly to memory.low (sharing the same hierarchical behavior), except that it's not disabled if there is no more reclaimable memory in the system. If cgroup is not populated, its memory.min is ignored, because otherwise even the OOM killer wouldn't be able to reclaim the protected memory, and the system can stall. [guro@fb.com: s/low/min/ in docs] Link: http://lkml.kernel.org/r/20180510130758.GA9129@castle.DHCP.thefacebook.com Link: http://lkml.kernel.org/r/20180509180734.GA4856@castle.DHCP.thefacebook.com Signed-off-by: Roman Gushchin <guro@fb.com> Reviewed-by: Randy Dunlap <rdunlap@infradead.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:07:46 +00:00
enum mem_cgroup_protection mem_cgroup_protected(struct mem_cgroup *root,
struct mem_cgroup *memcg)
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
{
mm: memcontrol: recursive memory.low protection Right now, the effective protection of any given cgroup is capped by its own explicit memory.low setting, regardless of what the parent says. The reasons for this are mostly historical and ease of implementation: to make delegation of memory.low safe, effective protection is the min() of all memory.low up the tree. Unfortunately, this limitation makes it impossible to protect an entire subtree from another without forcing the user to make explicit protection allocations all the way to the leaf cgroups - something that is highly undesirable in real life scenarios. Consider memory in a data center host. At the cgroup top level, we have a distinction between system management software and the actual workload the system is executing. Both branches are further subdivided into individual services, job components etc. We want to protect the workload as a whole from the system management software, but that doesn't mean we want to protect and prioritize individual workload wrt each other. Their memory demand can vary over time, and we'd want the VM to simply cache the hottest data within the workload subtree. Yet, the current memory.low limitations force us to allocate a fixed amount of protection to each workload component in order to get protection from system management software in general. This results in very inefficient resource distribution. Another concern with mandating downward allocation is that, as the complexity of the cgroup tree grows, it gets harder for the lower levels to be informed about decisions made at the host-level. Consider a container inside a namespace that in turn creates its own nested tree of cgroups to run multiple workloads. It'd be extremely difficult to configure memory.low parameters in those leaf cgroups that on one hand balance pressure among siblings as the container desires, while also reflecting the host-level protection from e.g. rpm upgrades, that lie beyond one or more delegation and namespacing points in the tree. It's highly unusual from a cgroup interface POV that nested levels have to be aware of and reflect decisions made at higher levels for them to be effective. To enable such use cases and scale configurability for complex trees, this patch implements a resource inheritance model for memory that is similar to how the CPU and the IO controller implement work-conserving resource allocations: a share of a resource allocated to a subree always applies to the entire subtree recursively, while allowing, but not mandating, children to further specify distribution rules. That means that if protection is explicitly allocated among siblings, those configured shares are being followed during page reclaim just like they are now. However, if the memory.low set at a higher level is not fully claimed by the children in that subtree, the "floating" remainder is applied to each cgroup in the tree in proportion to its size. Since reclaim pressure is applied in proportion to size as well, each child in that tree gets the same boost, and the effect is neutral among siblings - with respect to each other, they behave as if no memory control was enabled at all, and the VM simply balances the memory demands optimally within the subtree. But collectively those cgroups enjoy a boost over the cgroups in neighboring trees. E.g. a leaf cgroup with a memory.low setting of 0 no longer means that it's not getting a share of the hierarchically assigned resource, just that it doesn't claim a fixed amount of it to protect from its siblings. This allows us to recursively protect one subtree (workload) from another (system management), while letting subgroups compete freely among each other - without having to assign fixed shares to each leaf, and without nested groups having to echo higher-level settings. The floating protection composes naturally with fixed protection. Consider the following example tree: A A: low = 2G / \ A1: low = 1G A1 A2 A2: low = 0G As outside pressure is applied to this tree, A1 will enjoy a fixed protection from A2 of 1G, but the remaining, unclaimed 1G from A is split evenly among A1 and A2, coming out to 1.5G and 0.5G. There is a slight risk of regressing theoretical setups where the top-level cgroups don't know about the true budgeting and set bogusly high "bypass" values that are meaningfully allocated down the tree. Such setups would rely on unclaimed protection to be discarded, and distributing it would change the intended behavior. Be safe and hide the new behavior behind a mount option, 'memory_recursiveprot'. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Tejun Heo <tj@kernel.org> Acked-by: Roman Gushchin <guro@fb.com> Acked-by: Chris Down <chris@chrisdown.name> Cc: Michal Hocko <mhocko@suse.com> Cc: Michal Koutný <mkoutny@suse.com> Link: http://lkml.kernel.org/r/20200227195606.46212-4-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-04-02 04:07:07 +00:00
unsigned long usage, parent_usage;
mm: memory.low hierarchical behavior This patch aims to address an issue in current memory.low semantics, which makes it hard to use it in a hierarchy, where some leaf memory cgroups are more valuable than others. For example, there are memcgs A, A/B, A/C, A/D and A/E: A A/memory.low = 2G, A/memory.current = 6G //\\ BC DE B/memory.low = 3G B/memory.current = 2G C/memory.low = 1G C/memory.current = 2G D/memory.low = 0 D/memory.current = 2G E/memory.low = 10G E/memory.current = 0 If we apply memory pressure, B, C and D are reclaimed at the same pace while A's usage exceeds 2G. This is obviously wrong, as B's usage is fully below B's memory.low, and C has 1G of protection as well. Also, A is pushed to the size, which is less than A's 2G memory.low, which is also wrong. A simple bash script (provided below) can be used to reproduce the problem. Current results are: A: 1430097920 A/B: 711929856 A/C: 717426688 A/D: 741376 A/E: 0 To address the issue a concept of effective memory.low is introduced. Effective memory.low is always equal or less than original memory.low. In a case, when there is no memory.low overcommittment (and also for top-level cgroups), these two values are equal. Otherwise it's a part of parent's effective memory.low, calculated as a cgroup's memory.low usage divided by sum of sibling's memory.low usages (under memory.low usage I mean the size of actually protected memory: memory.current if memory.current < memory.low, 0 otherwise). It's necessary to track the actual usage, because otherwise an empty cgroup with memory.low set (A/E in my example) will affect actual memory distribution, which makes no sense. To avoid traversing the cgroup tree twice, page_counters code is reused. Calculating effective memory.low can be done in the reclaim path, as we conveniently traversing the cgroup tree from top to bottom and check memory.low on each level. So, it's a perfect place to calculate effective memory low and save it to use it for children cgroups. This also eliminates a need to traverse the cgroup tree from bottom to top each time to check if parent's guarantee is not exceeded. Setting/resetting effective memory.low is intentionally racy, but it's fine and shouldn't lead to any significant differences in actual memory distribution. With this patch applied results are matching the expectations: A: 2147930112 A/B: 1428721664 A/C: 718393344 A/D: 815104 A/E: 0 Test script: #!/bin/bash CGPATH="/sys/fs/cgroup" truncate /file1 --size 2G truncate /file2 --size 2G truncate /file3 --size 2G truncate /file4 --size 50G mkdir "${CGPATH}/A" echo "+memory" > "${CGPATH}/A/cgroup.subtree_control" mkdir "${CGPATH}/A/B" "${CGPATH}/A/C" "${CGPATH}/A/D" "${CGPATH}/A/E" echo 2G > "${CGPATH}/A/memory.low" echo 3G > "${CGPATH}/A/B/memory.low" echo 1G > "${CGPATH}/A/C/memory.low" echo 0 > "${CGPATH}/A/D/memory.low" echo 10G > "${CGPATH}/A/E/memory.low" echo $$ > "${CGPATH}/A/B/cgroup.procs" && vmtouch -qt /file1 echo $$ > "${CGPATH}/A/C/cgroup.procs" && vmtouch -qt /file2 echo $$ > "${CGPATH}/A/D/cgroup.procs" && vmtouch -qt /file3 echo $$ > "${CGPATH}/cgroup.procs" && vmtouch -qt /file4 echo "A: " `cat "${CGPATH}/A/memory.current"` echo "A/B: " `cat "${CGPATH}/A/B/memory.current"` echo "A/C: " `cat "${CGPATH}/A/C/memory.current"` echo "A/D: " `cat "${CGPATH}/A/D/memory.current"` echo "A/E: " `cat "${CGPATH}/A/E/memory.current"` rmdir "${CGPATH}/A/B" "${CGPATH}/A/C" "${CGPATH}/A/D" "${CGPATH}/A/E" rmdir "${CGPATH}/A" rm /file1 /file2 /file3 /file4 Link: http://lkml.kernel.org/r/20180405185921.4942-2-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:06:22 +00:00
struct mem_cgroup *parent;
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
if (mem_cgroup_disabled())
memcg: introduce memory.min Memory controller implements the memory.low best-effort memory protection mechanism, which works perfectly in many cases and allows protecting working sets of important workloads from sudden reclaim. But its semantics has a significant limitation: it works only as long as there is a supply of reclaimable memory. This makes it pretty useless against any sort of slow memory leaks or memory usage increases. This is especially true for swapless systems. If swap is enabled, memory soft protection effectively postpones problems, allowing a leaking application to fill all swap area, which makes no sense. The only effective way to guarantee the memory protection in this case is to invoke the OOM killer. It's possible to handle this case in userspace by reacting on MEMCG_LOW events; but there is still a place for a fail-safe in-kernel mechanism to provide stronger guarantees. This patch introduces the memory.min interface for cgroup v2 memory controller. It works very similarly to memory.low (sharing the same hierarchical behavior), except that it's not disabled if there is no more reclaimable memory in the system. If cgroup is not populated, its memory.min is ignored, because otherwise even the OOM killer wouldn't be able to reclaim the protected memory, and the system can stall. [guro@fb.com: s/low/min/ in docs] Link: http://lkml.kernel.org/r/20180510130758.GA9129@castle.DHCP.thefacebook.com Link: http://lkml.kernel.org/r/20180509180734.GA4856@castle.DHCP.thefacebook.com Signed-off-by: Roman Gushchin <guro@fb.com> Reviewed-by: Randy Dunlap <rdunlap@infradead.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:07:46 +00:00
return MEMCG_PROT_NONE;
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
mm/memcontrol: exclude @root from checks in mem_cgroup_low Make @root exclusive in mem_cgroup_low; it is never considered low when looked at directly and is not checked when traversing the tree. In effect, @root is handled identically to how root_mem_cgroup was previously handled by mem_cgroup_low. If @root is not excluded from the checks, a cgroup underneath @root will never be considered low during targeted reclaim of @root, e.g. due to memory.current > memory.high, unless @root is misconfigured to have memory.low > memory.high. Excluding @root enables using memory.low to prioritize memory usage between cgroups within a subtree of the hierarchy that is limited by memory.high or memory.max, e.g. when ROOT owns @root's controls but delegates the @root directory to a USER so that USER can create and administer children of @root. For example, given cgroup A with children B and C: A / \ B C and 1. A/memory.current > A/memory.high 2. A/B/memory.current < A/B/memory.low 3. A/C/memory.current >= A/C/memory.low As 'A' is high, i.e. triggers reclaim from 'A', and 'B' is low, we should reclaim from 'C' until 'A' is no longer high or until we can no longer reclaim from 'C'. If 'A', i.e. @root, isn't excluded by mem_cgroup_low when reclaming from 'A', then 'B' won't be considered low and we will reclaim indiscriminately from both 'B' and 'C'. Here is the test I used to confirm the bug and the patch. 20:00:55@sjchrist-vm ? ~ $ cat ~/.bin/memcg_low_test #!/bin/bash x62mb=$((62<<20)) x66mb=$((66<<20)) x94mb=$((94<<20)) x98mb=$((98<<20)) setup() { set -e if [[ -n $DEBUG ]]; then set -x fi trap teardown EXIT HUP INT TERM if [[ ! -e /mnt/1gb.swap ]]; then sudo fallocate -l 1G /mnt/1gb.swap > /dev/null sudo mkswap /mnt/1gb.swap > /dev/null fi if ! swapon --show=NAME | grep -q "/mnt/1gb.swap"; then sudo swapon /mnt/1gb.swap fi if [[ ! -e /cgroup/cgroup.controllers ]]; then sudo mount -t cgroup2 none /cgroup fi grep -q memory /cgroup/cgroup.controllers sudo sh -c "echo '+memory' > /cgroup/cgroup.subtree_control" sudo mkdir /cgroup/A && sudo chown $USER:$USER /cgroup/A sudo sh -c "echo '+memory' > /cgroup/A/cgroup.subtree_control" sudo sh -c "echo '96m' > /cgroup/A/memory.high" mkdir /cgroup/A/0 mkdir /cgroup/A/1 echo 64m > /cgroup/A/0/memory.low } teardown() { set +e trap - EXIT HUP INT TERM if [[ -z $1 ]]; then printf "\n" printf "%0.s*" {1..35} printf "\nFAILED!\n\n" tail /cgroup/A/**/memory.current printf "%0.s*" {1..35} printf "\n\n" fi ps | grep stress | tr -s ' ' | cut -f 2 -d ' ' | xargs -I % kill % sleep 2 if [[ -e /cgroup/A/0 ]]; then rmdir /cgroup/A/0 fi if [[ -e /cgroup/A/1 ]]; then rmdir /cgroup/A/1 fi if [[ -e /cgroup/A ]]; then sudo rmdir /cgroup/A fi } stress_test() { sudo sh -c "echo $$ > /cgroup/A/$1/cgroup.procs" stress --vm 1 --vm-bytes 64M --vm-keep > /dev/null & sudo sh -c "echo $$ > /cgroup/A/$2/cgroup.procs" stress --vm 1 --vm-bytes 64M --vm-keep > /dev/null & sudo sh -c "echo $$ > /cgroup/cgroup.procs" sleep 1 # A/0 should be consuming more memory than A/1 [[ $(cat /cgroup/A/0/memory.current) -ge $(cat /cgroup/A/1/memory.current) ]] # A/0 should be consuming ~64mb [[ $(cat /cgroup/A/0/memory.current) -ge $x62mb ]] && [[ $(cat /cgroup/A/0/memory.current) -le $x66mb ]] # A should cumulatively be consuming ~96mb [[ $(cat /cgroup/A/memory.current) -ge $x94mb ]] && [[ $(cat /cgroup/A/memory.current) -le $x98mb ]] # Stop the stressors ps | grep stress | tr -s ' ' | cut -f 2 -d ' ' | xargs -I % kill % } teardown 1 setup for ((i=1;i<=$1;i++)); do printf "ITERATION $i of $1 - stress_test 0 1" stress_test 0 1 printf "\x1b[2K\r" printf "ITERATION $i of $1 - stress_test 1 0" stress_test 1 0 printf "\x1b[2K\r" printf "ITERATION $i of $1 - PASSED\n" done teardown 1 echo PASSED! 20:11:26@sjchrist-vm ? ~ $ memcg_low_test 10 Link: http://lkml.kernel.org/r/1496434412-21005-1-git-send-email-sean.j.christopherson@intel.com Signed-off-by: Sean Christopherson <sean.j.christopherson@intel.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Acked-by: Balbir Singh <bsingharora@gmail.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-07-10 22:48:05 +00:00
if (!root)
root = root_mem_cgroup;
if (memcg == root)
memcg: introduce memory.min Memory controller implements the memory.low best-effort memory protection mechanism, which works perfectly in many cases and allows protecting working sets of important workloads from sudden reclaim. But its semantics has a significant limitation: it works only as long as there is a supply of reclaimable memory. This makes it pretty useless against any sort of slow memory leaks or memory usage increases. This is especially true for swapless systems. If swap is enabled, memory soft protection effectively postpones problems, allowing a leaking application to fill all swap area, which makes no sense. The only effective way to guarantee the memory protection in this case is to invoke the OOM killer. It's possible to handle this case in userspace by reacting on MEMCG_LOW events; but there is still a place for a fail-safe in-kernel mechanism to provide stronger guarantees. This patch introduces the memory.min interface for cgroup v2 memory controller. It works very similarly to memory.low (sharing the same hierarchical behavior), except that it's not disabled if there is no more reclaimable memory in the system. If cgroup is not populated, its memory.min is ignored, because otherwise even the OOM killer wouldn't be able to reclaim the protected memory, and the system can stall. [guro@fb.com: s/low/min/ in docs] Link: http://lkml.kernel.org/r/20180510130758.GA9129@castle.DHCP.thefacebook.com Link: http://lkml.kernel.org/r/20180509180734.GA4856@castle.DHCP.thefacebook.com Signed-off-by: Roman Gushchin <guro@fb.com> Reviewed-by: Randy Dunlap <rdunlap@infradead.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:07:46 +00:00
return MEMCG_PROT_NONE;
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
mm: memory.low hierarchical behavior This patch aims to address an issue in current memory.low semantics, which makes it hard to use it in a hierarchy, where some leaf memory cgroups are more valuable than others. For example, there are memcgs A, A/B, A/C, A/D and A/E: A A/memory.low = 2G, A/memory.current = 6G //\\ BC DE B/memory.low = 3G B/memory.current = 2G C/memory.low = 1G C/memory.current = 2G D/memory.low = 0 D/memory.current = 2G E/memory.low = 10G E/memory.current = 0 If we apply memory pressure, B, C and D are reclaimed at the same pace while A's usage exceeds 2G. This is obviously wrong, as B's usage is fully below B's memory.low, and C has 1G of protection as well. Also, A is pushed to the size, which is less than A's 2G memory.low, which is also wrong. A simple bash script (provided below) can be used to reproduce the problem. Current results are: A: 1430097920 A/B: 711929856 A/C: 717426688 A/D: 741376 A/E: 0 To address the issue a concept of effective memory.low is introduced. Effective memory.low is always equal or less than original memory.low. In a case, when there is no memory.low overcommittment (and also for top-level cgroups), these two values are equal. Otherwise it's a part of parent's effective memory.low, calculated as a cgroup's memory.low usage divided by sum of sibling's memory.low usages (under memory.low usage I mean the size of actually protected memory: memory.current if memory.current < memory.low, 0 otherwise). It's necessary to track the actual usage, because otherwise an empty cgroup with memory.low set (A/E in my example) will affect actual memory distribution, which makes no sense. To avoid traversing the cgroup tree twice, page_counters code is reused. Calculating effective memory.low can be done in the reclaim path, as we conveniently traversing the cgroup tree from top to bottom and check memory.low on each level. So, it's a perfect place to calculate effective memory low and save it to use it for children cgroups. This also eliminates a need to traverse the cgroup tree from bottom to top each time to check if parent's guarantee is not exceeded. Setting/resetting effective memory.low is intentionally racy, but it's fine and shouldn't lead to any significant differences in actual memory distribution. With this patch applied results are matching the expectations: A: 2147930112 A/B: 1428721664 A/C: 718393344 A/D: 815104 A/E: 0 Test script: #!/bin/bash CGPATH="/sys/fs/cgroup" truncate /file1 --size 2G truncate /file2 --size 2G truncate /file3 --size 2G truncate /file4 --size 50G mkdir "${CGPATH}/A" echo "+memory" > "${CGPATH}/A/cgroup.subtree_control" mkdir "${CGPATH}/A/B" "${CGPATH}/A/C" "${CGPATH}/A/D" "${CGPATH}/A/E" echo 2G > "${CGPATH}/A/memory.low" echo 3G > "${CGPATH}/A/B/memory.low" echo 1G > "${CGPATH}/A/C/memory.low" echo 0 > "${CGPATH}/A/D/memory.low" echo 10G > "${CGPATH}/A/E/memory.low" echo $$ > "${CGPATH}/A/B/cgroup.procs" && vmtouch -qt /file1 echo $$ > "${CGPATH}/A/C/cgroup.procs" && vmtouch -qt /file2 echo $$ > "${CGPATH}/A/D/cgroup.procs" && vmtouch -qt /file3 echo $$ > "${CGPATH}/cgroup.procs" && vmtouch -qt /file4 echo "A: " `cat "${CGPATH}/A/memory.current"` echo "A/B: " `cat "${CGPATH}/A/B/memory.current"` echo "A/C: " `cat "${CGPATH}/A/C/memory.current"` echo "A/D: " `cat "${CGPATH}/A/D/memory.current"` echo "A/E: " `cat "${CGPATH}/A/E/memory.current"` rmdir "${CGPATH}/A/B" "${CGPATH}/A/C" "${CGPATH}/A/D" "${CGPATH}/A/E" rmdir "${CGPATH}/A" rm /file1 /file2 /file3 /file4 Link: http://lkml.kernel.org/r/20180405185921.4942-2-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:06:22 +00:00
usage = page_counter_read(&memcg->memory);
memcg: introduce memory.min Memory controller implements the memory.low best-effort memory protection mechanism, which works perfectly in many cases and allows protecting working sets of important workloads from sudden reclaim. But its semantics has a significant limitation: it works only as long as there is a supply of reclaimable memory. This makes it pretty useless against any sort of slow memory leaks or memory usage increases. This is especially true for swapless systems. If swap is enabled, memory soft protection effectively postpones problems, allowing a leaking application to fill all swap area, which makes no sense. The only effective way to guarantee the memory protection in this case is to invoke the OOM killer. It's possible to handle this case in userspace by reacting on MEMCG_LOW events; but there is still a place for a fail-safe in-kernel mechanism to provide stronger guarantees. This patch introduces the memory.min interface for cgroup v2 memory controller. It works very similarly to memory.low (sharing the same hierarchical behavior), except that it's not disabled if there is no more reclaimable memory in the system. If cgroup is not populated, its memory.min is ignored, because otherwise even the OOM killer wouldn't be able to reclaim the protected memory, and the system can stall. [guro@fb.com: s/low/min/ in docs] Link: http://lkml.kernel.org/r/20180510130758.GA9129@castle.DHCP.thefacebook.com Link: http://lkml.kernel.org/r/20180509180734.GA4856@castle.DHCP.thefacebook.com Signed-off-by: Roman Gushchin <guro@fb.com> Reviewed-by: Randy Dunlap <rdunlap@infradead.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:07:46 +00:00
if (!usage)
return MEMCG_PROT_NONE;
parent = parent_mem_cgroup(memcg);
mm: fix null pointer dereference in mem_cgroup_protected Shakeel reported a crash in mem_cgroup_protected(), which can be triggered by memcg reclaim if the legacy cgroup v1 use_hierarchy=0 mode is used: BUG: unable to handle kernel NULL pointer dereference at 0000000000000120 PGD 8000001ff55da067 P4D 8000001ff55da067 PUD 1fdc7df067 PMD 0 Oops: 0000 [#4] SMP PTI CPU: 0 PID: 15581 Comm: bash Tainted: G D 4.17.0-smp-clean #5 Hardware name: ... RIP: 0010:mem_cgroup_protected+0x54/0x130 Code: 4c 8b 8e 00 01 00 00 4c 8b 86 08 01 00 00 48 8d 8a 08 ff ff ff 48 85 d2 ba 00 00 00 00 48 0f 44 ca 48 39 c8 0f 84 cf 00 00 00 <48> 8b 81 20 01 00 00 4d 89 ca 4c 39 c8 4c 0f 46 d0 4d 85 d2 74 05 RSP: 0000:ffffabe64dfafa58 EFLAGS: 00010286 RAX: ffff9fb6ff03d000 RBX: ffff9fb6f5b1b000 RCX: 0000000000000000 RDX: 0000000000000000 RSI: ffff9fb6f5b1b000 RDI: ffff9fb6f5b1b000 RBP: ffffabe64dfafb08 R08: 0000000000000000 R09: 0000000000000000 R10: 0000000000000000 R11: 000000000000c800 R12: ffffabe64dfafb88 R13: ffff9fb6f5b1b000 R14: ffffabe64dfafb88 R15: ffff9fb77fffe000 FS: 00007fed1f8ac700(0000) GS:ffff9fb6ff400000(0000) knlGS:0000000000000000 CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 CR2: 0000000000000120 CR3: 0000001fdcf86003 CR4: 00000000001606f0 Call Trace: ? shrink_node+0x194/0x510 do_try_to_free_pages+0xfd/0x390 try_to_free_mem_cgroup_pages+0x123/0x210 try_charge+0x19e/0x700 mem_cgroup_try_charge+0x10b/0x1a0 wp_page_copy+0x134/0x5b0 do_wp_page+0x90/0x460 __handle_mm_fault+0x8e3/0xf30 handle_mm_fault+0xfe/0x220 __do_page_fault+0x262/0x500 do_page_fault+0x28/0xd0 ? page_fault+0x8/0x30 page_fault+0x1e/0x30 RIP: 0033:0x485b72 The problem happens because parent_mem_cgroup() returns a NULL pointer, which is dereferenced later without a check. As cgroup v1 has no memory guarantee support, let's make mem_cgroup_protected() immediately return MEMCG_PROT_NONE, if the given cgroup has no parent (non-hierarchical mode is used). Link: http://lkml.kernel.org/r/20180611175418.7007-2-guro@fb.com Fixes: bf8d5d52ffe8 ("memcg: introduce memory.min") Signed-off-by: Roman Gushchin <guro@fb.com> Reported-by: Shakeel Butt <shakeelb@google.com> Tested-by: Shakeel Butt <shakeelb@google.com> Tested-by: John Stultz <john.stultz@linaro.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-14 22:26:17 +00:00
/* No parent means a non-hierarchical mode on v1 memcg */
if (!parent)
return MEMCG_PROT_NONE;
2020-04-02 04:07:03 +00:00
if (parent == root) {
memcg->memory.emin = READ_ONCE(memcg->memory.min);
2020-04-02 04:07:03 +00:00
memcg->memory.elow = memcg->memory.low;
goto out;
memcg: introduce memory.min Memory controller implements the memory.low best-effort memory protection mechanism, which works perfectly in many cases and allows protecting working sets of important workloads from sudden reclaim. But its semantics has a significant limitation: it works only as long as there is a supply of reclaimable memory. This makes it pretty useless against any sort of slow memory leaks or memory usage increases. This is especially true for swapless systems. If swap is enabled, memory soft protection effectively postpones problems, allowing a leaking application to fill all swap area, which makes no sense. The only effective way to guarantee the memory protection in this case is to invoke the OOM killer. It's possible to handle this case in userspace by reacting on MEMCG_LOW events; but there is still a place for a fail-safe in-kernel mechanism to provide stronger guarantees. This patch introduces the memory.min interface for cgroup v2 memory controller. It works very similarly to memory.low (sharing the same hierarchical behavior), except that it's not disabled if there is no more reclaimable memory in the system. If cgroup is not populated, its memory.min is ignored, because otherwise even the OOM killer wouldn't be able to reclaim the protected memory, and the system can stall. [guro@fb.com: s/low/min/ in docs] Link: http://lkml.kernel.org/r/20180510130758.GA9129@castle.DHCP.thefacebook.com Link: http://lkml.kernel.org/r/20180509180734.GA4856@castle.DHCP.thefacebook.com Signed-off-by: Roman Gushchin <guro@fb.com> Reviewed-by: Randy Dunlap <rdunlap@infradead.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:07:46 +00:00
}
mm: memcontrol: recursive memory.low protection Right now, the effective protection of any given cgroup is capped by its own explicit memory.low setting, regardless of what the parent says. The reasons for this are mostly historical and ease of implementation: to make delegation of memory.low safe, effective protection is the min() of all memory.low up the tree. Unfortunately, this limitation makes it impossible to protect an entire subtree from another without forcing the user to make explicit protection allocations all the way to the leaf cgroups - something that is highly undesirable in real life scenarios. Consider memory in a data center host. At the cgroup top level, we have a distinction between system management software and the actual workload the system is executing. Both branches are further subdivided into individual services, job components etc. We want to protect the workload as a whole from the system management software, but that doesn't mean we want to protect and prioritize individual workload wrt each other. Their memory demand can vary over time, and we'd want the VM to simply cache the hottest data within the workload subtree. Yet, the current memory.low limitations force us to allocate a fixed amount of protection to each workload component in order to get protection from system management software in general. This results in very inefficient resource distribution. Another concern with mandating downward allocation is that, as the complexity of the cgroup tree grows, it gets harder for the lower levels to be informed about decisions made at the host-level. Consider a container inside a namespace that in turn creates its own nested tree of cgroups to run multiple workloads. It'd be extremely difficult to configure memory.low parameters in those leaf cgroups that on one hand balance pressure among siblings as the container desires, while also reflecting the host-level protection from e.g. rpm upgrades, that lie beyond one or more delegation and namespacing points in the tree. It's highly unusual from a cgroup interface POV that nested levels have to be aware of and reflect decisions made at higher levels for them to be effective. To enable such use cases and scale configurability for complex trees, this patch implements a resource inheritance model for memory that is similar to how the CPU and the IO controller implement work-conserving resource allocations: a share of a resource allocated to a subree always applies to the entire subtree recursively, while allowing, but not mandating, children to further specify distribution rules. That means that if protection is explicitly allocated among siblings, those configured shares are being followed during page reclaim just like they are now. However, if the memory.low set at a higher level is not fully claimed by the children in that subtree, the "floating" remainder is applied to each cgroup in the tree in proportion to its size. Since reclaim pressure is applied in proportion to size as well, each child in that tree gets the same boost, and the effect is neutral among siblings - with respect to each other, they behave as if no memory control was enabled at all, and the VM simply balances the memory demands optimally within the subtree. But collectively those cgroups enjoy a boost over the cgroups in neighboring trees. E.g. a leaf cgroup with a memory.low setting of 0 no longer means that it's not getting a share of the hierarchically assigned resource, just that it doesn't claim a fixed amount of it to protect from its siblings. This allows us to recursively protect one subtree (workload) from another (system management), while letting subgroups compete freely among each other - without having to assign fixed shares to each leaf, and without nested groups having to echo higher-level settings. The floating protection composes naturally with fixed protection. Consider the following example tree: A A: low = 2G / \ A1: low = 1G A1 A2 A2: low = 0G As outside pressure is applied to this tree, A1 will enjoy a fixed protection from A2 of 1G, but the remaining, unclaimed 1G from A is split evenly among A1 and A2, coming out to 1.5G and 0.5G. There is a slight risk of regressing theoretical setups where the top-level cgroups don't know about the true budgeting and set bogusly high "bypass" values that are meaningfully allocated down the tree. Such setups would rely on unclaimed protection to be discarded, and distributing it would change the intended behavior. Be safe and hide the new behavior behind a mount option, 'memory_recursiveprot'. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Tejun Heo <tj@kernel.org> Acked-by: Roman Gushchin <guro@fb.com> Acked-by: Chris Down <chris@chrisdown.name> Cc: Michal Hocko <mhocko@suse.com> Cc: Michal Koutný <mkoutny@suse.com> Link: http://lkml.kernel.org/r/20200227195606.46212-4-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-04-02 04:07:07 +00:00
parent_usage = page_counter_read(&parent->memory);
WRITE_ONCE(memcg->memory.emin, effective_protection(usage, parent_usage,
READ_ONCE(memcg->memory.min),
READ_ONCE(parent->memory.emin),
atomic_long_read(&parent->memory.children_min_usage)));
