Patch series "codingstyle: avoid unused parameters for a function-like
macro", v7.
A function-like macro could result in build warnings such as "unused
variable." This patchset updates the guidance to recommend always using a
static inline function instead and also provides checkpatch support for
this new rule.
This patch (of 2):
Recent commit 77292bb8ca ("crypto: scomp - remove memcpy if
sg_nents is 1 and pages are lowmem") leads to warnings on xtensa
and loongarch,
In file included from crypto/scompress.c:12:
include/crypto/scatterwalk.h: In function 'scatterwalk_pagedone':
include/crypto/scatterwalk.h:76:30: warning: variable 'page' set but not used [-Wunused-but-set-variable]
76 | struct page *page;
| ^~~~
crypto/scompress.c: In function 'scomp_acomp_comp_decomp':
>> crypto/scompress.c:174:38: warning: unused variable 'dst_page' [-Wunused-variable]
174 | struct page *dst_page = sg_page(req->dst);
|
The reason is that flush_dcache_page() is implemented as a noop
macro on these platforms as below,
#define flush_dcache_page(page) do { } while (0)
The driver code, for itself, seems be quite innocent and placing
maybe_unused seems pointless,
struct page *dst_page = sg_page(req->dst);
for (i = 0; i < nr_pages; i++)
flush_dcache_page(dst_page + i);
And it should be independent of architectural implementation
differences.
Let's provide guidance on coding style for requesting parameter
evaluation or proposing the migration to a static inline
function.
Link: https://lkml.kernel.org/r/20240507032757.146386-1-21cnbao@gmail.com
Link: https://lkml.kernel.org/r/20240507032757.146386-2-21cnbao@gmail.com
Signed-off-by: Barry Song <v-songbaohua@oppo.com>
Suggested-by: Max Filippov <jcmvbkbc@gmail.com>
Reviewed-by: Mark Brown <broonie@kernel.org>
Acked-by: Joe Perches <joe@perches.com>
Cc: Chris Zankel <chris@zankel.net>
Cc: Huacai Chen <chenhuacai@loongson.cn>
Cc: Herbert Xu <herbert@gondor.apana.org.au>
Cc: Guenter Roeck <linux@roeck-us.net>
Cc: Stephen Rothwell <sfr@canb.auug.org.au>
Cc: Andy Whitcroft <apw@canonical.com>
Cc: Dwaipayan Ray <dwaipayanray1@gmail.com>
Cc: Joe Perches <joe@perches.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Lukas Bulwahn <lukas.bulwahn@gmail.com>
Cc: Xining Xu <mac.xxn@outlook.com>
Cc: Charlemagne Lasse <charlemagnelasse@gmail.com>
Cc: Jeff Johnson <quic_jjohnson@quicinc.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
As one can see in include/trace/stages/stage4_event_fields.h, the
implementation of __field() uses the is_signed_type() macro. As one can
see in commit dcf8e5633e ("tracing: Define the is_signed_type() macro
once"), there has been an attempt to not make is_signed_type() trigger
sparse warnings for bitwise types.
Despite that change, sparse complains when passing a bitwise type to
is_signed_type(). The reason is that in its definition below, an
inequality comparison will be made against bitwise types, which are random
collections of bits (the casts to bitwise types themselves are
semantically valid and not problematic):
#define is_signed_type(type) (((type)(-1)) < (__force type)1)
So, as a workaround, follow the example of <trace/events/initcall.h> and
suppress the following sparse warnings by changing __field() into
__field_struct() that doesn't use is_signed_type():
fs/nilfs2/segment.c: note: in included file (through
include/trace/trace_events.h, include/trace/define_trace.h,
include/trace/events/nilfs2.h):
./include/trace/events/nilfs2.h:191:1: warning: cast to restricted
blk_opf_t
./include/trace/events/nilfs2.h:191:1: warning: restricted blk_opf_t
degrades to integer
./include/trace/events/nilfs2.h:191:1: warning: restricted blk_opf_t
degrades to integer
[konishi.ryusuke: describe the reason for the warnings per Linus's explanation]
Link: https://lkml.kernel.org/r/20240507222041.4876-1-konishi.ryusuke@gmail.com
Link: https://lkml.kernel.org/r/20240507142454.3344-1-konishi.ryusuke@gmail.com
Signed-off-by: Bart Van Assche <bvanassche@acm.org>
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
Reported-by: kernel test robot <lkp@intel.com>
Closes: https://lore.kernel.org/oe-kbuild-all/202401092241.I4mm9OWl-lkp@intel.com/
Reported-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
Closes: https://lore.kernel.org/all/20240430080019.4242-2-konishi.ryusuke@gmail.com/
Cc: Linus Torvalds <torvalds@linux-foundation.org>
Cc: Rasmus Villemoes <linux@rasmusvillemoes.dk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Android bionic warns that open modes are ignored if O_CREAT or O_TMPFILE
aren't specified. The permissions for the file are set above:
fd1 = open(kpath, O_RDWR | O_CREAT | O_TRUNC, 0644);
Link: https://lkml.kernel.org/r/20240429234610.191144-1-edliaw@google.com
Fixes: d97b46a646 ("syscalls, x86: add __NR_kcmp syscall")
Signed-off-by: Edward Liaw <edliaw@google.com>
Reviewed-by: Cyrill Gorcunov <gorcunov@gmail.com>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Shuah Khan <shuah@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
It is unconventional to have a blank line between name-of-function and
description-of-args.
