linux/fs/nfs/write.c

2215 lines
58 KiB
C
Raw Normal View History

// SPDX-License-Identifier: GPL-2.0-only
/*
* linux/fs/nfs/write.c
*
* Write file data over NFS.
*
* Copyright (C) 1996, 1997, Olaf Kirch <okir@monad.swb.de>
*/
#include <linux/types.h>
#include <linux/slab.h>
#include <linux/mm.h>
#include <linux/pagemap.h>
#include <linux/file.h>
#include <linux/writeback.h>
#include <linux/swap.h>
#include <linux/migrate.h>
#include <linux/sunrpc/clnt.h>
#include <linux/nfs_fs.h>
#include <linux/nfs_mount.h>
#include <linux/nfs_page.h>
#include <linux/backing-dev.h>
#include <linux/export.h>
#include <linux/freezer.h>
#include <linux/wait.h>
#include <linux/iversion.h>
#include <linux/filelock.h>
#include <linux/uaccess.h>
#include <linux/sched/mm.h>
#include "delegation.h"
#include "internal.h"
#include "iostat.h"
#include "nfs4_fs.h"
#include "fscache.h"
#include "pnfs.h"
#include "nfstrace.h"
#define NFSDBG_FACILITY NFSDBG_PAGECACHE
#define MIN_POOL_WRITE (32)
#define MIN_POOL_COMMIT (4)
struct nfs_io_completion {
void (*complete)(void *data);
void *data;
struct kref refcount;
};
/*
* Local function declarations
*/
2008-03-19 15:24:39 +00:00
static void nfs_redirty_request(struct nfs_page *req);
static const struct rpc_call_ops nfs_commit_ops;
static const struct nfs_pgio_completion_ops nfs_async_write_completion_ops;
static const struct nfs_commit_completion_ops nfs_commit_completion_ops;
static const struct nfs_rw_ops nfs_rw_write_ops;
static void nfs_inode_remove_request(struct nfs_page *req);
static void nfs_clear_request_commit(struct nfs_page *req);
static void nfs_init_cinfo_from_inode(struct nfs_commit_info *cinfo,
struct inode *inode);
static struct nfs_page *
nfs_page_search_commits_for_head_request_locked(struct nfs_inode *nfsi,
struct folio *folio);
static struct kmem_cache *nfs_wdata_cachep;
static mempool_t *nfs_wdata_mempool;
static struct kmem_cache *nfs_cdata_cachep;
static mempool_t *nfs_commit_mempool;
struct nfs_commit_data *nfs_commitdata_alloc(void)
{
NFS: fix usage of mempools. When passed GFP flags that allow sleeping (such as GFP_NOIO), mempool_alloc() will never return NULL, it will wait until memory is available. This means that we don't need to handle failure, but that we do need to ensure one thread doesn't call mempool_alloc() twice on the one pool without queuing or freeing the first allocation. If multiple threads did this during times of high memory pressure, the pool could be exhausted and a deadlock could result. pnfs_generic_alloc_ds_commits() attempts to allocate from the nfs_commit_mempool while already holding an allocation from that pool. This is not safe. So change nfs_commitdata_alloc() to take a flag that indicates whether failure is acceptable. In pnfs_generic_alloc_ds_commits(), accept failure and handle it as we currently do. Else where, do not accept failure, and do not handle it. Even when failure is acceptable, we want to succeed if possible. That means both - using an entry from the pool if there is one - waiting for direct reclaim is there isn't. We call mempool_alloc(GFP_NOWAIT) to achieve the first, then kmem_cache_alloc(GFP_NOIO|__GFP_NORETRY) to achieve the second. Each of these can fail, but together they do the best they can without blocking indefinitely. The objects returned by kmem_cache_alloc() will still be freed by mempool_free(). This is safe as mempool_alloc() uses exactly the same function to allocate objects (since the mempool was created with mempool_create_slab_pool()). The object returned by mempool_alloc() and kmem_cache_alloc() are indistinguishable so mempool_free() will handle both identically, either adding to the pool or calling kmem_cache_free(). Also, don't test for failure when allocating from nfs_wdata_mempool. Signed-off-by: NeilBrown <neilb@suse.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-04-10 02:22:09 +00:00
struct nfs_commit_data *p;
p = kmem_cache_zalloc(nfs_cdata_cachep, nfs_io_gfp_mask());
if (!p) {
NFS: fix usage of mempools. When passed GFP flags that allow sleeping (such as GFP_NOIO), mempool_alloc() will never return NULL, it will wait until memory is available. This means that we don't need to handle failure, but that we do need to ensure one thread doesn't call mempool_alloc() twice on the one pool without queuing or freeing the first allocation. If multiple threads did this during times of high memory pressure, the pool could be exhausted and a deadlock could result. pnfs_generic_alloc_ds_commits() attempts to allocate from the nfs_commit_mempool while already holding an allocation from that pool. This is not safe. So change nfs_commitdata_alloc() to take a flag that indicates whether failure is acceptable. In pnfs_generic_alloc_ds_commits(), accept failure and handle it as we currently do. Else where, do not accept failure, and do not handle it. Even when failure is acceptable, we want to succeed if possible. That means both - using an entry from the pool if there is one - waiting for direct reclaim is there isn't. We call mempool_alloc(GFP_NOWAIT) to achieve the first, then kmem_cache_alloc(GFP_NOIO|__GFP_NORETRY) to achieve the second. Each of these can fail, but together they do the best they can without blocking indefinitely. The objects returned by kmem_cache_alloc() will still be freed by mempool_free(). This is safe as mempool_alloc() uses exactly the same function to allocate objects (since the mempool was created with mempool_create_slab_pool()). The object returned by mempool_alloc() and kmem_cache_alloc() are indistinguishable so mempool_free() will handle both identically, either adding to the pool or calling kmem_cache_free(). Also, don't test for failure when allocating from nfs_wdata_mempool. Signed-off-by: NeilBrown <neilb@suse.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-04-10 02:22:09 +00:00
p = mempool_alloc(nfs_commit_mempool, GFP_NOWAIT);
if (!p)
return NULL;
memset(p, 0, sizeof(*p));
}
NFS: fix usage of mempools. When passed GFP flags that allow sleeping (such as GFP_NOIO), mempool_alloc() will never return NULL, it will wait until memory is available. This means that we don't need to handle failure, but that we do need to ensure one thread doesn't call mempool_alloc() twice on the one pool without queuing or freeing the first allocation. If multiple threads did this during times of high memory pressure, the pool could be exhausted and a deadlock could result. pnfs_generic_alloc_ds_commits() attempts to allocate from the nfs_commit_mempool while already holding an allocation from that pool. This is not safe. So change nfs_commitdata_alloc() to take a flag that indicates whether failure is acceptable. In pnfs_generic_alloc_ds_commits(), accept failure and handle it as we currently do. Else where, do not accept failure, and do not handle it. Even when failure is acceptable, we want to succeed if possible. That means both - using an entry from the pool if there is one - waiting for direct reclaim is there isn't. We call mempool_alloc(GFP_NOWAIT) to achieve the first, then kmem_cache_alloc(GFP_NOIO|__GFP_NORETRY) to achieve the second. Each of these can fail, but together they do the best they can without blocking indefinitely. The objects returned by kmem_cache_alloc() will still be freed by mempool_free(). This is safe as mempool_alloc() uses exactly the same function to allocate objects (since the mempool was created with mempool_create_slab_pool()). The object returned by mempool_alloc() and kmem_cache_alloc() are indistinguishable so mempool_free() will handle both identically, either adding to the pool or calling kmem_cache_free(). Also, don't test for failure when allocating from nfs_wdata_mempool. Signed-off-by: NeilBrown <neilb@suse.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-04-10 02:22:09 +00:00
INIT_LIST_HEAD(&p->pages);
return p;
}
EXPORT_SYMBOL_GPL(nfs_commitdata_alloc);
void nfs_commit_free(struct nfs_commit_data *p)
{
mempool_free(p, nfs_commit_mempool);
}
EXPORT_SYMBOL_GPL(nfs_commit_free);
static struct nfs_pgio_header *nfs_writehdr_alloc(void)
{
struct nfs_pgio_header *p;
p = kmem_cache_zalloc(nfs_wdata_cachep, nfs_io_gfp_mask());
if (!p) {
p = mempool_alloc(nfs_wdata_mempool, GFP_NOWAIT);
if (!p)
return NULL;
memset(p, 0, sizeof(*p));
}
p->rw_mode = FMODE_WRITE;
return p;
}
static void nfs_writehdr_free(struct nfs_pgio_header *hdr)
{
mempool_free(hdr, nfs_wdata_mempool);
}
static struct nfs_io_completion *nfs_io_completion_alloc(gfp_t gfp_flags)
{
return kmalloc(sizeof(struct nfs_io_completion), gfp_flags);
}
static void nfs_io_completion_init(struct nfs_io_completion *ioc,
void (*complete)(void *), void *data)
{
ioc->complete = complete;
ioc->data = data;
kref_init(&ioc->refcount);
}
static void nfs_io_completion_release(struct kref *kref)
{
struct nfs_io_completion *ioc = container_of(kref,
struct nfs_io_completion, refcount);
ioc->complete(ioc->data);
kfree(ioc);
}
static void nfs_io_completion_get(struct nfs_io_completion *ioc)
{
if (ioc != NULL)
kref_get(&ioc->refcount);
}
static void nfs_io_completion_put(struct nfs_io_completion *ioc)
{
if (ioc != NULL)
kref_put(&ioc->refcount, nfs_io_completion_release);
}
static void
nfs_page_set_inode_ref(struct nfs_page *req, struct inode *inode)
{
if (!test_and_set_bit(PG_INODE_REF, &req->wb_flags)) {
kref_get(&req->wb_kref);
atomic_long_inc(&NFS_I(inode)->nrequests);
}
}
static int
nfs_cancel_remove_inode(struct nfs_page *req, struct inode *inode)
{
int ret;
if (!test_bit(PG_REMOVE, &req->wb_flags))
return 0;
ret = nfs_page_group_lock(req);
if (ret)
return ret;
if (test_and_clear_bit(PG_REMOVE, &req->wb_flags))
nfs_page_set_inode_ref(req, inode);
nfs_page_group_unlock(req);
return 0;
}
static struct nfs_page *nfs_folio_private_request(struct folio *folio)
{
return folio_get_private(folio);
}
/**
* nfs_folio_find_private_request - find head request associated with a folio
* @folio: pointer to folio
*
* must be called while holding the inode lock.
*
* returns matching head request with reference held, or NULL if not found.
*/
static struct nfs_page *nfs_folio_find_private_request(struct folio *folio)
{
struct address_space *mapping = folio_file_mapping(folio);
struct nfs_page *req;
if (!folio_test_private(folio))
return NULL;
spin_lock(&mapping->private_lock);
req = nfs_folio_private_request(folio);
if (req) {
WARN_ON_ONCE(req->wb_head != req);
kref_get(&req->wb_kref);
}
spin_unlock(&mapping->private_lock);
return req;
}
static struct nfs_page *nfs_folio_find_swap_request(struct folio *folio)
{
struct inode *inode = folio_file_mapping(folio)->host;
struct nfs_inode *nfsi = NFS_I(inode);
struct nfs_page *req = NULL;
if (!folio_test_swapcache(folio))
return NULL;
mutex_lock(&nfsi->commit_mutex);
if (folio_test_swapcache(folio)) {
req = nfs_page_search_commits_for_head_request_locked(nfsi,
folio);
if (req) {
WARN_ON_ONCE(req->wb_head != req);
kref_get(&req->wb_kref);
}
}
mutex_unlock(&nfsi->commit_mutex);
return req;
}
/**
* nfs_folio_find_head_request - find head request associated with a folio
* @folio: pointer to folio
*
* returns matching head request with reference held, or NULL if not found.
*/
static struct nfs_page *nfs_folio_find_head_request(struct folio *folio)
{
struct nfs_page *req;
req = nfs_folio_find_private_request(folio);
if (!req)
req = nfs_folio_find_swap_request(folio);
return req;
}
static struct nfs_page *nfs_folio_find_and_lock_request(struct folio *folio)
{
struct inode *inode = folio_file_mapping(folio)->host;
struct nfs_page *req, *head;
int ret;
for (;;) {
req = nfs_folio_find_head_request(folio);
if (!req)
return req;
head = nfs_page_group_lock_head(req);
if (head != req)
nfs_release_request(req);
if (IS_ERR(head))
return head;
ret = nfs_cancel_remove_inode(head, inode);
if (ret < 0) {
nfs_unlock_and_release_request(head);
return ERR_PTR(ret);
}
/* Ensure that nobody removed the request before we locked it */
if (head == nfs_folio_private_request(folio))
break;
if (folio_test_swapcache(folio))
break;
nfs_unlock_and_release_request(head);
}
return head;
}
/* Adjust the file length if we're writing beyond the end */
static void nfs_grow_file(struct folio *folio, unsigned int offset,
unsigned int count)
{
struct inode *inode = folio_file_mapping(folio)->host;
loff_t end, i_size;
pgoff_t end_index;
spin_lock(&inode->i_lock);
i_size = i_size_read(inode);
end_index = ((i_size - 1) >> folio_shift(folio)) << folio_order(folio);
if (i_size > 0 && folio_index(folio) < end_index)
goto out;
end = folio_file_pos(folio) + (loff_t)offset + (loff_t)count;
if (i_size >= end)
goto out;
trace_nfs_size_grow(inode, end);
i_size_write(inode, end);
NFS_I(inode)->cache_validity &= ~NFS_INO_INVALID_SIZE;
nfs_inc_stats(inode, NFSIOS_EXTENDWRITE);
out:
spin_unlock(&inode->i_lock);
nfs: Convert to new fscache volume/cookie API Change the nfs filesystem to support fscache's indexing rewrite and reenable caching in nfs. The following changes have been made: (1) The fscache_netfs struct is no more, and there's no need to register the filesystem as a whole. (2) The session cookie is now an fscache_volume cookie, allocated with fscache_acquire_volume(). That takes three parameters: a string representing the "volume" in the index, a string naming the cache to use (or NULL) and a u64 that conveys coherency metadata for the volume. For nfs, I've made it render the volume name string as: "nfs,<ver>,<family>,<address>,<port>,<fsidH>,<fsidL>*<,param>[,<uniq>]" (3) The fscache_cookie_def is no more and needed information is passed directly to fscache_acquire_cookie(). The cache no longer calls back into the filesystem, but rather metadata changes are indicated at other times. fscache_acquire_cookie() is passed the same keying and coherency information as before. (4) fscache_enable/disable_cookie() have been removed. Call fscache_use_cookie() and fscache_unuse_cookie() when a file is opened or closed to prevent a cache file from being culled and to keep resources to hand that are needed to do I/O. If a file is opened for writing, we invalidate it with FSCACHE_INVAL_DIO_WRITE in lieu of doing writeback to the cache, thereby making it cease caching until all currently open files are closed. This should give the same behaviour as the uptream code. Making the cache store local modifications isn't straightforward for NFS, so that's left for future patches. (5) fscache_invalidate() now needs to be given uptodate auxiliary data and a file size. It also takes a flag to indicate if this was due to a DIO write. (6) Call nfs_fscache_invalidate() with FSCACHE_INVAL_DIO_WRITE on a file to which a DIO write is made. (7) Call fscache_note_page_release() from nfs_release_page(). (8) Use a killable wait in nfs_vm_page_mkwrite() when waiting for PG_fscache to be cleared. (9) The functions to read and write data to/from the cache are stubbed out pending a conversion to use netfslib. Changes ======= ver #3: - Added missing =n fallback for nfs_fscache_release_file()[1][2]. ver #2: - Use gfpflags_allow_blocking() rather than using flag directly. - fscache_acquire_volume() now returns errors. - Remove NFS_INO_FSCACHE as it's no longer used. - Need to unuse a cookie on file-release, not inode-clear. Signed-off-by: Dave Wysochanski <dwysocha@redhat.com> Co-developed-by: David Howells <dhowells@redhat.com> Signed-off-by: David Howells <dhowells@redhat.com> Tested-by: Dave Wysochanski <dwysocha@redhat.com> Acked-by: Jeff Layton <jlayton@kernel.org> cc: Trond Myklebust <trond.myklebust@hammerspace.com> cc: Anna Schumaker <anna.schumaker@netapp.com> cc: linux-nfs@vger.kernel.org cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/202112100804.nksO8K4u-lkp@intel.com/ [1] Link: https://lore.kernel.org/r/202112100957.2oEDT20W-lkp@intel.com/ [2] Link: https://lore.kernel.org/r/163819668938.215744.14448852181937731615.stgit@warthog.procyon.org.uk/ # v1 Link: https://lore.kernel.org/r/163906979003.143852.2601189243864854724.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967182112.1823006.7791504655391213379.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021575950.640689.12069642327533368467.stgit@warthog.procyon.org.uk/ # v4
2020-11-14 18:43:54 +00:00
nfs_fscache_invalidate(inode, 0);
}
/* A writeback failed: mark the page as bad, and invalidate the page cache */
static void nfs_set_pageerror(struct address_space *mapping)
{
struct inode *inode = mapping->host;
nfs_zap_mapping(mapping->host, mapping);
/* Force file size revalidation */
spin_lock(&inode->i_lock);
nfs_set_cache_invalid(inode, NFS_INO_REVAL_FORCED |
NFS_INO_INVALID_CHANGE |
NFS_INO_INVALID_SIZE);
spin_unlock(&inode->i_lock);
}
static void nfs_mapping_set_error(struct folio *folio, int error)
{
struct address_space *mapping = folio_file_mapping(folio);
folio_set_error(folio);
filemap_set_wb_err(mapping, error);
if (mapping->host)
errseq_set(&mapping->host->i_sb->s_wb_err,
error == -ENOSPC ? -ENOSPC : -EIO);
nfs_set_pageerror(mapping);
}
/*
* nfs_page_group_search_locked
* @head - head request of page group
* @page_offset - offset into page
*
* Search page group with head @head to find a request that contains the
* page offset @page_offset.