mm: memory.low hierarchical behavior This patch aims to address an issue in current memory.low semantics, which makes it hard to use it in a hierarchy, where some leaf memory cgroups are more valuable than others. For example, there are memcgs A, A/B, A/C, A/D and A/E: A A/memory.low = 2G, A/memory.current = 6G //\\ BC DE B/memory.low = 3G B/memory.current = 2G C/memory.low = 1G C/memory.current = 2G D/memory.low = 0 D/memory.current = 2G E/memory.low = 10G E/memory.current = 0 If we apply memory pressure, B, C and D are reclaimed at the same pace while A's usage exceeds 2G. This is obviously wrong, as B's usage is fully below B's memory.low, and C has 1G of protection as well. Also, A is pushed to the size, which is less than A's 2G memory.low, which is also wrong. A simple bash script (provided below) can be used to reproduce the problem. Current results are: A: 1430097920 A/B: 711929856 A/C: 717426688 A/D: 741376 A/E: 0 To address the issue a concept of effective memory.low is introduced. Effective memory.low is always equal or less than original memory.low. In a case, when there is no memory.low overcommittment (and also for top-level cgroups), these two values are equal. Otherwise it's a part of parent's effective memory.low, calculated as a cgroup's memory.low usage divided by sum of sibling's memory.low usages (under memory.low usage I mean the size of actually protected memory: memory.current if memory.current < memory.low, 0 otherwise). It's necessary to track the actual usage, because otherwise an empty cgroup with memory.low set (A/E in my example) will affect actual memory distribution, which makes no sense. To avoid traversing the cgroup tree twice, page_counters code is reused. Calculating effective memory.low can be done in the reclaim path, as we conveniently traversing the cgroup tree from top to bottom and check memory.low on each level. So, it's a perfect place to calculate effective memory low and save it to use it for children cgroups. This also eliminates a need to traverse the cgroup tree from bottom to top each time to check if parent's guarantee is not exceeded. Setting/resetting effective memory.low is intentionally racy, but it's fine and shouldn't lead to any significant differences in actual memory distribution. With this patch applied results are matching the expectations: A: 2147930112 A/B: 1428721664 A/C: 718393344 A/D: 815104 A/E: 0 Test script: #!/bin/bash CGPATH="/sys/fs/cgroup" truncate /file1 --size 2G truncate /file2 --size 2G truncate /file3 --size 2G truncate /file4 --size 50G mkdir "${CGPATH}/A" echo "+memory" > "${CGPATH}/A/cgroup.subtree_control" mkdir "${CGPATH}/A/B" "${CGPATH}/A/C" "${CGPATH}/A/D" "${CGPATH}/A/E" echo 2G > "${CGPATH}/A/memory.low" echo 3G > "${CGPATH}/A/B/memory.low" echo 1G > "${CGPATH}/A/C/memory.low" echo 0 > "${CGPATH}/A/D/memory.low" echo 10G > "${CGPATH}/A/E/memory.low" echo $$ > "${CGPATH}/A/B/cgroup.procs" && vmtouch -qt /file1 echo $$ > "${CGPATH}/A/C/cgroup.procs" && vmtouch -qt /file2 echo $$ > "${CGPATH}/A/D/cgroup.procs" && vmtouch -qt /file3 echo $$ > "${CGPATH}/cgroup.procs" && vmtouch -qt /file4 echo "A: " `cat "${CGPATH}/A/memory.current"` echo "A/B: " `cat "${CGPATH}/A/B/memory.current"` echo "A/C: " `cat "${CGPATH}/A/C/memory.current"` echo "A/D: " `cat "${CGPATH}/A/D/memory.current"` echo "A/E: " `cat "${CGPATH}/A/E/memory.current"` rmdir "${CGPATH}/A/B" "${CGPATH}/A/C" "${CGPATH}/A/D" "${CGPATH}/A/E" rmdir "${CGPATH}/A" rm /file1 /file2 /file3 /file4 Link: http://lkml.kernel.org/r/20180405185921.4942-2-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:06:22 +00:00
WRITE_ONCE(memcg->memory.elow, effective_protection(usage, parent_usage,
2020-04-02 04:07:03 +00:00
memcg->memory.low, READ_ONCE(parent->memory.elow),
atomic_long_read(&parent->memory.children_low_usage)));
mm: memory.low hierarchical behavior This patch aims to address an issue in current memory.low semantics, which makes it hard to use it in a hierarchy, where some leaf memory cgroups are more valuable than others. For example, there are memcgs A, A/B, A/C, A/D and A/E: A A/memory.low = 2G, A/memory.current = 6G //\\ BC DE B/memory.low = 3G B/memory.current = 2G C/memory.low = 1G C/memory.current = 2G D/memory.low = 0 D/memory.current = 2G E/memory.low = 10G E/memory.current = 0 If we apply memory pressure, B, C and D are reclaimed at the same pace while A's usage exceeds 2G. This is obviously wrong, as B's usage is fully below B's memory.low, and C has 1G of protection as well. Also, A is pushed to the size, which is less than A's 2G memory.low, which is also wrong. A simple bash script (provided below) can be used to reproduce the problem. Current results are: A: 1430097920 A/B: 711929856 A/C: 717426688 A/D: 741376 A/E: 0 To address the issue a concept of effective memory.low is introduced. Effective memory.low is always equal or less than original memory.low. In a case, when there is no memory.low overcommittment (and also for top-level cgroups), these two values are equal. Otherwise it's a part of parent's effective memory.low, calculated as a cgroup's memory.low usage divided by sum of sibling's memory.low usages (under memory.low usage I mean the size of actually protected memory: memory.current if memory.current < memory.low, 0 otherwise). It's necessary to track the actual usage, because otherwise an empty cgroup with memory.low set (A/E in my example) will affect actual memory distribution, which makes no sense. To avoid traversing the cgroup tree twice, page_counters code is reused. Calculating effective memory.low can be done in the reclaim path, as we conveniently traversing the cgroup tree from top to bottom and check memory.low on each level. So, it's a perfect place to calculate effective memory low and save it to use it for children cgroups. This also eliminates a need to traverse the cgroup tree from bottom to top each time to check if parent's guarantee is not exceeded. Setting/resetting effective memory.low is intentionally racy, but it's fine and shouldn't lead to any significant differences in actual memory distribution. With this patch applied results are matching the expectations: A: 2147930112 A/B: 1428721664 A/C: 718393344 A/D: 815104 A/E: 0 Test script: #!/bin/bash CGPATH="/sys/fs/cgroup" truncate /file1 --size 2G truncate /file2 --size 2G truncate /file3 --size 2G truncate /file4 --size 50G mkdir "${CGPATH}/A" echo "+memory" > "${CGPATH}/A/cgroup.subtree_control" mkdir "${CGPATH}/A/B" "${CGPATH}/A/C" "${CGPATH}/A/D" "${CGPATH}/A/E" echo 2G > "${CGPATH}/A/memory.low" echo 3G > "${CGPATH}/A/B/memory.low" echo 1G > "${CGPATH}/A/C/memory.low" echo 0 > "${CGPATH}/A/D/memory.low" echo 10G > "${CGPATH}/A/E/memory.low" echo $$ > "${CGPATH}/A/B/cgroup.procs" && vmtouch -qt /file1 echo $$ > "${CGPATH}/A/C/cgroup.procs" && vmtouch -qt /file2 echo $$ > "${CGPATH}/A/D/cgroup.procs" && vmtouch -qt /file3 echo $$ > "${CGPATH}/cgroup.procs" && vmtouch -qt /file4 echo "A: " `cat "${CGPATH}/A/memory.current"` echo "A/B: " `cat "${CGPATH}/A/B/memory.current"` echo "A/C: " `cat "${CGPATH}/A/C/memory.current"` echo "A/D: " `cat "${CGPATH}/A/D/memory.current"` echo "A/E: " `cat "${CGPATH}/A/E/memory.current"` rmdir "${CGPATH}/A/B" "${CGPATH}/A/C" "${CGPATH}/A/D" "${CGPATH}/A/E" rmdir "${CGPATH}/A" rm /file1 /file2 /file3 /file4 Link: http://lkml.kernel.org/r/20180405185921.4942-2-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:06:22 +00:00
2020-04-02 04:07:03 +00:00
out:
if (usage <= memcg->memory.emin)
memcg: introduce memory.min Memory controller implements the memory.low best-effort memory protection mechanism, which works perfectly in many cases and allows protecting working sets of important workloads from sudden reclaim. But its semantics has a significant limitation: it works only as long as there is a supply of reclaimable memory. This makes it pretty useless against any sort of slow memory leaks or memory usage increases. This is especially true for swapless systems. If swap is enabled, memory soft protection effectively postpones problems, allowing a leaking application to fill all swap area, which makes no sense. The only effective way to guarantee the memory protection in this case is to invoke the OOM killer. It's possible to handle this case in userspace by reacting on MEMCG_LOW events; but there is still a place for a fail-safe in-kernel mechanism to provide stronger guarantees. This patch introduces the memory.min interface for cgroup v2 memory controller. It works very similarly to memory.low (sharing the same hierarchical behavior), except that it's not disabled if there is no more reclaimable memory in the system. If cgroup is not populated, its memory.min is ignored, because otherwise even the OOM killer wouldn't be able to reclaim the protected memory, and the system can stall. [guro@fb.com: s/low/min/ in docs] Link: http://lkml.kernel.org/r/20180510130758.GA9129@castle.DHCP.thefacebook.com Link: http://lkml.kernel.org/r/20180509180734.GA4856@castle.DHCP.thefacebook.com Signed-off-by: Roman Gushchin <guro@fb.com> Reviewed-by: Randy Dunlap <rdunlap@infradead.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:07:46 +00:00
return MEMCG_PROT_MIN;
2020-04-02 04:07:03 +00:00
else if (usage <= memcg->memory.elow)
memcg: introduce memory.min Memory controller implements the memory.low best-effort memory protection mechanism, which works perfectly in many cases and allows protecting working sets of important workloads from sudden reclaim. But its semantics has a significant limitation: it works only as long as there is a supply of reclaimable memory. This makes it pretty useless against any sort of slow memory leaks or memory usage increases. This is especially true for swapless systems. If swap is enabled, memory soft protection effectively postpones problems, allowing a leaking application to fill all swap area, which makes no sense. The only effective way to guarantee the memory protection in this case is to invoke the OOM killer. It's possible to handle this case in userspace by reacting on MEMCG_LOW events; but there is still a place for a fail-safe in-kernel mechanism to provide stronger guarantees. This patch introduces the memory.min interface for cgroup v2 memory controller. It works very similarly to memory.low (sharing the same hierarchical behavior), except that it's not disabled if there is no more reclaimable memory in the system. If cgroup is not populated, its memory.min is ignored, because otherwise even the OOM killer wouldn't be able to reclaim the protected memory, and the system can stall. [guro@fb.com: s/low/min/ in docs] Link: http://lkml.kernel.org/r/20180510130758.GA9129@castle.DHCP.thefacebook.com Link: http://lkml.kernel.org/r/20180509180734.GA4856@castle.DHCP.thefacebook.com Signed-off-by: Roman Gushchin <guro@fb.com> Reviewed-by: Randy Dunlap <rdunlap@infradead.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:07:46 +00:00
return MEMCG_PROT_LOW;
else
return MEMCG_PROT_NONE;
mm: memcontrol: default hierarchy interface for memory Introduce the basic control files to account, partition, and limit memory using cgroups in default hierarchy mode. This interface versioning allows us to address fundamental design issues in the existing memory cgroup interface, further explained below. The old interface will be maintained indefinitely, but a clearer model and improved workload performance should encourage existing users to switch over to the new one eventually. The control files are thus: - memory.current shows the current consumption of the cgroup and its descendants, in bytes. - memory.low configures the lower end of the cgroup's expected memory consumption range. The kernel considers memory below that boundary to be a reserve - the minimum that the workload needs in order to make forward progress - and generally avoids reclaiming it, unless there is an imminent risk of entering an OOM situation. - memory.high configures the upper end of the cgroup's expected memory consumption range. A cgroup whose consumption grows beyond this threshold is forced into direct reclaim, to work off the excess and to throttle new allocations heavily, but is generally allowed to continue and the OOM killer is not invoked. - memory.max configures the hard maximum amount of memory that the cgroup is allowed to consume before the OOM killer is invoked. - memory.events shows event counters that indicate how often the cgroup was reclaimed while below memory.low, how often it was forced to reclaim excess beyond memory.high, how often it hit memory.max, and how often it entered OOM due to memory.max. This allows users to identify configuration problems when observing a degradation in workload performance. An overcommitted system will have an increased rate of low boundary breaches, whereas increased rates of high limit breaches, maximum hits, or even OOM situations will indicate internally overcommitted cgroups. For existing users of memory cgroups, the following deviations from the current interface are worth pointing out and explaining: - The original lower boundary, the soft limit, is defined as a limit that is per default unset. As a result, the set of cgroups that global reclaim prefers is opt-in, rather than opt-out. The costs for optimizing these mostly negative lookups are so high that the implementation, despite its enormous size, does not even provide the basic desirable behavior. First off, the soft limit has no hierarchical meaning. All configured groups are organized in a global rbtree and treated like equal peers, regardless where they are located in the hierarchy. This makes subtree delegation impossible. Second, the soft limit reclaim pass is so aggressive that it not just introduces high allocation latencies into the system, but also impacts system performance due to overreclaim, to the point where the feature becomes self-defeating. The memory.low boundary on the other hand is a top-down allocated reserve. A cgroup enjoys reclaim protection when it and all its ancestors are below their low boundaries, which makes delegation of subtrees possible. Secondly, new cgroups have no reserve per default and in the common case most cgroups are eligible for the preferred reclaim pass. This allows the new low boundary to be efficiently implemented with just a minor addition to the generic reclaim code, without the need for out-of-band data structures and reclaim passes. Because the generic reclaim code considers all cgroups except for the ones running low in the preferred first reclaim pass, overreclaim of individual groups is eliminated as well, resulting in much better overall workload performance. - The original high boundary, the hard limit, is defined as a strict limit that can not budge, even if the OOM killer has to be called. But this generally goes against the goal of making the most out of the available memory. The memory consumption of workloads varies during runtime, and that requires users to overcommit. But doing that with a strict upper limit requires either a fairly accurate prediction of the working set size or adding slack to the limit. Since working set size estimation is hard and error prone, and getting it wrong results in OOM kills, most users tend to err on the side of a looser limit and end up wasting precious resources. The memory.high boundary on the other hand can be set much more conservatively. When hit, it throttles allocations by forcing them into direct reclaim to work off the excess, but it never invokes the OOM killer. As a result, a high boundary that is chosen too aggressively will not terminate the processes, but instead it will lead to gradual performance degradation. The user can monitor this and make corrections until the minimal memory footprint that still gives acceptable performance is found. In extreme cases, with many concurrent allocations and a complete breakdown of reclaim progress within the group, the high boundary can be exceeded. But even then it's mostly better to satisfy the allocation from the slack available in other groups or the rest of the system than killing the group. Otherwise, memory.max is there to limit this type of spillover and ultimately contain buggy or even malicious applications. - The original control file names are unwieldy and inconsistent in many different ways. For example, the upper boundary hit count is exported in the memory.failcnt file, but an OOM event count has to be manually counted by listening to memory.oom_control events, and lower boundary / soft limit events have to be counted by first setting a threshold for that value and then counting those events. Also, usage and limit files encode their units in the filename. That makes the filenames very long, even though this is not information that a user needs to be reminded of every time they type out those names. To address these naming issues, as well as to signal clearly that the new interface carries a new configuration model, the naming conventions in it necessarily differ from the old interface. - The original limit files indicate the state of an unset limit with a very high number, and a configured limit can be unset by echoing -1 into those files. But that very high number is implementation and architecture dependent and not very descriptive. And while -1 can be understood as an underflow into the highest possible value, -2 or -10M etc. do not work, so it's not inconsistent. memory.low, memory.high, and memory.max will use the string "infinity" to indicate and set the highest possible value. [akpm@linux-foundation.org: use seq_puts() for basic strings] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Greg Thelen <gthelen@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-02-11 23:26:06 +00:00
}
mm: memcontrol: rewrite charge API These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 4): The memcg charge API charges pages before they are rmapped - i.e. have an actual "type" - and so every callsite needs its own set of charge and uncharge functions to know what type is being operated on. Worse, uncharge has to happen from a context that is still type-specific, rather than at the end of the page's lifetime with exclusive access, and so requires a lot of synchronization. Rewrite the charge API to provide a generic set of try_charge(), commit_charge() and cancel_charge() transaction operations, much like what's currently done for swap-in: mem_cgroup_try_charge() attempts to reserve a charge, reclaiming pages from the memcg if necessary. mem_cgroup_commit_charge() commits the page to the charge once it has a valid page->mapping and PageAnon() reliably tells the type. mem_cgroup_cancel_charge() aborts the transaction. This reduces the charge API and enables subsequent patches to drastically simplify uncharging. As pages need to be committed after rmap is established but before they are added to the LRU, page_add_new_anon_rmap() must stop doing LRU additions again. Revive lru_cache_add_active_or_unevictable(). [hughd@google.com: fix shmem_unuse] [hughd@google.com: Add comments on the private use of -EAGAIN] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Hugh Dickins <hughd@google.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:20 +00:00
/**
* mem_cgroup_charge - charge a newly allocated page to a cgroup
mm: memcontrol: rewrite charge API These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 4): The memcg charge API charges pages before they are rmapped - i.e. have an actual "type" - and so every callsite needs its own set of charge and uncharge functions to know what type is being operated on. Worse, uncharge has to happen from a context that is still type-specific, rather than at the end of the page's lifetime with exclusive access, and so requires a lot of synchronization. Rewrite the charge API to provide a generic set of try_charge(), commit_charge() and cancel_charge() transaction operations, much like what's currently done for swap-in: mem_cgroup_try_charge() attempts to reserve a charge, reclaiming pages from the memcg if necessary. mem_cgroup_commit_charge() commits the page to the charge once it has a valid page->mapping and PageAnon() reliably tells the type. mem_cgroup_cancel_charge() aborts the transaction. This reduces the charge API and enables subsequent patches to drastically simplify uncharging. As pages need to be committed after rmap is established but before they are added to the LRU, page_add_new_anon_rmap() must stop doing LRU additions again. Revive lru_cache_add_active_or_unevictable(). [hughd@google.com: fix shmem_unuse] [hughd@google.com: Add comments on the private use of -EAGAIN] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Hugh Dickins <hughd@google.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:20 +00:00
* @page: page to charge
* @mm: mm context of the victim
* @gfp_mask: reclaim mode
*
* Try to charge @page to the memcg that @mm belongs to, reclaiming
* pages according to @gfp_mask if necessary.
*
* Returns 0 on success. Otherwise, an error code is returned.
mm: memcontrol: rewrite charge API These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 4): The memcg charge API charges pages before they are rmapped - i.e. have an actual "type" - and so every callsite needs its own set of charge and uncharge functions to know what type is being operated on. Worse, uncharge has to happen from a context that is still type-specific, rather than at the end of the page's lifetime with exclusive access, and so requires a lot of synchronization. Rewrite the charge API to provide a generic set of try_charge(), commit_charge() and cancel_charge() transaction operations, much like what's currently done for swap-in: mem_cgroup_try_charge() attempts to reserve a charge, reclaiming pages from the memcg if necessary. mem_cgroup_commit_charge() commits the page to the charge once it has a valid page->mapping and PageAnon() reliably tells the type. mem_cgroup_cancel_charge() aborts the transaction. This reduces the charge API and enables subsequent patches to drastically simplify uncharging. As pages need to be committed after rmap is established but before they are added to the LRU, page_add_new_anon_rmap() must stop doing LRU additions again. Revive lru_cache_add_active_or_unevictable(). [hughd@google.com: fix shmem_unuse] [hughd@google.com: Add comments on the private use of -EAGAIN] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Hugh Dickins <hughd@google.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:20 +00:00
*/
int mem_cgroup_charge(struct page *page, struct mm_struct *mm, gfp_t gfp_mask)
mm: memcontrol: rewrite charge API These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 4): The memcg charge API charges pages before they are rmapped - i.e. have an actual "type" - and so every callsite needs its own set of charge and uncharge functions to know what type is being operated on. Worse, uncharge has to happen from a context that is still type-specific, rather than at the end of the page's lifetime with exclusive access, and so requires a lot of synchronization. Rewrite the charge API to provide a generic set of try_charge(), commit_charge() and cancel_charge() transaction operations, much like what's currently done for swap-in: mem_cgroup_try_charge() attempts to reserve a charge, reclaiming pages from the memcg if necessary. mem_cgroup_commit_charge() commits the page to the charge once it has a valid page->mapping and PageAnon() reliably tells the type. mem_cgroup_cancel_charge() aborts the transaction. This reduces the charge API and enables subsequent patches to drastically simplify uncharging. As pages need to be committed after rmap is established but before they are added to the LRU, page_add_new_anon_rmap() must stop doing LRU additions again. Revive lru_cache_add_active_or_unevictable(). [hughd@google.com: fix shmem_unuse] [hughd@google.com: Add comments on the private use of -EAGAIN] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Hugh Dickins <hughd@google.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:20 +00:00
{
mm: memcontrol: drop @compound parameter from memcg charging API The memcg charging API carries a boolean @compound parameter that tells whether the page we're dealing with is a hugepage. mem_cgroup_commit_charge() has another boolean @lrucare that indicates whether the page needs LRU locking or not while charging. The majority of callsites know those parameters at compile time, which results in a lot of naked "false, false" argument lists. This makes for cryptic code and is a breeding ground for subtle mistakes. Thankfully, the huge page state can be inferred from the page itself and doesn't need to be passed along. This is safe because charging completes before the page is published and somebody may split it. Simplify the callsites by removing @compound, and let memcg infer the state by using hpage_nr_pages() unconditionally. That function does PageTransHuge() to identify huge pages, which also helpfully asserts that nobody passes in tail pages by accident. The following patches will introduce a new charging API, best not to carry over unnecessary weight. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Alex Shi <alex.shi@linux.alibaba.com> Reviewed-by: Joonsoo Kim <iamjoonsoo.kim@lge.com> Reviewed-by: Shakeel Butt <shakeelb@google.com> Cc: Hugh Dickins <hughd@google.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.com> Cc: Roman Gushchin <guro@fb.com> Cc: Balbir Singh <bsingharora@gmail.com> Link: http://lkml.kernel.org/r/20200508183105.225460-4-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:01:31 +00:00
unsigned int nr_pages = hpage_nr_pages(page);
mm: memcontrol: rewrite charge API These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 4): The memcg charge API charges pages before they are rmapped - i.e. have an actual "type" - and so every callsite needs its own set of charge and uncharge functions to know what type is being operated on. Worse, uncharge has to happen from a context that is still type-specific, rather than at the end of the page's lifetime with exclusive access, and so requires a lot of synchronization. Rewrite the charge API to provide a generic set of try_charge(), commit_charge() and cancel_charge() transaction operations, much like what's currently done for swap-in: mem_cgroup_try_charge() attempts to reserve a charge, reclaiming pages from the memcg if necessary. mem_cgroup_commit_charge() commits the page to the charge once it has a valid page->mapping and PageAnon() reliably tells the type. mem_cgroup_cancel_charge() aborts the transaction. This reduces the charge API and enables subsequent patches to drastically simplify uncharging. As pages need to be committed after rmap is established but before they are added to the LRU, page_add_new_anon_rmap() must stop doing LRU additions again. Revive lru_cache_add_active_or_unevictable(). [hughd@google.com: fix shmem_unuse] [hughd@google.com: Add comments on the private use of -EAGAIN] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Hugh Dickins <hughd@google.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:20 +00:00
struct mem_cgroup *memcg = NULL;
int ret = 0;
if (mem_cgroup_disabled())
goto out;
if (PageSwapCache(page)) {
mm: memcontrol: make swap tracking an integral part of memory control Without swap page tracking, users that are otherwise memory controlled can easily escape their containment and allocate significant amounts of memory that they're not being charged for. That's because swap does readahead, but without the cgroup records of who owned the page at swapout, readahead pages don't get charged until somebody actually faults them into their page table and we can identify an owner task. This can be maliciously exploited with MADV_WILLNEED, which triggers arbitrary readahead allocations without charging the pages. Make swap swap page tracking an integral part of memcg and remove the Kconfig options. In the first place, it was only made configurable to allow users to save some memory. But the overhead of tracking cgroup ownership per swap page is minimal - 2 byte per page, or 512k per 1G of swap, or 0.04%. Saving that at the expense of broken containment semantics is not something we should present as a coequal option. The swapaccount=0 boot option will continue to exist, and it will eliminate the page_counter overhead and hide the swap control files, but it won't disable swap slot ownership tracking. This patch makes sure we always have the cgroup records at swapin time; the next patch will fix the actual bug by charging readahead swap pages at swapin time rather than at fault time. v2: fix double swap charge bug in cgroup1/cgroup2 code gating [hannes@cmpxchg.org: fix crash with cgroup_disable=memory] Link: http://lkml.kernel.org/r/20200521215855.GB815153@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Alex Shi <alex.shi@linux.alibaba.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Roman Gushchin <guro@fb.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Naresh Kamboju <naresh.kamboju@linaro.org> Link: http://lkml.kernel.org/r/20200508183105.225460-16-hannes@cmpxchg.org Debugged-by: Hugh Dickins <hughd@google.com> Debugged-by: Michal Hocko <mhocko@kernel.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:02:14 +00:00
swp_entry_t ent = { .val = page_private(page), };
unsigned short id;
mm: memcontrol: rewrite charge API These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 4): The memcg charge API charges pages before they are rmapped - i.e. have an actual "type" - and so every callsite needs its own set of charge and uncharge functions to know what type is being operated on. Worse, uncharge has to happen from a context that is still type-specific, rather than at the end of the page's lifetime with exclusive access, and so requires a lot of synchronization. Rewrite the charge API to provide a generic set of try_charge(), commit_charge() and cancel_charge() transaction operations, much like what's currently done for swap-in: mem_cgroup_try_charge() attempts to reserve a charge, reclaiming pages from the memcg if necessary. mem_cgroup_commit_charge() commits the page to the charge once it has a valid page->mapping and PageAnon() reliably tells the type. mem_cgroup_cancel_charge() aborts the transaction. This reduces the charge API and enables subsequent patches to drastically simplify uncharging. As pages need to be committed after rmap is established but before they are added to the LRU, page_add_new_anon_rmap() must stop doing LRU additions again. Revive lru_cache_add_active_or_unevictable(). [hughd@google.com: fix shmem_unuse] [hughd@google.com: Add comments on the private use of -EAGAIN] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Hugh Dickins <hughd@google.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:20 +00:00
/*
* Every swap fault against a single page tries to charge the
* page, bail as early as possible. shmem_unuse() encounters
* already charged pages, too. page->mem_cgroup is protected
* by the page lock, which serializes swap cache removal, which
mm: memcontrol: rewrite charge API These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 4): The memcg charge API charges pages before they are rmapped - i.e. have an actual "type" - and so every callsite needs its own set of charge and uncharge functions to know what type is being operated on. Worse, uncharge has to happen from a context that is still type-specific, rather than at the end of the page's lifetime with exclusive access, and so requires a lot of synchronization. Rewrite the charge API to provide a generic set of try_charge(), commit_charge() and cancel_charge() transaction operations, much like what's currently done for swap-in: mem_cgroup_try_charge() attempts to reserve a charge, reclaiming pages from the memcg if necessary. mem_cgroup_commit_charge() commits the page to the charge once it has a valid page->mapping and PageAnon() reliably tells the type. mem_cgroup_cancel_charge() aborts the transaction. This reduces the charge API and enables subsequent patches to drastically simplify uncharging. As pages need to be committed after rmap is established but before they are added to the LRU, page_add_new_anon_rmap() must stop doing LRU additions again. Revive lru_cache_add_active_or_unevictable(). [hughd@google.com: fix shmem_unuse] [hughd@google.com: Add comments on the private use of -EAGAIN] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Hugh Dickins <hughd@google.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:20 +00:00
* in turn serializes uncharging.