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: Song Liu <song@kernel.org>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Cc: Ryusuke Konishi <konishi.ryusuke@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
NMI watchdog permanently consumes one hardware counters per CPU on the
system. For systems that use many hardware counters, this causes more
aggressive time multiplexing of perf events.
OTOH, some CPUs (mostly Intel) support "ref-cycles" event, which is rarely
used. Add kernel cmdline arg nmi_watchdog=rNNN to configure the watchdog
to use raw event. For example, on Intel CPUs, we can use "r300" to
configure the watchdog to use ref-cycles event.
If the raw event does not work, fall back to use "cycles".
[akpm@linux-foundation.org: fix kerneldoc]
Link: https://lkml.kernel.org/r/20240430060236.1878002-2-song@kernel.org
Signed-off-by: Song Liu <song@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Cc: "Matthew Wilcox (Oracle)" <willy@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Per the document, the kernel can accept comma separated command line like
nmi_watchdog=nopanic,0. However, the code doesn't really handle it. Fix
the kernel to handle it properly.
Link: https://lkml.kernel.org/r/20240430060236.1878002-1-song@kernel.org
Signed-off-by: Song Liu <song@kernel.org>
Cc: Peter Zijlstra <peterz@infradead.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Upon running sparse, "warning: dubious: x & !y" is output at an array
index calculation within nilfs_load_super_block().
The calculation is not wrong, but to eliminate the sparse warning, replace
it with an equivalent calculation.
Also, add a comment to make it easier to understand what the unintuitive
array index calculation is doing and whether it's correct.
Link: https://lkml.kernel.org/r/20240430080019.4242-3-konishi.ryusuke@gmail.com
Fixes: e339ad31f5 ("nilfs2: introduce secondary super block")
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
Cc: Bart Van Assche <bvanassche@acm.org>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: kernel test robot <lkp@intel.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Nobody checks the error flag on squashfs folios, so stop setting it.
Link: https://lkml.kernel.org/r/20240420025029.2166544-24-willy@infradead.org
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Tested-by: Phillip Lougher <phillip@squashfs.org.uk>
Reviewed-by: Phillip Lougher <phillip@squashfs.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Remove use of page APIs, return the errno instead of 0, switch from
kmap_atomic to kmap_local and use folio_end_read() to unify the two exit
paths.
Link: https://lkml.kernel.org/r/20240420025029.2166544-23-willy@infradead.org
Signed-off-by: Matthew Wilcox (Oracle) <willy@infradead.org>
Tested-by: Phillip Lougher <phillip@squashfs.org.uk>
Reviewed-by: Phillip Lougher <phillip@squashfs.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Directly read the current CPU number from the kgdb_active variable.
Before, the active CPU was obtained through the current task, which
required searching the task list for the pid of GDB's selected thread.
Obtaining the pid was buggy: GDB may use selected_thread().ptid[1] (LWPID)
instead of .ptid[2] (TID) to store the threads pid; see
https://sourceware.org/gdb/current/onlinedocs/gdb.html/Threads-In-Python.html
As a result, the detection could return the wrong CPU number, leading to
incorrect results for $lx_per_cpu and $lx_current.
As a side effect, the patch significantly speeds up $lx_per_cpu and
$lx_current in KGDB by avoiding the task-list iteration.
Link: https://lkml.kernel.org/r/20240425153501.749966-5-mail@florommel.de
Signed-off-by: Florian Rommel <mail@florommel.de>
Cc: Andrew Jones <ajones@ventanamicro.com>
Cc: Deepak Gupta <debug@rivosinc.com>
Cc: Jan Kiszka <jan.kiszka@siemens.com>
Cc: Kieran Bingham <kbingham@kernel.org>
Cc: Kuan-Ying Lee <Kuan-Ying.Lee@mediatek.com>
Cc: Palmer Dabbelt <palmer@rivosinc.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
get_thread_info ($lx_thread_info) only accepted a dereferenced task
parameter. Passing a pointer to a task_struct (like $lx_per_cpu does with
KGDB) threw an exception.
With this patch, both (dereferenced values and pointers) are accepted.