*
* Returns a pointer to the first matching nfs request, or NULL if no
* match is found.
*
* Must be called with the page group lock held
*/
static struct nfs_page *
nfs_page_group_search_locked(struct nfs_page *head, unsigned int page_offset)
{
struct nfs_page *req;
req = head;
do {
if (page_offset >= req->wb_pgbase &&
page_offset < (req->wb_pgbase + req->wb_bytes))
return req;
req = req->wb_this_page;
} while (req != head);
return NULL;
}
/*
* nfs_page_group_covers_page
* @head - head request of page group
*
* Return true if the page group with head @head covers the whole page,
* returns false otherwise
*/
static bool nfs_page_group_covers_page(struct nfs_page *req)
{
unsigned int len = nfs_folio_length(nfs_page_to_folio(req));
struct nfs_page *tmp;
unsigned int pos = 0;
nfs_page_group_lock(req);
for (;;) {
tmp = nfs_page_group_search_locked(req->wb_head, pos);
if (!tmp)
break;
pos = tmp->wb_pgbase + tmp->wb_bytes;
}
nfs_page_group_unlock(req);
return pos >= len;
}
/* We can set the PG_uptodate flag if we see that a write request
* covers the full page.
*/
static void nfs_mark_uptodate(struct nfs_page *req)
{
struct folio *folio = nfs_page_to_folio(req);
if (folio_test_uptodate(folio))
return;
if (!nfs_page_group_covers_page(req))
return;
folio_mark_uptodate(folio);
}
static int wb_priority(struct writeback_control *wbc)
{
int ret = 0;
if (wbc->sync_mode == WB_SYNC_ALL)
ret = FLUSH_COND_STABLE;
return ret;
}
/*
* NFS congestion control
*/
int nfs_congestion_kb;
#define NFS_CONGESTION_ON_THRESH (nfs_congestion_kb >> (PAGE_SHIFT-10))
#define NFS_CONGESTION_OFF_THRESH \
(NFS_CONGESTION_ON_THRESH - (NFS_CONGESTION_ON_THRESH >> 2))
static void nfs_folio_set_writeback(struct folio *folio)
{
struct nfs_server *nfss = NFS_SERVER(folio_file_mapping(folio)->host);
folio_start_writeback(folio);
if (atomic_long_inc_return(&nfss->writeback) > NFS_CONGESTION_ON_THRESH)
nfs: remove reliance on bdi congestion The bdi congestion tracking in not widely used and will be removed. NFS is one of a small number of filesystems that uses it, setting just the async (write) congestion flag at what it determines are appropriate times. The only remaining effect of the async flag is to cause (some) WB_SYNC_NONE writes to be skipped. So instead of setting the flag, set an internal flag and change: - .writepages to do nothing if WB_SYNC_NONE and the flag is set - .writepage to return AOP_WRITEPAGE_ACTIVATE if WB_SYNC_NONE and the flag is set. The writepages change causes a behavioural change in that pageout() can now return PAGE_ACTIVATE instead of PAGE_KEEP, so SetPageActive() will be called on the page which (I think) wil further delay the next attempt at writeout. This might be a good thing. Link: https://lkml.kernel.org/r/164549983738.9187.3972219847989393182.stgit@noble.brown Signed-off-by: NeilBrown <neilb@suse.de> Cc: Anna Schumaker <Anna.Schumaker@Netapp.com> Cc: Chao Yu <chao@kernel.org> Cc: Darrick J. Wong <djwong@kernel.org> Cc: Ilya Dryomov <idryomov@gmail.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Jeff Layton <jlayton@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Lars Ellenberg <lars.ellenberg@linbit.com> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Paolo Valente <paolo.valente@linaro.org> Cc: Philipp Reisner <philipp.reisner@linbit.com> Cc: Ryusuke Konishi <konishi.ryusuke@gmail.com> Cc: Trond Myklebust <trond.myklebust@hammerspace.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2022-03-22 21:39:01 +00:00
nfss->write_congested = 1;
}
static void nfs_folio_end_writeback(struct folio *folio)
{
struct nfs_server *nfss = NFS_SERVER(folio_file_mapping(folio)->host);
folio_end_writeback(folio);
if (atomic_long_dec_return(&nfss->writeback) <
NFS_CONGESTION_OFF_THRESH)
nfs: remove reliance on bdi congestion The bdi congestion tracking in not widely used and will be removed. NFS is one of a small number of filesystems that uses it, setting just the async (write) congestion flag at what it determines are appropriate times. The only remaining effect of the async flag is to cause (some) WB_SYNC_NONE writes to be skipped. So instead of setting the flag, set an internal flag and change: - .writepages to do nothing if WB_SYNC_NONE and the flag is set - .writepage to return AOP_WRITEPAGE_ACTIVATE if WB_SYNC_NONE and the flag is set. The writepages change causes a behavioural change in that pageout() can now return PAGE_ACTIVATE instead of PAGE_KEEP, so SetPageActive() will be called on the page which (I think) wil further delay the next attempt at writeout. This might be a good thing. Link: https://lkml.kernel.org/r/164549983738.9187.3972219847989393182.stgit@noble.brown Signed-off-by: NeilBrown <neilb@suse.de> Cc: Anna Schumaker <Anna.Schumaker@Netapp.com> Cc: Chao Yu <chao@kernel.org> Cc: Darrick J. Wong <djwong@kernel.org> Cc: Ilya Dryomov <idryomov@gmail.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Jeff Layton <jlayton@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Lars Ellenberg <lars.ellenberg@linbit.com> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Paolo Valente <paolo.valente@linaro.org> Cc: Philipp Reisner <philipp.reisner@linbit.com> Cc: Ryusuke Konishi <konishi.ryusuke@gmail.com> Cc: Trond Myklebust <trond.myklebust@hammerspace.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2022-03-22 21:39:01 +00:00
nfss->write_congested = 0;
}
static void nfs_page_end_writeback(struct nfs_page *req)
{
if (nfs_page_group_sync_on_bit(req, PG_WB_END)) {
nfs_unlock_request(req);
nfs_folio_end_writeback(nfs_page_to_folio(req));
} else
nfs_unlock_request(req);
}
/*
* nfs_destroy_unlinked_subrequests - destroy recently unlinked subrequests
*
* @destroy_list - request list (using wb_this_page) terminated by @old_head
* @old_head - the old head of the list
*
* All subrequests must be locked and removed from all lists, so at this point
* they are only "active" in this function, and possibly in nfs_wait_on_request
* with a reference held by some other context.
*/
static void
nfs_destroy_unlinked_subrequests(struct nfs_page *destroy_list,
struct nfs_page *old_head,
struct inode *inode)
{
while (destroy_list) {
struct nfs_page *subreq = destroy_list;
destroy_list = (subreq->wb_this_page == old_head) ?
NULL : subreq->wb_this_page;
/* Note: lock subreq in order to change subreq->wb_head */
nfs_page_set_headlock(subreq);
WARN_ON_ONCE(old_head != subreq->wb_head);
/* make sure old group is not used */
subreq->wb_this_page = subreq;
subreq->wb_head = subreq;
clear_bit(PG_REMOVE, &subreq->wb_flags);
/* Note: races with nfs_page_group_destroy() */
if (!kref_read(&subreq->wb_kref)) {
/* Check if we raced with nfs_page_group_destroy() */
if (test_and_clear_bit(PG_TEARDOWN, &subreq->wb_flags)) {
nfs_page_clear_headlock(subreq);
nfs_free_request(subreq);
} else
nfs_page_clear_headlock(subreq);
continue;
}
nfs_page_clear_headlock(subreq);
nfs_release_request(old_head);
if (test_and_clear_bit(PG_INODE_REF, &subreq->wb_flags)) {
nfs_release_request(subreq);
atomic_long_dec(&NFS_I(inode)->nrequests);
}
/* subreq is now totally disconnected from page group or any
* write / commit lists. last chance to wake any waiters */
nfs_unlock_and_release_request(subreq);
}
}
/*
* nfs_join_page_group - destroy subrequests of the head req
* @head: the page used to lookup the "page group" of nfs_page structures
* @inode: Inode to which the request belongs.
*
* This function joins all sub requests to the head request by first
* locking all requests in the group, cancelling any pending operations
* and finally updating the head request to cover the whole range covered by
* the (former) group. All subrequests are removed from any write or commit
* lists, unlinked from the group and destroyed.
*/
void
nfs_join_page_group(struct nfs_page *head, struct inode *inode)
{
struct nfs_page *subreq;
struct nfs_page *destroy_list = NULL;
unsigned int pgbase, off, bytes;
pgbase = head->wb_pgbase;
bytes = head->wb_bytes;
off = head->wb_offset;
for (subreq = head->wb_this_page; subreq != head;
subreq = subreq->wb_this_page) {
/* Subrequests should always form a contiguous range */
if (pgbase > subreq->wb_pgbase) {
off -= pgbase - subreq->wb_pgbase;
bytes += pgbase - subreq->wb_pgbase;
pgbase = subreq->wb_pgbase;
}
bytes = max(subreq->wb_pgbase + subreq->wb_bytes
- pgbase, bytes);
}
/* Set the head request's range to cover the former page group */
head->wb_pgbase = pgbase;
head->wb_bytes = bytes;
head->wb_offset = off;
/* Now that all requests are locked, make sure they aren't on any list.
* Commit list removal accounting is done after locks are dropped */
subreq = head;
do {
nfs_clear_request_commit(subreq);
subreq = subreq->wb_this_page;
} while (subreq != head);
/* unlink subrequests from head, destroy them later */
if (head->wb_this_page != head) {
/* destroy list will be terminated by head */
destroy_list = head->wb_this_page;
head->wb_this_page = head;
}
nfs_destroy_unlinked_subrequests(destroy_list, head, inode);
}
/*
* nfs_lock_and_join_requests - join all subreqs to the head req
* @folio: the folio used to lookup the "page group" of nfs_page structures
*
* This function joins all sub requests to the head request by first
* locking all requests in the group, cancelling any pending operations
* and finally updating the head request to cover the whole range covered by
* the (former) group. All subrequests are removed from any write or commit
* lists, unlinked from the group and destroyed.
*
* Returns a locked, referenced pointer to the head request - which after
* this call is guaranteed to be the only request associated with the page.
* Returns NULL if no requests are found for @folio, or a ERR_PTR if an
* error was encountered.
*/
static struct nfs_page *nfs_lock_and_join_requests(struct folio *folio)
{
struct inode *inode = folio_file_mapping(folio)->host;
struct nfs_page *head;
int ret;
/*
* A reference is taken only on the head request which acts as a
* reference to the whole page group - the group will not be destroyed
* until the head reference is released.
*/
head = nfs_folio_find_and_lock_request(folio);
if (IS_ERR_OR_NULL(head))
return head;
/* lock each request in the page group */
ret = nfs_page_group_lock_subrequests(head);
if (ret < 0) {
nfs_unlock_and_release_request(head);
return ERR_PTR(ret);
}
nfs_join_page_group(head, inode);
return head;
}
static void nfs_write_error(struct nfs_page *req, int error)
{
trace_nfs_write_error(nfs_page_to_inode(req), req, error);
nfs_mapping_set_error(nfs_page_to_folio(req), error);
nfs_inode_remove_request(req);
nfs_page_end_writeback(req);
nfs_release_request(req);
}
/*
* Find an associated nfs write request, and prepare to flush it out
* May return an error if the user signalled nfs_wait_on_request().
*/
static int nfs_page_async_flush(struct folio *folio,
struct writeback_control *wbc,
struct nfs_pageio_descriptor *pgio)
{
struct nfs_page *req;
int ret = 0;
req = nfs_lock_and_join_requests(folio);
if (!req)
goto out;
ret = PTR_ERR(req);
if (IS_ERR(req))
goto out;
nfs_folio_set_writeback(folio);
WARN_ON_ONCE(test_bit(PG_CLEAN, &req->wb_flags));
/* If there is a fatal error that covers this write, just exit */
ret = pgio->pg_error;
if (nfs_error_is_fatal_on_server(ret))
goto out_launder;
ret = 0;
2008-03-19 15:24:39 +00:00
if (!nfs_pageio_add_request(pgio, req)) {
ret = pgio->pg_error;
/*
* Remove the problematic req upon fatal errors on the server
*/
if (nfs_error_is_fatal_on_server(ret))
goto out_launder;
if (wbc->sync_mode == WB_SYNC_NONE)
ret = AOP_WRITEPAGE_ACTIVATE;
folio_redirty_for_writepage(wbc, folio);
nfs_redirty_request(req);
pgio->pg_error = 0;
} else
nfs_add_stats(folio_file_mapping(folio)->host,
NFSIOS_WRITEPAGES, 1);
out:
return ret;
out_launder:
nfs_write_error(req, ret);
return 0;
}
static int nfs_do_writepage(struct folio *folio, struct writeback_control *wbc,
struct nfs_pageio_descriptor *pgio)
{
nfs_pageio_cond_complete(pgio, folio_index(folio));
return nfs_page_async_flush(folio, wbc, pgio);
}
/*
* Write an mmapped page to the server.