*/
VM_BUG_ON_PAGE(!PageLocked(page), page);
if (compound_head(page)->mem_cgroup)
mm: memcontrol: rewrite charge API These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 4): The memcg charge API charges pages before they are rmapped - i.e. have an actual "type" - and so every callsite needs its own set of charge and uncharge functions to know what type is being operated on. Worse, uncharge has to happen from a context that is still type-specific, rather than at the end of the page's lifetime with exclusive access, and so requires a lot of synchronization. Rewrite the charge API to provide a generic set of try_charge(), commit_charge() and cancel_charge() transaction operations, much like what's currently done for swap-in: mem_cgroup_try_charge() attempts to reserve a charge, reclaiming pages from the memcg if necessary. mem_cgroup_commit_charge() commits the page to the charge once it has a valid page->mapping and PageAnon() reliably tells the type. mem_cgroup_cancel_charge() aborts the transaction. This reduces the charge API and enables subsequent patches to drastically simplify uncharging. As pages need to be committed after rmap is established but before they are added to the LRU, page_add_new_anon_rmap() must stop doing LRU additions again. Revive lru_cache_add_active_or_unevictable(). [hughd@google.com: fix shmem_unuse] [hughd@google.com: Add comments on the private use of -EAGAIN] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Hugh Dickins <hughd@google.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:20 +00:00
goto out;
mm: memcontrol: make swap tracking an integral part of memory control Without swap page tracking, users that are otherwise memory controlled can easily escape their containment and allocate significant amounts of memory that they're not being charged for. That's because swap does readahead, but without the cgroup records of who owned the page at swapout, readahead pages don't get charged until somebody actually faults them into their page table and we can identify an owner task. This can be maliciously exploited with MADV_WILLNEED, which triggers arbitrary readahead allocations without charging the pages. Make swap swap page tracking an integral part of memcg and remove the Kconfig options. In the first place, it was only made configurable to allow users to save some memory. But the overhead of tracking cgroup ownership per swap page is minimal - 2 byte per page, or 512k per 1G of swap, or 0.04%. Saving that at the expense of broken containment semantics is not something we should present as a coequal option. The swapaccount=0 boot option will continue to exist, and it will eliminate the page_counter overhead and hide the swap control files, but it won't disable swap slot ownership tracking. This patch makes sure we always have the cgroup records at swapin time; the next patch will fix the actual bug by charging readahead swap pages at swapin time rather than at fault time. v2: fix double swap charge bug in cgroup1/cgroup2 code gating [hannes@cmpxchg.org: fix crash with cgroup_disable=memory] Link: http://lkml.kernel.org/r/20200521215855.GB815153@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Alex Shi <alex.shi@linux.alibaba.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Roman Gushchin <guro@fb.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Naresh Kamboju <naresh.kamboju@linaro.org> Link: http://lkml.kernel.org/r/20200508183105.225460-16-hannes@cmpxchg.org Debugged-by: Hugh Dickins <hughd@google.com> Debugged-by: Michal Hocko <mhocko@kernel.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:02:14 +00:00
id = lookup_swap_cgroup_id(ent);
rcu_read_lock();
memcg = mem_cgroup_from_id(id);
if (memcg && !css_tryget_online(&memcg->css))
memcg = NULL;
rcu_read_unlock();
mm: memcontrol: rewrite charge API These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 4): The memcg charge API charges pages before they are rmapped - i.e. have an actual "type" - and so every callsite needs its own set of charge and uncharge functions to know what type is being operated on. Worse, uncharge has to happen from a context that is still type-specific, rather than at the end of the page's lifetime with exclusive access, and so requires a lot of synchronization. Rewrite the charge API to provide a generic set of try_charge(), commit_charge() and cancel_charge() transaction operations, much like what's currently done for swap-in: mem_cgroup_try_charge() attempts to reserve a charge, reclaiming pages from the memcg if necessary. mem_cgroup_commit_charge() commits the page to the charge once it has a valid page->mapping and PageAnon() reliably tells the type. mem_cgroup_cancel_charge() aborts the transaction. This reduces the charge API and enables subsequent patches to drastically simplify uncharging. As pages need to be committed after rmap is established but before they are added to the LRU, page_add_new_anon_rmap() must stop doing LRU additions again. Revive lru_cache_add_active_or_unevictable(). [hughd@google.com: fix shmem_unuse] [hughd@google.com: Add comments on the private use of -EAGAIN] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Hugh Dickins <hughd@google.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:20 +00:00
}
if (!memcg)
memcg = get_mem_cgroup_from_mm(mm);
ret = try_charge(memcg, gfp_mask, nr_pages);
if (ret)
goto out_put;
mm: memcontrol: rewrite charge API These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 4): The memcg charge API charges pages before they are rmapped - i.e. have an actual "type" - and so every callsite needs its own set of charge and uncharge functions to know what type is being operated on. Worse, uncharge has to happen from a context that is still type-specific, rather than at the end of the page's lifetime with exclusive access, and so requires a lot of synchronization. Rewrite the charge API to provide a generic set of try_charge(), commit_charge() and cancel_charge() transaction operations, much like what's currently done for swap-in: mem_cgroup_try_charge() attempts to reserve a charge, reclaiming pages from the memcg if necessary. mem_cgroup_commit_charge() commits the page to the charge once it has a valid page->mapping and PageAnon() reliably tells the type. mem_cgroup_cancel_charge() aborts the transaction. This reduces the charge API and enables subsequent patches to drastically simplify uncharging. As pages need to be committed after rmap is established but before they are added to the LRU, page_add_new_anon_rmap() must stop doing LRU additions again. Revive lru_cache_add_active_or_unevictable(). [hughd@google.com: fix shmem_unuse] [hughd@google.com: Add comments on the private use of -EAGAIN] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Hugh Dickins <hughd@google.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:20 +00:00
commit_charge(page, memcg);
local_irq_disable();
mm: memcontrol: drop @compound parameter from memcg charging API The memcg charging API carries a boolean @compound parameter that tells whether the page we're dealing with is a hugepage. mem_cgroup_commit_charge() has another boolean @lrucare that indicates whether the page needs LRU locking or not while charging. The majority of callsites know those parameters at compile time, which results in a lot of naked "false, false" argument lists. This makes for cryptic code and is a breeding ground for subtle mistakes. Thankfully, the huge page state can be inferred from the page itself and doesn't need to be passed along. This is safe because charging completes before the page is published and somebody may split it. Simplify the callsites by removing @compound, and let memcg infer the state by using hpage_nr_pages() unconditionally. That function does PageTransHuge() to identify huge pages, which also helpfully asserts that nobody passes in tail pages by accident. The following patches will introduce a new charging API, best not to carry over unnecessary weight. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Alex Shi <alex.shi@linux.alibaba.com> Reviewed-by: Joonsoo Kim <iamjoonsoo.kim@lge.com> Reviewed-by: Shakeel Butt <shakeelb@google.com> Cc: Hugh Dickins <hughd@google.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.com> Cc: Roman Gushchin <guro@fb.com> Cc: Balbir Singh <bsingharora@gmail.com> Link: http://lkml.kernel.org/r/20200508183105.225460-4-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:01:31 +00:00
mem_cgroup_charge_statistics(memcg, page, nr_pages);
memcg_check_events(memcg, page);
local_irq_enable();
mm: memcontrol: rewrite charge API These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 4): The memcg charge API charges pages before they are rmapped - i.e. have an actual "type" - and so every callsite needs its own set of charge and uncharge functions to know what type is being operated on. Worse, uncharge has to happen from a context that is still type-specific, rather than at the end of the page's lifetime with exclusive access, and so requires a lot of synchronization. Rewrite the charge API to provide a generic set of try_charge(), commit_charge() and cancel_charge() transaction operations, much like what's currently done for swap-in: mem_cgroup_try_charge() attempts to reserve a charge, reclaiming pages from the memcg if necessary. mem_cgroup_commit_charge() commits the page to the charge once it has a valid page->mapping and PageAnon() reliably tells the type. mem_cgroup_cancel_charge() aborts the transaction. This reduces the charge API and enables subsequent patches to drastically simplify uncharging. As pages need to be committed after rmap is established but before they are added to the LRU, page_add_new_anon_rmap() must stop doing LRU additions again. Revive lru_cache_add_active_or_unevictable(). [hughd@google.com: fix shmem_unuse] [hughd@google.com: Add comments on the private use of -EAGAIN] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Hugh Dickins <hughd@google.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:20 +00:00
mm: memcontrol: make swap tracking an integral part of memory control Without swap page tracking, users that are otherwise memory controlled can easily escape their containment and allocate significant amounts of memory that they're not being charged for. That's because swap does readahead, but without the cgroup records of who owned the page at swapout, readahead pages don't get charged until somebody actually faults them into their page table and we can identify an owner task. This can be maliciously exploited with MADV_WILLNEED, which triggers arbitrary readahead allocations without charging the pages. Make swap swap page tracking an integral part of memcg and remove the Kconfig options. In the first place, it was only made configurable to allow users to save some memory. But the overhead of tracking cgroup ownership per swap page is minimal - 2 byte per page, or 512k per 1G of swap, or 0.04%. Saving that at the expense of broken containment semantics is not something we should present as a coequal option. The swapaccount=0 boot option will continue to exist, and it will eliminate the page_counter overhead and hide the swap control files, but it won't disable swap slot ownership tracking. This patch makes sure we always have the cgroup records at swapin time; the next patch will fix the actual bug by charging readahead swap pages at swapin time rather than at fault time. v2: fix double swap charge bug in cgroup1/cgroup2 code gating [hannes@cmpxchg.org: fix crash with cgroup_disable=memory] Link: http://lkml.kernel.org/r/20200521215855.GB815153@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Alex Shi <alex.shi@linux.alibaba.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Roman Gushchin <guro@fb.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Naresh Kamboju <naresh.kamboju@linaro.org> Link: http://lkml.kernel.org/r/20200508183105.225460-16-hannes@cmpxchg.org Debugged-by: Hugh Dickins <hughd@google.com> Debugged-by: Michal Hocko <mhocko@kernel.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:02:14 +00:00
if (PageSwapCache(page)) {
mm: memcontrol: rewrite charge API These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 4): The memcg charge API charges pages before they are rmapped - i.e. have an actual "type" - and so every callsite needs its own set of charge and uncharge functions to know what type is being operated on. Worse, uncharge has to happen from a context that is still type-specific, rather than at the end of the page's lifetime with exclusive access, and so requires a lot of synchronization. Rewrite the charge API to provide a generic set of try_charge(), commit_charge() and cancel_charge() transaction operations, much like what's currently done for swap-in: mem_cgroup_try_charge() attempts to reserve a charge, reclaiming pages from the memcg if necessary. mem_cgroup_commit_charge() commits the page to the charge once it has a valid page->mapping and PageAnon() reliably tells the type. mem_cgroup_cancel_charge() aborts the transaction. This reduces the charge API and enables subsequent patches to drastically simplify uncharging. As pages need to be committed after rmap is established but before they are added to the LRU, page_add_new_anon_rmap() must stop doing LRU additions again. Revive lru_cache_add_active_or_unevictable(). [hughd@google.com: fix shmem_unuse] [hughd@google.com: Add comments on the private use of -EAGAIN] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Hugh Dickins <hughd@google.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:20 +00:00
swp_entry_t entry = { .val = page_private(page) };
/*
* The swap entry might not get freed for a long time,
* let's not wait for it. The page already received a
* memory+swap charge, drop the swap entry duplicate.
*/
mm, THP, swap: delay splitting THP during swap out Patch series "THP swap: Delay splitting THP during swapping out", v11. This patchset is to optimize the performance of Transparent Huge Page (THP) swap. Recently, the performance of the storage devices improved so fast that we cannot saturate the disk bandwidth with single logical CPU when do page swap out even on a high-end server machine. Because the performance of the storage device improved faster than that of single logical CPU. And it seems that the trend will not change in the near future. On the other hand, the THP becomes more and more popular because of increased memory size. So it becomes necessary to optimize THP swap performance. The advantages of the THP swap support include: - Batch the swap operations for the THP to reduce lock acquiring/releasing, including allocating/freeing the swap space, adding/deleting to/from the swap cache, and writing/reading the swap space, etc. This will help improve the performance of the THP swap. - The THP swap space read/write will be 2M sequential IO. It is particularly helpful for the swap read, which are usually 4k random IO. This will improve the performance of the THP swap too. - It will help the memory fragmentation, especially when the THP is heavily used by the applications. The 2M continuous pages will be free up after THP swapping out. - It will improve the THP utilization on the system with the swap turned on. Because the speed for khugepaged to collapse the normal pages into the THP is quite slow. After the THP is split during the swapping out, it will take quite long time for the normal pages to collapse back into the THP after being swapped in. The high THP utilization helps the efficiency of the page based memory management too. There are some concerns regarding THP swap in, mainly because possible enlarged read/write IO size (for swap in/out) may put more overhead on the storage device. To deal with that, the THP swap in should be turned on only when necessary. For example, it can be selected via "always/never/madvise" logic, to be turned on globally, turned off globally, or turned on only for VMA with MADV_HUGEPAGE, etc. This patchset is the first step for the THP swap support. The plan is to delay splitting THP step by step, finally avoid splitting THP during the THP swapping out and swap out/in the THP as a whole. As the first step, in this patchset, the splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP and adding the THP into the swap cache. This will reduce lock acquiring/releasing for the locks used for the swap cache management. With the patchset, the swap out throughput improves 15.5% (from about 3.73GB/s to about 4.31GB/s) in the vm-scalability swap-w-seq test case with 8 processes. The test is done on a Xeon E5 v3 system. The swap device used is a RAM simulated PMEM (persistent memory) device. To test the sequential swapping out, the test case creates 8 processes, which sequentially allocate and write to the anonymous pages until the RAM and part of the swap device is used up. This patch (of 5): In this patch, splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP (Transparent Huge Page) and adding the THP into the swap cache. This will batch the corresponding operation, thus improve THP swap out throughput. This is the first step for the THP swap optimization. The plan is to delay splitting the THP step by step and avoid splitting the THP finally. In this patch, one swap cluster is used to hold the contents of each THP swapped out. So, the size of the swap cluster is changed to that of the THP (Transparent Huge Page) on x86_64 architecture (512). For other architectures which want such THP swap optimization, ARCH_USES_THP_SWAP_CLUSTER needs to be selected in the Kconfig file for the architecture. In effect, this will enlarge swap cluster size by 2 times on x86_64. Which may make it harder to find a free cluster when the swap space becomes fragmented. So that, this may reduce the continuous swap space allocation and sequential write in theory. The performance test in 0day shows no regressions caused by this. In the future of THP swap optimization, some information of the swapped out THP (such as compound map count) will be recorded in the swap_cluster_info data structure. The mem cgroup swap accounting functions are enhanced to support charge or uncharge a swap cluster backing a THP as a whole. The swap cluster allocate/free functions are added to allocate/free a swap cluster for a THP. A fair simple algorithm is used for swap cluster allocation, that is, only the first swap device in priority list will be tried to allocate the swap cluster. The function will fail if the trying is not successful, and the caller will fallback to allocate a single swap slot instead. This works good enough for normal cases. If the difference of the number of the free swap clusters among multiple swap devices is significant, it is possible that some THPs are split earlier than necessary. For example, this could be caused by big size difference among multiple swap devices. The swap cache functions is enhanced to support add/delete THP to/from the swap cache as a set of (HPAGE_PMD_NR) sub-pages. This may be enhanced in the future with multi-order radix tree. But because we will split the THP soon during swapping out, that optimization doesn't make much sense for this first step. The THP splitting functions are enhanced to support to split THP in swap cache during swapping out. The page lock will be held during allocating the swap cluster, adding the THP into the swap cache and splitting the THP. So in the code path other than swapping out, if the THP need to be split, the PageSwapCache(THP) will be always false. The swap cluster is only available for SSD, so the THP swap optimization in this patchset has no effect for HDD. [ying.huang@intel.com: fix two issues in THP optimize patch] Link: http://lkml.kernel.org/r/87k25ed8zo.fsf@yhuang-dev.intel.com [hannes@cmpxchg.org: extensive cleanups and simplifications, reduce code size] Link: http://lkml.kernel.org/r/20170515112522.32457-2-ying.huang@intel.com Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Suggested-by: Andrew Morton <akpm@linux-foundation.org> [for config option] Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> [for changes in huge_memory.c and huge_mm.h] Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Ebru Akagunduz <ebru.akagunduz@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Cc: Hugh Dickins <hughd@google.com> Cc: Shaohua Li <shli@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-07-06 22:37:18 +00:00
mem_cgroup_uncharge_swap(entry, nr_pages);
mm: memcontrol: rewrite charge API These patches rework memcg charge lifetime to integrate more naturally with the lifetime of user pages. This drastically simplifies the code and reduces charging and uncharging overhead. The most expensive part of charging and uncharging is the page_cgroup bit spinlock, which is removed entirely after this series. Here are the top-10 profile entries of a stress test that reads a 128G sparse file on a freshly booted box, without even a dedicated cgroup (i.e. executing in the root memcg). Before: 15.36% cat [kernel.kallsyms] [k] copy_user_generic_string 13.31% cat [kernel.kallsyms] [k] memset 11.48% cat [kernel.kallsyms] [k] do_mpage_readpage 4.23% cat [kernel.kallsyms] [k] get_page_from_freelist 2.38% cat [kernel.kallsyms] [k] put_page 2.32% cat [kernel.kallsyms] [k] __mem_cgroup_commit_charge 2.18% kswapd0 [kernel.kallsyms] [k] __mem_cgroup_uncharge_common 1.92% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.86% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.62% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn After: 15.67% cat [kernel.kallsyms] [k] copy_user_generic_string 13.48% cat [kernel.kallsyms] [k] memset 11.42% cat [kernel.kallsyms] [k] do_mpage_readpage 3.98% cat [kernel.kallsyms] [k] get_page_from_freelist 2.46% cat [kernel.kallsyms] [k] put_page 2.13% kswapd0 [kernel.kallsyms] [k] shrink_page_list 1.88% cat [kernel.kallsyms] [k] __radix_tree_lookup 1.67% cat [kernel.kallsyms] [k] __pagevec_lru_add_fn 1.39% kswapd0 [kernel.kallsyms] [k] free_pcppages_bulk 1.30% cat [kernel.kallsyms] [k] kfree As you can see, the memcg footprint has shrunk quite a bit. text data bss dec hex filename 37970 9892 400 48262 bc86 mm/memcontrol.o.old 35239 9892 400 45531 b1db mm/memcontrol.o This patch (of 4): The memcg charge API charges pages before they are rmapped - i.e. have an actual "type" - and so every callsite needs its own set of charge and uncharge functions to know what type is being operated on. Worse, uncharge has to happen from a context that is still type-specific, rather than at the end of the page's lifetime with exclusive access, and so requires a lot of synchronization. Rewrite the charge API to provide a generic set of try_charge(), commit_charge() and cancel_charge() transaction operations, much like what's currently done for swap-in: mem_cgroup_try_charge() attempts to reserve a charge, reclaiming pages from the memcg if necessary. mem_cgroup_commit_charge() commits the page to the charge once it has a valid page->mapping and PageAnon() reliably tells the type. mem_cgroup_cancel_charge() aborts the transaction. This reduces the charge API and enables subsequent patches to drastically simplify uncharging. As pages need to be committed after rmap is established but before they are added to the LRU, page_add_new_anon_rmap() must stop doing LRU additions again. Revive lru_cache_add_active_or_unevictable(). [hughd@google.com: fix shmem_unuse] [hughd@google.com: Add comments on the private use of -EAGAIN] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Hugh Dickins <hughd@google.com> Cc: Naoya Horiguchi <n-horiguchi@ah.jp.nec.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:20 +00:00
}
out_put:
css_put(&memcg->css);
out:
return ret;
mm: memcontrol: convert page cache to a new mem_cgroup_charge() API The try/commit/cancel protocol that memcg uses dates back to when pages used to be uncharged upon removal from the page cache, and thus couldn't be committed before the insertion had succeeded. Nowadays, pages are uncharged when they are physically freed; it doesn't matter whether the insertion was successful or not. For the page cache, the transaction dance has become unnecessary. Introduce a mem_cgroup_charge() function that simply charges a newly allocated page to a cgroup and sets up page->mem_cgroup in one single step. If the insertion fails, the caller doesn't have to do anything but free/put the page. Then switch the page cache over to this new API. Subsequent patches will also convert anon pages, but it needs a bit more prep work. Right now, memcg depends on page->mapping being already set up at the time of charging, so that it can maintain its own MEMCG_CACHE and MEMCG_RSS counters. For anon, page->mapping is set under the same pte lock under which the page is publishd, so a single charge point that can block doesn't work there just yet. The following prep patches will replace the private memcg counters with the generic vmstat counters, thus removing the page->mapping dependency, then complete the transition to the new single-point charge API and delete the old transactional scheme. v2: leave shmem swapcache when charging fails to avoid double IO (Joonsoo) v3: rebase on preceeding shmem simplification patch Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Alex Shi <alex.shi@linux.alibaba.com> Cc: Hugh Dickins <hughd@google.com> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.com> Cc: Roman Gushchin <guro@fb.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Balbir Singh <bsingharora@gmail.com> Link: http://lkml.kernel.org/r/20200508183105.225460-6-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:01:41 +00:00
}
struct uncharge_gather {
struct mem_cgroup *memcg;
unsigned long nr_pages;
unsigned long pgpgout;
unsigned long nr_kmem;
struct page *dummy_page;
};
static inline void uncharge_gather_clear(struct uncharge_gather *ug)
{
memset(ug, 0, sizeof(*ug));
}
static void uncharge_batch(const struct uncharge_gather *ug)
{
unsigned long flags;
if (!mem_cgroup_is_root(ug->memcg)) {
page_counter_uncharge(&ug->memcg->memory, ug->nr_pages);
if (do_memsw_account())
page_counter_uncharge(&ug->memcg->memsw, ug->nr_pages);
if (!cgroup_subsys_on_dfl(memory_cgrp_subsys) && ug->nr_kmem)
page_counter_uncharge(&ug->memcg->kmem, ug->nr_kmem);
memcg_oom_recover(ug->memcg);
}
local_irq_save(flags);
__count_memcg_events(ug->memcg, PGPGOUT, ug->pgpgout);
__this_cpu_add(ug->memcg->vmstats_percpu->nr_page_events, ug->nr_pages);
memcg_check_events(ug->memcg, ug->dummy_page);
local_irq_restore(flags);
if (!mem_cgroup_is_root(ug->memcg))
css_put_many(&ug->memcg->css, ug->nr_pages);
}
static void uncharge_page(struct page *page, struct uncharge_gather *ug)
{
unsigned long nr_pages;
VM_BUG_ON_PAGE(PageLRU(page), page);
if (!page->mem_cgroup)
return;
/*
* Nobody should be changing or seriously looking at
* page->mem_cgroup at this point, we have fully
* exclusive access to the page.
*/
if (ug->memcg != page->mem_cgroup) {
if (ug->memcg) {
uncharge_batch(ug);
uncharge_gather_clear(ug);
}
ug->memcg = page->mem_cgroup;
}
nr_pages = compound_nr(page);
ug->nr_pages += nr_pages;
if (!PageKmemcg(page)) {
ug->pgpgout++;
} else {
ug->nr_kmem += nr_pages;
__ClearPageKmemcg(page);
}
ug->dummy_page = page;
page->mem_cgroup = NULL;
}
static void uncharge_list(struct list_head *page_list)
{
struct uncharge_gather ug;
struct list_head *next;
uncharge_gather_clear(&ug);
/*
* Note that the list can be a single page->lru; hence the
* do-while loop instead of a simple list_for_each_entry().
*/
next = page_list->next;
do {
struct page *page;
page = list_entry(next, struct page, lru);
next = page->lru.next;
uncharge_page(page, &ug);
} while (next != page_list);
if (ug.memcg)
uncharge_batch(&ug);
}
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
/**
* mem_cgroup_uncharge - uncharge a page
* @page: page to uncharge
*
* Uncharge a page previously charged with mem_cgroup_charge().
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
*/
void mem_cgroup_uncharge(struct page *page)
{
struct uncharge_gather ug;
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
if (mem_cgroup_disabled())
return;
/* Don't touch page->lru of any random page, pre-check: */
mm: embed the memcg pointer directly into struct page Memory cgroups used to have 5 per-page pointers. To allow users to disable that amount of overhead during runtime, those pointers were allocated in a separate array, with a translation layer between them and struct page. There is now only one page pointer remaining: the memcg pointer, that indicates which cgroup the page is associated with when charged. The complexity of runtime allocation and the runtime translation overhead is no longer justified to save that *potential* 0.19% of memory. With CONFIG_SLUB, page->mem_cgroup actually sits in the doubleword padding after the page->private member and doesn't even increase struct page, and then this patch actually saves space. Remaining users that care can still compile their kernels without CONFIG_MEMCG. text data bss dec hex filename 8828345 1725264 983040 11536649 b00909 vmlinux.old 8827425 1725264 966656 11519345 afc571 vmlinux.new [mhocko@suse.cz: update Documentation/cgroups/memory.txt] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: David S. Miller <davem@davemloft.net> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Acked-by: Konstantin Khlebnikov <koct9i@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:44:52 +00:00
if (!page->mem_cgroup)
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
return;
uncharge_gather_clear(&ug);
uncharge_page(page, &ug);
uncharge_batch(&ug);
}
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
/**
* mem_cgroup_uncharge_list - uncharge a list of page
* @page_list: list of pages to uncharge
*
* Uncharge a list of pages previously charged with
* mem_cgroup_charge().
*/
void mem_cgroup_uncharge_list(struct list_head *page_list)
{
if (mem_cgroup_disabled())
return;
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
if (!list_empty(page_list))
uncharge_list(page_list);
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
}
/**
* mem_cgroup_migrate - charge a page's replacement
* @oldpage: currently circulating page
* @newpage: replacement page
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
*
* Charge @newpage as a replacement page for @oldpage. @oldpage will
* be uncharged upon free.
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
*
* Both pages must be locked, @newpage->mapping must be set up.