Before (on x86, KGDB):
>>> p $lx_per_cpu(cpu_info)
Traceback (most recent call last):
File "./scripts/gdb/linux/cpus.py", line 158, in invoke
return per_cpu(var_ptr, cpu)
^^^^^^^^^^^^^^^^^^^^^
File "./scripts/gdb/linux/cpus.py", line 42, in per_cpu
cpu = get_current_cpu()
^^^^^^^^^^^^^^^^^
File "./scripts/gdb/linux/cpus.py", line 33, in get_current_cpu
return tasks.get_thread_info(tasks.get_task_by_pid(tid))['cpu']
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
File "./scripts/gdb/linux/tasks.py", line 88, in get_thread_info
if task.type.fields()[0].type == thread_info_type.get_type():
~~~~~~~~~~~~~~~~~~^^^
IndexError: list index out of range
Link: https://lkml.kernel.org/r/20240425153501.749966-4-mail@florommel.de
Signed-off-by: Florian Rommel <mail@florommel.de>
Cc: Andrew Jones <ajones@ventanamicro.com>
Cc: Deepak Gupta <debug@rivosinc.com>
Cc: Jan Kiszka <jan.kiszka@siemens.com>
Cc: Kieran Bingham <kbingham@kernel.org>
Cc: Kuan-Ying Lee <Kuan-Ying.Lee@mediatek.com>
Cc: Palmer Dabbelt <palmer@rivosinc.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Before, the script tried to get the address by constructing a pointer to
the parameter (by name). However, since GDB now passes the parameter as a
GdbValue, we cannot get its name. Instead, we retrieve the address
through GdbValue's address attribute.
Before:
>>> p $lx_per_cpu(cpu_info)
Traceback (most recent call last):
File "./scripts/gdb/linux/cpus.py", line 152, in invoke
var_ptr = gdb.parse_and_eval("&" + var_name.string())
^^^^^^^^^^^^^^^^^
gdb.error: Trying to read string with inappropriate type `struct cpuinfo_x86'.
Link: https://lkml.kernel.org/r/20240425153501.749966-3-mail@florommel.de
Signed-off-by: Florian Rommel <mail@florommel.de>
Cc: Andrew Jones <ajones@ventanamicro.com>
Cc: Deepak Gupta <debug@rivosinc.com>
Cc: Jan Kiszka <jan.kiszka@siemens.com>
Cc: Kieran Bingham <kbingham@kernel.org>
Cc: Kuan-Ying Lee <Kuan-Ying.Lee@mediatek.com>
Cc: Palmer Dabbelt <palmer@rivosinc.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "scripts/gdb: Fixes for $lx_current and $lx_per_cpu".
This series fixes several bugs in the GDB scripts related to the
$lx_current and $lx_per_cpu functions. The changes were tested with GDB
10, 11, 12, 13, and 14.
Patch 1 fixes false-negative results when probing for KGDB
Patch 2 fixes the $lx_per_cpu function, which is currently non-functional
in QEMU-GDB and KGDB.
Patch 3 fixes an additional bug in $lx_per_cpu that occurs with KGDB.
Patch 4 fixes the incorrect detection of the current CPU number in KGDB,
which silently breaks $lx_per_cpu and $lx_current.
This patch (of 4):
The KGDB probe function sometimes failed to detect KGDB for SMP machines
as it assumed that task 2 (kthreadd) is running on CPU 0, which is not
necessarily the case. Now, the detection is agnostic to kthreadd's CPU.
Link: https://lkml.kernel.org/r/20240425153501.749966-1-mail@florommel.de
Link: https://lkml.kernel.org/r/20240425153501.749966-2-mail@florommel.de
Signed-off-by: Florian Rommel <mail@florommel.de>
Cc: Andrew Jones <ajones@ventanamicro.com>
Cc: Deepak Gupta <debug@rivosinc.com>
Cc: Jan Kiszka <jan.kiszka@siemens.com>
Cc: Kieran Bingham <kbingham@kernel.org>
Cc: Kuan-Ying Lee <Kuan-Ying.Lee@mediatek.com>
Cc: Palmer Dabbelt <palmer@rivosinc.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Update header inclusions to follow IWYU (Include What You Use) principle.
Link: https://lkml.kernel.org/r/20240423192529.3249134-4-andriy.shevchenko@linux.intel.com
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Alain Volmat <alain.volmat@foss.st.com>
Cc: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Cc: Chen-Yu Tsai <wens@csie.org>
Cc: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Cc: Jernej Skrabec <jernej.skrabec@gmail.com>
Cc: Matthias Brugger <matthias.bgg@gmail.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Patrice Chotard <patrice.chotard@foss.st.com>
Cc: Rob Herring <robh@kernel.org>
Cc: Samuel Holland <samuel@sholland.org>
Cc: Sean Wang <sean.wang@mediatek.com>
Cc: Sean Young <sean@mess.org>
Cc: Stefani Seibold <stefani@seibold.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
In the driver the io.h is implied by others. This is not good as it
prevents from cleanups done in other headers. Add missing include.