*/
static int nfs_writepage_locked(struct folio *folio,
struct writeback_control *wbc)
{
struct nfs_pageio_descriptor pgio;
struct inode *inode = folio_file_mapping(folio)->host;
int err;
nfs: remove reliance on bdi congestion The bdi congestion tracking in not widely used and will be removed. NFS is one of a small number of filesystems that uses it, setting just the async (write) congestion flag at what it determines are appropriate times. The only remaining effect of the async flag is to cause (some) WB_SYNC_NONE writes to be skipped. So instead of setting the flag, set an internal flag and change: - .writepages to do nothing if WB_SYNC_NONE and the flag is set - .writepage to return AOP_WRITEPAGE_ACTIVATE if WB_SYNC_NONE and the flag is set. The writepages change causes a behavioural change in that pageout() can now return PAGE_ACTIVATE instead of PAGE_KEEP, so SetPageActive() will be called on the page which (I think) wil further delay the next attempt at writeout. This might be a good thing. Link: https://lkml.kernel.org/r/164549983738.9187.3972219847989393182.stgit@noble.brown Signed-off-by: NeilBrown <neilb@suse.de> Cc: Anna Schumaker <Anna.Schumaker@Netapp.com> Cc: Chao Yu <chao@kernel.org> Cc: Darrick J. Wong <djwong@kernel.org> Cc: Ilya Dryomov <idryomov@gmail.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Jeff Layton <jlayton@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Lars Ellenberg <lars.ellenberg@linbit.com> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Paolo Valente <paolo.valente@linaro.org> Cc: Philipp Reisner <philipp.reisner@linbit.com> Cc: Ryusuke Konishi <konishi.ryusuke@gmail.com> Cc: Trond Myklebust <trond.myklebust@hammerspace.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2022-03-22 21:39:01 +00:00
if (wbc->sync_mode == WB_SYNC_NONE &&
NFS_SERVER(inode)->write_congested)
return AOP_WRITEPAGE_ACTIVATE;
nfs_inc_stats(inode, NFSIOS_VFSWRITEPAGE);
nfs_pageio_init_write(&pgio, inode, 0, false,
&nfs_async_write_completion_ops);
err = nfs_do_writepage(folio, wbc, &pgio);
pgio.pg_error = 0;
nfs_pageio_complete(&pgio);
return err;
}
int nfs_writepage(struct page *page, struct writeback_control *wbc)
{
struct folio *folio = page_folio(page);
int ret;
ret = nfs_writepage_locked(folio, wbc);
if (ret != AOP_WRITEPAGE_ACTIVATE)
unlock_page(page);
return ret;
}
static int nfs_writepages_callback(struct folio *folio,
struct writeback_control *wbc, void *data)
{
int ret;
ret = nfs_do_writepage(folio, wbc, data);
if (ret != AOP_WRITEPAGE_ACTIVATE)
folio_unlock(folio);
return ret;
}
static void nfs_io_completion_commit(void *inode)
{
nfs_commit_inode(inode, 0);
}
int nfs_writepages(struct address_space *mapping, struct writeback_control *wbc)
{
struct inode *inode = mapping->host;
struct nfs_pageio_descriptor pgio;
struct nfs_io_completion *ioc = NULL;
unsigned int mntflags = NFS_SERVER(inode)->flags;
int priority = 0;
int err;
nfs: remove reliance on bdi congestion The bdi congestion tracking in not widely used and will be removed. NFS is one of a small number of filesystems that uses it, setting just the async (write) congestion flag at what it determines are appropriate times. The only remaining effect of the async flag is to cause (some) WB_SYNC_NONE writes to be skipped. So instead of setting the flag, set an internal flag and change: - .writepages to do nothing if WB_SYNC_NONE and the flag is set - .writepage to return AOP_WRITEPAGE_ACTIVATE if WB_SYNC_NONE and the flag is set. The writepages change causes a behavioural change in that pageout() can now return PAGE_ACTIVATE instead of PAGE_KEEP, so SetPageActive() will be called on the page which (I think) wil further delay the next attempt at writeout. This might be a good thing. Link: https://lkml.kernel.org/r/164549983738.9187.3972219847989393182.stgit@noble.brown Signed-off-by: NeilBrown <neilb@suse.de> Cc: Anna Schumaker <Anna.Schumaker@Netapp.com> Cc: Chao Yu <chao@kernel.org> Cc: Darrick J. Wong <djwong@kernel.org> Cc: Ilya Dryomov <idryomov@gmail.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Jeff Layton <jlayton@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Lars Ellenberg <lars.ellenberg@linbit.com> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Paolo Valente <paolo.valente@linaro.org> Cc: Philipp Reisner <philipp.reisner@linbit.com> Cc: Ryusuke Konishi <konishi.ryusuke@gmail.com> Cc: Trond Myklebust <trond.myklebust@hammerspace.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2022-03-22 21:39:01 +00:00
if (wbc->sync_mode == WB_SYNC_NONE &&
NFS_SERVER(inode)->write_congested)
return 0;
nfs_inc_stats(inode, NFSIOS_VFSWRITEPAGES);
if (!(mntflags & NFS_MOUNT_WRITE_EAGER) || wbc->for_kupdate ||
wbc->for_background || wbc->for_sync || wbc->for_reclaim) {
ioc = nfs_io_completion_alloc(GFP_KERNEL);
if (ioc)
nfs_io_completion_init(ioc, nfs_io_completion_commit,
inode);
priority = wb_priority(wbc);
}
do {
nfs_pageio_init_write(&pgio, inode, priority, false,
&nfs_async_write_completion_ops);
pgio.pg_io_completion = ioc;
err = write_cache_pages(mapping, wbc, nfs_writepages_callback,
&pgio);
pgio.pg_error = 0;
nfs_pageio_complete(&pgio);
} while (err < 0 && !nfs_error_is_fatal(err));
nfs_io_completion_put(ioc);
2009-03-11 18:10:30 +00:00
if (err < 0)
2009-03-11 18:10:30 +00:00
goto out_err;
return 0;
2009-03-11 18:10:30 +00:00
out_err:
return err;
}
/*
* Insert a write request into an inode
*/
static void nfs_inode_add_request(struct nfs_page *req)
{
struct folio *folio = nfs_page_to_folio(req);
struct address_space *mapping = folio_file_mapping(folio);
struct nfs_inode *nfsi = NFS_I(mapping->host);
nfs: add support for multiple nfs reqs per page Add "page groups" - a circular list of nfs requests (struct nfs_page) that all reference the same page. This gives nfs read and write paths the ability to account for sub-page regions independently. This somewhat follows the design of struct buffer_head's sub-page accounting. Only "head" requests are ever added/removed from the inode list in the buffered write path. "head" and "sub" requests are treated the same through the read path and the rest of the write/commit path. Requests are given an extra reference across the life of the list. Page groups are never rejoined after being split. If the read/write request fails and the client falls back to another path (ie revert to MDS in PNFS case), the already split requests are pushed through the recoalescing code again, which may split them further and then coalesce them into properly sized requests on the wire. Fragmentation shouldn't be a problem with the current design, because we flush all requests in page group when a non-contiguous request is added, so the only time resplitting should occur is on a resend of a read or write. This patch lays the groundwork for sub-page splitting, but does not actually do any splitting. For now all page groups have one request as pg_test functions don't yet split pages. There are several related patches that are needed support multiple requests per page group. Signed-off-by: Weston Andros Adamson <dros@primarydata.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2014-05-15 15:56:45 +00:00
WARN_ON_ONCE(req->wb_this_page != req);
/* Lock the request! */
nfs_lock_request(req);
/*
* Swap-space should not get truncated. Hence no need to plug the race
* with invalidate/truncate.
*/
spin_lock(&mapping->private_lock);
if (likely(!folio_test_swapcache(folio))) {
set_bit(PG_MAPPED, &req->wb_flags);
folio_set_private(folio);
folio->private = req;
}
spin_unlock(&mapping->private_lock);
atomic_long_inc(&nfsi->nrequests);
/* this a head request for a page group - mark it as having an
* extra reference so sub groups can follow suit.
* This flag also informs pgio layer when to bump nrequests when
* adding subrequests. */
WARN_ON(test_and_set_bit(PG_INODE_REF, &req->wb_flags));
kref_get(&req->wb_kref);
}
/*
* Remove a write request from an inode
*/
static void nfs_inode_remove_request(struct nfs_page *req)
{
if (nfs_page_group_sync_on_bit(req, PG_REMOVE)) {
struct folio *folio = nfs_page_to_folio(req->wb_head);
struct address_space *mapping = folio_file_mapping(folio);
spin_lock(&mapping->private_lock);
if (likely(folio && !folio_test_swapcache(folio))) {
folio->private = NULL;
folio_clear_private(folio);
clear_bit(PG_MAPPED, &req->wb_head->wb_flags);
}
spin_unlock(&mapping->private_lock);
}
nfs: Fix nfsi->nrequests count error on nfs_inode_remove_request When xfstests testing, there are some WARNING as below: WARNING: CPU: 0 PID: 6235 at fs/nfs/inode.c:122 nfs_clear_inode+0x9c/0xd8 Modules linked in: CPU: 0 PID: 6235 Comm: umount.nfs Hardware name: linux,dummy-virt (DT) pstate: 60000005 (nZCv daif -PAN -UAO) pc : nfs_clear_inode+0x9c/0xd8 lr : nfs_evict_inode+0x60/0x78 sp : fffffc000f68fc00 x29: fffffc000f68fc00 x28: fffffe00c53155c0 x27: fffffe00c5315000 x26: fffffc0009a63748 x25: fffffc000f68fd18 x24: fffffc000bfaaf40 x23: fffffc000936d3c0 x22: fffffe00c4ff5e20 x21: fffffc000bfaaf40 x20: fffffe00c4ff5d10 x19: fffffc000c056000 x18: 000000000000003c x17: 0000000000000000 x16: 0000000000000000 x15: 0000000000000040 x14: 0000000000000228 x13: fffffc000c3a2000 x12: 0000000000000045 x11: 0000000000000000 x10: 0000000000000000 x9 : 0000000000000000 x8 : 0000000000000000 x7 : 0000000000000000 x6 : fffffc00084b027c x5 : fffffc0009a64000 x4 : fffffe00c0e77400 x3 : fffffc000c0563a8 x2 : fffffffffffffffb x1 : 000000000000764e x0 : 0000000000000001 Call trace: nfs_clear_inode+0x9c/0xd8 nfs_evict_inode+0x60/0x78 evict+0x108/0x380 dispose_list+0x70/0xa0 evict_inodes+0x194/0x210 generic_shutdown_super+0xb0/0x220 nfs_kill_super+0x40/0x88 deactivate_locked_super+0xb4/0x120 deactivate_super+0x144/0x160 cleanup_mnt+0x98/0x148 __cleanup_mnt+0x38/0x50 task_work_run+0x114/0x160 do_notify_resume+0x2f8/0x308 work_pending+0x8/0x14 The nrequest should be increased/decreased only if PG_INODE_REF flag was setted. But in the nfs_inode_remove_request function, it maybe decrease when no PG_INODE_REF flag, this maybe lead nrequests count error. Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: ZhangXiaoxu <zhangxiaoxu5@huawei.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-09-26 06:29:38 +00:00
if (test_and_clear_bit(PG_INODE_REF, &req->wb_flags)) {
nfs_release_request(req);
atomic_long_dec(&NFS_I(nfs_page_to_inode(req))->nrequests);
nfs: Fix nfsi->nrequests count error on nfs_inode_remove_request When xfstests testing, there are some WARNING as below: WARNING: CPU: 0 PID: 6235 at fs/nfs/inode.c:122 nfs_clear_inode+0x9c/0xd8 Modules linked in: CPU: 0 PID: 6235 Comm: umount.nfs Hardware name: linux,dummy-virt (DT) pstate: 60000005 (nZCv daif -PAN -UAO) pc : nfs_clear_inode+0x9c/0xd8 lr : nfs_evict_inode+0x60/0x78 sp : fffffc000f68fc00 x29: fffffc000f68fc00 x28: fffffe00c53155c0 x27: fffffe00c5315000 x26: fffffc0009a63748 x25: fffffc000f68fd18 x24: fffffc000bfaaf40 x23: fffffc000936d3c0 x22: fffffe00c4ff5e20 x21: fffffc000bfaaf40 x20: fffffe00c4ff5d10 x19: fffffc000c056000 x18: 000000000000003c x17: 0000000000000000 x16: 0000000000000000 x15: 0000000000000040 x14: 0000000000000228 x13: fffffc000c3a2000 x12: 0000000000000045 x11: 0000000000000000 x10: 0000000000000000 x9 : 0000000000000000 x8 : 0000000000000000 x7 : 0000000000000000 x6 : fffffc00084b027c x5 : fffffc0009a64000 x4 : fffffe00c0e77400 x3 : fffffc000c0563a8 x2 : fffffffffffffffb x1 : 000000000000764e x0 : 0000000000000001 Call trace: nfs_clear_inode+0x9c/0xd8 nfs_evict_inode+0x60/0x78 evict+0x108/0x380 dispose_list+0x70/0xa0 evict_inodes+0x194/0x210 generic_shutdown_super+0xb0/0x220 nfs_kill_super+0x40/0x88 deactivate_locked_super+0xb4/0x120 deactivate_super+0x144/0x160 cleanup_mnt+0x98/0x148 __cleanup_mnt+0x38/0x50 task_work_run+0x114/0x160 do_notify_resume+0x2f8/0x308 work_pending+0x8/0x14 The nrequest should be increased/decreased only if PG_INODE_REF flag was setted. But in the nfs_inode_remove_request function, it maybe decrease when no PG_INODE_REF flag, this maybe lead nrequests count error. Reported-by: Hulk Robot <hulkci@huawei.com> Signed-off-by: ZhangXiaoxu <zhangxiaoxu5@huawei.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2019-09-26 06:29:38 +00:00
}
}
static void nfs_mark_request_dirty(struct nfs_page *req)
{
struct folio *folio = nfs_page_to_folio(req);
if (folio)
filemap_dirty_folio(folio_mapping(folio), folio);
}
/*
* nfs_page_search_commits_for_head_request_locked
*
* Search through commit lists on @inode for the head request for @folio.
* Must be called while holding the inode (which is cinfo) lock.
*
* Returns the head request if found, or NULL if not found.
*/
static struct nfs_page *
nfs_page_search_commits_for_head_request_locked(struct nfs_inode *nfsi,
struct folio *folio)
{
struct nfs_page *freq, *t;
struct nfs_commit_info cinfo;
struct inode *inode = &nfsi->vfs_inode;
nfs_init_cinfo_from_inode(&cinfo, inode);
/* search through pnfs commit lists */
freq = pnfs_search_commit_reqs(inode, &cinfo, folio);
if (freq)
return freq->wb_head;
/* Linearly search the commit list for the correct request */
list_for_each_entry_safe(freq, t, &cinfo.mds->list, wb_list) {
if (nfs_page_to_folio(freq) == folio)
return freq->wb_head;
}
return NULL;
}
/**
* nfs_request_add_commit_list_locked - add request to a commit list
* @req: pointer to a struct nfs_page
* @dst: commit list head
* @cinfo: holds list lock and accounting info
*
* This sets the PG_CLEAN bit, updates the cinfo count of
* number of outstanding requests requiring a commit as well as
* the MM page stats.
*
* The caller must hold NFS_I(cinfo->inode)->commit_mutex, and the
* nfs_page lock.
*/
void
nfs_request_add_commit_list_locked(struct nfs_page *req, struct list_head *dst,
struct nfs_commit_info *cinfo)
{
set_bit(PG_CLEAN, &req->wb_flags);
nfs_list_add_request(req, dst);
atomic_long_inc(&cinfo->mds->ncommit);
}
EXPORT_SYMBOL_GPL(nfs_request_add_commit_list_locked);
/**
* nfs_request_add_commit_list - add request to a commit list
* @req: pointer to a struct nfs_page
* @cinfo: holds list lock and accounting info
*
* This sets the PG_CLEAN bit, updates the cinfo count of
* number of outstanding requests requiring a commit as well as
* the MM page stats.
*
* The caller must _not_ hold the cinfo->lock, but must be
* holding the nfs_page lock.
*/
void
nfs_request_add_commit_list(struct nfs_page *req, struct nfs_commit_info *cinfo)
{
mutex_lock(&NFS_I(cinfo->inode)->commit_mutex);
nfs_request_add_commit_list_locked(req, &cinfo->mds->list, cinfo);
mutex_unlock(&NFS_I(cinfo->inode)->commit_mutex);
nfs_folio_mark_unstable(nfs_page_to_folio(req), cinfo);
}
EXPORT_SYMBOL_GPL(nfs_request_add_commit_list);
/**
* nfs_request_remove_commit_list - Remove request from a commit list
* @req: pointer to a nfs_page
* @cinfo: holds list lock and accounting info
*
* This clears the PG_CLEAN bit, and updates the cinfo's count of
* number of outstanding requests requiring a commit
* It does not update the MM page stats.