*/
void mem_cgroup_migrate(struct page *oldpage, struct page *newpage)
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
{
struct mem_cgroup *memcg;
unsigned int nr_pages;
memcg: mem_cgroup_migrate() may be called with irq disabled mem_cgroup_migrate() uses local_irq_disable/enable() but can be called with irq disabled from migrate_page_copy(). This ends up enabling irq while holding a irq context lock triggering the following lockdep warning. Fix it by using irq_save/restore instead. ================================= [ INFO: inconsistent lock state ] 4.7.0-rc1+ #52 Tainted: G W --------------------------------- inconsistent {IN-SOFTIRQ-W} -> {SOFTIRQ-ON-W} usage. kcompactd0/151 [HC0[0]:SC0[0]:HE1:SE1] takes: (&(&ctx->completion_lock)->rlock){+.?.-.}, at: [<000000000038fd96>] aio_migratepage+0x156/0x1e8 {IN-SOFTIRQ-W} state was registered at: __lock_acquire+0x5b6/0x1930 lock_acquire+0xee/0x270 _raw_spin_lock_irqsave+0x66/0xb0 aio_complete+0x98/0x328 dio_complete+0xe4/0x1e0 blk_update_request+0xd4/0x450 scsi_end_request+0x48/0x1c8 scsi_io_completion+0x272/0x698 blk_done_softirq+0xca/0xe8 __do_softirq+0xc8/0x518 irq_exit+0xee/0x110 do_IRQ+0x6a/0x88 io_int_handler+0x11a/0x25c __mutex_unlock_slowpath+0x144/0x1d8 __mutex_unlock_slowpath+0x140/0x1d8 kernfs_iop_permission+0x64/0x80 __inode_permission+0x9e/0xf0 link_path_walk+0x6e/0x510 path_lookupat+0xc4/0x1a8 filename_lookup+0x9c/0x160 user_path_at_empty+0x5c/0x70 SyS_readlinkat+0x68/0x140 system_call+0xd6/0x270 irq event stamp: 971410 hardirqs last enabled at (971409): migrate_page_move_mapping+0x3ea/0x588 hardirqs last disabled at (971410): _raw_spin_lock_irqsave+0x3c/0xb0 softirqs last enabled at (970526): __do_softirq+0x460/0x518 softirqs last disabled at (970519): irq_exit+0xee/0x110 other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock(&(&ctx->completion_lock)->rlock); <Interrupt> lock(&(&ctx->completion_lock)->rlock); *** DEADLOCK *** 3 locks held by kcompactd0/151: #0: (&(&mapping->private_lock)->rlock){+.+.-.}, at: aio_migratepage+0x42/0x1e8 #1: (&ctx->ring_lock){+.+.+.}, at: aio_migratepage+0x5a/0x1e8 #2: (&(&ctx->completion_lock)->rlock){+.?.-.}, at: aio_migratepage+0x156/0x1e8 stack backtrace: CPU: 20 PID: 151 Comm: kcompactd0 Tainted: G W 4.7.0-rc1+ #52 Call Trace: show_trace+0xea/0xf0 show_stack+0x72/0xf0 dump_stack+0x9a/0xd8 print_usage_bug.part.27+0x2d4/0x2e8 mark_lock+0x17e/0x758 mark_held_locks+0xa2/0xd0 trace_hardirqs_on_caller+0x140/0x1c0 mem_cgroup_migrate+0x266/0x370 aio_migratepage+0x16a/0x1e8 move_to_new_page+0xb0/0x260 migrate_pages+0x8f4/0x9f0 compact_zone+0x4dc/0xdc8 kcompactd_do_work+0x1aa/0x358 kcompactd+0xba/0x2c8 kthread+0x10a/0x110 kernel_thread_starter+0x6/0xc kernel_thread_starter+0x0/0xc INFO: lockdep is turned off. Link: http://lkml.kernel.org/r/20160620184158.GO3262@mtj.duckdns.org Link: http://lkml.kernel.org/g/5767CFE5.7080904@de.ibm.com Fixes: 74485cf2bc85 ("mm: migrate: consolidate mem_cgroup_migrate() calls") Signed-off-by: Tejun Heo <tj@kernel.org> Reported-by: Christian Borntraeger <borntraeger@de.ibm.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: <stable@vger.kernel.org> [4.5+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-06-24 21:49:54 +00:00
unsigned long flags;
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
VM_BUG_ON_PAGE(!PageLocked(oldpage), oldpage);
VM_BUG_ON_PAGE(!PageLocked(newpage), newpage);
VM_BUG_ON_PAGE(PageAnon(oldpage) != PageAnon(newpage), newpage);
VM_BUG_ON_PAGE(PageTransHuge(oldpage) != PageTransHuge(newpage),
newpage);
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
if (mem_cgroup_disabled())
return;
/* Page cache replacement: new page already charged? */
mm: embed the memcg pointer directly into struct page Memory cgroups used to have 5 per-page pointers. To allow users to disable that amount of overhead during runtime, those pointers were allocated in a separate array, with a translation layer between them and struct page. There is now only one page pointer remaining: the memcg pointer, that indicates which cgroup the page is associated with when charged. The complexity of runtime allocation and the runtime translation overhead is no longer justified to save that *potential* 0.19% of memory. With CONFIG_SLUB, page->mem_cgroup actually sits in the doubleword padding after the page->private member and doesn't even increase struct page, and then this patch actually saves space. Remaining users that care can still compile their kernels without CONFIG_MEMCG. text data bss dec hex filename 8828345 1725264 983040 11536649 b00909 vmlinux.old 8827425 1725264 966656 11519345 afc571 vmlinux.new [mhocko@suse.cz: update Documentation/cgroups/memory.txt] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: David S. Miller <davem@davemloft.net> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Acked-by: Konstantin Khlebnikov <koct9i@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:44:52 +00:00
if (newpage->mem_cgroup)
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
return;
/* Swapcache readahead pages can get replaced before being charged */
mm: embed the memcg pointer directly into struct page Memory cgroups used to have 5 per-page pointers. To allow users to disable that amount of overhead during runtime, those pointers were allocated in a separate array, with a translation layer between them and struct page. There is now only one page pointer remaining: the memcg pointer, that indicates which cgroup the page is associated with when charged. The complexity of runtime allocation and the runtime translation overhead is no longer justified to save that *potential* 0.19% of memory. With CONFIG_SLUB, page->mem_cgroup actually sits in the doubleword padding after the page->private member and doesn't even increase struct page, and then this patch actually saves space. Remaining users that care can still compile their kernels without CONFIG_MEMCG. text data bss dec hex filename 8828345 1725264 983040 11536649 b00909 vmlinux.old 8827425 1725264 966656 11519345 afc571 vmlinux.new [mhocko@suse.cz: update Documentation/cgroups/memory.txt] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.cz> Acked-by: Vladimir Davydov <vdavydov@parallels.com> Acked-by: David S. Miller <davem@davemloft.net> Acked-by: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.cz> Cc: Vladimir Davydov <vdavydov@parallels.com> Cc: Tejun Heo <tj@kernel.org> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Acked-by: Konstantin Khlebnikov <koct9i@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-12-10 23:44:52 +00:00
memcg = oldpage->mem_cgroup;
if (!memcg)
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
return;
/* Force-charge the new page. The old one will be freed soon */
nr_pages = hpage_nr_pages(newpage);
page_counter_charge(&memcg->memory, nr_pages);
if (do_memsw_account())
page_counter_charge(&memcg->memsw, nr_pages);
css_get_many(&memcg->css, nr_pages);
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
commit_charge(newpage, memcg);
memcg: mem_cgroup_migrate() may be called with irq disabled mem_cgroup_migrate() uses local_irq_disable/enable() but can be called with irq disabled from migrate_page_copy(). This ends up enabling irq while holding a irq context lock triggering the following lockdep warning. Fix it by using irq_save/restore instead. ================================= [ INFO: inconsistent lock state ] 4.7.0-rc1+ #52 Tainted: G W --------------------------------- inconsistent {IN-SOFTIRQ-W} -> {SOFTIRQ-ON-W} usage. kcompactd0/151 [HC0[0]:SC0[0]:HE1:SE1] takes: (&(&ctx->completion_lock)->rlock){+.?.-.}, at: [<000000000038fd96>] aio_migratepage+0x156/0x1e8 {IN-SOFTIRQ-W} state was registered at: __lock_acquire+0x5b6/0x1930 lock_acquire+0xee/0x270 _raw_spin_lock_irqsave+0x66/0xb0 aio_complete+0x98/0x328 dio_complete+0xe4/0x1e0 blk_update_request+0xd4/0x450 scsi_end_request+0x48/0x1c8 scsi_io_completion+0x272/0x698 blk_done_softirq+0xca/0xe8 __do_softirq+0xc8/0x518 irq_exit+0xee/0x110 do_IRQ+0x6a/0x88 io_int_handler+0x11a/0x25c __mutex_unlock_slowpath+0x144/0x1d8 __mutex_unlock_slowpath+0x140/0x1d8 kernfs_iop_permission+0x64/0x80 __inode_permission+0x9e/0xf0 link_path_walk+0x6e/0x510 path_lookupat+0xc4/0x1a8 filename_lookup+0x9c/0x160 user_path_at_empty+0x5c/0x70 SyS_readlinkat+0x68/0x140 system_call+0xd6/0x270 irq event stamp: 971410 hardirqs last enabled at (971409): migrate_page_move_mapping+0x3ea/0x588 hardirqs last disabled at (971410): _raw_spin_lock_irqsave+0x3c/0xb0 softirqs last enabled at (970526): __do_softirq+0x460/0x518 softirqs last disabled at (970519): irq_exit+0xee/0x110 other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock(&(&ctx->completion_lock)->rlock); <Interrupt> lock(&(&ctx->completion_lock)->rlock); *** DEADLOCK *** 3 locks held by kcompactd0/151: #0: (&(&mapping->private_lock)->rlock){+.+.-.}, at: aio_migratepage+0x42/0x1e8 #1: (&ctx->ring_lock){+.+.+.}, at: aio_migratepage+0x5a/0x1e8 #2: (&(&ctx->completion_lock)->rlock){+.?.-.}, at: aio_migratepage+0x156/0x1e8 stack backtrace: CPU: 20 PID: 151 Comm: kcompactd0 Tainted: G W 4.7.0-rc1+ #52 Call Trace: show_trace+0xea/0xf0 show_stack+0x72/0xf0 dump_stack+0x9a/0xd8 print_usage_bug.part.27+0x2d4/0x2e8 mark_lock+0x17e/0x758 mark_held_locks+0xa2/0xd0 trace_hardirqs_on_caller+0x140/0x1c0 mem_cgroup_migrate+0x266/0x370 aio_migratepage+0x16a/0x1e8 move_to_new_page+0xb0/0x260 migrate_pages+0x8f4/0x9f0 compact_zone+0x4dc/0xdc8 kcompactd_do_work+0x1aa/0x358 kcompactd+0xba/0x2c8 kthread+0x10a/0x110 kernel_thread_starter+0x6/0xc kernel_thread_starter+0x0/0xc INFO: lockdep is turned off. Link: http://lkml.kernel.org/r/20160620184158.GO3262@mtj.duckdns.org Link: http://lkml.kernel.org/g/5767CFE5.7080904@de.ibm.com Fixes: 74485cf2bc85 ("mm: migrate: consolidate mem_cgroup_migrate() calls") Signed-off-by: Tejun Heo <tj@kernel.org> Reported-by: Christian Borntraeger <borntraeger@de.ibm.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: <stable@vger.kernel.org> [4.5+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-06-24 21:49:54 +00:00
local_irq_save(flags);
mm: memcontrol: drop @compound parameter from memcg charging API The memcg charging API carries a boolean @compound parameter that tells whether the page we're dealing with is a hugepage. mem_cgroup_commit_charge() has another boolean @lrucare that indicates whether the page needs LRU locking or not while charging. The majority of callsites know those parameters at compile time, which results in a lot of naked "false, false" argument lists. This makes for cryptic code and is a breeding ground for subtle mistakes. Thankfully, the huge page state can be inferred from the page itself and doesn't need to be passed along. This is safe because charging completes before the page is published and somebody may split it. Simplify the callsites by removing @compound, and let memcg infer the state by using hpage_nr_pages() unconditionally. That function does PageTransHuge() to identify huge pages, which also helpfully asserts that nobody passes in tail pages by accident. The following patches will introduce a new charging API, best not to carry over unnecessary weight. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Alex Shi <alex.shi@linux.alibaba.com> Reviewed-by: Joonsoo Kim <iamjoonsoo.kim@lge.com> Reviewed-by: Shakeel Butt <shakeelb@google.com> Cc: Hugh Dickins <hughd@google.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.com> Cc: Roman Gushchin <guro@fb.com> Cc: Balbir Singh <bsingharora@gmail.com> Link: http://lkml.kernel.org/r/20200508183105.225460-4-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:01:31 +00:00
mem_cgroup_charge_statistics(memcg, newpage, nr_pages);
memcg_check_events(memcg, newpage);
memcg: mem_cgroup_migrate() may be called with irq disabled mem_cgroup_migrate() uses local_irq_disable/enable() but can be called with irq disabled from migrate_page_copy(). This ends up enabling irq while holding a irq context lock triggering the following lockdep warning. Fix it by using irq_save/restore instead. ================================= [ INFO: inconsistent lock state ] 4.7.0-rc1+ #52 Tainted: G W --------------------------------- inconsistent {IN-SOFTIRQ-W} -> {SOFTIRQ-ON-W} usage. kcompactd0/151 [HC0[0]:SC0[0]:HE1:SE1] takes: (&(&ctx->completion_lock)->rlock){+.?.-.}, at: [<000000000038fd96>] aio_migratepage+0x156/0x1e8 {IN-SOFTIRQ-W} state was registered at: __lock_acquire+0x5b6/0x1930 lock_acquire+0xee/0x270 _raw_spin_lock_irqsave+0x66/0xb0 aio_complete+0x98/0x328 dio_complete+0xe4/0x1e0 blk_update_request+0xd4/0x450 scsi_end_request+0x48/0x1c8 scsi_io_completion+0x272/0x698 blk_done_softirq+0xca/0xe8 __do_softirq+0xc8/0x518 irq_exit+0xee/0x110 do_IRQ+0x6a/0x88 io_int_handler+0x11a/0x25c __mutex_unlock_slowpath+0x144/0x1d8 __mutex_unlock_slowpath+0x140/0x1d8 kernfs_iop_permission+0x64/0x80 __inode_permission+0x9e/0xf0 link_path_walk+0x6e/0x510 path_lookupat+0xc4/0x1a8 filename_lookup+0x9c/0x160 user_path_at_empty+0x5c/0x70 SyS_readlinkat+0x68/0x140 system_call+0xd6/0x270 irq event stamp: 971410 hardirqs last enabled at (971409): migrate_page_move_mapping+0x3ea/0x588 hardirqs last disabled at (971410): _raw_spin_lock_irqsave+0x3c/0xb0 softirqs last enabled at (970526): __do_softirq+0x460/0x518 softirqs last disabled at (970519): irq_exit+0xee/0x110 other info that might help us debug this: Possible unsafe locking scenario: CPU0 ---- lock(&(&ctx->completion_lock)->rlock); <Interrupt> lock(&(&ctx->completion_lock)->rlock); *** DEADLOCK *** 3 locks held by kcompactd0/151: #0: (&(&mapping->private_lock)->rlock){+.+.-.}, at: aio_migratepage+0x42/0x1e8 #1: (&ctx->ring_lock){+.+.+.}, at: aio_migratepage+0x5a/0x1e8 #2: (&(&ctx->completion_lock)->rlock){+.?.-.}, at: aio_migratepage+0x156/0x1e8 stack backtrace: CPU: 20 PID: 151 Comm: kcompactd0 Tainted: G W 4.7.0-rc1+ #52 Call Trace: show_trace+0xea/0xf0 show_stack+0x72/0xf0 dump_stack+0x9a/0xd8 print_usage_bug.part.27+0x2d4/0x2e8 mark_lock+0x17e/0x758 mark_held_locks+0xa2/0xd0 trace_hardirqs_on_caller+0x140/0x1c0 mem_cgroup_migrate+0x266/0x370 aio_migratepage+0x16a/0x1e8 move_to_new_page+0xb0/0x260 migrate_pages+0x8f4/0x9f0 compact_zone+0x4dc/0xdc8 kcompactd_do_work+0x1aa/0x358 kcompactd+0xba/0x2c8 kthread+0x10a/0x110 kernel_thread_starter+0x6/0xc kernel_thread_starter+0x0/0xc INFO: lockdep is turned off. Link: http://lkml.kernel.org/r/20160620184158.GO3262@mtj.duckdns.org Link: http://lkml.kernel.org/g/5767CFE5.7080904@de.ibm.com Fixes: 74485cf2bc85 ("mm: migrate: consolidate mem_cgroup_migrate() calls") Signed-off-by: Tejun Heo <tj@kernel.org> Reported-by: Christian Borntraeger <borntraeger@de.ibm.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Cc: <stable@vger.kernel.org> [4.5+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-06-24 21:49:54 +00:00
local_irq_restore(flags);
mm: memcontrol: rewrite uncharge API The memcg uncharging code that is involved towards the end of a page's lifetime - truncation, reclaim, swapout, migration - is impressively complicated and fragile. Because anonymous and file pages were always charged before they had their page->mapping established, uncharges had to happen when the page type could still be known from the context; as in unmap for anonymous, page cache removal for file and shmem pages, and swap cache truncation for swap pages. However, these operations happen well before the page is actually freed, and so a lot of synchronization is necessary: - Charging, uncharging, page migration, and charge migration all need to take a per-page bit spinlock as they could race with uncharging. - Swap cache truncation happens during both swap-in and swap-out, and possibly repeatedly before the page is actually freed. This means that the memcg swapout code is called from many contexts that make no sense and it has to figure out the direction from page state to make sure memory and memory+swap are always correctly charged. - On page migration, the old page might be unmapped but then reused, so memcg code has to prevent untimely uncharging in that case. Because this code - which should be a simple charge transfer - is so special-cased, it is not reusable for replace_page_cache(). But now that charged pages always have a page->mapping, introduce mem_cgroup_uncharge(), which is called after the final put_page(), when we know for sure that nobody is looking at the page anymore. For page migration, introduce mem_cgroup_migrate(), which is called after the migration is successful and the new page is fully rmapped. Because the old page is no longer uncharged after migration, prevent double charges by decoupling the page's memcg association (PCG_USED and pc->mem_cgroup) from the page holding an actual charge. The new bits PCG_MEM and PCG_MEMSW represent the respective charges and are transferred to the new page during migration. mem_cgroup_migrate() is suitable for replace_page_cache() as well, which gets rid of mem_cgroup_replace_page_cache(). However, care needs to be taken because both the source and the target page can already be charged and on the LRU when fuse is splicing: grab the page lock on the charge moving side to prevent changing pc->mem_cgroup of a page under migration. Also, the lruvecs of both pages change as we uncharge the old and charge the new during migration, and putback may race with us, so grab the lru lock and isolate the pages iff on LRU to prevent races and ensure the pages are on the right lruvec afterward. Swap accounting is massively simplified: because the page is no longer uncharged as early as swap cache deletion, a new mem_cgroup_swapout() can transfer the page's memory+swap charge (PCG_MEMSW) to the swap entry before the final put_page() in page reclaim. Finally, page_cgroup changes are now protected by whatever protection the page itself offers: anonymous pages are charged under the page table lock, whereas page cache insertions, swapin, and migration hold the page lock. Uncharging happens under full exclusion with no outstanding references. Charging and uncharging also ensure that the page is off-LRU, which serializes against charge migration. Remove the very costly page_cgroup lock and set pc->flags non-atomically. [mhocko@suse.cz: mem_cgroup_charge_statistics needs preempt_disable] [vdavydov@parallels.com: fix flags definition] Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Hugh Dickins <hughd@google.com> Cc: Tejun Heo <tj@kernel.org> Cc: Vladimir Davydov <vdavydov@parallels.com> Tested-by: Jet Chen <jet.chen@intel.com> Acked-by: Michal Hocko <mhocko@suse.cz> Tested-by: Felipe Balbi <balbi@ti.com> Signed-off-by: Vladimir Davydov <vdavydov@parallels.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-08-08 21:19:22 +00:00
}
DEFINE_STATIC_KEY_FALSE(memcg_sockets_enabled_key);
EXPORT_SYMBOL(memcg_sockets_enabled_key);
void mem_cgroup_sk_alloc(struct sock *sk)
{
struct mem_cgroup *memcg;
if (!mem_cgroup_sockets_enabled)
return;
cgroup: memcg: net: do not associate sock with unrelated cgroup We are testing network memory accounting in our setup and noticed inconsistent network memory usage and often unrelated cgroups network usage correlates with testing workload. On further inspection, it seems like mem_cgroup_sk_alloc() and cgroup_sk_alloc() are broken in irq context specially for cgroup v1. mem_cgroup_sk_alloc() and cgroup_sk_alloc() can be called in irq context and kind of assumes that this can only happen from sk_clone_lock() and the source sock object has already associated cgroup. However in cgroup v1, where network memory accounting is opt-in, the source sock can be unassociated with any cgroup and the new cloned sock can get associated with unrelated interrupted cgroup. Cgroup v2 can also suffer if the source sock object was created by process in the root cgroup or if sk_alloc() is called in irq context. The fix is to just do nothing in interrupt. WARNING: Please note that about half of the TCP sockets are allocated from the IRQ context, so, memory used by such sockets will not be accouted by the memcg. The stack trace of mem_cgroup_sk_alloc() from IRQ-context: CPU: 70 PID: 12720 Comm: ssh Tainted: 5.6.0-smp-DEV #1 Hardware name: ... Call Trace: <IRQ> dump_stack+0x57/0x75 mem_cgroup_sk_alloc+0xe9/0xf0 sk_clone_lock+0x2a7/0x420 inet_csk_clone_lock+0x1b/0x110 tcp_create_openreq_child+0x23/0x3b0 tcp_v6_syn_recv_sock+0x88/0x730 tcp_check_req+0x429/0x560 tcp_v6_rcv+0x72d/0xa40 ip6_protocol_deliver_rcu+0xc9/0x400 ip6_input+0x44/0xd0 ? ip6_protocol_deliver_rcu+0x400/0x400 ip6_rcv_finish+0x71/0x80 ipv6_rcv+0x5b/0xe0 ? ip6_sublist_rcv+0x2e0/0x2e0 process_backlog+0x108/0x1e0 net_rx_action+0x26b/0x460 __do_softirq+0x104/0x2a6 do_softirq_own_stack+0x2a/0x40 </IRQ> do_softirq.part.19+0x40/0x50 __local_bh_enable_ip+0x51/0x60 ip6_finish_output2+0x23d/0x520 ? ip6table_mangle_hook+0x55/0x160 __ip6_finish_output+0xa1/0x100 ip6_finish_output+0x30/0xd0 ip6_output+0x73/0x120 ? __ip6_finish_output+0x100/0x100 ip6_xmit+0x2e3/0x600 ? ipv6_anycast_cleanup+0x50/0x50 ? inet6_csk_route_socket+0x136/0x1e0 ? skb_free_head+0x1e/0x30 inet6_csk_xmit+0x95/0xf0 __tcp_transmit_skb+0x5b4/0xb20 __tcp_send_ack.part.60+0xa3/0x110 tcp_send_ack+0x1d/0x20 tcp_rcv_state_process+0xe64/0xe80 ? tcp_v6_connect+0x5d1/0x5f0 tcp_v6_do_rcv+0x1b1/0x3f0 ? tcp_v6_do_rcv+0x1b1/0x3f0 __release_sock+0x7f/0xd0 release_sock+0x30/0xa0 __inet_stream_connect+0x1c3/0x3b0 ? prepare_to_wait+0xb0/0xb0 inet_stream_connect+0x3b/0x60 __sys_connect+0x101/0x120 ? __sys_getsockopt+0x11b/0x140 __x64_sys_connect+0x1a/0x20 do_syscall_64+0x51/0x200 entry_SYSCALL_64_after_hwframe+0x44/0xa9 The stack trace of mem_cgroup_sk_alloc() from IRQ-context: Fixes: 2d7580738345 ("mm: memcontrol: consolidate cgroup socket tracking") Fixes: d979a39d7242 ("cgroup: duplicate cgroup reference when cloning sockets") Signed-off-by: Shakeel Butt <shakeelb@google.com> Reviewed-by: Roman Gushchin <guro@fb.com> Signed-off-by: David S. Miller <davem@davemloft.net>
2020-03-10 05:16:05 +00:00
/* Do not associate the sock with unrelated interrupted task's memcg. */
if (in_interrupt())
return;
rcu_read_lock();
memcg = mem_cgroup_from_task(current);
if (memcg == root_mem_cgroup)
goto out;
if (!cgroup_subsys_on_dfl(memory_cgrp_subsys) && !memcg->tcpmem_active)
goto out;
memcg: css_tryget_online cleanups Currently multiple locations in memcg code, css_tryget_online() is being used. However it doesn't matter whether the cgroup is online for the callers. Online used to matter when we had reparenting on offlining and we needed a way to prevent new ones from showing up. The failure case for couple of these css_tryget_online usage is to fallback to root_mem_cgroup which kind of make bypassing the memcg limits possible for some workloads. For example creating an inotify group in a subcontainer and then deleting that container after moving the process to a different container will make all the event objects allocated for that group to the root_mem_cgroup. So, using css_tryget_online() is dangerous for such cases. Two locations still use the online version. The swapin of offlined memcg's pages and the memcg kmem cache creation. The kmem cache indeed needs the online version as the kernel does the reparenting of memcg kmem caches. For the swapin case, it has been left for later as the fallback is not really that concerning. With swap accounting enabled, if the memcg of the swapped out page is not online then the memcg extracted from the given 'mm' will be charged and if 'mm' is NULL then root memcg will be charged. However I could not find a code path where the given 'mm' will be NULL for swap-in case. Signed-off-by: Shakeel Butt <shakeelb@google.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Cc: Roman Gushchin <guro@fb.com> Link: http://lkml.kernel.org/r/20200302203109.179417-1-shakeelb@google.com Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-04-02 04:07:10 +00:00
if (css_tryget(&memcg->css))
sk->sk_memcg = memcg;
out:
rcu_read_unlock();
}
void mem_cgroup_sk_free(struct sock *sk)
{
if (sk->sk_memcg)
css_put(&sk->sk_memcg->css);
}
/**
* mem_cgroup_charge_skmem - charge socket memory
* @memcg: memcg to charge
* @nr_pages: number of pages to charge
*
* Charges @nr_pages to @memcg. Returns %true if the charge fit within
* @memcg's configured limit, %false if the charge had to be forced.
*/
bool mem_cgroup_charge_skmem(struct mem_cgroup *memcg, unsigned int nr_pages)
{
gfp_t gfp_mask = GFP_KERNEL;
if (!cgroup_subsys_on_dfl(memory_cgrp_subsys)) {
struct page_counter *fail;
if (page_counter_try_charge(&memcg->tcpmem, nr_pages, &fail)) {
memcg->tcpmem_pressure = 0;
return true;
}
page_counter_charge(&memcg->tcpmem, nr_pages);
memcg->tcpmem_pressure = 1;
return false;
}
/* Don't block in the packet receive path */
if (in_softirq())
gfp_mask = GFP_NOWAIT;
mod_memcg_state(memcg, MEMCG_SOCK, nr_pages);
if (try_charge(memcg, gfp_mask, nr_pages) == 0)
return true;
try_charge(memcg, gfp_mask|__GFP_NOFAIL, nr_pages);
return false;
}
/**
* mem_cgroup_uncharge_skmem - uncharge socket memory
* @memcg: memcg to uncharge
* @nr_pages: number of pages to uncharge
*/
void mem_cgroup_uncharge_skmem(struct mem_cgroup *memcg, unsigned int nr_pages)
{
if (!cgroup_subsys_on_dfl(memory_cgrp_subsys)) {
page_counter_uncharge(&memcg->tcpmem, nr_pages);
return;
}
mod_memcg_state(memcg, MEMCG_SOCK, -nr_pages);
mm: memcontrol: use per-cpu stocks for socket memory uncharging We've noticed a quite noticeable performance overhead on some hosts with significant network traffic when socket memory accounting is enabled. Perf top shows that socket memory uncharging path is hot: 2.13% [kernel] [k] page_counter_cancel 1.14% [kernel] [k] __sk_mem_reduce_allocated 1.14% [kernel] [k] _raw_spin_lock 0.87% [kernel] [k] _raw_spin_lock_irqsave 0.84% [kernel] [k] tcp_ack 0.84% [kernel] [k] ixgbe_poll 0.83% < workload > 0.82% [kernel] [k] enqueue_entity 0.68% [kernel] [k] __fget 0.68% [kernel] [k] tcp_delack_timer_handler 0.67% [kernel] [k] __schedule 0.60% < workload > 0.59% [kernel] [k] __inet6_lookup_established 0.55% [kernel] [k] __switch_to 0.55% [kernel] [k] menu_select 0.54% libc-2.20.so [.] __memcpy_avx_unaligned To address this issue, the existing per-cpu stock infrastructure can be used. refill_stock() can be called from mem_cgroup_uncharge_skmem() to move charge to a per-cpu stock instead of calling atomic page_counter_uncharge(). To prevent the uncontrolled growth of per-cpu stocks, refill_stock() will explicitly drain the cached charge, if the cached value exceeds CHARGE_BATCH. This allows significantly optimize the load: 1.21% [kernel] [k] _raw_spin_lock 1.01% [kernel] [k] ixgbe_poll 0.92% [kernel] [k] _raw_spin_lock_irqsave 0.90% [kernel] [k] enqueue_entity 0.86% [kernel] [k] tcp_ack 0.85% < workload > 0.74% perf-11120.map [.] 0x000000000061bf24 0.73% [kernel] [k] __schedule 0.67% [kernel] [k] __fget 0.63% [kernel] [k] __inet6_lookup_established 0.62% [kernel] [k] menu_select 0.59% < workload > 0.59% [kernel] [k] __switch_to 0.57% libc-2.20.so [.] __memcpy_avx_unaligned Link: http://lkml.kernel.org/r/20170829100150.4580-1-guro@fb.com Signed-off-by: Roman Gushchin <guro@fb.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Vladimir Davydov <vdavydov.dev@gmail.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-09-08 23:13:09 +00:00
refill_stock(memcg, nr_pages);
}
static int __init cgroup_memory(char *s)
{
char *token;
while ((token = strsep(&s, ",")) != NULL) {
if (!*token)
continue;
if (!strcmp(token, "nosocket"))
cgroup_memory_nosocket = true;
if (!strcmp(token, "nokmem"))
cgroup_memory_nokmem = true;
}
return 0;
}
__setup("cgroup.memory=", cgroup_memory);
/*
* subsys_initcall() for memory controller.