Link: https://lkml.kernel.org/r/20240423192529.3249134-3-andriy.shevchenko@linux.intel.com
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Alain Volmat <alain.volmat@foss.st.com>
Cc: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Cc: Chen-Yu Tsai <wens@csie.org>
Cc: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Cc: Jernej Skrabec <jernej.skrabec@gmail.com>
Cc: Matthias Brugger <matthias.bgg@gmail.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Patrice Chotard <patrice.chotard@foss.st.com>
Cc: Rob Herring <robh@kernel.org>
Cc: Samuel Holland <samuel@sholland.org>
Cc: Sean Wang <sean.wang@mediatek.com>
Cc: Sean Young <sean@mess.org>
Cc: Stefani Seibold <stefani@seibold.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "kfifo: Clean up kfifo.h", v2.
To reduce dependency hell a degree, clean up kfifo.h (mainly getting rid
of kernel.h in the global header).
This patch (of 3):
In many remote control drivers the io.h is implied by others. This is not
good as it prevents from cleanups done in other headers. Add missing
include.
Link: https://lkml.kernel.org/r/20240423192529.3249134-1-andriy.shevchenko@linux.intel.com
Link: https://lkml.kernel.org/r/20240423192529.3249134-2-andriy.shevchenko@linux.intel.com
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Cc: Alain Volmat <alain.volmat@foss.st.com>
Cc: AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
Cc: Chen-Yu Tsai <wens@csie.org>
Cc: Hans Verkuil <hverkuil-cisco@xs4all.nl>
Cc: Jernej Skrabec <jernej.skrabec@gmail.com>
Cc: Matthias Brugger <matthias.bgg@gmail.com>
Cc: Mauro Carvalho Chehab <mchehab@kernel.org>
Cc: Patrice Chotard <patrice.chotard@foss.st.com>
Cc: Rob Herring <robh@kernel.org>
Cc: Samuel Holland <samuel@sholland.org>
Cc: Sean Wang <sean.wang@mediatek.com>
Cc: Sean Young <sean@mess.org>
Cc: Stefani Seibold <stefani@seibold.net>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
The tools/lib/rbtree.c code came from the kernel. Remove the
EXPORT_SYMBOL() that make sense only there. Unfortunately it is not being
checked with tools/perf/check_headers.sh. Will try to remedy this. Until
then pick the improvements from:
b0687c1119 ("lib/rbtree: use '+' instead of '|' for setting color.")
That I noticed by doing:
diff -u tools/lib/rbtree.c lib/rbtree.c
diff -u tools/include/linux/rbtree_augmented.h include/linux/rbtree_augmented.h
There is one other cases, but lets pick it in separate patches.
Link: https://lkml.kernel.org/r/ZigZzeFoukzRKG1Q@x1
Signed-off-by: Arnaldo Carvalho de Melo <acme@redhat.com>
Cc: Adrian Hunter <adrian.hunter@intel.com>
Cc: Ian Rogers <irogers@google.com>
Cc: Jiri Olsa <jolsa@kernel.org>
Cc: Kan Liang <kan.liang@linux.intel.com>
Cc: Namhyung Kim <namhyung@kernel.org>
Cc: Noah Goldstein <goldstein.w.n@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Variable status is being assigned and error code that is never read, it is
being assigned inside of a do-while loop. The assignment is redundant and
can be removed.
Cleans up clang scan build warning:
fs/ocfs2/dlm/dlmdomain.c:1530:2: warning: Value stored to 'status' is never
read [deadcode.DeadStores]
Link: https://lkml.kernel.org/r/20240423223018.1573213-1-colin.i.king@gmail.com
Signed-off-by: Colin Ian King <colin.i.king@gmail.com>
Acked-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Changwei Ge <gechangwei@live.cn>
Cc: Gang He <ghe@suse.com>
Cc: Jun Piao <piaojun@huawei.com>
Cc: Heming Zhao <heming.zhao@suse.com>
Cc: Dan Carpenter <dan.carpenter@linaro.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Jiri reported that the current kexec_dprintk() always prints out debugging
message whenever kexec/kdmmp loading is triggered. That is not wanted.
The debugging message is supposed to be printed out when 'kexec -s -d' is
specified for kexec/kdump loading.
After investigating, the reason is the current kexec_dprintk() takes
printk(KERN_INFO) or printk(KERN_DEBUG) depending on whether '-d' is
specified. However, distros usually have defaulg log level like below:
[~]# cat /proc/sys/kernel/printk
7 4 1 7
So, even though '-d' is not specified, printk(KERN_DEBUG) also always
prints out. I thought printk(KERN_DEBUG) is equal to pr_debug(), it's
not.
Fix it by changing to use pr_info() instead which are expected to work.