*
* The caller _must_ hold the cinfo->lock and the nfs_page lock.
*/
void
nfs_request_remove_commit_list(struct nfs_page *req,
struct nfs_commit_info *cinfo)
{
if (!test_and_clear_bit(PG_CLEAN, &(req)->wb_flags))
return;
nfs_list_remove_request(req);
atomic_long_dec(&cinfo->mds->ncommit);
}
EXPORT_SYMBOL_GPL(nfs_request_remove_commit_list);
static void nfs_init_cinfo_from_inode(struct nfs_commit_info *cinfo,
struct inode *inode)
{
cinfo->inode = inode;
cinfo->mds = &NFS_I(inode)->commit_info;
cinfo->ds = pnfs_get_ds_info(inode);
cinfo->dreq = NULL;
cinfo->completion_ops = &nfs_commit_completion_ops;
}
void nfs_init_cinfo(struct nfs_commit_info *cinfo,
struct inode *inode,
struct nfs_direct_req *dreq)
{
if (dreq)
nfs_init_cinfo_from_dreq(cinfo, dreq);
else
nfs_init_cinfo_from_inode(cinfo, inode);
}
EXPORT_SYMBOL_GPL(nfs_init_cinfo);
/*
* Add a request to the inode's commit list.
*/
void
nfs_mark_request_commit(struct nfs_page *req, struct pnfs_layout_segment *lseg,
struct nfs_commit_info *cinfo, u32 ds_commit_idx)
{
if (pnfs_mark_request_commit(req, lseg, cinfo, ds_commit_idx))
return;
nfs_request_add_commit_list(req, cinfo);
}
static void nfs_folio_clear_commit(struct folio *folio)
{
if (folio) {
long nr = folio_nr_pages(folio);
node_stat_mod_folio(folio, NR_WRITEBACK, -nr);
wb_stat_mod(&inode_to_bdi(folio_file_mapping(folio)->host)->wb,
WB_WRITEBACK, -nr);
}
}
/* Called holding the request lock on @req */
static void
nfs_clear_request_commit(struct nfs_page *req)
{
if (test_bit(PG_CLEAN, &req->wb_flags)) {
struct nfs_open_context *ctx = nfs_req_openctx(req);
struct inode *inode = d_inode(ctx->dentry);
struct nfs_commit_info cinfo;
nfs_init_cinfo_from_inode(&cinfo, inode);
mutex_lock(&NFS_I(inode)->commit_mutex);
if (!pnfs_clear_request_commit(req, &cinfo)) {
nfs_request_remove_commit_list(req, &cinfo);
}
mutex_unlock(&NFS_I(inode)->commit_mutex);
nfs_folio_clear_commit(nfs_page_to_folio(req));
}
}
int nfs_write_need_commit(struct nfs_pgio_header *hdr)
{
if (hdr->verf.committed == NFS_DATA_SYNC)
return hdr->lseg == NULL;
return hdr->verf.committed != NFS_FILE_SYNC;
}
static void nfs_async_write_init(struct nfs_pgio_header *hdr)
{
nfs_io_completion_get(hdr->io_completion);
}
static void nfs_write_completion(struct nfs_pgio_header *hdr)
{
struct nfs_commit_info cinfo;
unsigned long bytes = 0;
if (test_bit(NFS_IOHDR_REDO, &hdr->flags))
goto out;
nfs_init_cinfo_from_inode(&cinfo, hdr->inode);
while (!list_empty(&hdr->pages)) {
struct nfs_page *req = nfs_list_entry(hdr->pages.next);
bytes += req->wb_bytes;
nfs_list_remove_request(req);
if (test_bit(NFS_IOHDR_ERROR, &hdr->flags) &&
(hdr->good_bytes < bytes)) {
trace_nfs_comp_error(hdr->inode, req, hdr->error);
nfs_mapping_set_error(nfs_page_to_folio(req),
hdr->error);
goto remove_req;
}
if (nfs_write_need_commit(hdr)) {
/* Reset wb_nio, since the write was successful. */
req->wb_nio = 0;
memcpy(&req->wb_verf, &hdr->verf.verifier, sizeof(req->wb_verf));
nfs_mark_request_commit(req, hdr->lseg, &cinfo,
hdr->pgio_mirror_idx);
goto next;
}
remove_req:
nfs_inode_remove_request(req);
next:
nfs_page_end_writeback(req);
nfs_release_request(req);
}
out:
nfs_io_completion_put(hdr->io_completion);
hdr->release(hdr);
}
unsigned long
nfs_reqs_to_commit(struct nfs_commit_info *cinfo)
{
return atomic_long_read(&cinfo->mds->ncommit);
}
/* NFS_I(cinfo->inode)->commit_mutex held by caller */
int
nfs_scan_commit_list(struct list_head *src, struct list_head *dst,
struct nfs_commit_info *cinfo, int max)
{
struct nfs_page *req, *tmp;
int ret = 0;
list_for_each_entry_safe(req, tmp, src, wb_list) {
kref_get(&req->wb_kref);
if (!nfs_lock_request(req)) {
nfs_release_request(req);
continue;
}
nfs_request_remove_commit_list(req, cinfo);
clear_bit(PG_COMMIT_TO_DS, &req->wb_flags);
nfs_list_add_request(req, dst);
ret++;
if ((ret == max) && !cinfo->dreq)
break;
cond_resched();
}
return ret;
}
EXPORT_SYMBOL_GPL(nfs_scan_commit_list);
/*
* nfs_scan_commit - Scan an inode for commit requests
* @inode: NFS inode to scan
* @dst: mds destination list
* @cinfo: mds and ds lists of reqs ready to commit
*
* Moves requests from the inode's 'commit' request list.
* The requests are *not* checked to ensure that they form a contiguous set.
*/
int
nfs_scan_commit(struct inode *inode, struct list_head *dst,
struct nfs_commit_info *cinfo)
{
int ret = 0;
if (!atomic_long_read(&cinfo->mds->ncommit))
return 0;
mutex_lock(&NFS_I(cinfo->inode)->commit_mutex);
if (atomic_long_read(&cinfo->mds->ncommit) > 0) {
const int max = INT_MAX;
ret = nfs_scan_commit_list(&cinfo->mds->list, dst,
cinfo, max);
ret += pnfs_scan_commit_lists(inode, cinfo, max - ret);
}
mutex_unlock(&NFS_I(cinfo->inode)->commit_mutex);
return ret;
}
/*
* Search for an existing write request, and attempt to update
* it to reflect a new dirty region on a given page.
*
* If the attempt fails, then the existing request is flushed out
* to disk.
*/
static struct nfs_page *nfs_try_to_update_request(struct folio *folio,
unsigned int offset,
unsigned int bytes)
{
struct nfs_page *req;
unsigned int rqend;
unsigned int end;
int error;
end = offset + bytes;
req = nfs_lock_and_join_requests(folio);
if (IS_ERR_OR_NULL(req))
return req;
rqend = req->wb_offset + req->wb_bytes;
/*
* Tell the caller to flush out the request if
* the offsets are non-contiguous.
* Note: nfs_flush_incompatible() will already
* have flushed out requests having wrong owners.
*/
if (offset > rqend || end < req->wb_offset)
goto out_flushme;
/* Okay, the request matches. Update the region */
if (offset < req->wb_offset) {
req->wb_offset = offset;
req->wb_pgbase = offset;
}
if (end > rqend)
req->wb_bytes = end - req->wb_offset;
else
req->wb_bytes = rqend - req->wb_offset;
req->wb_nio = 0;
return req;
out_flushme:
/*
* Note: we mark the request dirty here because
* nfs_lock_and_join_requests() cannot preserve
* commit flags, so we have to replay the write.
*/
nfs_mark_request_dirty(req);
nfs_unlock_and_release_request(req);
error = nfs_wb_folio(folio_file_mapping(folio)->host, folio);
return (error < 0) ? ERR_PTR(error) : NULL;
}
/*
* Try to update an existing write request, or create one if there is none.
*
* Note: Should always be called with the Page Lock held to prevent races
* if we have to add a new request. Also assumes that the caller has
* already called nfs_flush_incompatible() if necessary.
*/
static struct nfs_page *nfs_setup_write_request(struct nfs_open_context *ctx,
struct folio *folio,
unsigned int offset,
unsigned int bytes)
{
struct nfs_page *req;
req = nfs_try_to_update_request(folio, offset, bytes);
if (req != NULL)
goto out;
req = nfs_page_create_from_folio(ctx, folio, offset, bytes);
if (IS_ERR(req))
goto out;
nfs_inode_add_request(req);
out:
return req;
}
static int nfs_writepage_setup(struct nfs_open_context *ctx,
struct folio *folio, unsigned int offset,
unsigned int count)
{
struct nfs_page *req;
req = nfs_setup_write_request(ctx, folio, offset, count);
if (IS_ERR(req))
return PTR_ERR(req);
/* Update file length */
nfs_grow_file(folio, offset, count);
nfs_mark_uptodate(req);
nfs_mark_request_dirty(req);
nfs_unlock_and_release_request(req);
return 0;
}
int nfs_flush_incompatible(struct file *file, struct folio *folio)
{
struct nfs_open_context *ctx = nfs_file_open_context(file);
struct nfs_lock_context *l_ctx;
struct file_lock_context *flctx = locks_inode_context(file_inode(file));
struct nfs_page *req;
int do_flush, status;
/*
* Look for a request corresponding to this page. If there
* is one, and it belongs to another file, we flush it out
* before we try to copy anything into the page. Do this
* due to the lack of an ACCESS-type call in NFSv2.
* Also do the same if we find a request from an existing
* dropped page.
*/
do {
req = nfs_folio_find_head_request(folio);
if (req == NULL)
return 0;
l_ctx = req->wb_lock_context;
do_flush = nfs_page_to_folio(req) != folio ||
!nfs_match_open_context(nfs_req_openctx(req), ctx);
if (l_ctx && flctx &&
!(list_empty_careful(&flctx->flc_posix) &&
list_empty_careful(&flctx->flc_flock))) {
do_flush |= l_ctx->lockowner != current->files;
}
nfs_release_request(req);
if (!do_flush)
return 0;
status = nfs_wb_folio(folio_file_mapping(folio)->host, folio);
} while (status == 0);
return status;
}
/*
* Avoid buffered writes when a open context credential's key would
* expire soon.
*
* Returns -EACCES if the key will expire within RPC_KEY_EXPIRE_FAIL.
*
* Return 0 and set a credential flag which triggers the inode to flush
* and performs NFS_FILE_SYNC writes if the key will expired within
* RPC_KEY_EXPIRE_TIMEO.
*/
int
nfs_key_timeout_notify(struct file *filp, struct inode *inode)
{
struct nfs_open_context *ctx = nfs_file_open_context(filp);
if (nfs_ctx_key_to_expire(ctx, inode) &&
!rcu_access_pointer(ctx->ll_cred))
/* Already expired! */
return -EACCES;
return 0;
}
/*
* Test if the open context credential key is marked to expire soon.
*/
sunrpc: move NO_CRKEY_TIMEOUT to the auth->au_flags A generic_cred can be used to look up a unx_cred or a gss_cred, so it's not really safe to use the the generic_cred->acred->ac_flags to store the NO_CRKEY_TIMEOUT flag. A lookup for a unx_cred triggered while the KEY_EXPIRE_SOON flag is already set will cause both NO_CRKEY_TIMEOUT and KEY_EXPIRE_SOON to be set in the ac_flags, leaving the user associated with the auth_cred to be in a state where they're perpetually doing 4K NFS_FILE_SYNC writes. This can be reproduced as follows: 1. Mount two NFS filesystems, one with sec=krb5 and one with sec=sys. They do not need to be the same export, nor do they even need to be from the same NFS server. Also, v3 is fine. $ sudo mount -o v3,sec=krb5 server1:/export /mnt/krb5 $ sudo mount -o v3,sec=sys server2:/export /mnt/sys 2. As the normal user, before accessing the kerberized mount, kinit with a short lifetime (but not so short that renewing the ticket would leave you within the 4-minute window again by the time the original ticket expires), e.g. $ kinit -l 10m -r 60m 3. Do some I/O to the kerberized mount and verify that the writes are wsize, UNSTABLE: $ dd if=/dev/zero of=/mnt/krb5/file bs=1M count=1 4. Wait until you're within 4 minutes of key expiry, then do some more I/O to the kerberized mount to ensure that RPC_CRED_KEY_EXPIRE_SOON gets set. Verify that the writes are 4K, FILE_SYNC: $ dd if=/dev/zero of=/mnt/krb5/file bs=1M count=1 5. Now do some I/O to the sec=sys mount. This will cause RPC_CRED_NO_CRKEY_TIMEOUT to be set: $ dd if=/dev/zero of=/mnt/sys/file bs=1M count=1 6. Writes for that user will now be permanently 4K, FILE_SYNC for that user, regardless of which mount is being written to, until you reboot the client. Renewing the kerberos ticket (assuming it hasn't already expired) will have no effect. Grabbing a new kerberos ticket at this point will have no effect either. Move the flag to the auth->au_flags field (which is currently unused) and rename it slightly to reflect that it's no longer associated with the auth_cred->ac_flags. Add the rpc_auth to the arg list of rpcauth_cred_key_to_expire and check the au_flags there too. Finally, add the inode to the arg list of nfs_ctx_key_to_expire so we can determine the rpc_auth to pass to rpcauth_cred_key_to_expire. Signed-off-by: Scott Mayhew <smayhew@redhat.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2016-06-07 19:14:48 +00:00
bool nfs_ctx_key_to_expire(struct nfs_open_context *ctx, struct inode *inode)
{
sunrpc: move NO_CRKEY_TIMEOUT to the auth->au_flags A generic_cred can be used to look up a unx_cred or a gss_cred, so it's not really safe to use the the generic_cred->acred->ac_flags to store the NO_CRKEY_TIMEOUT flag. A lookup for a unx_cred triggered while the KEY_EXPIRE_SOON flag is already set will cause both NO_CRKEY_TIMEOUT and KEY_EXPIRE_SOON to be set in the ac_flags, leaving the user associated with the auth_cred to be in a state where they're perpetually doing 4K NFS_FILE_SYNC writes. This can be reproduced as follows: 1. Mount two NFS filesystems, one with sec=krb5 and one with sec=sys. They do not need to be the same export, nor do they even need to be from the same NFS server. Also, v3 is fine. $ sudo mount -o v3,sec=krb5 server1:/export /mnt/krb5 $ sudo mount -o v3,sec=sys server2:/export /mnt/sys 2. As the normal user, before accessing the kerberized mount, kinit with a short lifetime (but not so short that renewing the ticket would leave you within the 4-minute window again by the time the original ticket expires), e.g. $ kinit -l 10m -r 60m 3. Do some I/O to the kerberized mount and verify that the writes are wsize, UNSTABLE: $ dd if=/dev/zero of=/mnt/krb5/file bs=1M count=1 4. Wait until you're within 4 minutes of key expiry, then do some more I/O to the kerberized mount to ensure that RPC_CRED_KEY_EXPIRE_SOON gets set. Verify that the writes are 4K, FILE_SYNC: $ dd if=/dev/zero of=/mnt/krb5/file bs=1M count=1 5. Now do some I/O to the sec=sys mount. This will cause RPC_CRED_NO_CRKEY_TIMEOUT to be set: $ dd if=/dev/zero of=/mnt/sys/file bs=1M count=1 6. Writes for that user will now be permanently 4K, FILE_SYNC for that user, regardless of which mount is being written to, until you reboot the client. Renewing the kerberos ticket (assuming it hasn't already expired) will have no effect. Grabbing a new kerberos ticket at this point will have no effect either. Move the flag to the auth->au_flags field (which is currently unused) and rename it slightly to reflect that it's no longer associated with the auth_cred->ac_flags. Add the rpc_auth to the arg list of rpcauth_cred_key_to_expire and check the au_flags there too. Finally, add the inode to the arg list of nfs_ctx_key_to_expire so we can determine the rpc_auth to pass to rpcauth_cred_key_to_expire. Signed-off-by: Scott Mayhew <smayhew@redhat.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2016-06-07 19:14:48 +00:00
struct rpc_auth *auth = NFS_SERVER(inode)->client->cl_auth;
struct rpc_cred *cred, *new, *old = NULL;
struct auth_cred acred = {
.cred = ctx->cred,
};
bool ret = false;
rcu_read_lock();
cred = rcu_dereference(ctx->ll_cred);
if (cred && !(cred->cr_ops->crkey_timeout &&
cred->cr_ops->crkey_timeout(cred)))
goto out;
rcu_read_unlock();
new = auth->au_ops->lookup_cred(auth, &acred, 0);
if (new == cred) {
put_rpccred(new);
return true;
}
if (IS_ERR_OR_NULL(new)) {
new = NULL;
ret = true;
} else if (new->cr_ops->crkey_timeout &&
new->cr_ops->crkey_timeout(new))
ret = true;
rcu_read_lock();
old = rcu_dereference_protected(xchg(&ctx->ll_cred,
RCU_INITIALIZER(new)), 1);
out:
rcu_read_unlock();
put_rpccred(old);
return ret;
}
/*
* If the page cache is marked as unsafe or invalid, then we can't rely on
* the PageUptodate() flag. In this case, we will need to turn off
* write optimisations that depend on the page contents being correct.