*
* Some parts like memcg_hotplug_cpu_dead() have to be initialized from this
* context because of lock dependencies (cgroup_lock -> cpu hotplug) but
* basically everything that doesn't depend on a specific mem_cgroup structure
* should be initialized from here.
*/
static int __init mem_cgroup_init(void)
{
int cpu, node;
mm: introduce CONFIG_MEMCG_KMEM as combination of CONFIG_MEMCG && !CONFIG_SLOB Introduce new config option, which is used to replace repeating CONFIG_MEMCG && !CONFIG_SLOB pattern. Next patches add a little more memcg+kmem related code, so let's keep the defines more clearly. Link: http://lkml.kernel.org/r/153063053670.1818.15013136946600481138.stgit@localhost.localdomain Signed-off-by: Kirill Tkhai <ktkhai@virtuozzo.com> Acked-by: Vladimir Davydov <vdavydov.dev@gmail.com> Tested-by: Shakeel Butt <shakeelb@google.com> Cc: Al Viro <viro@zeniv.linux.org.uk> Cc: Andrey Ryabinin <aryabinin@virtuozzo.com> Cc: Chris Wilson <chris@chris-wilson.co.uk> Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org> Cc: Guenter Roeck <linux@roeck-us.net> Cc: "Huang, Ying" <ying.huang@intel.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Josef Bacik <jbacik@fb.com> Cc: Li RongQing <lirongqing@baidu.com> Cc: Matthew Wilcox <willy@infradead.org> Cc: Matthias Kaehlcke <mka@chromium.org> Cc: Mel Gorman <mgorman@techsingularity.net> Cc: Michal Hocko <mhocko@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Philippe Ombredanne <pombredanne@nexb.com> Cc: Roman Gushchin <guro@fb.com> Cc: Sahitya Tummala <stummala@codeaurora.org> Cc: Stephen Rothwell <sfr@canb.auug.org.au> Cc: Tetsuo Handa <penguin-kernel@I-love.SAKURA.ne.jp> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Waiman Long <longman@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-08-17 22:47:25 +00:00
#ifdef CONFIG_MEMCG_KMEM
mm: memcontrol: use special workqueue for creating per-memcg caches Creating a lot of cgroups at the same time might stall all worker threads with kmem cache creation works, because kmem cache creation is done with the slab_mutex held. The problem was amplified by commits 801faf0db894 ("mm/slab: lockless decision to grow cache") in case of SLAB and 81ae6d03952c ("mm/slub.c: replace kick_all_cpus_sync() with synchronize_sched() in kmem_cache_shrink()") in case of SLUB, which increased the maximal time the slab_mutex can be held. To prevent that from happening, let's use a special ordered single threaded workqueue for kmem cache creation. This shouldn't introduce any functional changes regarding how kmem caches are created, as the work function holds the global slab_mutex during its whole runtime anyway, making it impossible to run more than one work at a time. By using a single threaded workqueue, we just avoid creating a thread per each work. Ordering is required to avoid a situation when a cgroup's work is put off indefinitely because there are other cgroups to serve, in other words to guarantee fairness. Link: https://bugzilla.kernel.org/show_bug.cgi?id=172981 Link: http://lkml.kernel.org/r/20161004131417.GC1862@esperanza Signed-off-by: Vladimir Davydov <vdavydov.dev@gmail.com> Reported-by: Doug Smythies <dsmythies@telus.net> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Pekka Enberg <penberg@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-12-13 00:41:29 +00:00
/*
* Kmem cache creation is mostly done with the slab_mutex held,
* so use a workqueue with limited concurrency to avoid stalling
* all worker threads in case lots of cgroups are created and
* destroyed simultaneously.
mm: memcontrol: use special workqueue for creating per-memcg caches Creating a lot of cgroups at the same time might stall all worker threads with kmem cache creation works, because kmem cache creation is done with the slab_mutex held. The problem was amplified by commits 801faf0db894 ("mm/slab: lockless decision to grow cache") in case of SLAB and 81ae6d03952c ("mm/slub.c: replace kick_all_cpus_sync() with synchronize_sched() in kmem_cache_shrink()") in case of SLUB, which increased the maximal time the slab_mutex can be held. To prevent that from happening, let's use a special ordered single threaded workqueue for kmem cache creation. This shouldn't introduce any functional changes regarding how kmem caches are created, as the work function holds the global slab_mutex during its whole runtime anyway, making it impossible to run more than one work at a time. By using a single threaded workqueue, we just avoid creating a thread per each work. Ordering is required to avoid a situation when a cgroup's work is put off indefinitely because there are other cgroups to serve, in other words to guarantee fairness. Link: https://bugzilla.kernel.org/show_bug.cgi?id=172981 Link: http://lkml.kernel.org/r/20161004131417.GC1862@esperanza Signed-off-by: Vladimir Davydov <vdavydov.dev@gmail.com> Reported-by: Doug Smythies <dsmythies@telus.net> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Pekka Enberg <penberg@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-12-13 00:41:29 +00:00
*/
memcg_kmem_cache_wq = alloc_workqueue("memcg_kmem_cache", 0, 1);
BUG_ON(!memcg_kmem_cache_wq);
mm: memcontrol: use special workqueue for creating per-memcg caches Creating a lot of cgroups at the same time might stall all worker threads with kmem cache creation works, because kmem cache creation is done with the slab_mutex held. The problem was amplified by commits 801faf0db894 ("mm/slab: lockless decision to grow cache") in case of SLAB and 81ae6d03952c ("mm/slub.c: replace kick_all_cpus_sync() with synchronize_sched() in kmem_cache_shrink()") in case of SLUB, which increased the maximal time the slab_mutex can be held. To prevent that from happening, let's use a special ordered single threaded workqueue for kmem cache creation. This shouldn't introduce any functional changes regarding how kmem caches are created, as the work function holds the global slab_mutex during its whole runtime anyway, making it impossible to run more than one work at a time. By using a single threaded workqueue, we just avoid creating a thread per each work. Ordering is required to avoid a situation when a cgroup's work is put off indefinitely because there are other cgroups to serve, in other words to guarantee fairness. Link: https://bugzilla.kernel.org/show_bug.cgi?id=172981 Link: http://lkml.kernel.org/r/20161004131417.GC1862@esperanza Signed-off-by: Vladimir Davydov <vdavydov.dev@gmail.com> Reported-by: Doug Smythies <dsmythies@telus.net> Acked-by: Michal Hocko <mhocko@suse.com> Cc: Christoph Lameter <cl@linux.com> Cc: David Rientjes <rientjes@google.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Pekka Enberg <penberg@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-12-13 00:41:29 +00:00
#endif
cpuhp_setup_state_nocalls(CPUHP_MM_MEMCQ_DEAD, "mm/memctrl:dead", NULL,
memcg_hotplug_cpu_dead);
for_each_possible_cpu(cpu)
INIT_WORK(&per_cpu_ptr(&memcg_stock, cpu)->work,
drain_local_stock);
for_each_node(node) {
struct mem_cgroup_tree_per_node *rtpn;
rtpn = kzalloc_node(sizeof(*rtpn), GFP_KERNEL,
node_online(node) ? node : NUMA_NO_NODE);
rtpn->rb_root = RB_ROOT;
rtpn->rb_rightmost = NULL;
spin_lock_init(&rtpn->lock);
soft_limit_tree.rb_tree_per_node[node] = rtpn;
}
return 0;
}
subsys_initcall(mem_cgroup_init);
#ifdef CONFIG_MEMCG_SWAP
static struct mem_cgroup *mem_cgroup_id_get_online(struct mem_cgroup *memcg)
{
while (!refcount_inc_not_zero(&memcg->id.ref)) {
/*
* The root cgroup cannot be destroyed, so it's refcount must
* always be >= 1.
*/
if (WARN_ON_ONCE(memcg == root_mem_cgroup)) {
VM_BUG_ON(1);
break;
}
memcg = parent_mem_cgroup(memcg);
if (!memcg)
memcg = root_mem_cgroup;
}
return memcg;
}
/**
* mem_cgroup_swapout - transfer a memsw charge to swap
* @page: page whose memsw charge to transfer
* @entry: swap entry to move the charge to
*
* Transfer the memsw charge of @page to @entry.
*/
void mem_cgroup_swapout(struct page *page, swp_entry_t entry)
{
struct mem_cgroup *memcg, *swap_memcg;
unsigned int nr_entries;
unsigned short oldid;
VM_BUG_ON_PAGE(PageLRU(page), page);
VM_BUG_ON_PAGE(page_count(page), page);
mm: memcontrol: make swap tracking an integral part of memory control Without swap page tracking, users that are otherwise memory controlled can easily escape their containment and allocate significant amounts of memory that they're not being charged for. That's because swap does readahead, but without the cgroup records of who owned the page at swapout, readahead pages don't get charged until somebody actually faults them into their page table and we can identify an owner task. This can be maliciously exploited with MADV_WILLNEED, which triggers arbitrary readahead allocations without charging the pages. Make swap swap page tracking an integral part of memcg and remove the Kconfig options. In the first place, it was only made configurable to allow users to save some memory. But the overhead of tracking cgroup ownership per swap page is minimal - 2 byte per page, or 512k per 1G of swap, or 0.04%. Saving that at the expense of broken containment semantics is not something we should present as a coequal option. The swapaccount=0 boot option will continue to exist, and it will eliminate the page_counter overhead and hide the swap control files, but it won't disable swap slot ownership tracking. This patch makes sure we always have the cgroup records at swapin time; the next patch will fix the actual bug by charging readahead swap pages at swapin time rather than at fault time. v2: fix double swap charge bug in cgroup1/cgroup2 code gating [hannes@cmpxchg.org: fix crash with cgroup_disable=memory] Link: http://lkml.kernel.org/r/20200521215855.GB815153@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Alex Shi <alex.shi@linux.alibaba.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Roman Gushchin <guro@fb.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Naresh Kamboju <naresh.kamboju@linaro.org> Link: http://lkml.kernel.org/r/20200508183105.225460-16-hannes@cmpxchg.org Debugged-by: Hugh Dickins <hughd@google.com> Debugged-by: Michal Hocko <mhocko@kernel.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:02:14 +00:00
if (cgroup_subsys_on_dfl(memory_cgrp_subsys))
return;
memcg = page->mem_cgroup;
/* Readahead page, never charged */
if (!memcg)
return;
/*
* In case the memcg owning these pages has been offlined and doesn't
* have an ID allocated to it anymore, charge the closest online
* ancestor for the swap instead and transfer the memory+swap charge.
*/
swap_memcg = mem_cgroup_id_get_online(memcg);
nr_entries = hpage_nr_pages(page);
/* Get references for the tail pages, too */
if (nr_entries > 1)
mem_cgroup_id_get_many(swap_memcg, nr_entries - 1);
oldid = swap_cgroup_record(entry, mem_cgroup_id(swap_memcg),
nr_entries);
VM_BUG_ON_PAGE(oldid, page);
mod_memcg_state(swap_memcg, MEMCG_SWAP, nr_entries);
page->mem_cgroup = NULL;
if (!mem_cgroup_is_root(memcg))
page_counter_uncharge(&memcg->memory, nr_entries);
mm: memcontrol: make swap tracking an integral part of memory control Without swap page tracking, users that are otherwise memory controlled can easily escape their containment and allocate significant amounts of memory that they're not being charged for. That's because swap does readahead, but without the cgroup records of who owned the page at swapout, readahead pages don't get charged until somebody actually faults them into their page table and we can identify an owner task. This can be maliciously exploited with MADV_WILLNEED, which triggers arbitrary readahead allocations without charging the pages. Make swap swap page tracking an integral part of memcg and remove the Kconfig options. In the first place, it was only made configurable to allow users to save some memory. But the overhead of tracking cgroup ownership per swap page is minimal - 2 byte per page, or 512k per 1G of swap, or 0.04%. Saving that at the expense of broken containment semantics is not something we should present as a coequal option. The swapaccount=0 boot option will continue to exist, and it will eliminate the page_counter overhead and hide the swap control files, but it won't disable swap slot ownership tracking. This patch makes sure we always have the cgroup records at swapin time; the next patch will fix the actual bug by charging readahead swap pages at swapin time rather than at fault time. v2: fix double swap charge bug in cgroup1/cgroup2 code gating [hannes@cmpxchg.org: fix crash with cgroup_disable=memory] Link: http://lkml.kernel.org/r/20200521215855.GB815153@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Alex Shi <alex.shi@linux.alibaba.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Roman Gushchin <guro@fb.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Naresh Kamboju <naresh.kamboju@linaro.org> Link: http://lkml.kernel.org/r/20200508183105.225460-16-hannes@cmpxchg.org Debugged-by: Hugh Dickins <hughd@google.com> Debugged-by: Michal Hocko <mhocko@kernel.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:02:14 +00:00
if (!cgroup_memory_noswap && memcg != swap_memcg) {
if (!mem_cgroup_is_root(swap_memcg))
page_counter_charge(&swap_memcg->memsw, nr_entries);
page_counter_uncharge(&memcg->memsw, nr_entries);
}
mm: memcontrol: bring back the VM_BUG_ON() in mem_cgroup_swapout() Clark stumbled over a VM_BUG_ON() in -RT which was then was removed by Johannes in commit f371763a79d ("mm: memcontrol: fix false-positive VM_BUG_ON() on -rt"). The comment before that patch was a tiny bit better than it is now. While the patch claimed to fix a false-postive on -RT this was not the case. None of the -RT folks ACKed it and it was not a false positive report. That was a *real* problem. This patch updates the comment that is improper because it refers to "disabled preemption" as a consequence of that lock being taken. A spin_lock() disables preemption, true, but in this case the code relies on the fact that the lock _also_ disables interrupts once it is acquired. And this is the important detail (which was checked the VM_BUG_ON()) which needs to be pointed out. This is the hint one needs while looking at the code. It was explained by Johannes on the list that the per-CPU variables are protected by local_irq_save(). The BUG_ON() was helpful. This code has been workarounded in -RT in the meantime. I wouldn't mind running into more of those if the code in question uses *special* kind of locking since now there is no verification (in terms of lockdep or BUG_ON()) and therefore I bring the VM_BUG_ON() check back in. The two functions after the comment could also have a "local_irq_save()" dance around them in order to serialize access to the per-CPU variables. This has been avoided because the interrupts should be off. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Clark Williams <williams@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-09-04 22:47:50 +00:00
/*
* Interrupts should be disabled here because the caller holds the
* i_pages lock which is taken with interrupts-off. It is
mm: memcontrol: bring back the VM_BUG_ON() in mem_cgroup_swapout() Clark stumbled over a VM_BUG_ON() in -RT which was then was removed by Johannes in commit f371763a79d ("mm: memcontrol: fix false-positive VM_BUG_ON() on -rt"). The comment before that patch was a tiny bit better than it is now. While the patch claimed to fix a false-postive on -RT this was not the case. None of the -RT folks ACKed it and it was not a false positive report. That was a *real* problem. This patch updates the comment that is improper because it refers to "disabled preemption" as a consequence of that lock being taken. A spin_lock() disables preemption, true, but in this case the code relies on the fact that the lock _also_ disables interrupts once it is acquired. And this is the important detail (which was checked the VM_BUG_ON()) which needs to be pointed out. This is the hint one needs while looking at the code. It was explained by Johannes on the list that the per-CPU variables are protected by local_irq_save(). The BUG_ON() was helpful. This code has been workarounded in -RT in the meantime. I wouldn't mind running into more of those if the code in question uses *special* kind of locking since now there is no verification (in terms of lockdep or BUG_ON()) and therefore I bring the VM_BUG_ON() check back in. The two functions after the comment could also have a "local_irq_save()" dance around them in order to serialize access to the per-CPU variables. This has been avoided because the interrupts should be off. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Clark Williams <williams@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-09-04 22:47:50 +00:00
* important here to have the interrupts disabled because it is the
* only synchronisation we have for updating the per-CPU variables.
mm: memcontrol: bring back the VM_BUG_ON() in mem_cgroup_swapout() Clark stumbled over a VM_BUG_ON() in -RT which was then was removed by Johannes in commit f371763a79d ("mm: memcontrol: fix false-positive VM_BUG_ON() on -rt"). The comment before that patch was a tiny bit better than it is now. While the patch claimed to fix a false-postive on -RT this was not the case. None of the -RT folks ACKed it and it was not a false positive report. That was a *real* problem. This patch updates the comment that is improper because it refers to "disabled preemption" as a consequence of that lock being taken. A spin_lock() disables preemption, true, but in this case the code relies on the fact that the lock _also_ disables interrupts once it is acquired. And this is the important detail (which was checked the VM_BUG_ON()) which needs to be pointed out. This is the hint one needs while looking at the code. It was explained by Johannes on the list that the per-CPU variables are protected by local_irq_save(). The BUG_ON() was helpful. This code has been workarounded in -RT in the meantime. I wouldn't mind running into more of those if the code in question uses *special* kind of locking since now there is no verification (in terms of lockdep or BUG_ON()) and therefore I bring the VM_BUG_ON() check back in. The two functions after the comment could also have a "local_irq_save()" dance around them in order to serialize access to the per-CPU variables. This has been avoided because the interrupts should be off. Signed-off-by: Sebastian Andrzej Siewior <bigeasy@linutronix.de> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Thomas Gleixner <tglx@linutronix.de> Cc: Steven Rostedt <rostedt@goodmis.org> Cc: Clark Williams <williams@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2015-09-04 22:47:50 +00:00
*/
VM_BUG_ON(!irqs_disabled());
mm: memcontrol: drop @compound parameter from memcg charging API The memcg charging API carries a boolean @compound parameter that tells whether the page we're dealing with is a hugepage. mem_cgroup_commit_charge() has another boolean @lrucare that indicates whether the page needs LRU locking or not while charging. The majority of callsites know those parameters at compile time, which results in a lot of naked "false, false" argument lists. This makes for cryptic code and is a breeding ground for subtle mistakes. Thankfully, the huge page state can be inferred from the page itself and doesn't need to be passed along. This is safe because charging completes before the page is published and somebody may split it. Simplify the callsites by removing @compound, and let memcg infer the state by using hpage_nr_pages() unconditionally. That function does PageTransHuge() to identify huge pages, which also helpfully asserts that nobody passes in tail pages by accident. The following patches will introduce a new charging API, best not to carry over unnecessary weight. Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Alex Shi <alex.shi@linux.alibaba.com> Reviewed-by: Joonsoo Kim <iamjoonsoo.kim@lge.com> Reviewed-by: Shakeel Butt <shakeelb@google.com> Cc: Hugh Dickins <hughd@google.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Michal Hocko <mhocko@suse.com> Cc: Roman Gushchin <guro@fb.com> Cc: Balbir Singh <bsingharora@gmail.com> Link: http://lkml.kernel.org/r/20200508183105.225460-4-hannes@cmpxchg.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:01:31 +00:00
mem_cgroup_charge_statistics(memcg, page, -nr_entries);
memcg_check_events(memcg, page);
mm: memcontrol: fix cgroup creation failure after many small jobs The memory controller has quite a bit of state that usually outlives the cgroup and pins its CSS until said state disappears. At the same time it imposes a 16-bit limit on the CSS ID space to economically store IDs in the wild. Consequently, when we use cgroups to contain frequent but small and short-lived jobs that leave behind some page cache, we quickly run into the 64k limitations of outstanding CSSs. Creating a new cgroup fails with -ENOSPC while there are only a few, or even no user-visible cgroups in existence. Although pinning CSSs past cgroup removal is common, there are only two instances that actually need an ID after a cgroup is deleted: cache shadow entries and swapout records. Cache shadow entries reference the ID weakly and can deal with the CSS having disappeared when it's looked up later. They pose no hurdle. Swap-out records do need to pin the css to hierarchically attribute swapins after the cgroup has been deleted; though the only pages that remain swapped out after offlining are tmpfs/shmem pages. And those references are under the user's control, so they are manageable. This patch introduces a private 16-bit memcg ID and switches swap and cache shadow entries over to using that. This ID can then be recycled after offlining when the CSS remains pinned only by objects that don't specifically need it. This script demonstrates the problem by faulting one cache page in a new cgroup and deleting it again: set -e mkdir -p pages for x in `seq 128000`; do [ $((x % 1000)) -eq 0 ] && echo $x mkdir /cgroup/foo echo $$ >/cgroup/foo/cgroup.procs echo trex >pages/$x echo $$ >/cgroup/cgroup.procs rmdir /cgroup/foo done When run on an unpatched kernel, we eventually run out of possible IDs even though there are no visible cgroups: [root@ham ~]# ./cssidstress.sh [...] 65000 mkdir: cannot create directory '/cgroup/foo': No space left on device After this patch, the IDs get released upon cgroup destruction and the cache and css objects get released once memory reclaim kicks in. [hannes@cmpxchg.org: init the IDR] Link: http://lkml.kernel.org/r/20160621154601.GA22431@cmpxchg.org Fixes: b2052564e66d ("mm: memcontrol: continue cache reclaim from offlined groups") Link: http://lkml.kernel.org/r/20160617162516.GD19084@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Reported-by: John Garcia <john.garcia@mesosphere.io> Reviewed-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Tejun Heo <tj@kernel.org> Cc: Nikolay Borisov <kernel@kyup.com> Cc: <stable@vger.kernel.org> [3.19+] Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-07-20 22:44:57 +00:00
if (!mem_cgroup_is_root(memcg))
css_put_many(&memcg->css, nr_entries);
}
mm, THP, swap: delay splitting THP during swap out Patch series "THP swap: Delay splitting THP during swapping out", v11. This patchset is to optimize the performance of Transparent Huge Page (THP) swap. Recently, the performance of the storage devices improved so fast that we cannot saturate the disk bandwidth with single logical CPU when do page swap out even on a high-end server machine. Because the performance of the storage device improved faster than that of single logical CPU. And it seems that the trend will not change in the near future. On the other hand, the THP becomes more and more popular because of increased memory size. So it becomes necessary to optimize THP swap performance. The advantages of the THP swap support include: - Batch the swap operations for the THP to reduce lock acquiring/releasing, including allocating/freeing the swap space, adding/deleting to/from the swap cache, and writing/reading the swap space, etc. This will help improve the performance of the THP swap. - The THP swap space read/write will be 2M sequential IO. It is particularly helpful for the swap read, which are usually 4k random IO. This will improve the performance of the THP swap too. - It will help the memory fragmentation, especially when the THP is heavily used by the applications. The 2M continuous pages will be free up after THP swapping out. - It will improve the THP utilization on the system with the swap turned on. Because the speed for khugepaged to collapse the normal pages into the THP is quite slow. After the THP is split during the swapping out, it will take quite long time for the normal pages to collapse back into the THP after being swapped in. The high THP utilization helps the efficiency of the page based memory management too. There are some concerns regarding THP swap in, mainly because possible enlarged read/write IO size (for swap in/out) may put more overhead on the storage device. To deal with that, the THP swap in should be turned on only when necessary. For example, it can be selected via "always/never/madvise" logic, to be turned on globally, turned off globally, or turned on only for VMA with MADV_HUGEPAGE, etc. This patchset is the first step for the THP swap support. The plan is to delay splitting THP step by step, finally avoid splitting THP during the THP swapping out and swap out/in the THP as a whole. As the first step, in this patchset, the splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP and adding the THP into the swap cache. This will reduce lock acquiring/releasing for the locks used for the swap cache management. With the patchset, the swap out throughput improves 15.5% (from about 3.73GB/s to about 4.31GB/s) in the vm-scalability swap-w-seq test case with 8 processes. The test is done on a Xeon E5 v3 system. The swap device used is a RAM simulated PMEM (persistent memory) device. To test the sequential swapping out, the test case creates 8 processes, which sequentially allocate and write to the anonymous pages until the RAM and part of the swap device is used up. This patch (of 5): In this patch, splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP (Transparent Huge Page) and adding the THP into the swap cache. This will batch the corresponding operation, thus improve THP swap out throughput. This is the first step for the THP swap optimization. The plan is to delay splitting the THP step by step and avoid splitting the THP finally. In this patch, one swap cluster is used to hold the contents of each THP swapped out. So, the size of the swap cluster is changed to that of the THP (Transparent Huge Page) on x86_64 architecture (512). For other architectures which want such THP swap optimization, ARCH_USES_THP_SWAP_CLUSTER needs to be selected in the Kconfig file for the architecture. In effect, this will enlarge swap cluster size by 2 times on x86_64. Which may make it harder to find a free cluster when the swap space becomes fragmented. So that, this may reduce the continuous swap space allocation and sequential write in theory. The performance test in 0day shows no regressions caused by this. In the future of THP swap optimization, some information of the swapped out THP (such as compound map count) will be recorded in the swap_cluster_info data structure. The mem cgroup swap accounting functions are enhanced to support charge or uncharge a swap cluster backing a THP as a whole. The swap cluster allocate/free functions are added to allocate/free a swap cluster for a THP. A fair simple algorithm is used for swap cluster allocation, that is, only the first swap device in priority list will be tried to allocate the swap cluster. The function will fail if the trying is not successful, and the caller will fallback to allocate a single swap slot instead. This works good enough for normal cases. If the difference of the number of the free swap clusters among multiple swap devices is significant, it is possible that some THPs are split earlier than necessary. For example, this could be caused by big size difference among multiple swap devices. The swap cache functions is enhanced to support add/delete THP to/from the swap cache as a set of (HPAGE_PMD_NR) sub-pages. This may be enhanced in the future with multi-order radix tree. But because we will split the THP soon during swapping out, that optimization doesn't make much sense for this first step. The THP splitting functions are enhanced to support to split THP in swap cache during swapping out. The page lock will be held during allocating the swap cluster, adding the THP into the swap cache and splitting the THP. So in the code path other than swapping out, if the THP need to be split, the PageSwapCache(THP) will be always false. The swap cluster is only available for SSD, so the THP swap optimization in this patchset has no effect for HDD. [ying.huang@intel.com: fix two issues in THP optimize patch] Link: http://lkml.kernel.org/r/87k25ed8zo.fsf@yhuang-dev.intel.com [hannes@cmpxchg.org: extensive cleanups and simplifications, reduce code size] Link: http://lkml.kernel.org/r/20170515112522.32457-2-ying.huang@intel.com Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Suggested-by: Andrew Morton <akpm@linux-foundation.org> [for config option] Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> [for changes in huge_memory.c and huge_mm.h] Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Ebru Akagunduz <ebru.akagunduz@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Cc: Hugh Dickins <hughd@google.com> Cc: Shaohua Li <shli@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-07-06 22:37:18 +00:00
/**
* mem_cgroup_try_charge_swap - try charging swap space for a page
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
* @page: page being added to swap
* @entry: swap entry to charge
*
mm, THP, swap: delay splitting THP during swap out Patch series "THP swap: Delay splitting THP during swapping out", v11. This patchset is to optimize the performance of Transparent Huge Page (THP) swap. Recently, the performance of the storage devices improved so fast that we cannot saturate the disk bandwidth with single logical CPU when do page swap out even on a high-end server machine. Because the performance of the storage device improved faster than that of single logical CPU. And it seems that the trend will not change in the near future. On the other hand, the THP becomes more and more popular because of increased memory size. So it becomes necessary to optimize THP swap performance. The advantages of the THP swap support include: - Batch the swap operations for the THP to reduce lock acquiring/releasing, including allocating/freeing the swap space, adding/deleting to/from the swap cache, and writing/reading the swap space, etc. This will help improve the performance of the THP swap. - The THP swap space read/write will be 2M sequential IO. It is particularly helpful for the swap read, which are usually 4k random IO. This will improve the performance of the THP swap too. - It will help the memory fragmentation, especially when the THP is heavily used by the applications. The 2M continuous pages will be free up after THP swapping out. - It will improve the THP utilization on the system with the swap turned on. Because the speed for khugepaged to collapse the normal pages into the THP is quite slow. After the THP is split during the swapping out, it will take quite long time for the normal pages to collapse back into the THP after being swapped in. The high THP utilization helps the efficiency of the page based memory management too. There are some concerns regarding THP swap in, mainly because possible enlarged read/write IO size (for swap in/out) may put more overhead on the storage device. To deal with that, the THP swap in should be turned on only when necessary. For example, it can be selected via "always/never/madvise" logic, to be turned on globally, turned off globally, or turned on only for VMA with MADV_HUGEPAGE, etc. This patchset is the first step for the THP swap support. The plan is to delay splitting THP step by step, finally avoid splitting THP during the THP swapping out and swap out/in the THP as a whole. As the first step, in this patchset, the splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP and adding the THP into the swap cache. This will reduce lock acquiring/releasing for the locks used for the swap cache management. With the patchset, the swap out throughput improves 15.5% (from about 3.73GB/s to about 4.31GB/s) in the vm-scalability swap-w-seq test case with 8 processes. The test is done on a Xeon E5 v3 system. The swap device used is a RAM simulated PMEM (persistent memory) device. To test the sequential swapping out, the test case creates 8 processes, which sequentially allocate and write to the anonymous pages until the RAM and part of the swap device is used up. This patch (of 5): In this patch, splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP (Transparent Huge Page) and adding the THP into the swap cache. This will batch the corresponding operation, thus improve THP swap out throughput. This is the first step for the THP swap optimization. The plan is to delay splitting the THP step by step and avoid splitting the THP finally. In this patch, one swap cluster is used to hold the contents of each THP swapped out. So, the size of the swap cluster is changed to that of the THP (Transparent Huge Page) on x86_64 architecture (512). For other architectures which want such THP swap optimization, ARCH_USES_THP_SWAP_CLUSTER needs to be selected in the Kconfig file for the architecture. In effect, this will enlarge swap cluster size by 2 times on x86_64. Which may make it harder to find a free cluster when the swap space becomes fragmented. So that, this may reduce the continuous swap space allocation and sequential write in theory. The performance test in 0day shows no regressions caused by this. In the future of THP swap optimization, some information of the swapped out THP (such as compound map count) will be recorded in the swap_cluster_info data structure. The mem cgroup swap accounting functions are enhanced to support charge or uncharge a swap cluster backing a THP as a whole. The swap cluster allocate/free functions are added to allocate/free a swap cluster for a THP. A fair simple algorithm is used for swap cluster allocation, that is, only the first swap device in priority list will be tried to allocate the swap cluster. The function will fail if the trying is not successful, and the caller will fallback to allocate a single swap slot instead. This works good enough for normal cases. If the difference of the number of the free swap clusters among multiple swap devices is significant, it is possible that some THPs are split earlier than necessary. For example, this could be caused by big size difference among multiple swap devices. The swap cache functions is enhanced to support add/delete THP to/from the swap cache as a set of (HPAGE_PMD_NR) sub-pages. This may be enhanced in the future with multi-order radix tree. But because we will split the THP soon during swapping out, that optimization doesn't make much sense for this first step. The THP splitting functions are enhanced to support to split THP in swap cache during swapping out. The page lock will be held during allocating the swap cluster, adding the THP into the swap cache and splitting the THP. So in the code path other than swapping out, if the THP need to be split, the PageSwapCache(THP) will be always false. The swap cluster is only available for SSD, so the THP swap optimization in this patchset has no effect for HDD. [ying.huang@intel.com: fix two issues in THP optimize patch] Link: http://lkml.kernel.org/r/87k25ed8zo.fsf@yhuang-dev.intel.com [hannes@cmpxchg.org: extensive cleanups and simplifications, reduce code size] Link: http://lkml.kernel.org/r/20170515112522.32457-2-ying.huang@intel.com Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Suggested-by: Andrew Morton <akpm@linux-foundation.org> [for config option] Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> [for changes in huge_memory.c and huge_mm.h] Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Ebru Akagunduz <ebru.akagunduz@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Cc: Hugh Dickins <hughd@google.com> Cc: Shaohua Li <shli@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-07-06 22:37:18 +00:00
* Try to charge @page's memcg for the swap space at @entry.