Link: https://lkml.kernel.org/r/20240409042238.1240462-1-bhe@redhat.com
Fixes: cbc2fe9d9c ("kexec_file: add kexec_file flag to control debug printing")
Signed-off-by: Baoquan He <bhe@redhat.com>
Reported-by: Jiri Slaby <jirislaby@kernel.org>
Closes: https://lore.kernel.org/all/4c775fca-5def-4a2d-8437-7130b02722a2@kernel.org
Reviewed-by: Dave Young <dyoung@redhat.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Add pr_fmt() to kernel/crash_core.c to add the module name to debugging
message printed as prefix.
And also add prefix 'crashkernel:' to two lines of message printing code
in kernel/crash_reserve.c. In kernel/crash_reserve.c, almost all
debugging messages have 'crashkernel:' prefix or there's keyword
crashkernel at the beginning or in the middle, adding pr_fmt() makes it
redundant.
Link: https://lkml.kernel.org/r/20240418035843.1562887-1-bhe@redhat.com
Signed-off-by: Baoquan He <bhe@redhat.com>
Cc: Dave Young <dyoung@redhat.com>
Cc: Jiri Slaby <jirislaby@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
ida_alloc() and ida_free() should be preferred to the deprecated
ida_simple_get() and ida_simple_remove().
This is less verbose.
Link: https://lkml.kernel.org/r/f82e013abe4c71f1c7d06819f96472f298acdcf3.1713089554.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Cc: Peter Rosin <peda@axentia.se>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
ida_alloc() and ida_free() should be preferred to the deprecated
ida_simple_get() and ida_simple_remove().
This is less verbose.
Link: https://lkml.kernel.org/r/9f681747d446b874952a892491387d79ffe565a9.1713089394.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Cc: Rodolfo Giometti <giometti@enneenne.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
ida_alloc() and ida_free() should be preferred to the deprecated
ida_simple_get() and ida_simple_remove().
This is less verbose.
Link: https://lkml.kernel.org/r/2aca50a9d061faecfd4ded80b5874cd3be9b855d.1713086613.git.christophe.jaillet@wanadoo.fr
Signed-off-by: Christophe JAILLET <christophe.jaillet@wanadoo.fr>
Cc: Alexander Shishkin <alexander.shishkin@linux.intel.com>
Cc: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
This commit adds kernel-doc style comments with complete parameter
descriptions for the function nilfs_remove_all_gcinodes.
Link: https://lkml.kernel.org/r/20240410075629.3441-4-konishi.ryusuke@gmail.com
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
This commit adds kernel-doc style comments with complete parameter
descriptions for the function nilfs_btree_convert_and_insert.
Link: https://lkml.kernel.org/r/20240410075629.3441-3-konishi.ryusuke@gmail.com
Signed-off-by: Yang Li <yang.lee@linux.alibaba.com>
Signed-off-by: Ryusuke Konishi <konishi.ryusuke@gmail.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
gcc-9 warns about a possibly non-terminated string copy:
kernel/trace/blktrace.c: In function 'do_blk_trace_setup':
kernel/trace/blktrace.c:527:2: error: 'strncpy' specified bound 32 equals destination size [-Werror=stringop-truncation]
Newer versions are fine here because they see the following explicit
nul-termination. Using strscpy_pad() avoids the warning and
simplifies the code a little. The padding helps give a clean
buffer to userspace.
Link: https://lkml.kernel.org/r/20240409140059.3806717-5-arnd@kernel.org
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Justin Stitt <justinstitt@google.com>
Cc: Alexey Starikovskiy <astarikovskiy@suse.de>
Cc: Bob Moore <robert.moore@intel.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Len Brown <lenb@kernel.org>
Cc: Lin Ming <ming.m.lin@intel.com>
Cc: Masahiro Yamada <masahiroy@kernel.org>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Nathan Chancellor <nathan@kernel.org>
Cc: Nicolas Schier <nicolas@fjasle.eu>
Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: "Richard Russon (FlatCap)" <ldm@flatcap.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
The strncpy() here can cause a non-terminated string, which older gcc
versions such as gcc-9 warn about:
In function 'ldm_parse_tocblock',
inlined from 'ldm_validate_tocblocks' at block/partitions/ldm.c:386:7,
inlined from 'ldm_partition' at block/partitions/ldm.c:1457:7:
block/partitions/ldm.c:134:2: error: 'strncpy' specified bound 16 equals destination size [-Werror=stringop-truncation]
134 | strncpy (toc->bitmap1_name, data + 0x24, sizeof (toc->bitmap1_name));
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
block/partitions/ldm.c:145:2: error: 'strncpy' specified bound 16 equals destination size [-Werror=stringop-truncation]
145 | strncpy (toc->bitmap2_name, data + 0x46, sizeof (toc->bitmap2_name));
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
New versions notice that the code is correct after all because of the
following termination, but replacing the strncpy() with strscpy_pad()
or strcpy() avoids the warning and simplifies the code at the same time.