*/
static bool nfs_folio_write_uptodate(struct folio *folio, unsigned int pagelen)
{
struct inode *inode = folio_file_mapping(folio)->host;
NFS: fix the handling of NFS_INO_INVALID_DATA flag in nfs_revalidate_mapping There is a possible race in how the nfs_invalidate_mapping function is handled. Currently, we go and invalidate the pages in the file and then clear NFS_INO_INVALID_DATA. The problem is that it's possible for a stale page to creep into the mapping after the page was invalidated (i.e., via readahead). If another writer comes along and sets the flag after that happens but before invalidate_inode_pages2 returns then we could clear the flag without the cache having been properly invalidated. So, we must clear the flag first and then invalidate the pages. Doing this however, opens another race: It's possible to have two concurrent read() calls that end up in nfs_revalidate_mapping at the same time. The first one clears the NFS_INO_INVALID_DATA flag and then goes to call nfs_invalidate_mapping. Just before calling that though, the other task races in, checks the flag and finds it cleared. At that point, it trusts that the mapping is good and gets the lock on the page, allowing the read() to be satisfied from the cache even though the data is no longer valid. These effects are easily manifested by running diotest3 from the LTP test suite on NFS. That program does a series of DIO writes and buffered reads. The operations are serialized and page-aligned but the existing code fails the test since it occasionally allows a read to come out of the cache incorrectly. While mixing direct and buffered I/O isn't recommended, I believe it's possible to hit this in other ways that just use buffered I/O, though that situation is much harder to reproduce. The problem is that the checking/clearing of that flag and the invalidation of the mapping really need to be atomic. Fix this by serializing concurrent invalidations with a bitlock. At the same time, we also need to allow other places that check NFS_INO_INVALID_DATA to check whether we might be in the middle of invalidating the file, so fix up a couple of places that do that to look for the new NFS_INO_INVALIDATING flag. Doing this requires us to be careful not to set the bitlock unnecessarily, so this code only does that if it believes it will be doing an invalidation. Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2014-01-27 18:46:15 +00:00
struct nfs_inode *nfsi = NFS_I(inode);
if (nfs_have_delegated_attributes(inode))
goto out;
if (nfsi->cache_validity &
(NFS_INO_INVALID_CHANGE | NFS_INO_INVALID_SIZE))
NFS: fix the handling of NFS_INO_INVALID_DATA flag in nfs_revalidate_mapping There is a possible race in how the nfs_invalidate_mapping function is handled. Currently, we go and invalidate the pages in the file and then clear NFS_INO_INVALID_DATA. The problem is that it's possible for a stale page to creep into the mapping after the page was invalidated (i.e., via readahead). If another writer comes along and sets the flag after that happens but before invalidate_inode_pages2 returns then we could clear the flag without the cache having been properly invalidated. So, we must clear the flag first and then invalidate the pages. Doing this however, opens another race: It's possible to have two concurrent read() calls that end up in nfs_revalidate_mapping at the same time. The first one clears the NFS_INO_INVALID_DATA flag and then goes to call nfs_invalidate_mapping. Just before calling that though, the other task races in, checks the flag and finds it cleared. At that point, it trusts that the mapping is good and gets the lock on the page, allowing the read() to be satisfied from the cache even though the data is no longer valid. These effects are easily manifested by running diotest3 from the LTP test suite on NFS. That program does a series of DIO writes and buffered reads. The operations are serialized and page-aligned but the existing code fails the test since it occasionally allows a read to come out of the cache incorrectly. While mixing direct and buffered I/O isn't recommended, I believe it's possible to hit this in other ways that just use buffered I/O, though that situation is much harder to reproduce. The problem is that the checking/clearing of that flag and the invalidation of the mapping really need to be atomic. Fix this by serializing concurrent invalidations with a bitlock. At the same time, we also need to allow other places that check NFS_INO_INVALID_DATA to check whether we might be in the middle of invalidating the file, so fix up a couple of places that do that to look for the new NFS_INO_INVALIDATING flag. Doing this requires us to be careful not to set the bitlock unnecessarily, so this code only does that if it believes it will be doing an invalidation. Signed-off-by: Jeff Layton <jlayton@redhat.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2014-01-27 18:46:15 +00:00
return false;
smp_rmb();
if (test_bit(NFS_INO_INVALIDATING, &nfsi->flags) && pagelen != 0)
return false;
out:
if (nfsi->cache_validity & NFS_INO_INVALID_DATA && pagelen != 0)
return false;
return folio_test_uptodate(folio) != 0;
}
static bool
is_whole_file_wrlock(struct file_lock *fl)
{
return fl->fl_start == 0 && fl->fl_end == OFFSET_MAX &&
fl->fl_type == F_WRLCK;
}
/* If we know the page is up to date, and we're not using byte range locks (or
* if we have the whole file locked for writing), it may be more efficient to
* extend the write to cover the entire page in order to avoid fragmentation
* inefficiencies.
*
* If the file is opened for synchronous writes then we can just skip the rest
* of the checks.
*/
static int nfs_can_extend_write(struct file *file, struct folio *folio,
unsigned int pagelen)
{
struct inode *inode = file_inode(file);
struct file_lock_context *flctx = locks_inode_context(inode);
struct file_lock *fl;
int ret;
if (file->f_flags & O_DSYNC)
return 0;
if (!nfs_folio_write_uptodate(folio, pagelen))
return 0;
if (NFS_PROTO(inode)->have_delegation(inode, FMODE_WRITE))
return 1;
if (!flctx || (list_empty_careful(&flctx->flc_flock) &&
list_empty_careful(&flctx->flc_posix)))
return 1;
/* Check to see if there are whole file write locks */
ret = 0;
spin_lock(&flctx->flc_lock);
if (!list_empty(&flctx->flc_posix)) {
fl = list_first_entry(&flctx->flc_posix, struct file_lock,
fl_list);
if (is_whole_file_wrlock(fl))
ret = 1;
} else if (!list_empty(&flctx->flc_flock)) {
fl = list_first_entry(&flctx->flc_flock, struct file_lock,
fl_list);
if (fl->fl_type == F_WRLCK)
ret = 1;
}
spin_unlock(&flctx->flc_lock);
return ret;
}
/*
* Update and possibly write a cached page of an NFS file.
*
* XXX: Keep an eye on generic_file_read to make sure it doesn't do bad
* things with a page scheduled for an RPC call (e.g. invalidate it).
*/
int nfs_update_folio(struct file *file, struct folio *folio,
unsigned int offset, unsigned int count)
{
struct nfs_open_context *ctx = nfs_file_open_context(file);
struct address_space *mapping = folio_file_mapping(folio);
struct inode *inode = mapping->host;
unsigned int pagelen = nfs_folio_length(folio);
int status = 0;
nfs_inc_stats(inode, NFSIOS_VFSUPDATEPAGE);
dprintk("NFS: nfs_update_folio(%pD2 %d@%lld)\n", file, count,
(long long)(folio_file_pos(folio) + offset));
if (!count)
goto out;
if (nfs_can_extend_write(file, folio, pagelen)) {
count = max(count + offset, pagelen);
offset = 0;
}
status = nfs_writepage_setup(ctx, folio, offset, count);
if (status < 0)
nfs_set_pageerror(mapping);
out:
dprintk("NFS: nfs_update_folio returns %d (isize %lld)\n",
status, (long long)i_size_read(inode));
return status;
}
static int flush_task_priority(int how)
{
switch (how & (FLUSH_HIGHPRI|FLUSH_LOWPRI)) {
case FLUSH_HIGHPRI:
return RPC_PRIORITY_HIGH;
case FLUSH_LOWPRI:
return RPC_PRIORITY_LOW;
}
return RPC_PRIORITY_NORMAL;
}
static void nfs_initiate_write(struct nfs_pgio_header *hdr,
struct rpc_message *msg,
const struct nfs_rpc_ops *rpc_ops,
struct rpc_task_setup *task_setup_data, int how)
{
int priority = flush_task_priority(how);
SUNRPC: improve 'swap' handling: scheduling and PF_MEMALLOC rpc tasks can be marked as RPC_TASK_SWAPPER. This causes GFP_MEMALLOC to be used for some allocations. This is needed in some cases, but not in all where it is currently provided, and in some where it isn't provided. Currently *all* tasks associated with a rpc_client on which swap is enabled get the flag and hence some GFP_MEMALLOC support. GFP_MEMALLOC is provided for ->buf_alloc() but only swap-writes need it. However xdr_alloc_bvec does not get GFP_MEMALLOC - though it often does need it. xdr_alloc_bvec is called while the XPRT_LOCK is held. If this blocks, then it blocks all other queued tasks. So this allocation needs GFP_MEMALLOC for *all* requests, not just writes, when the xprt is used for any swap writes. Similarly, if the transport is not connected, that will block all requests including swap writes, so memory allocations should get GFP_MEMALLOC if swap writes are possible. So with this patch: 1/ we ONLY set RPC_TASK_SWAPPER for swap writes. 2/ __rpc_execute() sets PF_MEMALLOC while handling any task with RPC_TASK_SWAPPER set, or when handling any task that holds the XPRT_LOCKED lock on an xprt used for swap. This removes the need for the RPC_IS_SWAPPER() test in ->buf_alloc handlers. 3/ xprt_prepare_transmit() sets PF_MEMALLOC after locking any task to a swapper xprt. __rpc_execute() will clear it. 3/ PF_MEMALLOC is set for all the connect workers. Reviewed-by: Chuck Lever <chuck.lever@oracle.com> (for xprtrdma parts) Signed-off-by: NeilBrown <neilb@suse.de> Signed-off-by: Trond Myklebust <trond.myklebust@hammerspace.com>
2022-03-06 23:41:44 +00:00
if (IS_SWAPFILE(hdr->inode))
task_setup_data->flags |= RPC_TASK_SWAPPER;
task_setup_data->priority = priority;
rpc_ops->write_setup(hdr, msg, &task_setup_data->rpc_client);
trace_nfs_initiate_write(hdr);
}
/* If a nfs_flush_* function fails, it should remove reqs from @head and
* call this on each, which will prepare them to be retried on next
* writeback using standard nfs.
*/
static void nfs_redirty_request(struct nfs_page *req)
{
struct nfs_inode *nfsi = NFS_I(nfs_page_to_inode(req));
/* Bump the transmission count */
req->wb_nio++;
nfs_mark_request_dirty(req);
atomic_long_inc(&nfsi->redirtied_pages);
nfs_page_end_writeback(req);
nfs_release_request(req);
}
static void nfs_async_write_error(struct list_head *head, int error)
{
struct nfs_page *req;
while (!list_empty(head)) {
req = nfs_list_entry(head->next);
nfs_list_remove_request(req);
if (nfs_error_is_fatal_on_server(error))
nfs_write_error(req, error);
else
nfs_redirty_request(req);
}
}
static void nfs_async_write_reschedule_io(struct nfs_pgio_header *hdr)
{
nfs_async_write_error(&hdr->pages, 0);
}
static const struct nfs_pgio_completion_ops nfs_async_write_completion_ops = {
.init_hdr = nfs_async_write_init,
.error_cleanup = nfs_async_write_error,
.completion = nfs_write_completion,
.reschedule_io = nfs_async_write_reschedule_io,
};
void nfs_pageio_init_write(struct nfs_pageio_descriptor *pgio,
struct inode *inode, int ioflags, bool force_mds,
const struct nfs_pgio_completion_ops *compl_ops)
{
struct nfs_server *server = NFS_SERVER(inode);
const struct nfs_pageio_ops *pg_ops = &nfs_pgio_rw_ops;
#ifdef CONFIG_NFS_V4_1
if (server->pnfs_curr_ld && !force_mds)
pg_ops = server->pnfs_curr_ld->pg_write_ops;
#endif
nfs_pageio_init(pgio, inode, pg_ops, compl_ops, &nfs_rw_write_ops,
server->wsize, ioflags);
}
EXPORT_SYMBOL_GPL(nfs_pageio_init_write);
void nfs_pageio_reset_write_mds(struct nfs_pageio_descriptor *pgio)
{
struct nfs_pgio_mirror *mirror;
if (pgio->pg_ops && pgio->pg_ops->pg_cleanup)
pgio->pg_ops->pg_cleanup(pgio);
pgio->pg_ops = &nfs_pgio_rw_ops;
nfs_pageio_stop_mirroring(pgio);
mirror = &pgio->pg_mirrors[0];
mirror->pg_bsize = NFS_SERVER(pgio->pg_inode)->wsize;
}
EXPORT_SYMBOL_GPL(nfs_pageio_reset_write_mds);
void nfs_commit_prepare(struct rpc_task *task, void *calldata)
{
struct nfs_commit_data *data = calldata;
NFS_PROTO(data->inode)->commit_rpc_prepare(task, data);
}
static void nfs_writeback_check_extend(struct nfs_pgio_header *hdr,
struct nfs_fattr *fattr)
{
struct nfs_pgio_args *argp = &hdr->args;
struct nfs_pgio_res *resp = &hdr->res;
u64 size = argp->offset + resp->count;
if (!(fattr->valid & NFS_ATTR_FATTR_SIZE))
fattr->size = size;
if (nfs_size_to_loff_t(fattr->size) < i_size_read(hdr->inode)) {
fattr->valid &= ~NFS_ATTR_FATTR_SIZE;
return;
}
if (size != fattr->size)
return;
/* Set attribute barrier */
nfs_fattr_set_barrier(fattr);
/* ...and update size */
fattr->valid |= NFS_ATTR_FATTR_SIZE;
}
void nfs_writeback_update_inode(struct nfs_pgio_header *hdr)
{
struct nfs_fattr *fattr = &hdr->fattr;
struct inode *inode = hdr->inode;
spin_lock(&inode->i_lock);
nfs_writeback_check_extend(hdr, fattr);
nfs_post_op_update_inode_force_wcc_locked(inode, fattr);
spin_unlock(&inode->i_lock);
}
EXPORT_SYMBOL_GPL(nfs_writeback_update_inode);
/*
* This function is called when the WRITE call is complete.
*/
static int nfs_writeback_done(struct rpc_task *task,
struct nfs_pgio_header *hdr,
struct inode *inode)
{
int status;
/*
* ->write_done will attempt to use post-op attributes to detect
* conflicting writes by other clients. A strict interpretation
* of close-to-open would allow us to continue caching even if
* another writer had changed the file, but some applications
* depend on tighter cache coherency when writing.
*/
status = NFS_PROTO(inode)->write_done(task, hdr);
if (status != 0)
return status;
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
nfs_add_stats(inode, NFSIOS_SERVERWRITTENBYTES, hdr->res.count);
trace_nfs_writeback_done(task, hdr);
if (task->tk_status >= 0) {
enum nfs3_stable_how committed = hdr->res.verf->committed;
if (committed == NFS_UNSTABLE) {
/*
* We have some uncommitted data on the server at
* this point, so ensure that we keep track of that
* fact irrespective of what later writes do.