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
*
* Returns 0 on success, -ENOMEM on failure.
*/
int mem_cgroup_try_charge_swap(struct page *page, swp_entry_t entry)
{
mm, THP, swap: delay splitting THP during swap out Patch series "THP swap: Delay splitting THP during swapping out", v11. This patchset is to optimize the performance of Transparent Huge Page (THP) swap. Recently, the performance of the storage devices improved so fast that we cannot saturate the disk bandwidth with single logical CPU when do page swap out even on a high-end server machine. Because the performance of the storage device improved faster than that of single logical CPU. And it seems that the trend will not change in the near future. On the other hand, the THP becomes more and more popular because of increased memory size. So it becomes necessary to optimize THP swap performance. The advantages of the THP swap support include: - Batch the swap operations for the THP to reduce lock acquiring/releasing, including allocating/freeing the swap space, adding/deleting to/from the swap cache, and writing/reading the swap space, etc. This will help improve the performance of the THP swap. - The THP swap space read/write will be 2M sequential IO. It is particularly helpful for the swap read, which are usually 4k random IO. This will improve the performance of the THP swap too. - It will help the memory fragmentation, especially when the THP is heavily used by the applications. The 2M continuous pages will be free up after THP swapping out. - It will improve the THP utilization on the system with the swap turned on. Because the speed for khugepaged to collapse the normal pages into the THP is quite slow. After the THP is split during the swapping out, it will take quite long time for the normal pages to collapse back into the THP after being swapped in. The high THP utilization helps the efficiency of the page based memory management too. There are some concerns regarding THP swap in, mainly because possible enlarged read/write IO size (for swap in/out) may put more overhead on the storage device. To deal with that, the THP swap in should be turned on only when necessary. For example, it can be selected via "always/never/madvise" logic, to be turned on globally, turned off globally, or turned on only for VMA with MADV_HUGEPAGE, etc. This patchset is the first step for the THP swap support. The plan is to delay splitting THP step by step, finally avoid splitting THP during the THP swapping out and swap out/in the THP as a whole. As the first step, in this patchset, the splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP and adding the THP into the swap cache. This will reduce lock acquiring/releasing for the locks used for the swap cache management. With the patchset, the swap out throughput improves 15.5% (from about 3.73GB/s to about 4.31GB/s) in the vm-scalability swap-w-seq test case with 8 processes. The test is done on a Xeon E5 v3 system. The swap device used is a RAM simulated PMEM (persistent memory) device. To test the sequential swapping out, the test case creates 8 processes, which sequentially allocate and write to the anonymous pages until the RAM and part of the swap device is used up. This patch (of 5): In this patch, splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP (Transparent Huge Page) and adding the THP into the swap cache. This will batch the corresponding operation, thus improve THP swap out throughput. This is the first step for the THP swap optimization. The plan is to delay splitting the THP step by step and avoid splitting the THP finally. In this patch, one swap cluster is used to hold the contents of each THP swapped out. So, the size of the swap cluster is changed to that of the THP (Transparent Huge Page) on x86_64 architecture (512). For other architectures which want such THP swap optimization, ARCH_USES_THP_SWAP_CLUSTER needs to be selected in the Kconfig file for the architecture. In effect, this will enlarge swap cluster size by 2 times on x86_64. Which may make it harder to find a free cluster when the swap space becomes fragmented. So that, this may reduce the continuous swap space allocation and sequential write in theory. The performance test in 0day shows no regressions caused by this. In the future of THP swap optimization, some information of the swapped out THP (such as compound map count) will be recorded in the swap_cluster_info data structure. The mem cgroup swap accounting functions are enhanced to support charge or uncharge a swap cluster backing a THP as a whole. The swap cluster allocate/free functions are added to allocate/free a swap cluster for a THP. A fair simple algorithm is used for swap cluster allocation, that is, only the first swap device in priority list will be tried to allocate the swap cluster. The function will fail if the trying is not successful, and the caller will fallback to allocate a single swap slot instead. This works good enough for normal cases. If the difference of the number of the free swap clusters among multiple swap devices is significant, it is possible that some THPs are split earlier than necessary. For example, this could be caused by big size difference among multiple swap devices. The swap cache functions is enhanced to support add/delete THP to/from the swap cache as a set of (HPAGE_PMD_NR) sub-pages. This may be enhanced in the future with multi-order radix tree. But because we will split the THP soon during swapping out, that optimization doesn't make much sense for this first step. The THP splitting functions are enhanced to support to split THP in swap cache during swapping out. The page lock will be held during allocating the swap cluster, adding the THP into the swap cache and splitting the THP. So in the code path other than swapping out, if the THP need to be split, the PageSwapCache(THP) will be always false. The swap cluster is only available for SSD, so the THP swap optimization in this patchset has no effect for HDD. [ying.huang@intel.com: fix two issues in THP optimize patch] Link: http://lkml.kernel.org/r/87k25ed8zo.fsf@yhuang-dev.intel.com [hannes@cmpxchg.org: extensive cleanups and simplifications, reduce code size] Link: http://lkml.kernel.org/r/20170515112522.32457-2-ying.huang@intel.com Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Suggested-by: Andrew Morton <akpm@linux-foundation.org> [for config option] Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> [for changes in huge_memory.c and huge_mm.h] Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Ebru Akagunduz <ebru.akagunduz@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Cc: Hugh Dickins <hughd@google.com> Cc: Shaohua Li <shli@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-07-06 22:37:18 +00:00
unsigned int nr_pages = hpage_nr_pages(page);
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
struct page_counter *counter;
mm, THP, swap: delay splitting THP during swap out Patch series "THP swap: Delay splitting THP during swapping out", v11. This patchset is to optimize the performance of Transparent Huge Page (THP) swap. Recently, the performance of the storage devices improved so fast that we cannot saturate the disk bandwidth with single logical CPU when do page swap out even on a high-end server machine. Because the performance of the storage device improved faster than that of single logical CPU. And it seems that the trend will not change in the near future. On the other hand, the THP becomes more and more popular because of increased memory size. So it becomes necessary to optimize THP swap performance. The advantages of the THP swap support include: - Batch the swap operations for the THP to reduce lock acquiring/releasing, including allocating/freeing the swap space, adding/deleting to/from the swap cache, and writing/reading the swap space, etc. This will help improve the performance of the THP swap. - The THP swap space read/write will be 2M sequential IO. It is particularly helpful for the swap read, which are usually 4k random IO. This will improve the performance of the THP swap too. - It will help the memory fragmentation, especially when the THP is heavily used by the applications. The 2M continuous pages will be free up after THP swapping out. - It will improve the THP utilization on the system with the swap turned on. Because the speed for khugepaged to collapse the normal pages into the THP is quite slow. After the THP is split during the swapping out, it will take quite long time for the normal pages to collapse back into the THP after being swapped in. The high THP utilization helps the efficiency of the page based memory management too. There are some concerns regarding THP swap in, mainly because possible enlarged read/write IO size (for swap in/out) may put more overhead on the storage device. To deal with that, the THP swap in should be turned on only when necessary. For example, it can be selected via "always/never/madvise" logic, to be turned on globally, turned off globally, or turned on only for VMA with MADV_HUGEPAGE, etc. This patchset is the first step for the THP swap support. The plan is to delay splitting THP step by step, finally avoid splitting THP during the THP swapping out and swap out/in the THP as a whole. As the first step, in this patchset, the splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP and adding the THP into the swap cache. This will reduce lock acquiring/releasing for the locks used for the swap cache management. With the patchset, the swap out throughput improves 15.5% (from about 3.73GB/s to about 4.31GB/s) in the vm-scalability swap-w-seq test case with 8 processes. The test is done on a Xeon E5 v3 system. The swap device used is a RAM simulated PMEM (persistent memory) device. To test the sequential swapping out, the test case creates 8 processes, which sequentially allocate and write to the anonymous pages until the RAM and part of the swap device is used up. This patch (of 5): In this patch, splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP (Transparent Huge Page) and adding the THP into the swap cache. This will batch the corresponding operation, thus improve THP swap out throughput. This is the first step for the THP swap optimization. The plan is to delay splitting the THP step by step and avoid splitting the THP finally. In this patch, one swap cluster is used to hold the contents of each THP swapped out. So, the size of the swap cluster is changed to that of the THP (Transparent Huge Page) on x86_64 architecture (512). For other architectures which want such THP swap optimization, ARCH_USES_THP_SWAP_CLUSTER needs to be selected in the Kconfig file for the architecture. In effect, this will enlarge swap cluster size by 2 times on x86_64. Which may make it harder to find a free cluster when the swap space becomes fragmented. So that, this may reduce the continuous swap space allocation and sequential write in theory. The performance test in 0day shows no regressions caused by this. In the future of THP swap optimization, some information of the swapped out THP (such as compound map count) will be recorded in the swap_cluster_info data structure. The mem cgroup swap accounting functions are enhanced to support charge or uncharge a swap cluster backing a THP as a whole. The swap cluster allocate/free functions are added to allocate/free a swap cluster for a THP. A fair simple algorithm is used for swap cluster allocation, that is, only the first swap device in priority list will be tried to allocate the swap cluster. The function will fail if the trying is not successful, and the caller will fallback to allocate a single swap slot instead. This works good enough for normal cases. If the difference of the number of the free swap clusters among multiple swap devices is significant, it is possible that some THPs are split earlier than necessary. For example, this could be caused by big size difference among multiple swap devices. The swap cache functions is enhanced to support add/delete THP to/from the swap cache as a set of (HPAGE_PMD_NR) sub-pages. This may be enhanced in the future with multi-order radix tree. But because we will split the THP soon during swapping out, that optimization doesn't make much sense for this first step. The THP splitting functions are enhanced to support to split THP in swap cache during swapping out. The page lock will be held during allocating the swap cluster, adding the THP into the swap cache and splitting the THP. So in the code path other than swapping out, if the THP need to be split, the PageSwapCache(THP) will be always false. The swap cluster is only available for SSD, so the THP swap optimization in this patchset has no effect for HDD. [ying.huang@intel.com: fix two issues in THP optimize patch] Link: http://lkml.kernel.org/r/87k25ed8zo.fsf@yhuang-dev.intel.com [hannes@cmpxchg.org: extensive cleanups and simplifications, reduce code size] Link: http://lkml.kernel.org/r/20170515112522.32457-2-ying.huang@intel.com Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Suggested-by: Andrew Morton <akpm@linux-foundation.org> [for config option] Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> [for changes in huge_memory.c and huge_mm.h] Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Ebru Akagunduz <ebru.akagunduz@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Cc: Hugh Dickins <hughd@google.com> Cc: Shaohua Li <shli@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-07-06 22:37:18 +00:00
struct mem_cgroup *memcg;
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
unsigned short oldid;
mm: memcontrol: make swap tracking an integral part of memory control Without swap page tracking, users that are otherwise memory controlled can easily escape their containment and allocate significant amounts of memory that they're not being charged for. That's because swap does readahead, but without the cgroup records of who owned the page at swapout, readahead pages don't get charged until somebody actually faults them into their page table and we can identify an owner task. This can be maliciously exploited with MADV_WILLNEED, which triggers arbitrary readahead allocations without charging the pages. Make swap swap page tracking an integral part of memcg and remove the Kconfig options. In the first place, it was only made configurable to allow users to save some memory. But the overhead of tracking cgroup ownership per swap page is minimal - 2 byte per page, or 512k per 1G of swap, or 0.04%. Saving that at the expense of broken containment semantics is not something we should present as a coequal option. The swapaccount=0 boot option will continue to exist, and it will eliminate the page_counter overhead and hide the swap control files, but it won't disable swap slot ownership tracking. This patch makes sure we always have the cgroup records at swapin time; the next patch will fix the actual bug by charging readahead swap pages at swapin time rather than at fault time. v2: fix double swap charge bug in cgroup1/cgroup2 code gating [hannes@cmpxchg.org: fix crash with cgroup_disable=memory] Link: http://lkml.kernel.org/r/20200521215855.GB815153@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Alex Shi <alex.shi@linux.alibaba.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Roman Gushchin <guro@fb.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Naresh Kamboju <naresh.kamboju@linaro.org> Link: http://lkml.kernel.org/r/20200508183105.225460-16-hannes@cmpxchg.org Debugged-by: Hugh Dickins <hughd@google.com> Debugged-by: Michal Hocko <mhocko@kernel.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:02:14 +00:00
if (!cgroup_subsys_on_dfl(memory_cgrp_subsys))
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
return 0;
memcg = page->mem_cgroup;
/* Readahead page, never charged */
if (!memcg)
return 0;
if (!entry.val) {
memcg_memory_event(memcg, MEMCG_SWAP_FAIL);
return 0;
}
memcg = mem_cgroup_id_get_online(memcg);
mm: memcontrol: make swap tracking an integral part of memory control Without swap page tracking, users that are otherwise memory controlled can easily escape their containment and allocate significant amounts of memory that they're not being charged for. That's because swap does readahead, but without the cgroup records of who owned the page at swapout, readahead pages don't get charged until somebody actually faults them into their page table and we can identify an owner task. This can be maliciously exploited with MADV_WILLNEED, which triggers arbitrary readahead allocations without charging the pages. Make swap swap page tracking an integral part of memcg and remove the Kconfig options. In the first place, it was only made configurable to allow users to save some memory. But the overhead of tracking cgroup ownership per swap page is minimal - 2 byte per page, or 512k per 1G of swap, or 0.04%. Saving that at the expense of broken containment semantics is not something we should present as a coequal option. The swapaccount=0 boot option will continue to exist, and it will eliminate the page_counter overhead and hide the swap control files, but it won't disable swap slot ownership tracking. This patch makes sure we always have the cgroup records at swapin time; the next patch will fix the actual bug by charging readahead swap pages at swapin time rather than at fault time. v2: fix double swap charge bug in cgroup1/cgroup2 code gating [hannes@cmpxchg.org: fix crash with cgroup_disable=memory] Link: http://lkml.kernel.org/r/20200521215855.GB815153@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Alex Shi <alex.shi@linux.alibaba.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Roman Gushchin <guro@fb.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Naresh Kamboju <naresh.kamboju@linaro.org> Link: http://lkml.kernel.org/r/20200508183105.225460-16-hannes@cmpxchg.org Debugged-by: Hugh Dickins <hughd@google.com> Debugged-by: Michal Hocko <mhocko@kernel.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:02:14 +00:00
if (!cgroup_memory_noswap && !mem_cgroup_is_root(memcg) &&
mm, THP, swap: delay splitting THP during swap out Patch series "THP swap: Delay splitting THP during swapping out", v11. This patchset is to optimize the performance of Transparent Huge Page (THP) swap. Recently, the performance of the storage devices improved so fast that we cannot saturate the disk bandwidth with single logical CPU when do page swap out even on a high-end server machine. Because the performance of the storage device improved faster than that of single logical CPU. And it seems that the trend will not change in the near future. On the other hand, the THP becomes more and more popular because of increased memory size. So it becomes necessary to optimize THP swap performance. The advantages of the THP swap support include: - Batch the swap operations for the THP to reduce lock acquiring/releasing, including allocating/freeing the swap space, adding/deleting to/from the swap cache, and writing/reading the swap space, etc. This will help improve the performance of the THP swap. - The THP swap space read/write will be 2M sequential IO. It is particularly helpful for the swap read, which are usually 4k random IO. This will improve the performance of the THP swap too. - It will help the memory fragmentation, especially when the THP is heavily used by the applications. The 2M continuous pages will be free up after THP swapping out. - It will improve the THP utilization on the system with the swap turned on. Because the speed for khugepaged to collapse the normal pages into the THP is quite slow. After the THP is split during the swapping out, it will take quite long time for the normal pages to collapse back into the THP after being swapped in. The high THP utilization helps the efficiency of the page based memory management too. There are some concerns regarding THP swap in, mainly because possible enlarged read/write IO size (for swap in/out) may put more overhead on the storage device. To deal with that, the THP swap in should be turned on only when necessary. For example, it can be selected via "always/never/madvise" logic, to be turned on globally, turned off globally, or turned on only for VMA with MADV_HUGEPAGE, etc. This patchset is the first step for the THP swap support. The plan is to delay splitting THP step by step, finally avoid splitting THP during the THP swapping out and swap out/in the THP as a whole. As the first step, in this patchset, the splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP and adding the THP into the swap cache. This will reduce lock acquiring/releasing for the locks used for the swap cache management. With the patchset, the swap out throughput improves 15.5% (from about 3.73GB/s to about 4.31GB/s) in the vm-scalability swap-w-seq test case with 8 processes. The test is done on a Xeon E5 v3 system. The swap device used is a RAM simulated PMEM (persistent memory) device. To test the sequential swapping out, the test case creates 8 processes, which sequentially allocate and write to the anonymous pages until the RAM and part of the swap device is used up. This patch (of 5): In this patch, splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP (Transparent Huge Page) and adding the THP into the swap cache. This will batch the corresponding operation, thus improve THP swap out throughput. This is the first step for the THP swap optimization. The plan is to delay splitting the THP step by step and avoid splitting the THP finally. In this patch, one swap cluster is used to hold the contents of each THP swapped out. So, the size of the swap cluster is changed to that of the THP (Transparent Huge Page) on x86_64 architecture (512). For other architectures which want such THP swap optimization, ARCH_USES_THP_SWAP_CLUSTER needs to be selected in the Kconfig file for the architecture. In effect, this will enlarge swap cluster size by 2 times on x86_64. Which may make it harder to find a free cluster when the swap space becomes fragmented. So that, this may reduce the continuous swap space allocation and sequential write in theory. The performance test in 0day shows no regressions caused by this. In the future of THP swap optimization, some information of the swapped out THP (such as compound map count) will be recorded in the swap_cluster_info data structure. The mem cgroup swap accounting functions are enhanced to support charge or uncharge a swap cluster backing a THP as a whole. The swap cluster allocate/free functions are added to allocate/free a swap cluster for a THP. A fair simple algorithm is used for swap cluster allocation, that is, only the first swap device in priority list will be tried to allocate the swap cluster. The function will fail if the trying is not successful, and the caller will fallback to allocate a single swap slot instead. This works good enough for normal cases. If the difference of the number of the free swap clusters among multiple swap devices is significant, it is possible that some THPs are split earlier than necessary. For example, this could be caused by big size difference among multiple swap devices. The swap cache functions is enhanced to support add/delete THP to/from the swap cache as a set of (HPAGE_PMD_NR) sub-pages. This may be enhanced in the future with multi-order radix tree. But because we will split the THP soon during swapping out, that optimization doesn't make much sense for this first step. The THP splitting functions are enhanced to support to split THP in swap cache during swapping out. The page lock will be held during allocating the swap cluster, adding the THP into the swap cache and splitting the THP. So in the code path other than swapping out, if the THP need to be split, the PageSwapCache(THP) will be always false. The swap cluster is only available for SSD, so the THP swap optimization in this patchset has no effect for HDD. [ying.huang@intel.com: fix two issues in THP optimize patch] Link: http://lkml.kernel.org/r/87k25ed8zo.fsf@yhuang-dev.intel.com [hannes@cmpxchg.org: extensive cleanups and simplifications, reduce code size] Link: http://lkml.kernel.org/r/20170515112522.32457-2-ying.huang@intel.com Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Suggested-by: Andrew Morton <akpm@linux-foundation.org> [for config option] Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> [for changes in huge_memory.c and huge_mm.h] Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Ebru Akagunduz <ebru.akagunduz@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Cc: Hugh Dickins <hughd@google.com> Cc: Shaohua Li <shli@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-07-06 22:37:18 +00:00
!page_counter_try_charge(&memcg->swap, nr_pages, &counter)) {
memcg_memory_event(memcg, MEMCG_SWAP_MAX);
memcg_memory_event(memcg, MEMCG_SWAP_FAIL);
mem_cgroup_id_put(memcg);
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
return -ENOMEM;
}
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
mm, THP, swap: delay splitting THP during swap out Patch series "THP swap: Delay splitting THP during swapping out", v11. This patchset is to optimize the performance of Transparent Huge Page (THP) swap. Recently, the performance of the storage devices improved so fast that we cannot saturate the disk bandwidth with single logical CPU when do page swap out even on a high-end server machine. Because the performance of the storage device improved faster than that of single logical CPU. And it seems that the trend will not change in the near future. On the other hand, the THP becomes more and more popular because of increased memory size. So it becomes necessary to optimize THP swap performance. The advantages of the THP swap support include: - Batch the swap operations for the THP to reduce lock acquiring/releasing, including allocating/freeing the swap space, adding/deleting to/from the swap cache, and writing/reading the swap space, etc. This will help improve the performance of the THP swap. - The THP swap space read/write will be 2M sequential IO. It is particularly helpful for the swap read, which are usually 4k random IO. This will improve the performance of the THP swap too. - It will help the memory fragmentation, especially when the THP is heavily used by the applications. The 2M continuous pages will be free up after THP swapping out. - It will improve the THP utilization on the system with the swap turned on. Because the speed for khugepaged to collapse the normal pages into the THP is quite slow. After the THP is split during the swapping out, it will take quite long time for the normal pages to collapse back into the THP after being swapped in. The high THP utilization helps the efficiency of the page based memory management too. There are some concerns regarding THP swap in, mainly because possible enlarged read/write IO size (for swap in/out) may put more overhead on the storage device. To deal with that, the THP swap in should be turned on only when necessary. For example, it can be selected via "always/never/madvise" logic, to be turned on globally, turned off globally, or turned on only for VMA with MADV_HUGEPAGE, etc. This patchset is the first step for the THP swap support. The plan is to delay splitting THP step by step, finally avoid splitting THP during the THP swapping out and swap out/in the THP as a whole. As the first step, in this patchset, the splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP and adding the THP into the swap cache. This will reduce lock acquiring/releasing for the locks used for the swap cache management. With the patchset, the swap out throughput improves 15.5% (from about 3.73GB/s to about 4.31GB/s) in the vm-scalability swap-w-seq test case with 8 processes. The test is done on a Xeon E5 v3 system. The swap device used is a RAM simulated PMEM (persistent memory) device. To test the sequential swapping out, the test case creates 8 processes, which sequentially allocate and write to the anonymous pages until the RAM and part of the swap device is used up. This patch (of 5): In this patch, splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP (Transparent Huge Page) and adding the THP into the swap cache. This will batch the corresponding operation, thus improve THP swap out throughput. This is the first step for the THP swap optimization. The plan is to delay splitting the THP step by step and avoid splitting the THP finally. In this patch, one swap cluster is used to hold the contents of each THP swapped out. So, the size of the swap cluster is changed to that of the THP (Transparent Huge Page) on x86_64 architecture (512). For other architectures which want such THP swap optimization, ARCH_USES_THP_SWAP_CLUSTER needs to be selected in the Kconfig file for the architecture. In effect, this will enlarge swap cluster size by 2 times on x86_64. Which may make it harder to find a free cluster when the swap space becomes fragmented. So that, this may reduce the continuous swap space allocation and sequential write in theory. The performance test in 0day shows no regressions caused by this. In the future of THP swap optimization, some information of the swapped out THP (such as compound map count) will be recorded in the swap_cluster_info data structure. The mem cgroup swap accounting functions are enhanced to support charge or uncharge a swap cluster backing a THP as a whole. The swap cluster allocate/free functions are added to allocate/free a swap cluster for a THP. A fair simple algorithm is used for swap cluster allocation, that is, only the first swap device in priority list will be tried to allocate the swap cluster. The function will fail if the trying is not successful, and the caller will fallback to allocate a single swap slot instead. This works good enough for normal cases. If the difference of the number of the free swap clusters among multiple swap devices is significant, it is possible that some THPs are split earlier than necessary. For example, this could be caused by big size difference among multiple swap devices. The swap cache functions is enhanced to support add/delete THP to/from the swap cache as a set of (HPAGE_PMD_NR) sub-pages. This may be enhanced in the future with multi-order radix tree. But because we will split the THP soon during swapping out, that optimization doesn't make much sense for this first step. The THP splitting functions are enhanced to support to split THP in swap cache during swapping out. The page lock will be held during allocating the swap cluster, adding the THP into the swap cache and splitting the THP. So in the code path other than swapping out, if the THP need to be split, the PageSwapCache(THP) will be always false. The swap cluster is only available for SSD, so the THP swap optimization in this patchset has no effect for HDD. [ying.huang@intel.com: fix two issues in THP optimize patch] Link: http://lkml.kernel.org/r/87k25ed8zo.fsf@yhuang-dev.intel.com [hannes@cmpxchg.org: extensive cleanups and simplifications, reduce code size] Link: http://lkml.kernel.org/r/20170515112522.32457-2-ying.huang@intel.com Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Suggested-by: Andrew Morton <akpm@linux-foundation.org> [for config option] Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> [for changes in huge_memory.c and huge_mm.h] Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Ebru Akagunduz <ebru.akagunduz@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Cc: Hugh Dickins <hughd@google.com> Cc: Shaohua Li <shli@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-07-06 22:37:18 +00:00
/* Get references for the tail pages, too */
if (nr_pages > 1)
mem_cgroup_id_get_many(memcg, nr_pages - 1);
oldid = swap_cgroup_record(entry, mem_cgroup_id(memcg), nr_pages);
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
VM_BUG_ON_PAGE(oldid, page);
mod_memcg_state(memcg, MEMCG_SWAP, nr_pages);
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
return 0;
}
/**