Use the padding version here to keep the existing behavior, in case
the code relies on not including uninitialized data.
Link: https://lkml.kernel.org/r/20240409140059.3806717-4-arnd@kernel.org
Reviewed-by: Justin Stitt <justinstitt@google.com>
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Cc: Alexey Starikovskiy <astarikovskiy@suse.de>
Cc: Bob Moore <robert.moore@intel.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Len Brown <lenb@kernel.org>
Cc: Lin Ming <ming.m.lin@intel.com>
Cc: Masahiro Yamada <masahiroy@kernel.org>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Nathan Chancellor <nathan@kernel.org>
Cc: Nicolas Schier <nicolas@fjasle.eu>
Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: "Richard Russon (FlatCap)" <ldm@flatcap.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
gcc can warn when a string is too long to fit into the strncpy()
destination buffer, as it is here depending on the function arguments:
inlined from 'test_hexdump_prepare_test.constprop' at /home/arnd/arm-soc/lib/test_hexdump.c:116:3:
include/linux/fortify-string.h:108:33: error: '__builtin_strncpy' output truncated copying between 0 and 32 bytes from a string of length 32 [-Werror=stringop-truncation]
108 | #define __underlying_strncpy __builtin_strncpy
| ^
include/linux/fortify-string.h:187:16: note: in expansion of macro '__underlying_strncpy'
187 | return __underlying_strncpy(p, q, size);
| ^~~~~~~~~~~~~~~~~~~~
The intention here is to copy exactly 'l' bytes without any padding or
NUL-termination, so the most logical change is to use memcpy(), just as
a previous change adapted the other output from strncpy() to memcpy().
Link: https://lkml.kernel.org/r/20240409140059.3806717-2-arnd@kernel.org
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Acked-by: Justin Stitt <justinstitt@google.com>
Cc: Alexey Starikovskiy <astarikovskiy@suse.de>
Cc: Bob Moore <robert.moore@intel.com>
Cc: Jens Axboe <axboe@kernel.dk>
Cc: Len Brown <lenb@kernel.org>
Cc: Lin Ming <ming.m.lin@intel.com>
Cc: Masahiro Yamada <masahiroy@kernel.org>
Cc: Masami Hiramatsu <mhiramat@kernel.org>
Cc: Mathieu Desnoyers <mathieu.desnoyers@efficios.com>
Cc: Nathan Chancellor <nathan@kernel.org>
Cc: Nicolas Schier <nicolas@fjasle.eu>
Cc: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
Cc: "Richard Russon (FlatCap)" <ldm@flatcap.org>
Cc: Steven Rostedt <rostedt@goodmis.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
The default atime related mount option is '-o realtime' which means file
atime should be updated if atime <= ctime or atime <= mtime. atime should
be updated in the following scenario, but it is not:
==========================================================
$ rm /mnt/testfile;
$ echo test > /mnt/testfile
$ stat -c "%X %Y %Z" /mnt/testfile
1711881646 1711881646 1711881646
$ sleep 5
$ cat /mnt/testfile > /dev/null
$ stat -c "%X %Y %Z" /mnt/testfile
1711881646 1711881646 1711881646
==========================================================
And the reason the atime in the test is not updated is that ocfs2 calls
ktime_get_real_ts64() in __ocfs2_mknod_locked during file creation. Then
inode_set_ctime_current() is called in inode_set_ctime_current() calls
ktime_get_coarse_real_ts64() to get current time.
ktime_get_real_ts64() is more accurate than ktime_get_coarse_real_ts64().
In my test box, I saw ctime set by ktime_get_coarse_real_ts64() is less
than ktime_get_real_ts64() even ctime is set later. The ctime of the new
inode is smaller than atime.
The call trace is like:
ocfs2_create
ocfs2_mknod
__ocfs2_mknod_locked
....
ktime_get_real_ts64 <------- set atime,ctime,mtime, more accurate
ocfs2_populate_inode
...
ocfs2_init_acl
ocfs2_acl_set_mode
inode_set_ctime_current
current_time
ktime_get_coarse_real_ts64 <-------less accurate
ocfs2_file_read_iter
ocfs2_inode_lock_atime
ocfs2_should_update_atime
atime <= ctime ? <-------- false, ctime < atime due to accuracy
So here call ktime_get_coarse_real_ts64 to set inode time coarser while
creating new files. It may lower the accuracy of file times. But it's
not a big deal since we already use coarse time in other places like
ocfs2_update_inode_atime and inode_set_ctime_current.
Link: https://lkml.kernel.org/r/20240408082041.20925-5-glass.su@suse.com
Fixes: c62c38f6b9 ("ocfs2: replace CURRENT_TIME macro")
Signed-off-by: Su Yue <glass.su@suse.com>
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Changwei Ge <gechangwei@live.cn>
Cc: Gang He <ghe@suse.com>
Cc: Jun Piao <piaojun@huawei.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
transaction id should be updated in ocfs2_unlink and ocfs2_link.