*/
set_bit(NFS_IOHDR_UNSTABLE_WRITES, &hdr->flags);
}
if (committed < hdr->args.stable) {
/* We tried a write call, but the server did not
* commit data to stable storage even though we
* requested it.
* Note: There is a known bug in Tru64 < 5.0 in which
* the server reports NFS_DATA_SYNC, but performs
* NFS_FILE_SYNC. We therefore implement this checking
* as a dprintk() in order to avoid filling syslog.
*/
static unsigned long complain;
/* Note this will print the MDS for a DS write */
if (time_before(complain, jiffies)) {
dprintk("NFS: faulty NFS server %s:"
" (committed = %d) != (stable = %d)\n",
NFS_SERVER(inode)->nfs_client->cl_hostname,
committed, hdr->args.stable);
complain = jiffies + 300 * HZ;
}
}
}
/* Deal with the suid/sgid bit corner case */
if (nfs_should_remove_suid(inode)) {
spin_lock(&inode->i_lock);
nfs_set_cache_invalid(inode, NFS_INO_INVALID_MODE);
spin_unlock(&inode->i_lock);
}
return 0;
}
/*
* This function is called when the WRITE call is complete.
*/
static void nfs_writeback_result(struct rpc_task *task,
struct nfs_pgio_header *hdr)
{
struct nfs_pgio_args *argp = &hdr->args;
struct nfs_pgio_res *resp = &hdr->res;
if (resp->count < argp->count) {
static unsigned long complain;
/* This a short write! */
nfs_inc_stats(hdr->inode, NFSIOS_SHORTWRITE);
/* Has the server at least made some progress? */
if (resp->count == 0) {
if (time_before(complain, jiffies)) {
printk(KERN_WARNING
"NFS: Server wrote zero bytes, expected %u.\n",
argp->count);
complain = jiffies + 300 * HZ;
}
nfs_set_pgio_error(hdr, -EIO, argp->offset);
task->tk_status = -EIO;
return;
}
NFSv4.1/pnfs: Retry through MDS when getting bad length of data If non rpc-based layout driver return bad length of data, nfs retries by calling rpc_restart_call_prepare() that cause an NULL reference panic. This patch lets nfs retry through MDS for non rpc-based layout driver return bad length of data. [13034.883329] BUG: unable to handle kernel NULL pointer dereference at (null) [13034.884902] IP: [<ffffffffa00db372>] rpc_restart_call_prepare+0x62/0x90 [sunrpc] [13034.886558] PGD 0 [13034.888126] Oops: 0000 [#1] KASAN [13034.889710] Modules linked in: blocklayoutdriver(OE) nfsv4(OE) nfs(OE) fscache(E) nfsd(OE) xfs libcrc32c coretemp btrfs crct10dif_pclmul crc32_pclmul crc32c_intel ghash_clmulni_intel ppdev vmw_balloon auth_rpcgss shpchp nfs_acl lockd vmw_vmci parport_pc xor raid6_pq grace parport sunrpc i2c_piix4 vmwgfx drm_kms_helper ttm drm mptspi e1000 serio_raw scsi_transport_spi mptscsih mptbase ata_generic pata_acpi [last unloaded: fscache] [13034.898260] CPU: 0 PID: 10112 Comm: kworker/0:1 Tainted: G OE 4.3.0-rc5+ #279 [13034.899932] Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 07/02/2015 [13034.903342] Workqueue: events bl_read_cleanup [blocklayoutdriver] [13034.905059] task: ffff88006a9148c0 ti: ffff880035e90000 task.ti: ffff880035e90000 [13034.906827] RIP: 0010:[<ffffffffa00db372>] [<ffffffffa00db372>] rpc_restart_call_prepare+0x62/0x90 [sunrpc] [13034.910522] RSP: 0018:ffff880035e97b58 EFLAGS: 00010282 [13034.912378] RAX: fffffbfff04a5a94 RBX: ffff880068fe4858 RCX: 0000000000000003 [13034.914339] RDX: dffffc0000000000 RSI: 0000000000000003 RDI: 0000000000000282 [13034.916236] RBP: ffff880035e97b68 R08: 0000000000000001 R09: 0000000000000001 [13034.918229] R10: 0000000000000000 R11: 0000000000000001 R12: 0000000000000000 [13034.920007] R13: ffff880068fe4858 R14: ffff880068fe4a60 R15: 0000000000001000 [13034.921845] FS: 0000000000000000(0000) GS:ffffffff82247000(0000) knlGS:0000000000000000 [13034.923645] CS: 0010 DS: 0000 ES: 0000 CR0: 0000000080050033 [13034.925525] CR2: 0000000000000000 CR3: 00000000063dd000 CR4: 00000000001406f0 [13034.932808] Stack: [13034.934813] ffff880068fe4780 0000000000001000 ffff880035e97ba8 ffffffffa08800d2 [13034.936675] ffffffffa088029d ffff880068fe4780 ffff880068fe4858 ffffffffa089c0a0 [13034.938593] ffff880068fe47e0 ffff88005d59faf0 ffff880035e97be0 ffffffffa087e08f [13034.940454] Call Trace: [13034.942388] [<ffffffffa08800d2>] nfs_readpage_result+0x112/0x200 [nfs] [13034.944317] [<ffffffffa088029d>] ? nfs_readpage_done+0xdd/0x160 [nfs] [13034.946267] [<ffffffffa087e08f>] nfs_pgio_result+0x9f/0x120 [nfs] [13034.948166] [<ffffffffa09266cc>] pnfs_ld_read_done+0x7c/0x1e0 [nfsv4] [13034.950247] [<ffffffffa03b07ee>] bl_read_cleanup+0x2e/0x60 [blocklayoutdriver] [13034.952156] [<ffffffff810ebf62>] process_one_work+0x412/0x870 [13034.954102] [<ffffffff810ebe84>] ? process_one_work+0x334/0x870 [13034.955949] [<ffffffff810ebb50>] ? queue_delayed_work_on+0x40/0x40 [13034.957985] [<ffffffff810ec441>] worker_thread+0x81/0x6a0 [13034.959817] [<ffffffff810ec3c0>] ? process_one_work+0x870/0x870 [13034.961785] [<ffffffff810f43bd>] kthread+0x17d/0x1a0 [13034.963544] [<ffffffff810f4240>] ? kthread_create_on_node+0x330/0x330 [13034.965479] [<ffffffff81100428>] ? finish_task_switch+0x88/0x220 [13034.967223] [<ffffffff810f4240>] ? kthread_create_on_node+0x330/0x330 [13034.968929] [<ffffffff81b6ae5f>] ret_from_fork+0x3f/0x70 [13034.970534] [<ffffffff810f4240>] ? kthread_create_on_node+0x330/0x330 [13034.972176] Code: c7 43 50 40 84 0d a0 e8 3d fe 1c e1 48 8d 7b 58 c7 83 e4 00 00 00 00 00 00 00 e8 ca fe 1c e1 4c 8b 63 58 4c 89 e7 e8 be fe 1c e1 <49> 83 3c 24 00 74 12 48 c7 43 50 f0 a2 0e a0 b8 01 00 00 00 5b [13034.977148] RIP [<ffffffffa00db372>] rpc_restart_call_prepare+0x62/0x90 [sunrpc] [13034.978780] RSP <ffff880035e97b58> [13034.980399] CR2: 0000000000000000 Signed-off-by: Kinglong Mee <kinglongmee@gmail.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2015-10-16 09:23:29 +00:00
/* For non rpc-based layout drivers, retry-through-MDS */
if (!task->tk_ops) {
hdr->pnfs_error = -EAGAIN;
return;
}
/* Was this an NFSv2 write or an NFSv3 stable write? */
if (resp->verf->committed != NFS_UNSTABLE) {
/* Resend from where the server left off */
hdr->mds_offset += resp->count;
argp->offset += resp->count;
argp->pgbase += resp->count;
argp->count -= resp->count;
} else {
/* Resend as a stable write in order to avoid
* headaches in the case of a server crash.
*/
argp->stable = NFS_FILE_SYNC;
}
resp->count = 0;
resp->verf->committed = 0;
rpc_restart_call_prepare(task);
}
}
static int wait_on_commit(struct nfs_mds_commit_info *cinfo)
{
return wait_var_event_killable(&cinfo->rpcs_out,
!atomic_read(&cinfo->rpcs_out));
}
static void nfs_commit_begin(struct nfs_mds_commit_info *cinfo)
{
atomic_inc(&cinfo->rpcs_out);
}
bool nfs_commit_end(struct nfs_mds_commit_info *cinfo)
{
if (atomic_dec_and_test(&cinfo->rpcs_out)) {
wake_up_var(&cinfo->rpcs_out);
return true;
}
return false;
}
void nfs_commitdata_release(struct nfs_commit_data *data)
{
put_nfs_open_context(data->context);
nfs_commit_free(data);
}
EXPORT_SYMBOL_GPL(nfs_commitdata_release);
int nfs_initiate_commit(struct rpc_clnt *clnt, struct nfs_commit_data *data,
const struct nfs_rpc_ops *nfs_ops,
const struct rpc_call_ops *call_ops,
int how, int flags)
{
struct rpc_task *task;
int priority = flush_task_priority(how);
struct rpc_message msg = {
.rpc_argp = &data->args,
.rpc_resp = &data->res,
.rpc_cred = data->cred,
};
struct rpc_task_setup task_setup_data = {
.task = &data->task,
.rpc_client = clnt,
.rpc_message = &msg,
.callback_ops = call_ops,
.callback_data = data,
.workqueue = nfsiod_workqueue,
.flags = RPC_TASK_ASYNC | flags,
.priority = priority,
};
if (nfs_server_capable(data->inode, NFS_CAP_MOVEABLE))
task_setup_data.flags |= RPC_TASK_MOVEABLE;
/* Set up the initial task struct. */
nfs_ops->commit_setup(data, &msg, &task_setup_data.rpc_client);
NFS: Add static NFS I/O tracepoints Tools like tcpdump and rpcdebug can be very useful. But there are plenty of environments where they are difficult or impossible to use. For example, we've had customers report I/O failures during workloads so heavy that collecting network traffic or enabling RPC debugging are themselves onerous. The kernel's static tracepoints are lightweight (less likely to introduce timing changes) and efficient (the trace data is compact). They also work in scenarios where capturing network traffic is not possible due to lack of hardware support (some InfiniBand HCAs) or where data or network privacy is a concern. Introduce tracepoints that show when an NFS READ, WRITE, or COMMIT is initiated, and when it completes. Record the arguments and results of each operation, which are not shown by existing sunrpc module's tracepoints. For instance, the recorded offset and count can be used to match an "initiate" event to a "done" event. If an NFS READ result returns fewer bytes than requested or zero, seeing the EOF flag can be probative. Seeing an NFS4ERR_BAD_STATEID result is also indication of a particular class of problems. The timing information attached to each event record can often be useful as well. Usage example: [root@manet tmp]# trace-cmd record -e nfs:*initiate* -e nfs:*done /sys/kernel/debug/tracing/events/nfs/*initiate*/filter /sys/kernel/debug/tracing/events/nfs/*done/filter Hit Ctrl^C to stop recording ^CKernel buffer statistics: Note: "entries" are the entries left in the kernel ring buffer and are not recorded in the trace data. They should all be zero. CPU: 0 entries: 0 overrun: 0 commit overrun: 0 bytes: 3680 oldest event ts: 78.367422 now ts: 100.124419 dropped events: 0 read events: 74 ... and so on. Signed-off-by: Chuck Lever <chuck.lever@oracle.com> Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>
2017-08-21 19:00:49 +00:00
trace_nfs_initiate_commit(data);
dprintk("NFS: initiated commit call\n");
task = rpc_run_task(&task_setup_data);
if (IS_ERR(task))
return PTR_ERR(task);
if (how & FLUSH_SYNC)
rpc_wait_for_completion_task(task);
rpc_put_task(task);
return 0;
}
EXPORT_SYMBOL_GPL(nfs_initiate_commit);
static loff_t nfs_get_lwb(struct list_head *head)
{
loff_t lwb = 0;
struct nfs_page *req;
list_for_each_entry(req, head, wb_list)
if (lwb < (req_offset(req) + req->wb_bytes))
lwb = req_offset(req) + req->wb_bytes;
return lwb;
}
/*
* Set up the argument/result storage required for the RPC call.