mm, THP, swap: delay splitting THP during swap out Patch series "THP swap: Delay splitting THP during swapping out", v11. This patchset is to optimize the performance of Transparent Huge Page (THP) swap. Recently, the performance of the storage devices improved so fast that we cannot saturate the disk bandwidth with single logical CPU when do page swap out even on a high-end server machine. Because the performance of the storage device improved faster than that of single logical CPU. And it seems that the trend will not change in the near future. On the other hand, the THP becomes more and more popular because of increased memory size. So it becomes necessary to optimize THP swap performance. The advantages of the THP swap support include: - Batch the swap operations for the THP to reduce lock acquiring/releasing, including allocating/freeing the swap space, adding/deleting to/from the swap cache, and writing/reading the swap space, etc. This will help improve the performance of the THP swap. - The THP swap space read/write will be 2M sequential IO. It is particularly helpful for the swap read, which are usually 4k random IO. This will improve the performance of the THP swap too. - It will help the memory fragmentation, especially when the THP is heavily used by the applications. The 2M continuous pages will be free up after THP swapping out. - It will improve the THP utilization on the system with the swap turned on. Because the speed for khugepaged to collapse the normal pages into the THP is quite slow. After the THP is split during the swapping out, it will take quite long time for the normal pages to collapse back into the THP after being swapped in. The high THP utilization helps the efficiency of the page based memory management too. There are some concerns regarding THP swap in, mainly because possible enlarged read/write IO size (for swap in/out) may put more overhead on the storage device. To deal with that, the THP swap in should be turned on only when necessary. For example, it can be selected via "always/never/madvise" logic, to be turned on globally, turned off globally, or turned on only for VMA with MADV_HUGEPAGE, etc. This patchset is the first step for the THP swap support. The plan is to delay splitting THP step by step, finally avoid splitting THP during the THP swapping out and swap out/in the THP as a whole. As the first step, in this patchset, the splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP and adding the THP into the swap cache. This will reduce lock acquiring/releasing for the locks used for the swap cache management. With the patchset, the swap out throughput improves 15.5% (from about 3.73GB/s to about 4.31GB/s) in the vm-scalability swap-w-seq test case with 8 processes. The test is done on a Xeon E5 v3 system. The swap device used is a RAM simulated PMEM (persistent memory) device. To test the sequential swapping out, the test case creates 8 processes, which sequentially allocate and write to the anonymous pages until the RAM and part of the swap device is used up. This patch (of 5): In this patch, splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP (Transparent Huge Page) and adding the THP into the swap cache. This will batch the corresponding operation, thus improve THP swap out throughput. This is the first step for the THP swap optimization. The plan is to delay splitting the THP step by step and avoid splitting the THP finally. In this patch, one swap cluster is used to hold the contents of each THP swapped out. So, the size of the swap cluster is changed to that of the THP (Transparent Huge Page) on x86_64 architecture (512). For other architectures which want such THP swap optimization, ARCH_USES_THP_SWAP_CLUSTER needs to be selected in the Kconfig file for the architecture. In effect, this will enlarge swap cluster size by 2 times on x86_64. Which may make it harder to find a free cluster when the swap space becomes fragmented. So that, this may reduce the continuous swap space allocation and sequential write in theory. The performance test in 0day shows no regressions caused by this. In the future of THP swap optimization, some information of the swapped out THP (such as compound map count) will be recorded in the swap_cluster_info data structure. The mem cgroup swap accounting functions are enhanced to support charge or uncharge a swap cluster backing a THP as a whole. The swap cluster allocate/free functions are added to allocate/free a swap cluster for a THP. A fair simple algorithm is used for swap cluster allocation, that is, only the first swap device in priority list will be tried to allocate the swap cluster. The function will fail if the trying is not successful, and the caller will fallback to allocate a single swap slot instead. This works good enough for normal cases. If the difference of the number of the free swap clusters among multiple swap devices is significant, it is possible that some THPs are split earlier than necessary. For example, this could be caused by big size difference among multiple swap devices. The swap cache functions is enhanced to support add/delete THP to/from the swap cache as a set of (HPAGE_PMD_NR) sub-pages. This may be enhanced in the future with multi-order radix tree. But because we will split the THP soon during swapping out, that optimization doesn't make much sense for this first step. The THP splitting functions are enhanced to support to split THP in swap cache during swapping out. The page lock will be held during allocating the swap cluster, adding the THP into the swap cache and splitting the THP. So in the code path other than swapping out, if the THP need to be split, the PageSwapCache(THP) will be always false. The swap cluster is only available for SSD, so the THP swap optimization in this patchset has no effect for HDD. [ying.huang@intel.com: fix two issues in THP optimize patch] Link: http://lkml.kernel.org/r/87k25ed8zo.fsf@yhuang-dev.intel.com [hannes@cmpxchg.org: extensive cleanups and simplifications, reduce code size] Link: http://lkml.kernel.org/r/20170515112522.32457-2-ying.huang@intel.com Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Suggested-by: Andrew Morton <akpm@linux-foundation.org> [for config option] Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> [for changes in huge_memory.c and huge_mm.h] Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Ebru Akagunduz <ebru.akagunduz@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Cc: Hugh Dickins <hughd@google.com> Cc: Shaohua Li <shli@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-07-06 22:37:18 +00:00
* mem_cgroup_uncharge_swap - uncharge swap space
* @entry: swap entry to uncharge
mm, THP, swap: delay splitting THP during swap out Patch series "THP swap: Delay splitting THP during swapping out", v11. This patchset is to optimize the performance of Transparent Huge Page (THP) swap. Recently, the performance of the storage devices improved so fast that we cannot saturate the disk bandwidth with single logical CPU when do page swap out even on a high-end server machine. Because the performance of the storage device improved faster than that of single logical CPU. And it seems that the trend will not change in the near future. On the other hand, the THP becomes more and more popular because of increased memory size. So it becomes necessary to optimize THP swap performance. The advantages of the THP swap support include: - Batch the swap operations for the THP to reduce lock acquiring/releasing, including allocating/freeing the swap space, adding/deleting to/from the swap cache, and writing/reading the swap space, etc. This will help improve the performance of the THP swap. - The THP swap space read/write will be 2M sequential IO. It is particularly helpful for the swap read, which are usually 4k random IO. This will improve the performance of the THP swap too. - It will help the memory fragmentation, especially when the THP is heavily used by the applications. The 2M continuous pages will be free up after THP swapping out. - It will improve the THP utilization on the system with the swap turned on. Because the speed for khugepaged to collapse the normal pages into the THP is quite slow. After the THP is split during the swapping out, it will take quite long time for the normal pages to collapse back into the THP after being swapped in. The high THP utilization helps the efficiency of the page based memory management too. There are some concerns regarding THP swap in, mainly because possible enlarged read/write IO size (for swap in/out) may put more overhead on the storage device. To deal with that, the THP swap in should be turned on only when necessary. For example, it can be selected via "always/never/madvise" logic, to be turned on globally, turned off globally, or turned on only for VMA with MADV_HUGEPAGE, etc. This patchset is the first step for the THP swap support. The plan is to delay splitting THP step by step, finally avoid splitting THP during the THP swapping out and swap out/in the THP as a whole. As the first step, in this patchset, the splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP and adding the THP into the swap cache. This will reduce lock acquiring/releasing for the locks used for the swap cache management. With the patchset, the swap out throughput improves 15.5% (from about 3.73GB/s to about 4.31GB/s) in the vm-scalability swap-w-seq test case with 8 processes. The test is done on a Xeon E5 v3 system. The swap device used is a RAM simulated PMEM (persistent memory) device. To test the sequential swapping out, the test case creates 8 processes, which sequentially allocate and write to the anonymous pages until the RAM and part of the swap device is used up. This patch (of 5): In this patch, splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP (Transparent Huge Page) and adding the THP into the swap cache. This will batch the corresponding operation, thus improve THP swap out throughput. This is the first step for the THP swap optimization. The plan is to delay splitting the THP step by step and avoid splitting the THP finally. In this patch, one swap cluster is used to hold the contents of each THP swapped out. So, the size of the swap cluster is changed to that of the THP (Transparent Huge Page) on x86_64 architecture (512). For other architectures which want such THP swap optimization, ARCH_USES_THP_SWAP_CLUSTER needs to be selected in the Kconfig file for the architecture. In effect, this will enlarge swap cluster size by 2 times on x86_64. Which may make it harder to find a free cluster when the swap space becomes fragmented. So that, this may reduce the continuous swap space allocation and sequential write in theory. The performance test in 0day shows no regressions caused by this. In the future of THP swap optimization, some information of the swapped out THP (such as compound map count) will be recorded in the swap_cluster_info data structure. The mem cgroup swap accounting functions are enhanced to support charge or uncharge a swap cluster backing a THP as a whole. The swap cluster allocate/free functions are added to allocate/free a swap cluster for a THP. A fair simple algorithm is used for swap cluster allocation, that is, only the first swap device in priority list will be tried to allocate the swap cluster. The function will fail if the trying is not successful, and the caller will fallback to allocate a single swap slot instead. This works good enough for normal cases. If the difference of the number of the free swap clusters among multiple swap devices is significant, it is possible that some THPs are split earlier than necessary. For example, this could be caused by big size difference among multiple swap devices. The swap cache functions is enhanced to support add/delete THP to/from the swap cache as a set of (HPAGE_PMD_NR) sub-pages. This may be enhanced in the future with multi-order radix tree. But because we will split the THP soon during swapping out, that optimization doesn't make much sense for this first step. The THP splitting functions are enhanced to support to split THP in swap cache during swapping out. The page lock will be held during allocating the swap cluster, adding the THP into the swap cache and splitting the THP. So in the code path other than swapping out, if the THP need to be split, the PageSwapCache(THP) will be always false. The swap cluster is only available for SSD, so the THP swap optimization in this patchset has no effect for HDD. [ying.huang@intel.com: fix two issues in THP optimize patch] Link: http://lkml.kernel.org/r/87k25ed8zo.fsf@yhuang-dev.intel.com [hannes@cmpxchg.org: extensive cleanups and simplifications, reduce code size] Link: http://lkml.kernel.org/r/20170515112522.32457-2-ying.huang@intel.com Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Suggested-by: Andrew Morton <akpm@linux-foundation.org> [for config option] Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> [for changes in huge_memory.c and huge_mm.h] Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Ebru Akagunduz <ebru.akagunduz@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Cc: Hugh Dickins <hughd@google.com> Cc: Shaohua Li <shli@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-07-06 22:37:18 +00:00
* @nr_pages: the amount of swap space to uncharge
*/
mm, THP, swap: delay splitting THP during swap out Patch series "THP swap: Delay splitting THP during swapping out", v11. This patchset is to optimize the performance of Transparent Huge Page (THP) swap. Recently, the performance of the storage devices improved so fast that we cannot saturate the disk bandwidth with single logical CPU when do page swap out even on a high-end server machine. Because the performance of the storage device improved faster than that of single logical CPU. And it seems that the trend will not change in the near future. On the other hand, the THP becomes more and more popular because of increased memory size. So it becomes necessary to optimize THP swap performance. The advantages of the THP swap support include: - Batch the swap operations for the THP to reduce lock acquiring/releasing, including allocating/freeing the swap space, adding/deleting to/from the swap cache, and writing/reading the swap space, etc. This will help improve the performance of the THP swap. - The THP swap space read/write will be 2M sequential IO. It is particularly helpful for the swap read, which are usually 4k random IO. This will improve the performance of the THP swap too. - It will help the memory fragmentation, especially when the THP is heavily used by the applications. The 2M continuous pages will be free up after THP swapping out. - It will improve the THP utilization on the system with the swap turned on. Because the speed for khugepaged to collapse the normal pages into the THP is quite slow. After the THP is split during the swapping out, it will take quite long time for the normal pages to collapse back into the THP after being swapped in. The high THP utilization helps the efficiency of the page based memory management too. There are some concerns regarding THP swap in, mainly because possible enlarged read/write IO size (for swap in/out) may put more overhead on the storage device. To deal with that, the THP swap in should be turned on only when necessary. For example, it can be selected via "always/never/madvise" logic, to be turned on globally, turned off globally, or turned on only for VMA with MADV_HUGEPAGE, etc. This patchset is the first step for the THP swap support. The plan is to delay splitting THP step by step, finally avoid splitting THP during the THP swapping out and swap out/in the THP as a whole. As the first step, in this patchset, the splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP and adding the THP into the swap cache. This will reduce lock acquiring/releasing for the locks used for the swap cache management. With the patchset, the swap out throughput improves 15.5% (from about 3.73GB/s to about 4.31GB/s) in the vm-scalability swap-w-seq test case with 8 processes. The test is done on a Xeon E5 v3 system. The swap device used is a RAM simulated PMEM (persistent memory) device. To test the sequential swapping out, the test case creates 8 processes, which sequentially allocate and write to the anonymous pages until the RAM and part of the swap device is used up. This patch (of 5): In this patch, splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP (Transparent Huge Page) and adding the THP into the swap cache. This will batch the corresponding operation, thus improve THP swap out throughput. This is the first step for the THP swap optimization. The plan is to delay splitting the THP step by step and avoid splitting the THP finally. In this patch, one swap cluster is used to hold the contents of each THP swapped out. So, the size of the swap cluster is changed to that of the THP (Transparent Huge Page) on x86_64 architecture (512). For other architectures which want such THP swap optimization, ARCH_USES_THP_SWAP_CLUSTER needs to be selected in the Kconfig file for the architecture. In effect, this will enlarge swap cluster size by 2 times on x86_64. Which may make it harder to find a free cluster when the swap space becomes fragmented. So that, this may reduce the continuous swap space allocation and sequential write in theory. The performance test in 0day shows no regressions caused by this. In the future of THP swap optimization, some information of the swapped out THP (such as compound map count) will be recorded in the swap_cluster_info data structure. The mem cgroup swap accounting functions are enhanced to support charge or uncharge a swap cluster backing a THP as a whole. The swap cluster allocate/free functions are added to allocate/free a swap cluster for a THP. A fair simple algorithm is used for swap cluster allocation, that is, only the first swap device in priority list will be tried to allocate the swap cluster. The function will fail if the trying is not successful, and the caller will fallback to allocate a single swap slot instead. This works good enough for normal cases. If the difference of the number of the free swap clusters among multiple swap devices is significant, it is possible that some THPs are split earlier than necessary. For example, this could be caused by big size difference among multiple swap devices. The swap cache functions is enhanced to support add/delete THP to/from the swap cache as a set of (HPAGE_PMD_NR) sub-pages. This may be enhanced in the future with multi-order radix tree. But because we will split the THP soon during swapping out, that optimization doesn't make much sense for this first step. The THP splitting functions are enhanced to support to split THP in swap cache during swapping out. The page lock will be held during allocating the swap cluster, adding the THP into the swap cache and splitting the THP. So in the code path other than swapping out, if the THP need to be split, the PageSwapCache(THP) will be always false. The swap cluster is only available for SSD, so the THP swap optimization in this patchset has no effect for HDD. [ying.huang@intel.com: fix two issues in THP optimize patch] Link: http://lkml.kernel.org/r/87k25ed8zo.fsf@yhuang-dev.intel.com [hannes@cmpxchg.org: extensive cleanups and simplifications, reduce code size] Link: http://lkml.kernel.org/r/20170515112522.32457-2-ying.huang@intel.com Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Suggested-by: Andrew Morton <akpm@linux-foundation.org> [for config option] Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> [for changes in huge_memory.c and huge_mm.h] Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Ebru Akagunduz <ebru.akagunduz@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Cc: Hugh Dickins <hughd@google.com> Cc: Shaohua Li <shli@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-07-06 22:37:18 +00:00
void mem_cgroup_uncharge_swap(swp_entry_t entry, unsigned int nr_pages)
{
struct mem_cgroup *memcg;
unsigned short id;
mm, THP, swap: delay splitting THP during swap out Patch series "THP swap: Delay splitting THP during swapping out", v11. This patchset is to optimize the performance of Transparent Huge Page (THP) swap. Recently, the performance of the storage devices improved so fast that we cannot saturate the disk bandwidth with single logical CPU when do page swap out even on a high-end server machine. Because the performance of the storage device improved faster than that of single logical CPU. And it seems that the trend will not change in the near future. On the other hand, the THP becomes more and more popular because of increased memory size. So it becomes necessary to optimize THP swap performance. The advantages of the THP swap support include: - Batch the swap operations for the THP to reduce lock acquiring/releasing, including allocating/freeing the swap space, adding/deleting to/from the swap cache, and writing/reading the swap space, etc. This will help improve the performance of the THP swap. - The THP swap space read/write will be 2M sequential IO. It is particularly helpful for the swap read, which are usually 4k random IO. This will improve the performance of the THP swap too. - It will help the memory fragmentation, especially when the THP is heavily used by the applications. The 2M continuous pages will be free up after THP swapping out. - It will improve the THP utilization on the system with the swap turned on. Because the speed for khugepaged to collapse the normal pages into the THP is quite slow. After the THP is split during the swapping out, it will take quite long time for the normal pages to collapse back into the THP after being swapped in. The high THP utilization helps the efficiency of the page based memory management too. There are some concerns regarding THP swap in, mainly because possible enlarged read/write IO size (for swap in/out) may put more overhead on the storage device. To deal with that, the THP swap in should be turned on only when necessary. For example, it can be selected via "always/never/madvise" logic, to be turned on globally, turned off globally, or turned on only for VMA with MADV_HUGEPAGE, etc. This patchset is the first step for the THP swap support. The plan is to delay splitting THP step by step, finally avoid splitting THP during the THP swapping out and swap out/in the THP as a whole. As the first step, in this patchset, the splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP and adding the THP into the swap cache. This will reduce lock acquiring/releasing for the locks used for the swap cache management. With the patchset, the swap out throughput improves 15.5% (from about 3.73GB/s to about 4.31GB/s) in the vm-scalability swap-w-seq test case with 8 processes. The test is done on a Xeon E5 v3 system. The swap device used is a RAM simulated PMEM (persistent memory) device. To test the sequential swapping out, the test case creates 8 processes, which sequentially allocate and write to the anonymous pages until the RAM and part of the swap device is used up. This patch (of 5): In this patch, splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP (Transparent Huge Page) and adding the THP into the swap cache. This will batch the corresponding operation, thus improve THP swap out throughput. This is the first step for the THP swap optimization. The plan is to delay splitting the THP step by step and avoid splitting the THP finally. In this patch, one swap cluster is used to hold the contents of each THP swapped out. So, the size of the swap cluster is changed to that of the THP (Transparent Huge Page) on x86_64 architecture (512). For other architectures which want such THP swap optimization, ARCH_USES_THP_SWAP_CLUSTER needs to be selected in the Kconfig file for the architecture. In effect, this will enlarge swap cluster size by 2 times on x86_64. Which may make it harder to find a free cluster when the swap space becomes fragmented. So that, this may reduce the continuous swap space allocation and sequential write in theory. The performance test in 0day shows no regressions caused by this. In the future of THP swap optimization, some information of the swapped out THP (such as compound map count) will be recorded in the swap_cluster_info data structure. The mem cgroup swap accounting functions are enhanced to support charge or uncharge a swap cluster backing a THP as a whole. The swap cluster allocate/free functions are added to allocate/free a swap cluster for a THP. A fair simple algorithm is used for swap cluster allocation, that is, only the first swap device in priority list will be tried to allocate the swap cluster. The function will fail if the trying is not successful, and the caller will fallback to allocate a single swap slot instead. This works good enough for normal cases. If the difference of the number of the free swap clusters among multiple swap devices is significant, it is possible that some THPs are split earlier than necessary. For example, this could be caused by big size difference among multiple swap devices. The swap cache functions is enhanced to support add/delete THP to/from the swap cache as a set of (HPAGE_PMD_NR) sub-pages. This may be enhanced in the future with multi-order radix tree. But because we will split the THP soon during swapping out, that optimization doesn't make much sense for this first step. The THP splitting functions are enhanced to support to split THP in swap cache during swapping out. The page lock will be held during allocating the swap cluster, adding the THP into the swap cache and splitting the THP. So in the code path other than swapping out, if the THP need to be split, the PageSwapCache(THP) will be always false. The swap cluster is only available for SSD, so the THP swap optimization in this patchset has no effect for HDD. [ying.huang@intel.com: fix two issues in THP optimize patch] Link: http://lkml.kernel.org/r/87k25ed8zo.fsf@yhuang-dev.intel.com [hannes@cmpxchg.org: extensive cleanups and simplifications, reduce code size] Link: http://lkml.kernel.org/r/20170515112522.32457-2-ying.huang@intel.com Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Suggested-by: Andrew Morton <akpm@linux-foundation.org> [for config option] Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> [for changes in huge_memory.c and huge_mm.h] Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Ebru Akagunduz <ebru.akagunduz@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Cc: Hugh Dickins <hughd@google.com> Cc: Shaohua Li <shli@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-07-06 22:37:18 +00:00
id = swap_cgroup_record(entry, 0, nr_pages);
rcu_read_lock();
memcg = mem_cgroup_from_id(id);
if (memcg) {
mm: memcontrol: make swap tracking an integral part of memory control Without swap page tracking, users that are otherwise memory controlled can easily escape their containment and allocate significant amounts of memory that they're not being charged for. That's because swap does readahead, but without the cgroup records of who owned the page at swapout, readahead pages don't get charged until somebody actually faults them into their page table and we can identify an owner task. This can be maliciously exploited with MADV_WILLNEED, which triggers arbitrary readahead allocations without charging the pages. Make swap swap page tracking an integral part of memcg and remove the Kconfig options. In the first place, it was only made configurable to allow users to save some memory. But the overhead of tracking cgroup ownership per swap page is minimal - 2 byte per page, or 512k per 1G of swap, or 0.04%. Saving that at the expense of broken containment semantics is not something we should present as a coequal option. The swapaccount=0 boot option will continue to exist, and it will eliminate the page_counter overhead and hide the swap control files, but it won't disable swap slot ownership tracking. This patch makes sure we always have the cgroup records at swapin time; the next patch will fix the actual bug by charging readahead swap pages at swapin time rather than at fault time. v2: fix double swap charge bug in cgroup1/cgroup2 code gating [hannes@cmpxchg.org: fix crash with cgroup_disable=memory] Link: http://lkml.kernel.org/r/20200521215855.GB815153@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Alex Shi <alex.shi@linux.alibaba.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Roman Gushchin <guro@fb.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Naresh Kamboju <naresh.kamboju@linaro.org> Link: http://lkml.kernel.org/r/20200508183105.225460-16-hannes@cmpxchg.org Debugged-by: Hugh Dickins <hughd@google.com> Debugged-by: Michal Hocko <mhocko@kernel.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:02:14 +00:00
if (!cgroup_memory_noswap && !mem_cgroup_is_root(memcg)) {
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
if (cgroup_subsys_on_dfl(memory_cgrp_subsys))
mm, THP, swap: delay splitting THP during swap out Patch series "THP swap: Delay splitting THP during swapping out", v11. This patchset is to optimize the performance of Transparent Huge Page (THP) swap. Recently, the performance of the storage devices improved so fast that we cannot saturate the disk bandwidth with single logical CPU when do page swap out even on a high-end server machine. Because the performance of the storage device improved faster than that of single logical CPU. And it seems that the trend will not change in the near future. On the other hand, the THP becomes more and more popular because of increased memory size. So it becomes necessary to optimize THP swap performance. The advantages of the THP swap support include: - Batch the swap operations for the THP to reduce lock acquiring/releasing, including allocating/freeing the swap space, adding/deleting to/from the swap cache, and writing/reading the swap space, etc. This will help improve the performance of the THP swap. - The THP swap space read/write will be 2M sequential IO. It is particularly helpful for the swap read, which are usually 4k random IO. This will improve the performance of the THP swap too. - It will help the memory fragmentation, especially when the THP is heavily used by the applications. The 2M continuous pages will be free up after THP swapping out. - It will improve the THP utilization on the system with the swap turned on. Because the speed for khugepaged to collapse the normal pages into the THP is quite slow. After the THP is split during the swapping out, it will take quite long time for the normal pages to collapse back into the THP after being swapped in. The high THP utilization helps the efficiency of the page based memory management too. There are some concerns regarding THP swap in, mainly because possible enlarged read/write IO size (for swap in/out) may put more overhead on the storage device. To deal with that, the THP swap in should be turned on only when necessary. For example, it can be selected via "always/never/madvise" logic, to be turned on globally, turned off globally, or turned on only for VMA with MADV_HUGEPAGE, etc. This patchset is the first step for the THP swap support. The plan is to delay splitting THP step by step, finally avoid splitting THP during the THP swapping out and swap out/in the THP as a whole. As the first step, in this patchset, the splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP and adding the THP into the swap cache. This will reduce lock acquiring/releasing for the locks used for the swap cache management. With the patchset, the swap out throughput improves 15.5% (from about 3.73GB/s to about 4.31GB/s) in the vm-scalability swap-w-seq test case with 8 processes. The test is done on a Xeon E5 v3 system. The swap device used is a RAM simulated PMEM (persistent memory) device. To test the sequential swapping out, the test case creates 8 processes, which sequentially allocate and write to the anonymous pages until the RAM and part of the swap device is used up. This patch (of 5): In this patch, splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP (Transparent Huge Page) and adding the THP into the swap cache. This will batch the corresponding operation, thus improve THP swap out throughput. This is the first step for the THP swap optimization. The plan is to delay splitting the THP step by step and avoid splitting the THP finally. In this patch, one swap cluster is used to hold the contents of each THP swapped out. So, the size of the swap cluster is changed to that of the THP (Transparent Huge Page) on x86_64 architecture (512). For other architectures which want such THP swap optimization, ARCH_USES_THP_SWAP_CLUSTER needs to be selected in the Kconfig file for the architecture. In effect, this will enlarge swap cluster size by 2 times on x86_64. Which may make it harder to find a free cluster when the swap space becomes fragmented. So that, this may reduce the continuous swap space allocation and sequential write in theory. The performance test in 0day shows no regressions caused by this. In the future of THP swap optimization, some information of the swapped out THP (such as compound map count) will be recorded in the swap_cluster_info data structure. The mem cgroup swap accounting functions are enhanced to support charge or uncharge a swap cluster backing a THP as a whole. The swap cluster allocate/free functions are added to allocate/free a swap cluster for a THP. A fair simple algorithm is used for swap cluster allocation, that is, only the first swap device in priority list will be tried to allocate the swap cluster. The function will fail if the trying is not successful, and the caller will fallback to allocate a single swap slot instead. This works good enough for normal cases. If the difference of the number of the free swap clusters among multiple swap devices is significant, it is possible that some THPs are split earlier than necessary. For example, this could be caused by big size difference among multiple swap devices. The swap cache functions is enhanced to support add/delete THP to/from the swap cache as a set of (HPAGE_PMD_NR) sub-pages. This may be enhanced in the future with multi-order radix tree. But because we will split the THP soon during swapping out, that optimization doesn't make much sense for this first step. The THP splitting functions are enhanced to support to split THP in swap cache during swapping out. The page lock will be held during allocating the swap cluster, adding the THP into the swap cache and splitting the THP. So in the code path other than swapping out, if the THP need to be split, the PageSwapCache(THP) will be always false. The swap cluster is only available for SSD, so the THP swap optimization in this patchset has no effect for HDD. [ying.huang@intel.com: fix two issues in THP optimize patch] Link: http://lkml.kernel.org/r/87k25ed8zo.fsf@yhuang-dev.intel.com [hannes@cmpxchg.org: extensive cleanups and simplifications, reduce code size] Link: http://lkml.kernel.org/r/20170515112522.32457-2-ying.huang@intel.com Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Suggested-by: Andrew Morton <akpm@linux-foundation.org> [for config option] Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> [for changes in huge_memory.c and huge_mm.h] Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Ebru Akagunduz <ebru.akagunduz@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Cc: Hugh Dickins <hughd@google.com> Cc: Shaohua Li <shli@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-07-06 22:37:18 +00:00
page_counter_uncharge(&memcg->swap, nr_pages);
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
else