Otherwise, inode link will be wrong after journal replay even fsync was
called before power failure:
=======================================================================
$ touch testdir/bar
$ ln testdir/bar testdir/bar_link
$ fsync testdir/bar
$ stat -c %h $SCRATCH_MNT/testdir/bar
1
$ stat -c %h $SCRATCH_MNT/testdir/bar
1
=======================================================================
Link: https://lkml.kernel.org/r/20240408082041.20925-4-glass.su@suse.com
Fixes: ccd979bdbc ("[PATCH] OCFS2: The Second Oracle Cluster Filesystem")
Signed-off-by: Su Yue <glass.su@suse.com>
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: Changwei Ge <gechangwei@live.cn>
Cc: Gang He <ghe@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Jun Piao <piaojun@huawei.com>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
After commit "ocfs2: return real error code in ocfs2_dio_wr_get_block",
fstests/generic/300 become from always failed to sometimes failed:
========================================================================
[ 473.293420 ] run fstests generic/300
[ 475.296983 ] JBD2: Ignoring recovery information on journal
[ 475.302473 ] ocfs2: Mounting device (253,1) on (node local, slot 0) with ordered data mode.
[ 494.290998 ] OCFS2: ERROR (device dm-1): ocfs2_change_extent_flag: Owner 5668 has an extent at cpos 78723 which can no longer be found
[ 494.291609 ] On-disk corruption discovered. Please run fsck.ocfs2 once the filesystem is unmounted.
[ 494.292018 ] OCFS2: File system is now read-only.
[ 494.292224 ] (kworker/19:11,2628,19):ocfs2_mark_extent_written:5272 ERROR: status = -30
[ 494.292602 ] (kworker/19:11,2628,19):ocfs2_dio_end_io_write:2374 ERROR: status = -3
fio: io_u error on file /mnt/scratch/racer: Read-only file system: write offset=460849152, buflen=131072
=========================================================================
In __blockdev_direct_IO, ocfs2_dio_wr_get_block is called to add unwritten
extents to a list. extents are also inserted into extent tree in
ocfs2_write_begin_nolock. Then another thread call fallocate to puch a
hole at one of the unwritten extent. The extent at cpos was removed by
ocfs2_remove_extent(). At end io worker thread, ocfs2_search_extent_list
found there is no such extent at the cpos.
T1 T2 T3
inode lock
...
insert extents
...
inode unlock
ocfs2_fallocate
__ocfs2_change_file_space
inode lock
lock ip_alloc_sem
ocfs2_remove_inode_range inode
ocfs2_remove_btree_range
ocfs2_remove_extent
^---remove the extent at cpos 78723
...
unlock ip_alloc_sem
inode unlock
ocfs2_dio_end_io
ocfs2_dio_end_io_write
lock ip_alloc_sem
ocfs2_mark_extent_written
ocfs2_change_extent_flag
ocfs2_search_extent_list
^---failed to find extent
...
unlock ip_alloc_sem
In most filesystems, fallocate is not compatible with racing with AIO+DIO,
so fix it by adding to wait for all dio before fallocate/punch_hole like
ext4.
Link: https://lkml.kernel.org/r/20240408082041.20925-3-glass.su@suse.com
Fixes: b25801038d ("ocfs2: Support xfs style space reservation ioctls")
Signed-off-by: Su Yue <glass.su@suse.com>
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: Changwei Ge <gechangwei@live.cn>
Cc: Gang He <ghe@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Jun Piao <piaojun@huawei.com>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: <stable@vger.kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "ocfs2 bugs fixes exposed by fstests", v3.
The patchset is to fix some wrong behavior of ocfs2 exposed by fstests.
Patch 1 makes userspace happy when some error happens when doing direct
io. Before the patch, DIO always return -EIO in case of error. After the
patch, it returns real error code such like -ENOSPC, EDQUOT...
Patch 2 fixes an error case when doing AIO+DIO and hole punching at same
file position in parallel. generic/300
Patch 3 fixes inode link count mismatch after power failure. Without the
patch, inode link would be wrong even fync was called on the file.
tests/generic/040,041,104,107,336
patch 4 fixes wrong atime with mount option realtime. Without the patch,
atime of new created file won't be updated in right time.
tests/generic/192
For stable kernels, I added fixes to patch 2,3,4. The patch 1 is not
recommended to be backported since ocfs2_dio_wr_get_block calls too many
functions. It's diffcult to check every git history of ocfs2 for every
LTS kernel.
This patch (of 4):
ocfs2_dio_wr_get_block always returns -EIO in case of errors. However,
some programs expect right exit codes while doing dio. For example, tools
like fio treat -ENOSPC as expected code while doing stress jobs. And
quota tools expect -EDQUOT when disk quota exceeds.