*/
void nfs_init_commit(struct nfs_commit_data *data,
struct list_head *head,
struct pnfs_layout_segment *lseg,
struct nfs_commit_info *cinfo)
{
struct nfs_page *first;
struct nfs_open_context *ctx;
struct inode *inode;
/* Set up the RPC argument and reply structs
* NB: take care not to mess about with data->commit et al. */
if (head)
list_splice_init(head, &data->pages);
first = nfs_list_entry(data->pages.next);
ctx = nfs_req_openctx(first);
inode = d_inode(ctx->dentry);
data->inode = inode;
data->cred = ctx->cred;
data->lseg = lseg; /* reference transferred */
/* only set lwb for pnfs commit */
if (lseg)
data->lwb = nfs_get_lwb(&data->pages);
data->mds_ops = &nfs_commit_ops;
data->completion_ops = cinfo->completion_ops;
data->dreq = cinfo->dreq;
data->args.fh = NFS_FH(data->inode);
/* Note: we always request a commit of the entire inode */
data->args.offset = 0;
data->args.count = 0;
data->context = get_nfs_open_context(ctx);
data->res.fattr = &data->fattr;
data->res.verf = &data->verf;
nfs_fattr_init(&data->fattr);
nfs_commit_begin(cinfo->mds);
}
EXPORT_SYMBOL_GPL(nfs_init_commit);
void nfs_retry_commit(struct list_head *page_list,
struct pnfs_layout_segment *lseg,
struct nfs_commit_info *cinfo,
u32 ds_commit_idx)
{
struct nfs_page *req;
while (!list_empty(page_list)) {
req = nfs_list_entry(page_list->next);
nfs_list_remove_request(req);
nfs_mark_request_commit(req, lseg, cinfo, ds_commit_idx);
nfs_folio_clear_commit(nfs_page_to_folio(req));
nfs_unlock_and_release_request(req);
}
}
EXPORT_SYMBOL_GPL(nfs_retry_commit);
static void nfs_commit_resched_write(struct nfs_commit_info *cinfo,
struct nfs_page *req)
{
struct folio *folio = nfs_page_to_folio(req);
filemap_dirty_folio(folio_mapping(folio), folio);
}
/*
* Commit dirty pages
*/
static int
nfs_commit_list(struct inode *inode, struct list_head *head, int how,
struct nfs_commit_info *cinfo)
{
struct nfs_commit_data *data;
unsigned short task_flags = 0;
nfs: avoid race that crashes nfs_init_commit Since the patch "NFS: Allow multiple commit requests in flight per file" we can run multiple simultaneous commits on the same inode. This introduced a race over collecting pages to commit that made it possible to call nfs_init_commit() with an empty list - which causes crashes like the one below. The fix is to catch this race and avoid calling nfs_init_commit and initiate_commit when there is no work to do. Here is the crash: [600522.076832] BUG: unable to handle kernel NULL pointer dereference at 0000000000000040 [600522.078475] IP: [<ffffffffa0479e72>] nfs_init_commit+0x22/0x130 [nfs] [600522.078745] PGD 4272b1067 PUD 4272cb067 PMD 0 [600522.078972] Oops: 0000 [#1] SMP [600522.079204] Modules linked in: nfsv3 nfs_layout_flexfiles rpcsec_gss_krb5 nfsv4 dns_resolver nfs fscache dcdbas ip6t_rpfilter ip6t_REJECT nf_reject_ipv6 xt_conntrack ebtable_nat ebtable_broute bridge stp llc ebtable_filter ebtables ip6table_nat nf_conntrack_ipv6 nf_defrag_ipv6 nf_nat_ipv6 ip6table_mangle ip6table_security ip6table_raw ip6table_filter ip6_tables iptable_nat nf_conntrack_ipv4 nf_defrag_ipv4 nf_nat_ipv4 nf_nat nf_conntrack iptable_mangle iptable_security iptable_raw vmw_vsock_vmci_transport vsock bonding ipmi_devintf ipmi_msghandler coretemp crct10dif_pclmul crc32_pclmul ghash_clmulni_intel ppdev vmw_balloon parport_pc parport acpi_cpufreq vmw_vmci i2c_piix4 shpchp nfsd auth_rpcgss nfs_acl lockd grace sunrpc xfs libcrc32c vmwgfx drm_kms_helper ttm drm crc32c_intel serio_raw vmxnet3 [600522.081380] vmw_pvscsi ata_generic pata_acpi [600522.081809] CPU: 3 PID: 15667 Comm: /usr/bin/python Not tainted 4.1.9-100.pd.88.el7.x86_64 #1 [600522.082281] Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 09/30/2014 [600522.082814] task: ffff8800bbbfa780 ti: ffff88042ae84000 task.ti: ffff88042ae84000 [600522.083378] RIP: 0010:[<ffffffffa0479e72>] [<ffffffffa0479e72>] nfs_init_commit+0x22/0x130 [nfs] [600522.083973] RSP: 0018:ffff88042ae87438 EFLAGS: 00010246 [600522.084571] RAX: 0000000000000000 RBX: ffff880003485e40 RCX: ffff88042ae87588 [600522.085188] RDX: 0000000000000000 RSI: ffff88042ae874b0 RDI: ffff880003485e40 [600522.085756] RBP: ffff88042ae87448 R08: ffff880003486010 R09: ffff88042ae874b0 [600522.086332] R10: 0000000000000000 R11: 0000000000000005 R12: ffff88042ae872d0 [600522.086905] R13: ffff88042ae874b0 R14: ffff880003485e40 R15: ffff88042704c840 [600522.087484] FS: 00007f4728ff2740(0000) GS:ffff88043fd80000(0000) knlGS:0000000000000000 [600522.088070] CS: 0010 DS: 0000 ES: 0000 CR0: 000000008005003b [600522.088663] CR2: 0000000000000040 CR3: 000000042b6aa000 CR4: 00000000001406e0 [600522.089327] Stack: [600522.089926] 0000000000000001 ffff88042ae87588 ffff88042ae874f8 ffffffffa04f09fa [600522.090549] 0000000000017840 0000000000017840 ffff88042ae87588 ffff8803258d9930 [600522.091169] ffff88042ae87578 ffffffffa0563d80 0000000000000000 ffff88042704c840 [600522.091789] Call Trace: [600522.092420] [<ffffffffa04f09fa>] pnfs_generic_commit_pagelist+0x1da/0x320 [nfsv4] [600522.093052] [<ffffffffa0563d80>] ? ff_layout_commit_prepare_v3+0x30/0x30 [nfs_layout_flexfiles] [600522.093696] [<ffffffffa0562645>] ff_layout_commit_pagelist+0x15/0x20 [nfs_layout_flexfiles] [600522.094359] [<ffffffffa047bc78>] nfs_generic_commit_list+0xe8/0x120 [nfs] [600522.095032] [<ffffffffa047bd6a>] nfs_commit_inode+0xba/0x110 [nfs] [600522.095719] [<ffffffffa046ac54>] nfs_release_page+0x44/0xd0 [nfs] [600522.096410] [<ffffffff811a8122>] try_to_release_page+0x32/0x50 [600522.097109] [<ffffffff811bd4f1>] shrink_page_list+0x961/0xb30 [600522.097812] [<ffffffff811bdced>] shrink_inactive_list+0x1cd/0x550 [600522.098530] [<ffffffff811bea65>] shrink_lruvec+0x635/0x840 [600522.099250] [<ffffffff811bed60>] shrink_zone+0xf0/0x2f0 [600522.099974] [<ffffffff811bf312>] do_try_to_free_pages+0x192/0x470 [600522.100709] [<ffffffff811bf6ca>] try_to_free_pages+0xda/0x170 [600522.101464] [<ffffffff811b2198>] __alloc_pages_nodemask+0x588/0x970 [600522.102235] [<ffffffff811fbbd5>] alloc_pages_vma+0xb5/0x230 [600522.103000] [<ffffffff813a1589>] ? cpumask_any_but+0x39/0x50 [600522.103774] [<ffffffff811d6115>] wp_page_copy.isra.55+0x95/0x490 [600522.104558] [<ffffffff810e3438>] ? __wake_up+0x48/0x60 [600522.105357] [<ffffffff811d7d3b>] do_wp_page+0xab/0x4f0 [600522.106137] [<ffffffff810a1bbb>] ? release_task+0x36b/0x470 [600522.106902] [<ffffffff8126dbd7>] ? eventfd_ctx_read+0x67/0x1c0 [600522.107659] [<ffffffff811da2a8>] handle_mm_fault+0xc78/0x1900 [600522.108431] [<ffffffff81067ef1>] __do_page_fault+0x181/0x420 [600522.109173] [<ffffffff811446a6>] ? __audit_syscall_exit+0x1e6/0x280 [600522.109893] [<ffffffff810681c0>] do_page_fault+0x30/0x80 [600522.110594] [<ffffffff81024f36>] ? syscall_trace_leave+0xc6/0x120 [600522.111288] [<ffffffff81790a58>] page_fault+0x28/0x30 [600522.111947] Code: 5d c3 0f 1f 80 00 00 00 00 0f 1f 44 00 00 55 4c 8d 87 d0 01 00 00 48 89 e5 53 48 89 fb 48 83 ec 08 4c 8b 0e 49 8b 41 18 4c 39 ce <48> 8b 40 40 4c 8b 50 30 74 24 48 8b 87 d0 01 00 00 48 8b 7e 08 [600522.113343] RIP [<ffffffffa0479e72>] nfs_init_commit+0x22/0x130 [nfs] [600522.114003] RSP <ffff88042ae87438> [600522.114636] CR2: 0000000000000040 Fixes: af7cf057 (NFS: Allow multiple commit requests in flight per file) CC: stable@vger.kernel.org Signed-off-by: Weston Andros Adamson <dros@primarydata.com> Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2016-05-25 14:07:23 +00:00
/* another commit raced with us */
if (list_empty(head))
return 0;
data = nfs_commitdata_alloc();
if (!data) {
nfs_retry_commit(head, NULL, cinfo, -1);
return -ENOMEM;
}
/* Set up the argument struct */
nfs_init_commit(data, head, NULL, cinfo);
if (NFS_SERVER(inode)->nfs_client->cl_minorversion)
task_flags = RPC_TASK_MOVEABLE;
return nfs_initiate_commit(NFS_CLIENT(inode), data, NFS_PROTO(inode),
data->mds_ops, how,
RPC_TASK_CRED_NOREF | task_flags);
}
/*
* COMMIT call returned
*/
static void nfs_commit_done(struct rpc_task *task, void *calldata)
{
struct nfs_commit_data *data = calldata;
/* Call the NFS version-specific code */
NFS_PROTO(data->inode)->commit_done(task, data);
trace_nfs_commit_done(task, data);
}
static void nfs_commit_release_pages(struct nfs_commit_data *data)
{
const struct nfs_writeverf *verf = data->res.verf;
struct nfs_page *req;
int status = data->task.tk_status;
struct nfs_commit_info cinfo;
struct nfs_server *nfss;
struct folio *folio;
while (!list_empty(&data->pages)) {
req = nfs_list_entry(data->pages.next);
nfs_list_remove_request(req);
folio = nfs_page_to_folio(req);
nfs_folio_clear_commit(folio);
dprintk("NFS: commit (%s/%llu %d@%lld)",
nfs_req_openctx(req)->dentry->d_sb->s_id,
(unsigned long long)NFS_FILEID(d_inode(nfs_req_openctx(req)->dentry)),
req->wb_bytes,
(long long)req_offset(req));
if (status < 0) {
if (folio) {
trace_nfs_commit_error(data->inode, req,
status);
nfs_mapping_set_error(folio, status);
NFS: fix the fault nrequests decreasing for nfs_inode COPY The nfs_commit_file for NFSv4.2's COPY operation goes through the commit path for normal WRITE, but without increase nrequests, so, the nrequests decreased in nfs_commit_release_pages is fault. After that, the nrequests will be wrong. [ 5670.299881] ------------[ cut here ]------------ [ 5670.300295] WARNING: CPU: 0 PID: 27656 at fs/nfs/inode.c:127 nfs_clear_inode+0x66/0x90 [nfs] [ 5670.300558] Modules linked in: nfsv4(E) nfs(E) fscache(E) tun bridge stp llc fuse ip_set nfnetlink vmw_vsock_vmci_transport vsock snd_seq_midi snd_seq_midi_event ppdev f2fs coretemp crct10dif_pclmul crc32_pclmul ghash_clmulni_intel snd_ens1371 intel_rapl_perf gameport snd_ac97_codec vmw_balloon ac97_bus snd_seq snd_pcm joydev snd_rawmidi snd_timer snd_seq_device snd soundcore nfit parport_pc parport acpi_cpufreq tpm_tis tpm_tis_core tpm i2c_piix4 vmw_vmci shpchp nfsd auth_rpcgss nfs_acl lockd grace sunrpc xfs libcrc32c vmwgfx drm_kms_helper ttm drm e1000 crc32c_intel mptspi scsi_transport_spi serio_raw mptscsih mptbase ata_generic pata_acpi fjes [last unloaded: fscache] [ 5670.302925] CPU: 0 PID: 27656 Comm: umount.nfs4 Tainted: G W E 4.11.0-rc1+ #519 [ 5670.303292] Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 07/02/2015 [ 5670.304094] Call Trace: [ 5670.304510] dump_stack+0x63/0x86 [ 5670.304917] __warn+0xcb/0xf0 [ 5670.305276] warn_slowpath_null+0x1d/0x20 [ 5670.305661] nfs_clear_inode+0x66/0x90 [nfs] [ 5670.306093] nfs4_evict_inode+0x61/0x70 [nfsv4] [ 5670.306480] evict+0xbb/0x1c0 [ 5670.306888] dispose_list+0x4d/0x70 [ 5670.307233] evict_inodes+0x178/0x1a0 [ 5670.307579] generic_shutdown_super+0x44/0xf0 [ 5670.307985] nfs_kill_super+0x21/0x40 [nfs] [ 5670.308325] deactivate_locked_super+0x43/0x70 [ 5670.308698] deactivate_super+0x5a/0x60 [ 5670.309036] cleanup_mnt+0x3f/0x90 [ 5670.309407] __cleanup_mnt+0x12/0x20 [ 5670.309837] task_work_run+0x80/0xa0 [ 5670.310162] exit_to_usermode_loop+0x89/0x90 [ 5670.310497] syscall_return_slowpath+0xaa/0xb0 [ 5670.310875] entry_SYSCALL_64_fastpath+0xa7/0xa9 [ 5670.311197] RIP: 0033:0x7f1bb3617fe7 [ 5670.311545] RSP: 002b:00007ffecbabb828 EFLAGS: 00000206 ORIG_RAX: 00000000000000a6 [ 5670.311906] RAX: 0000000000000000 RBX: 0000000001dca1f0 RCX: 00007f1bb3617fe7 [ 5670.312239] RDX: 000000000000000c RSI: 0000000000000001 RDI: 0000000001dc83c0 [ 5670.312653] RBP: 0000000001dc83c0 R08: 0000000000000001 R09: 0000000000000000 [ 5670.312998] R10: 0000000000000755 R11: 0000000000000206 R12: 00007ffecbabc66a [ 5670.313335] R13: 0000000001dc83a0 R14: 0000000000000000 R15: 0000000000000000 [ 5670.313758] ---[ end trace bf4bfe7764e4eb40 ]--- Cc: linux-kernel@vger.kernel.org Fixes: 67911c8f18 ("NFS: Add nfs_commit_file()") Signed-off-by: Kinglong Mee <kinglongmee@gmail.com> Cc: stable@vger.kernel.org # 4.7+ Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2017-03-09 03:36:36 +00:00
nfs_inode_remove_request(req);
}
dprintk_cont(", error = %d\n", status);
goto next;
}
/* Okay, COMMIT succeeded, apparently. Check the verifier
* returned by the server against all stored verfs. */
if (nfs_write_match_verf(verf, req)) {
/* We have a match */
if (folio)
NFS: fix the fault nrequests decreasing for nfs_inode COPY The nfs_commit_file for NFSv4.2's COPY operation goes through the commit path for normal WRITE, but without increase nrequests, so, the nrequests decreased in nfs_commit_release_pages is fault. After that, the nrequests will be wrong. [ 5670.299881] ------------[ cut here ]------------ [ 5670.300295] WARNING: CPU: 0 PID: 27656 at fs/nfs/inode.c:127 nfs_clear_inode+0x66/0x90 [nfs] [ 5670.300558] Modules linked in: nfsv4(E) nfs(E) fscache(E) tun bridge stp llc fuse ip_set nfnetlink vmw_vsock_vmci_transport vsock snd_seq_midi snd_seq_midi_event ppdev f2fs coretemp crct10dif_pclmul crc32_pclmul ghash_clmulni_intel snd_ens1371 intel_rapl_perf gameport snd_ac97_codec vmw_balloon ac97_bus snd_seq snd_pcm joydev snd_rawmidi snd_timer snd_seq_device snd soundcore nfit parport_pc parport acpi_cpufreq tpm_tis tpm_tis_core tpm i2c_piix4 vmw_vmci shpchp nfsd auth_rpcgss nfs_acl lockd grace sunrpc xfs libcrc32c vmwgfx drm_kms_helper ttm drm e1000 crc32c_intel mptspi scsi_transport_spi serio_raw mptscsih mptbase ata_generic pata_acpi fjes [last unloaded: fscache] [ 5670.302925] CPU: 0 PID: 27656 Comm: umount.nfs4 Tainted: G W E 4.11.0-rc1+ #519 [ 5670.303292] Hardware name: VMware, Inc. VMware Virtual Platform/440BX Desktop Reference Platform, BIOS 6.00 07/02/2015 [ 5670.304094] Call Trace: [ 5670.304510] dump_stack+0x63/0x86 [ 5670.304917] __warn+0xcb/0xf0 [ 5670.305276] warn_slowpath_null+0x1d/0x20 [ 5670.305661] nfs_clear_inode+0x66/0x90 [nfs] [ 5670.306093] nfs4_evict_inode+0x61/0x70 [nfsv4] [ 5670.306480] evict+0xbb/0x1c0 [ 5670.306888] dispose_list+0x4d/0x70 [ 5670.307233] evict_inodes+0x178/0x1a0 [ 5670.307579] generic_shutdown_super+0x44/0xf0 [ 5670.307985] nfs_kill_super+0x21/0x40 [nfs] [ 5670.308325] deactivate_locked_super+0x43/0x70 [ 5670.308698] deactivate_super+0x5a/0x60 [ 5670.309036] cleanup_mnt+0x3f/0x90 [ 5670.309407] __cleanup_mnt+0x12/0x20 [ 5670.309837] task_work_run+0x80/0xa0 [ 5670.310162] exit_to_usermode_loop+0x89/0x90 [ 5670.310497] syscall_return_slowpath+0xaa/0xb0 [ 5670.310875] entry_SYSCALL_64_fastpath+0xa7/0xa9 [ 5670.311197] RIP: 0033:0x7f1bb3617fe7 [ 5670.311545] RSP: 002b:00007ffecbabb828 EFLAGS: 00000206 ORIG_RAX: 00000000000000a6 [ 5670.311906] RAX: 0000000000000000 RBX: 0000000001dca1f0 RCX: 00007f1bb3617fe7 [ 5670.312239] RDX: 000000000000000c RSI: 0000000000000001 RDI: 0000000001dc83c0 [ 5670.312653] RBP: 0000000001dc83c0 R08: 0000000000000001 R09: 0000000000000000 [ 5670.312998] R10: 0000000000000755 R11: 0000000000000206 R12: 00007ffecbabc66a [ 5670.313335] R13: 0000000001dc83a0 R14: 0000000000000000 R15: 0000000000000000 [ 5670.313758] ---[ end trace bf4bfe7764e4eb40 ]--- Cc: linux-kernel@vger.kernel.org Fixes: 67911c8f18 ("NFS: Add nfs_commit_file()") Signed-off-by: Kinglong Mee <kinglongmee@gmail.com> Cc: stable@vger.kernel.org # 4.7+ Signed-off-by: Anna Schumaker <Anna.Schumaker@Netapp.com>