mm, THP, swap: delay splitting THP during swap out Patch series "THP swap: Delay splitting THP during swapping out", v11. This patchset is to optimize the performance of Transparent Huge Page (THP) swap. Recently, the performance of the storage devices improved so fast that we cannot saturate the disk bandwidth with single logical CPU when do page swap out even on a high-end server machine. Because the performance of the storage device improved faster than that of single logical CPU. And it seems that the trend will not change in the near future. On the other hand, the THP becomes more and more popular because of increased memory size. So it becomes necessary to optimize THP swap performance. The advantages of the THP swap support include: - Batch the swap operations for the THP to reduce lock acquiring/releasing, including allocating/freeing the swap space, adding/deleting to/from the swap cache, and writing/reading the swap space, etc. This will help improve the performance of the THP swap. - The THP swap space read/write will be 2M sequential IO. It is particularly helpful for the swap read, which are usually 4k random IO. This will improve the performance of the THP swap too. - It will help the memory fragmentation, especially when the THP is heavily used by the applications. The 2M continuous pages will be free up after THP swapping out. - It will improve the THP utilization on the system with the swap turned on. Because the speed for khugepaged to collapse the normal pages into the THP is quite slow. After the THP is split during the swapping out, it will take quite long time for the normal pages to collapse back into the THP after being swapped in. The high THP utilization helps the efficiency of the page based memory management too. There are some concerns regarding THP swap in, mainly because possible enlarged read/write IO size (for swap in/out) may put more overhead on the storage device. To deal with that, the THP swap in should be turned on only when necessary. For example, it can be selected via "always/never/madvise" logic, to be turned on globally, turned off globally, or turned on only for VMA with MADV_HUGEPAGE, etc. This patchset is the first step for the THP swap support. The plan is to delay splitting THP step by step, finally avoid splitting THP during the THP swapping out and swap out/in the THP as a whole. As the first step, in this patchset, the splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP and adding the THP into the swap cache. This will reduce lock acquiring/releasing for the locks used for the swap cache management. With the patchset, the swap out throughput improves 15.5% (from about 3.73GB/s to about 4.31GB/s) in the vm-scalability swap-w-seq test case with 8 processes. The test is done on a Xeon E5 v3 system. The swap device used is a RAM simulated PMEM (persistent memory) device. To test the sequential swapping out, the test case creates 8 processes, which sequentially allocate and write to the anonymous pages until the RAM and part of the swap device is used up. This patch (of 5): In this patch, splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP (Transparent Huge Page) and adding the THP into the swap cache. This will batch the corresponding operation, thus improve THP swap out throughput. This is the first step for the THP swap optimization. The plan is to delay splitting the THP step by step and avoid splitting the THP finally. In this patch, one swap cluster is used to hold the contents of each THP swapped out. So, the size of the swap cluster is changed to that of the THP (Transparent Huge Page) on x86_64 architecture (512). For other architectures which want such THP swap optimization, ARCH_USES_THP_SWAP_CLUSTER needs to be selected in the Kconfig file for the architecture. In effect, this will enlarge swap cluster size by 2 times on x86_64. Which may make it harder to find a free cluster when the swap space becomes fragmented. So that, this may reduce the continuous swap space allocation and sequential write in theory. The performance test in 0day shows no regressions caused by this. In the future of THP swap optimization, some information of the swapped out THP (such as compound map count) will be recorded in the swap_cluster_info data structure. The mem cgroup swap accounting functions are enhanced to support charge or uncharge a swap cluster backing a THP as a whole. The swap cluster allocate/free functions are added to allocate/free a swap cluster for a THP. A fair simple algorithm is used for swap cluster allocation, that is, only the first swap device in priority list will be tried to allocate the swap cluster. The function will fail if the trying is not successful, and the caller will fallback to allocate a single swap slot instead. This works good enough for normal cases. If the difference of the number of the free swap clusters among multiple swap devices is significant, it is possible that some THPs are split earlier than necessary. For example, this could be caused by big size difference among multiple swap devices. The swap cache functions is enhanced to support add/delete THP to/from the swap cache as a set of (HPAGE_PMD_NR) sub-pages. This may be enhanced in the future with multi-order radix tree. But because we will split the THP soon during swapping out, that optimization doesn't make much sense for this first step. The THP splitting functions are enhanced to support to split THP in swap cache during swapping out. The page lock will be held during allocating the swap cluster, adding the THP into the swap cache and splitting the THP. So in the code path other than swapping out, if the THP need to be split, the PageSwapCache(THP) will be always false. The swap cluster is only available for SSD, so the THP swap optimization in this patchset has no effect for HDD. [ying.huang@intel.com: fix two issues in THP optimize patch] Link: http://lkml.kernel.org/r/87k25ed8zo.fsf@yhuang-dev.intel.com [hannes@cmpxchg.org: extensive cleanups and simplifications, reduce code size] Link: http://lkml.kernel.org/r/20170515112522.32457-2-ying.huang@intel.com Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Suggested-by: Andrew Morton <akpm@linux-foundation.org> [for config option] Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> [for changes in huge_memory.c and huge_mm.h] Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Ebru Akagunduz <ebru.akagunduz@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Cc: Hugh Dickins <hughd@google.com> Cc: Shaohua Li <shli@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-07-06 22:37:18 +00:00
page_counter_uncharge(&memcg->memsw, nr_pages);
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
}
mod_memcg_state(memcg, MEMCG_SWAP, -nr_pages);
mm, THP, swap: delay splitting THP during swap out Patch series "THP swap: Delay splitting THP during swapping out", v11. This patchset is to optimize the performance of Transparent Huge Page (THP) swap. Recently, the performance of the storage devices improved so fast that we cannot saturate the disk bandwidth with single logical CPU when do page swap out even on a high-end server machine. Because the performance of the storage device improved faster than that of single logical CPU. And it seems that the trend will not change in the near future. On the other hand, the THP becomes more and more popular because of increased memory size. So it becomes necessary to optimize THP swap performance. The advantages of the THP swap support include: - Batch the swap operations for the THP to reduce lock acquiring/releasing, including allocating/freeing the swap space, adding/deleting to/from the swap cache, and writing/reading the swap space, etc. This will help improve the performance of the THP swap. - The THP swap space read/write will be 2M sequential IO. It is particularly helpful for the swap read, which are usually 4k random IO. This will improve the performance of the THP swap too. - It will help the memory fragmentation, especially when the THP is heavily used by the applications. The 2M continuous pages will be free up after THP swapping out. - It will improve the THP utilization on the system with the swap turned on. Because the speed for khugepaged to collapse the normal pages into the THP is quite slow. After the THP is split during the swapping out, it will take quite long time for the normal pages to collapse back into the THP after being swapped in. The high THP utilization helps the efficiency of the page based memory management too. There are some concerns regarding THP swap in, mainly because possible enlarged read/write IO size (for swap in/out) may put more overhead on the storage device. To deal with that, the THP swap in should be turned on only when necessary. For example, it can be selected via "always/never/madvise" logic, to be turned on globally, turned off globally, or turned on only for VMA with MADV_HUGEPAGE, etc. This patchset is the first step for the THP swap support. The plan is to delay splitting THP step by step, finally avoid splitting THP during the THP swapping out and swap out/in the THP as a whole. As the first step, in this patchset, the splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP and adding the THP into the swap cache. This will reduce lock acquiring/releasing for the locks used for the swap cache management. With the patchset, the swap out throughput improves 15.5% (from about 3.73GB/s to about 4.31GB/s) in the vm-scalability swap-w-seq test case with 8 processes. The test is done on a Xeon E5 v3 system. The swap device used is a RAM simulated PMEM (persistent memory) device. To test the sequential swapping out, the test case creates 8 processes, which sequentially allocate and write to the anonymous pages until the RAM and part of the swap device is used up. This patch (of 5): In this patch, splitting huge page is delayed from almost the first step of swapping out to after allocating the swap space for the THP (Transparent Huge Page) and adding the THP into the swap cache. This will batch the corresponding operation, thus improve THP swap out throughput. This is the first step for the THP swap optimization. The plan is to delay splitting the THP step by step and avoid splitting the THP finally. In this patch, one swap cluster is used to hold the contents of each THP swapped out. So, the size of the swap cluster is changed to that of the THP (Transparent Huge Page) on x86_64 architecture (512). For other architectures which want such THP swap optimization, ARCH_USES_THP_SWAP_CLUSTER needs to be selected in the Kconfig file for the architecture. In effect, this will enlarge swap cluster size by 2 times on x86_64. Which may make it harder to find a free cluster when the swap space becomes fragmented. So that, this may reduce the continuous swap space allocation and sequential write in theory. The performance test in 0day shows no regressions caused by this. In the future of THP swap optimization, some information of the swapped out THP (such as compound map count) will be recorded in the swap_cluster_info data structure. The mem cgroup swap accounting functions are enhanced to support charge or uncharge a swap cluster backing a THP as a whole. The swap cluster allocate/free functions are added to allocate/free a swap cluster for a THP. A fair simple algorithm is used for swap cluster allocation, that is, only the first swap device in priority list will be tried to allocate the swap cluster. The function will fail if the trying is not successful, and the caller will fallback to allocate a single swap slot instead. This works good enough for normal cases. If the difference of the number of the free swap clusters among multiple swap devices is significant, it is possible that some THPs are split earlier than necessary. For example, this could be caused by big size difference among multiple swap devices. The swap cache functions is enhanced to support add/delete THP to/from the swap cache as a set of (HPAGE_PMD_NR) sub-pages. This may be enhanced in the future with multi-order radix tree. But because we will split the THP soon during swapping out, that optimization doesn't make much sense for this first step. The THP splitting functions are enhanced to support to split THP in swap cache during swapping out. The page lock will be held during allocating the swap cluster, adding the THP into the swap cache and splitting the THP. So in the code path other than swapping out, if the THP need to be split, the PageSwapCache(THP) will be always false. The swap cluster is only available for SSD, so the THP swap optimization in this patchset has no effect for HDD. [ying.huang@intel.com: fix two issues in THP optimize patch] Link: http://lkml.kernel.org/r/87k25ed8zo.fsf@yhuang-dev.intel.com [hannes@cmpxchg.org: extensive cleanups and simplifications, reduce code size] Link: http://lkml.kernel.org/r/20170515112522.32457-2-ying.huang@intel.com Signed-off-by: "Huang, Ying" <ying.huang@intel.com> Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Suggested-by: Andrew Morton <akpm@linux-foundation.org> [for config option] Acked-by: Kirill A. Shutemov <kirill.shutemov@linux.intel.com> [for changes in huge_memory.c and huge_mm.h] Cc: Andrea Arcangeli <aarcange@redhat.com> Cc: Ebru Akagunduz <ebru.akagunduz@gmail.com> Cc: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Cc: Hugh Dickins <hughd@google.com> Cc: Shaohua Li <shli@kernel.org> Cc: Minchan Kim <minchan@kernel.org> Cc: Rik van Riel <riel@redhat.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2017-07-06 22:37:18 +00:00
mem_cgroup_id_put_many(memcg, nr_pages);
}
rcu_read_unlock();
}
long mem_cgroup_get_nr_swap_pages(struct mem_cgroup *memcg)
{
long nr_swap_pages = get_nr_swap_pages();
if (cgroup_memory_noswap || !cgroup_subsys_on_dfl(memory_cgrp_subsys))
return nr_swap_pages;
for (; memcg != root_mem_cgroup; memcg = parent_mem_cgroup(memcg))
nr_swap_pages = min_t(long, nr_swap_pages,
READ_ONCE(memcg->swap.max) -
page_counter_read(&memcg->swap));
return nr_swap_pages;
}
bool mem_cgroup_swap_full(struct page *page)
{
struct mem_cgroup *memcg;
VM_BUG_ON_PAGE(!PageLocked(page), page);
if (vm_swap_full())
return true;
if (cgroup_memory_noswap || !cgroup_subsys_on_dfl(memory_cgrp_subsys))
return false;
memcg = page->mem_cgroup;
if (!memcg)
return false;
mm/memcg: automatically penalize tasks with high swap use Add a memory.swap.high knob, which can be used to protect the system from SWAP exhaustion. The mechanism used for penalizing is similar to memory.high penalty (sleep on return to user space). That is not to say that the knob itself is equivalent to memory.high. The objective is more to protect the system from potentially buggy tasks consuming a lot of swap and impacting other tasks, or even bringing the whole system to stand still with complete SWAP exhaustion. Hopefully without the need to find per-task hard limits. Slowing misbehaving tasks down gradually allows user space oom killers or other protection mechanisms to react. oomd and earlyoom already do killing based on swap exhaustion, and memory.swap.high protection will help implement such userspace oom policies more reliably. We can use one counter for number of pages allocated under pressure to save struct task space and avoid two separate hierarchy walks on the hot path. The exact overage is calculated on return to user space, anyway. Take the new high limit into account when determining if swap is "full". Borrowing the explanation from Johannes: The idea behind "swap full" is that as long as the workload has plenty of swap space available and it's not changing its memory contents, it makes sense to generously hold on to copies of data in the swap device, even after the swapin. A later reclaim cycle can drop the page without any IO. Trading disk space for IO. But the only two ways to reclaim a swap slot is when they're faulted in and the references go away, or by scanning the virtual address space like swapoff does - which is very expensive (one could argue it's too expensive even for swapoff, it's often more practical to just reboot). So at some point in the fill level, we have to start freeing up swap slots on fault/swapin. Otherwise we could eventually run out of swap slots while they're filled with copies of data that is also in RAM. We don't want to OOM a workload because its available swap space is filled with redundant cache. Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Chris Down <chris@chrisdown.name> Cc: Shakeel Butt <shakeelb@google.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Hugh Dickins <hughd@google.com> Link: http://lkml.kernel.org/r/20200527195846.102707-5-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:52 +00:00
for (; memcg != root_mem_cgroup; memcg = parent_mem_cgroup(memcg)) {
unsigned long usage = page_counter_read(&memcg->swap);
if (usage * 2 >= READ_ONCE(memcg->swap.high) ||
usage * 2 >= READ_ONCE(memcg->swap.max))
return true;
mm/memcg: automatically penalize tasks with high swap use Add a memory.swap.high knob, which can be used to protect the system from SWAP exhaustion. The mechanism used for penalizing is similar to memory.high penalty (sleep on return to user space). That is not to say that the knob itself is equivalent to memory.high. The objective is more to protect the system from potentially buggy tasks consuming a lot of swap and impacting other tasks, or even bringing the whole system to stand still with complete SWAP exhaustion. Hopefully without the need to find per-task hard limits. Slowing misbehaving tasks down gradually allows user space oom killers or other protection mechanisms to react. oomd and earlyoom already do killing based on swap exhaustion, and memory.swap.high protection will help implement such userspace oom policies more reliably. We can use one counter for number of pages allocated under pressure to save struct task space and avoid two separate hierarchy walks on the hot path. The exact overage is calculated on return to user space, anyway. Take the new high limit into account when determining if swap is "full". Borrowing the explanation from Johannes: The idea behind "swap full" is that as long as the workload has plenty of swap space available and it's not changing its memory contents, it makes sense to generously hold on to copies of data in the swap device, even after the swapin. A later reclaim cycle can drop the page without any IO. Trading disk space for IO. But the only two ways to reclaim a swap slot is when they're faulted in and the references go away, or by scanning the virtual address space like swapoff does - which is very expensive (one could argue it's too expensive even for swapoff, it's often more practical to just reboot). So at some point in the fill level, we have to start freeing up swap slots on fault/swapin. Otherwise we could eventually run out of swap slots while they're filled with copies of data that is also in RAM. We don't want to OOM a workload because its available swap space is filled with redundant cache. Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Chris Down <chris@chrisdown.name> Cc: Shakeel Butt <shakeelb@google.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Hugh Dickins <hughd@google.com> Link: http://lkml.kernel.org/r/20200527195846.102707-5-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:52 +00:00
}
return false;
}
static int __init setup_swap_account(char *s)
{
if (!strcmp(s, "1"))
cgroup_memory_noswap = 0;
else if (!strcmp(s, "0"))
cgroup_memory_noswap = 1;
return 1;
}
__setup("swapaccount=", setup_swap_account);
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
static u64 swap_current_read(struct cgroup_subsys_state *css,
struct cftype *cft)
{
struct mem_cgroup *memcg = mem_cgroup_from_css(css);
return (u64)page_counter_read(&memcg->swap) * PAGE_SIZE;
}
mm/memcg: automatically penalize tasks with high swap use Add a memory.swap.high knob, which can be used to protect the system from SWAP exhaustion. The mechanism used for penalizing is similar to memory.high penalty (sleep on return to user space). That is not to say that the knob itself is equivalent to memory.high. The objective is more to protect the system from potentially buggy tasks consuming a lot of swap and impacting other tasks, or even bringing the whole system to stand still with complete SWAP exhaustion. Hopefully without the need to find per-task hard limits. Slowing misbehaving tasks down gradually allows user space oom killers or other protection mechanisms to react. oomd and earlyoom already do killing based on swap exhaustion, and memory.swap.high protection will help implement such userspace oom policies more reliably. We can use one counter for number of pages allocated under pressure to save struct task space and avoid two separate hierarchy walks on the hot path. The exact overage is calculated on return to user space, anyway. Take the new high limit into account when determining if swap is "full". Borrowing the explanation from Johannes: The idea behind "swap full" is that as long as the workload has plenty of swap space available and it's not changing its memory contents, it makes sense to generously hold on to copies of data in the swap device, even after the swapin. A later reclaim cycle can drop the page without any IO. Trading disk space for IO. But the only two ways to reclaim a swap slot is when they're faulted in and the references go away, or by scanning the virtual address space like swapoff does - which is very expensive (one could argue it's too expensive even for swapoff, it's often more practical to just reboot). So at some point in the fill level, we have to start freeing up swap slots on fault/swapin. Otherwise we could eventually run out of swap slots while they're filled with copies of data that is also in RAM. We don't want to OOM a workload because its available swap space is filled with redundant cache. Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Chris Down <chris@chrisdown.name> Cc: Shakeel Butt <shakeelb@google.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Hugh Dickins <hughd@google.com> Link: http://lkml.kernel.org/r/20200527195846.102707-5-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:52 +00:00
static int swap_high_show(struct seq_file *m, void *v)
{
return seq_puts_memcg_tunable(m,
READ_ONCE(mem_cgroup_from_seq(m)->swap.high));
}
static ssize_t swap_high_write(struct kernfs_open_file *of,
char *buf, size_t nbytes, loff_t off)
{
struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of));
unsigned long high;
int err;
buf = strstrip(buf);
err = page_counter_memparse(buf, "max", &high);
if (err)
return err;
page_counter_set_high(&memcg->swap, high);
return nbytes;
}
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
static int swap_max_show(struct seq_file *m, void *v)
{
return seq_puts_memcg_tunable(m,
READ_ONCE(mem_cgroup_from_seq(m)->swap.max));
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
}
static ssize_t swap_max_write(struct kernfs_open_file *of,
char *buf, size_t nbytes, loff_t off)
{
struct mem_cgroup *memcg = mem_cgroup_from_css(of_css(of));
unsigned long max;
int err;
buf = strstrip(buf);
err = page_counter_memparse(buf, "max", &max);
if (err)
return err;
mm: memcg: allow lowering memory.swap.max below the current usage Currently an attempt to set swap.max into a value lower than the actual swap usage fails, which causes configuration problems as there's no way of lowering the configuration below the current usage short of turning off swap entirely. This makes swap.max difficult to use and allows delegatees to lock the delegator out of reducing swap allocation. This patch updates swap_max_write() so that the limit can be lowered below the current usage. It doesn't implement active reclaiming of swap entries for the following reasons. * mem_cgroup_swap_full() already tells the swap machinary to aggressively reclaim swap entries if the usage is above 50% of limit, so simply lowering the limit automatically triggers gradual reclaim. * Forcing back swapped out pages is likely to heavily impact the workload and mess up the working set. Given that swap usually is a lot less valuable and less scarce, letting the existing usage dissipate over time through the above gradual reclaim and as they're falted back in is likely the better behavior. Link: http://lkml.kernel.org/r/20180523185041.GR1718769@devbig577.frc2.facebook.com Signed-off-by: Tejun Heo <tj@kernel.org> Acked-by: Roman Gushchin <guro@fb.com> Acked-by: Rik van Riel <riel@surriel.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Shaohua Li <shli@fb.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2018-06-08 00:09:21 +00:00
xchg(&memcg->swap.max, max);
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
return nbytes;
}
static int swap_events_show(struct seq_file *m, void *v)
{
struct mem_cgroup *memcg = mem_cgroup_from_seq(m);
mm/memcg: automatically penalize tasks with high swap use Add a memory.swap.high knob, which can be used to protect the system from SWAP exhaustion. The mechanism used for penalizing is similar to memory.high penalty (sleep on return to user space). That is not to say that the knob itself is equivalent to memory.high. The objective is more to protect the system from potentially buggy tasks consuming a lot of swap and impacting other tasks, or even bringing the whole system to stand still with complete SWAP exhaustion. Hopefully without the need to find per-task hard limits. Slowing misbehaving tasks down gradually allows user space oom killers or other protection mechanisms to react. oomd and earlyoom already do killing based on swap exhaustion, and memory.swap.high protection will help implement such userspace oom policies more reliably. We can use one counter for number of pages allocated under pressure to save struct task space and avoid two separate hierarchy walks on the hot path. The exact overage is calculated on return to user space, anyway. Take the new high limit into account when determining if swap is "full". Borrowing the explanation from Johannes: The idea behind "swap full" is that as long as the workload has plenty of swap space available and it's not changing its memory contents, it makes sense to generously hold on to copies of data in the swap device, even after the swapin. A later reclaim cycle can drop the page without any IO. Trading disk space for IO. But the only two ways to reclaim a swap slot is when they're faulted in and the references go away, or by scanning the virtual address space like swapoff does - which is very expensive (one could argue it's too expensive even for swapoff, it's often more practical to just reboot). So at some point in the fill level, we have to start freeing up swap slots on fault/swapin. Otherwise we could eventually run out of swap slots while they're filled with copies of data that is also in RAM. We don't want to OOM a workload because its available swap space is filled with redundant cache. Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Chris Down <chris@chrisdown.name> Cc: Shakeel Butt <shakeelb@google.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Hugh Dickins <hughd@google.com> Link: http://lkml.kernel.org/r/20200527195846.102707-5-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:52 +00:00
seq_printf(m, "high %lu\n",
atomic_long_read(&memcg->memory_events[MEMCG_SWAP_HIGH]));
seq_printf(m, "max %lu\n",
atomic_long_read(&memcg->memory_events[MEMCG_SWAP_MAX]));
seq_printf(m, "fail %lu\n",
atomic_long_read(&memcg->memory_events[MEMCG_SWAP_FAIL]));
return 0;
}
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
static struct cftype swap_files[] = {
{
.name = "swap.current",
.flags = CFTYPE_NOT_ON_ROOT,
.read_u64 = swap_current_read,
},
mm/memcg: automatically penalize tasks with high swap use Add a memory.swap.high knob, which can be used to protect the system from SWAP exhaustion. The mechanism used for penalizing is similar to memory.high penalty (sleep on return to user space). That is not to say that the knob itself is equivalent to memory.high. The objective is more to protect the system from potentially buggy tasks consuming a lot of swap and impacting other tasks, or even bringing the whole system to stand still with complete SWAP exhaustion. Hopefully without the need to find per-task hard limits. Slowing misbehaving tasks down gradually allows user space oom killers or other protection mechanisms to react. oomd and earlyoom already do killing based on swap exhaustion, and memory.swap.high protection will help implement such userspace oom policies more reliably. We can use one counter for number of pages allocated under pressure to save struct task space and avoid two separate hierarchy walks on the hot path. The exact overage is calculated on return to user space, anyway. Take the new high limit into account when determining if swap is "full". Borrowing the explanation from Johannes: The idea behind "swap full" is that as long as the workload has plenty of swap space available and it's not changing its memory contents, it makes sense to generously hold on to copies of data in the swap device, even after the swapin. A later reclaim cycle can drop the page without any IO. Trading disk space for IO. But the only two ways to reclaim a swap slot is when they're faulted in and the references go away, or by scanning the virtual address space like swapoff does - which is very expensive (one could argue it's too expensive even for swapoff, it's often more practical to just reboot). So at some point in the fill level, we have to start freeing up swap slots on fault/swapin. Otherwise we could eventually run out of swap slots while they're filled with copies of data that is also in RAM. We don't want to OOM a workload because its available swap space is filled with redundant cache. Signed-off-by: Jakub Kicinski <kuba@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Tejun Heo <tj@kernel.org> Cc: Chris Down <chris@chrisdown.name> Cc: Shakeel Butt <shakeelb@google.com> Cc: Michal Hocko <mhocko@kernel.org> Cc: Hugh Dickins <hughd@google.com> Link: http://lkml.kernel.org/r/20200527195846.102707-5-kuba@kernel.org Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-02 04:49:52 +00:00
{
.name = "swap.high",
.flags = CFTYPE_NOT_ON_ROOT,
.seq_show = swap_high_show,
.write = swap_high_write,
},
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
{
.name = "swap.max",
.flags = CFTYPE_NOT_ON_ROOT,
.seq_show = swap_max_show,
.write = swap_max_write,
},
{
.name = "swap.events",
.flags = CFTYPE_NOT_ON_ROOT,
.file_offset = offsetof(struct mem_cgroup, swap_events_file),
.seq_show = swap_events_show,
},
mm: memcontrol: charge swap to cgroup2 This patchset introduces swap accounting to cgroup2. This patch (of 7): In the legacy hierarchy we charge memsw, which is dubious, because: - memsw.limit must be >= memory.limit, so it is impossible to limit swap usage less than memory usage. Taking into account the fact that the primary limiting mechanism in the unified hierarchy is memory.high while memory.limit is either left unset or set to a very large value, moving memsw.limit knob to the unified hierarchy would effectively make it impossible to limit swap usage according to the user preference. - memsw.usage != memory.usage + swap.usage, because a page occupying both swap entry and a swap cache page is charged only once to memsw counter. As a result, it is possible to effectively eat up to memory.limit of memory pages *and* memsw.limit of swap entries, which looks unexpected. That said, we should provide a different swap limiting mechanism for cgroup2. This patch adds mem_cgroup->swap counter, which charges the actual number of swap entries used by a cgroup. It is only charged in the unified hierarchy, while the legacy hierarchy memsw logic is left intact. The swap usage can be monitored using new memory.swap.current file and limited using memory.swap.max. Note, to charge swap resource properly in the unified hierarchy, we have to make swap_entry_free uncharge swap only when ->usage reaches zero, not just ->count, i.e. when all references to a swap entry, including the one taken by swap cache, are gone. This is necessary, because otherwise swap-in could result in uncharging swap even if the page is still in swap cache and hence still occupies a swap entry. At the same time, this shouldn't break memsw counter logic, where a page is never charged twice for using both memory and swap, because in case of legacy hierarchy we uncharge swap on commit (see mem_cgroup_commit_charge). Signed-off-by: Vladimir Davydov <vdavydov@virtuozzo.com> Acked-by: Johannes Weiner <hannes@cmpxchg.org> Cc: Michal Hocko <mhocko@kernel.org> Cc: Tejun Heo <tj@kernel.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2016-01-20 23:02:56 +00:00
{ } /* terminate */
};
static struct cftype memsw_files[] = {
{
.name = "memsw.usage_in_bytes",
.private = MEMFILE_PRIVATE(_MEMSWAP, RES_USAGE),
.read_u64 = mem_cgroup_read_u64,
},
{
.name = "memsw.max_usage_in_bytes",
.private = MEMFILE_PRIVATE(_MEMSWAP, RES_MAX_USAGE),
.write = mem_cgroup_reset,
.read_u64 = mem_cgroup_read_u64,
},
{
.name = "memsw.limit_in_bytes",
.private = MEMFILE_PRIVATE(_MEMSWAP, RES_LIMIT),
.write = mem_cgroup_write,
.read_u64 = mem_cgroup_read_u64,
},
{
.name = "memsw.failcnt",
.private = MEMFILE_PRIVATE(_MEMSWAP, RES_FAILCNT),
.write = mem_cgroup_reset,
.read_u64 = mem_cgroup_read_u64,
},
{ }, /* terminate */
};
static int __init mem_cgroup_swap_init(void)
{
mm: memcontrol: make swap tracking an integral part of memory control Without swap page tracking, users that are otherwise memory controlled can easily escape their containment and allocate significant amounts of memory that they're not being charged for. That's because swap does readahead, but without the cgroup records of who owned the page at swapout, readahead pages don't get charged until somebody actually faults them into their page table and we can identify an owner task. This can be maliciously exploited with MADV_WILLNEED, which triggers arbitrary readahead allocations without charging the pages. Make swap swap page tracking an integral part of memcg and remove the Kconfig options. In the first place, it was only made configurable to allow users to save some memory. But the overhead of tracking cgroup ownership per swap page is minimal - 2 byte per page, or 512k per 1G of swap, or 0.04%. Saving that at the expense of broken containment semantics is not something we should present as a coequal option. The swapaccount=0 boot option will continue to exist, and it will eliminate the page_counter overhead and hide the swap control files, but it won't disable swap slot ownership tracking. This patch makes sure we always have the cgroup records at swapin time; the next patch will fix the actual bug by charging readahead swap pages at swapin time rather than at fault time. v2: fix double swap charge bug in cgroup1/cgroup2 code gating [hannes@cmpxchg.org: fix crash with cgroup_disable=memory] Link: http://lkml.kernel.org/r/20200521215855.GB815153@cmpxchg.org Signed-off-by: Johannes Weiner <hannes@cmpxchg.org> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Reviewed-by: Joonsoo Kim <iamjoonsoo.kim@lge.com> Cc: Alex Shi <alex.shi@linux.alibaba.com> Cc: "Kirill A. Shutemov" <kirill@shutemov.name> Cc: Roman Gushchin <guro@fb.com> Cc: Shakeel Butt <shakeelb@google.com> Cc: Balbir Singh <bsingharora@gmail.com> Cc: Naresh Kamboju <naresh.kamboju@linaro.org> Link: http://lkml.kernel.org/r/20200508183105.225460-16-hannes@cmpxchg.org Debugged-by: Hugh Dickins <hughd@google.com> Debugged-by: Michal Hocko <mhocko@kernel.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2020-06-03 23:02:14 +00:00
/* No memory control -> no swap control */
if (mem_cgroup_disabled())
cgroup_memory_noswap = true;
if (cgroup_memory_noswap)
return 0;
WARN_ON(cgroup_add_dfl_cftypes(&memory_cgrp_subsys, swap_files));
WARN_ON(cgroup_add_legacy_cftypes(&memory_cgrp_subsys, memsw_files));
return 0;
}
subsys_initcall(mem_cgroup_swap_init);
#endif /* CONFIG_MEMCG_SWAP */