2017-03-09 03:36:36 +00:00
nfs_inode_remove_request(req);
dprintk_cont(" OK\n");
goto next;
}
/* We have a mismatch. Write the page again */
dprintk_cont(" mismatch\n");
nfs_mark_request_dirty(req);
atomic_long_inc(&NFS_I(data->inode)->redirtied_pages);
next:
nfs_unlock_and_release_request(req);
/* Latency breaker */
cond_resched();
}
nfss = NFS_SERVER(data->inode);
if (atomic_long_read(&nfss->writeback) < NFS_CONGESTION_OFF_THRESH)
nfs: remove reliance on bdi congestion The bdi congestion tracking in not widely used and will be removed. NFS is one of a small number of filesystems that uses it, setting just the async (write) congestion flag at what it determines are appropriate times. The only remaining effect of the async flag is to cause (some) WB_SYNC_NONE writes to be skipped. So instead of setting the flag, set an internal flag and change: - .writepages to do nothing if WB_SYNC_NONE and the flag is set - .writepage to return AOP_WRITEPAGE_ACTIVATE if WB_SYNC_NONE and the flag is set. The writepages change causes a behavioural change in that pageout() can now return PAGE_ACTIVATE instead of PAGE_KEEP, so SetPageActive() will be called on the page which (I think) wil further delay the next attempt at writeout. This might be a good thing. Link: https://lkml.kernel.org/r/164549983738.9187.3972219847989393182.stgit@noble.brown Signed-off-by: NeilBrown <neilb@suse.de> Cc: Anna Schumaker <Anna.Schumaker@Netapp.com> Cc: Chao Yu <chao@kernel.org> Cc: Darrick J. Wong <djwong@kernel.org> Cc: Ilya Dryomov <idryomov@gmail.com> Cc: Jaegeuk Kim <jaegeuk@kernel.org> Cc: Jan Kara <jack@suse.cz> Cc: Jeff Layton <jlayton@kernel.org> Cc: Jens Axboe <axboe@kernel.dk> Cc: Lars Ellenberg <lars.ellenberg@linbit.com> Cc: Miklos Szeredi <miklos@szeredi.hu> Cc: Paolo Valente <paolo.valente@linaro.org> Cc: Philipp Reisner <philipp.reisner@linbit.com> Cc: Ryusuke Konishi <konishi.ryusuke@gmail.com> Cc: Trond Myklebust <trond.myklebust@hammerspace.com> Cc: Wu Fengguang <fengguang.wu@intel.com> Signed-off-by: Andrew Morton <akpm@linux-foundation.org> Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2022-03-22 21:39:01 +00:00
nfss->write_congested = 0;
nfs_init_cinfo(&cinfo, data->inode, data->dreq);
nfs_commit_end(cinfo.mds);
}
static void nfs_commit_release(void *calldata)
{
struct nfs_commit_data *data = calldata;
data->completion_ops->completion(data);
nfs_commitdata_release(calldata);
}
static const struct rpc_call_ops nfs_commit_ops = {
.rpc_call_prepare = nfs_commit_prepare,
.rpc_call_done = nfs_commit_done,
.rpc_release = nfs_commit_release,
};
static const struct nfs_commit_completion_ops nfs_commit_completion_ops = {
.completion = nfs_commit_release_pages,
.resched_write = nfs_commit_resched_write,
};
int nfs_generic_commit_list(struct inode *inode, struct list_head *head,
int how, struct nfs_commit_info *cinfo)
{
int status;
status = pnfs_commit_list(inode, head, how, cinfo);
if (status == PNFS_NOT_ATTEMPTED)
status = nfs_commit_list(inode, head, how, cinfo);
return status;
}
static int __nfs_commit_inode(struct inode *inode, int how,
struct writeback_control *wbc)
{
LIST_HEAD(head);
struct nfs_commit_info cinfo;
int may_wait = how & FLUSH_SYNC;
int ret, nscan;
how &= ~FLUSH_SYNC;
nfs_init_cinfo_from_inode(&cinfo, inode);
nfs_commit_begin(cinfo.mds);
for (;;) {
ret = nscan = nfs_scan_commit(inode, &head, &cinfo);
if (ret <= 0)
break;
ret = nfs_generic_commit_list(inode, &head, how, &cinfo);
if (ret < 0)
break;
ret = 0;
if (wbc && wbc->sync_mode == WB_SYNC_NONE) {
if (nscan < wbc->nr_to_write)
wbc->nr_to_write -= nscan;
else
wbc->nr_to_write = 0;
}
if (nscan < INT_MAX)
break;
cond_resched();
}
nfs_commit_end(cinfo.mds);
if (ret || !may_wait)
return ret;
return wait_on_commit(cinfo.mds);
}
int nfs_commit_inode(struct inode *inode, int how)
{
return __nfs_commit_inode(inode, how, NULL);
}
EXPORT_SYMBOL_GPL(nfs_commit_inode);
int nfs_write_inode(struct inode *inode, struct writeback_control *wbc)
{
struct nfs_inode *nfsi = NFS_I(inode);
int flags = FLUSH_SYNC;
int ret = 0;
if (wbc->sync_mode == WB_SYNC_NONE) {
/* no commits means nothing needs to be done */
if (!atomic_long_read(&nfsi->commit_info.ncommit))
goto check_requests_outstanding;
/* Don't commit yet if this is a non-blocking flush and there
* are a lot of outstanding writes for this mapping.
*/
if (mapping_tagged(inode->i_mapping, PAGECACHE_TAG_WRITEBACK))
goto out_mark_dirty;
/* don't wait for the COMMIT response */
flags = 0;
}
ret = __nfs_commit_inode(inode, flags, wbc);
if (!ret) {
if (flags & FLUSH_SYNC)
return 0;
} else if (atomic_long_read(&nfsi->commit_info.ncommit))
goto out_mark_dirty;
check_requests_outstanding:
if (!atomic_read(&nfsi->commit_info.rpcs_out))
return ret;
out_mark_dirty:
__mark_inode_dirty(inode, I_DIRTY_DATASYNC);
return ret;
}
EXPORT_SYMBOL_GPL(nfs_write_inode);
/*
* Wrapper for filemap_write_and_wait_range()
*
* Needed for pNFS in order to ensure data becomes visible to the
* client.
*/
int nfs_filemap_write_and_wait_range(struct address_space *mapping,
loff_t lstart, loff_t lend)
{
int ret;
ret = filemap_write_and_wait_range(mapping, lstart, lend);
if (ret == 0)
ret = pnfs_sync_inode(mapping->host, true);
return ret;
}
EXPORT_SYMBOL_GPL(nfs_filemap_write_and_wait_range);
/*
* flush the inode to disk.
*/
int nfs_wb_all(struct inode *inode)
{
int ret;
trace_nfs_writeback_inode_enter(inode);
ret = filemap_write_and_wait(inode->i_mapping);
if (ret)
goto out;
ret = nfs_commit_inode(inode, FLUSH_SYNC);
if (ret < 0)
goto out;
pnfs_sync_inode(inode, true);
ret = 0;
out:
trace_nfs_writeback_inode_exit(inode, ret);
return ret;
}
EXPORT_SYMBOL_GPL(nfs_wb_all);
int nfs_wb_folio_cancel(struct inode *inode, struct folio *folio)
{
struct nfs_page *req;
int ret = 0;
folio_wait_writeback(folio);
/* blocking call to cancel all requests and join to a single (head)
* request */
req = nfs_lock_and_join_requests(folio);
if (IS_ERR(req)) {
ret = PTR_ERR(req);
} else if (req) {
/* all requests from this folio have been cancelled by
* nfs_lock_and_join_requests, so just remove the head
* request from the inode / page_private pointer and
* release it */
nfs_inode_remove_request(req);
nfs_unlock_and_release_request(req);
}
return ret;
}
/**
* nfs_wb_folio - Write back all requests on one page
* @inode: pointer to page
* @folio: pointer to folio
*
* Assumes that the folio has been locked by the caller, and will
* not unlock it.
*/
int nfs_wb_folio(struct inode *inode, struct folio *folio)
{
loff_t range_start = folio_file_pos(folio);
loff_t range_end = range_start + (loff_t)folio_size(folio) - 1;
struct writeback_control wbc = {
.sync_mode = WB_SYNC_ALL,
.nr_to_write = 0,
.range_start = range_start,
.range_end = range_end,
};
int ret;
trace_nfs_writeback_folio(inode, folio);
for (;;) {
folio_wait_writeback(folio);
if (folio_clear_dirty_for_io(folio)) {
ret = nfs_writepage_locked(folio, &wbc);
if (ret < 0)
goto out_error;
continue;
}
ret = 0;
if (!folio_test_private(folio))
break;
ret = nfs_commit_inode(inode, FLUSH_SYNC);
if (ret < 0)
goto out_error;
}
out_error:
trace_nfs_writeback_folio_done(inode, folio, ret);
return ret;
}
#ifdef CONFIG_MIGRATION
int nfs_migrate_folio(struct address_space *mapping, struct folio *dst,
struct folio *src, enum migrate_mode mode)
{
/*
* If the private flag is set, the folio is currently associated with
* an in-progress read or write request. Don't try to migrate it.
*
* FIXME: we could do this in principle, but we'll need a way to ensure
* that we can safely release the inode reference while holding
* the folio lock.
*/
if (folio_test_private(src))
return -EBUSY;
if (folio_test_fscache(src)) {
nfs: Implement cache I/O by accessing the cache directly Move NFS to using fscache DIO API instead of the old upstream I/O API as that has been removed. This is a stopgap solution as the intention is that at sometime in the future, the cache will move to using larger blocks and won't be able to store individual pages in order to deal with the potential for data corruption due to the backing filesystem being able insert/remove bridging blocks of zeros into its extent list[1]. NFS then reads and writes cache pages synchronously and one page at a time. The preferred change would be to use the netfs lib, but the new I/O API can be used directly. It's just that as the cache now needs to track data for itself, caching blocks may exceed page size... This code is somewhat borrowed from my "fallback I/O" patchset[2]. Changes ======= ver #3: - Restore lost =n fallback for nfs_fscache_release_page()[2]. Signed-off-by: David Howells <dhowells@redhat.com> Tested-by: Dave Wysochanski <dwysocha@redhat.com> Acked-by: Jeff Layton <jlayton@kernel.org> cc: Trond Myklebust <trond.myklebust@hammerspace.com> cc: Anna Schumaker <anna.schumaker@netapp.com> cc: linux-nfs@vger.kernel.org cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/YO17ZNOcq+9PajfQ@mit.edu [1] Link: https://lore.kernel.org/r/202112100957.2oEDT20W-lkp@intel.com/ [2] Link: https://lore.kernel.org/r/163189108292.2509237.12615909591150927232.stgit@warthog.procyon.org.uk/ [2] Link: https://lore.kernel.org/r/163906981318.143852.17220018647843475985.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967184451.1823006.6450645559828329590.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021577632.640689.11069627070150063812.stgit@warthog.procyon.org.uk/ # v4
2021-08-27 14:19:34 +00:00
if (mode == MIGRATE_ASYNC)
return -EBUSY;
folio_wait_fscache(src);
nfs: Implement cache I/O by accessing the cache directly Move NFS to using fscache DIO API instead of the old upstream I/O API as that has been removed. This is a stopgap solution as the intention is that at sometime in the future, the cache will move to using larger blocks and won't be able to store individual pages in order to deal with the potential for data corruption due to the backing filesystem being able insert/remove bridging blocks of zeros into its extent list[1]. NFS then reads and writes cache pages synchronously and one page at a time. The preferred change would be to use the netfs lib, but the new I/O API can be used directly. It's just that as the cache now needs to track data for itself, caching blocks may exceed page size... This code is somewhat borrowed from my "fallback I/O" patchset[2]. Changes ======= ver #3: - Restore lost =n fallback for nfs_fscache_release_page()[2]. Signed-off-by: David Howells <dhowells@redhat.com> Tested-by: Dave Wysochanski <dwysocha@redhat.com> Acked-by: Jeff Layton <jlayton@kernel.org> cc: Trond Myklebust <trond.myklebust@hammerspace.com> cc: Anna Schumaker <anna.schumaker@netapp.com> cc: linux-nfs@vger.kernel.org cc: linux-cachefs@redhat.com Link: https://lore.kernel.org/r/YO17ZNOcq+9PajfQ@mit.edu [1] Link: https://lore.kernel.org/r/202112100957.2oEDT20W-lkp@intel.com/ [2] Link: https://lore.kernel.org/r/163189108292.2509237.12615909591150927232.stgit@warthog.procyon.org.uk/ [2] Link: https://lore.kernel.org/r/163906981318.143852.17220018647843475985.stgit@warthog.procyon.org.uk/ # v2 Link: https://lore.kernel.org/r/163967184451.1823006.6450645559828329590.stgit@warthog.procyon.org.uk/ # v3 Link: https://lore.kernel.org/r/164021577632.640689.11069627070150063812.stgit@warthog.procyon.org.uk/ # v4
2021-08-27 14:19:34 +00:00
}
return migrate_folio(mapping, dst, src, mode);
}
#endif
int __init nfs_init_writepagecache(void)
{
nfs_wdata_cachep = kmem_cache_create("nfs_write_data",
sizeof(struct nfs_pgio_header),
0, SLAB_HWCACHE_ALIGN,
NULL);
if (nfs_wdata_cachep == NULL)
return -ENOMEM;
nfs_wdata_mempool = mempool_create_slab_pool(MIN_POOL_WRITE,
nfs_wdata_cachep);
if (nfs_wdata_mempool == NULL)
goto out_destroy_write_cache;
nfs_cdata_cachep = kmem_cache_create("nfs_commit_data",
sizeof(struct nfs_commit_data),
0, SLAB_HWCACHE_ALIGN,
NULL);
if (nfs_cdata_cachep == NULL)
goto out_destroy_write_mempool;
nfs_commit_mempool = mempool_create_slab_pool(MIN_POOL_COMMIT,
nfs_cdata_cachep);
if (nfs_commit_mempool == NULL)
goto out_destroy_commit_cache;
/*
* NFS congestion size, scale with available memory.
*
* 64MB: 8192k
* 128MB: 11585k
* 256MB: 16384k
* 512MB: 23170k
* 1GB: 32768k
* 2GB: 46340k
* 4GB: 65536k
* 8GB: 92681k
* 16GB: 131072k
*
* This allows larger machines to have larger/more transfers.
* Limit the default to 256M
*/
nfs_congestion_kb = (16*int_sqrt(totalram_pages())) << (PAGE_SHIFT-10);
if (nfs_congestion_kb > 256*1024)
nfs_congestion_kb = 256*1024;
return 0;
out_destroy_commit_cache:
kmem_cache_destroy(nfs_cdata_cachep);
out_destroy_write_mempool:
mempool_destroy(nfs_wdata_mempool);
out_destroy_write_cache:
kmem_cache_destroy(nfs_wdata_cachep);
return -ENOMEM;
}
void nfs_destroy_writepagecache(void)
{
mempool_destroy(nfs_commit_mempool);
kmem_cache_destroy(nfs_cdata_cachep);
mempool_destroy(nfs_wdata_mempool);
kmem_cache_destroy(nfs_wdata_cachep);
}
static const struct nfs_rw_ops nfs_rw_write_ops = {
.rw_alloc_header = nfs_writehdr_alloc,
.rw_free_header = nfs_writehdr_free,
.rw_done = nfs_writeback_done,
.rw_result = nfs_writeback_result,
.rw_initiate = nfs_initiate_write,
};