-EIO is too strong return code in the dio path. The caller of
ocfs2_dio_wr_get_block is __blockdev_direct_IO which is widely used and it
handles error codes well. I have checked functions called by
ocfs2_dio_wr_get_block and their return codes look good and clear. So I
think it's safe to let ocfs2_dio_wr_get_block return real error code.
Link: https://lkml.kernel.org/r/20240408082041.20925-1-glass.su@suse.com
Link: https://lkml.kernel.org/r/20240408082041.20925-2-glass.su@suse.com
Signed-off-by: Su Yue <glass.su@suse.com>
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: Changwei Ge <gechangwei@live.cn>
Cc: Gang He <ghe@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Jun Piao <piaojun@huawei.com>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Mark Fasheh <mark@fasheh.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
strncpy() is in the process of being replaced as it is deprecated [1].
We should move towards safer and less ambiguous string interfaces.
Looking at vmcoredd_header's definition:
| struct vmcoredd_header {
| __u32 n_namesz; /* Name size */
| __u32 n_descsz; /* Content size */
| __u32 n_type; /* NT_VMCOREDD */
| __u8 name[8]; /* LINUX\0\0\0 */
| __u8 dump_name[VMCOREDD_MAX_NAME_BYTES]; /* Device dump's name */
| };
.. we see that @name wants to be NUL-padded.
We're copying data->dump_name which is defined as:
| char dump_name[VMCOREDD_MAX_NAME_BYTES]; /* Unique name of the dump */
.. which shares the same size as vdd_hdr->dump_name. Let's make sure we
NUL-pad this as well.
Use strscpy_pad() which NUL-terminates and NUL-pads its destination
buffers. Specifically, use the new 2-argument version of strscpy_pad
introduced in Commit e6584c3964 ("string: Allow 2-argument
strscpy()").
Link: https://www.kernel.org/doc/html/latest/process/deprecated.html#strncpy-on-nul-terminated-strings [1]
Link: https://github.com/KSPP/linux/issues/90
Link: https://lkml.kernel.org/r/20240401-strncpy-fs-proc-vmcore-c-v2-1-dd0a73f42635@google.com
Signed-off-by: Justin Stitt <justinstitt@google.com>
Acked-by: Baoquan He <bhe@redhat.com>
Cc: Dave Young <dyoung@redhat.com>
Cc: Vivek Goyal <vgoyal@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Update header inclusions to follow IWYU (Include What You Use) principle.
Link: https://lkml.kernel.org/r/20240403104820.557487-3-andriy.shevchenko@linux.intel.com
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Philipp Stanner <pstanner@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "devres: A couple of cleanups".
A couple of ad-hoc cleanups. No functional changes intended.
This patch (of 2):
The devm_*() APIs are supposed to be called during the ->probe() stage.
Many drivers (especially new ones) have switched to use dev_err_probe()
for error messaging for the sake of unification. Let's do the same in the
devres APIs.
Link: https://lkml.kernel.org/r/20240403104820.557487-1-andriy.shevchenko@linux.intel.com
Link: https://lkml.kernel.org/r/20240403104820.557487-2-andriy.shevchenko@linux.intel.com
Signed-off-by: Andy Shevchenko <andriy.shevchenko@linux.intel.com>
Reviewed-by: Philipp Stanner <pstanner@redhat.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
In a future patch HAS_IOPORT=n will disable inb()/outb() and friends at
compile time. We thus need to add HAS_IOPORT as dependency for those
drivers using them.
Link: https://lkml.kernel.org/r/20240403132547.762429-2-schnelle@linux.ibm.com
Co-developed-by: Arnd Bergmann <arnd@kernel.org>
Signed-off-by: Arnd Bergmann <arnd@kernel.org>
Signed-off-by: Niklas Schnelle <schnelle@linux.ibm.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Squashfs copied the passed string (name) into a temporary buffer to ensure
it was NUL-terminated. This however is completely unnecessary as the
string is already NUL-terminated. So remove the deprecated strncpy() by
completely removing the string copy.
The background behind this unnecessary string copy is that it dates back
to the days when Squashfs was an out of kernel patch. The code
deliberately did not assume the string was NUL-terminated in case in
future this changed (due to kernel changes). This would mean the out of
tree patches would be broken but still compile OK.
Link: https://lkml.kernel.org/r/20240403183352.391308-1-phillip@squashfs.org.uk
Signed-off-by: Phillip Lougher <phillip@squashfs.org.uk>
Reviewed-by: Kees Cook <keescook@chromium.org>
Reviewed-by: Justin Stitt <justinstitt@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
The area_size is never larger than the maximum on 64-bit architectutes:
kernel/kcov.c:634:29: error: result of comparison of constant 1152921504606846975 with expression of type '__u32' (aka 'unsigned int') is always false [-Werror,-Wtautological-constant-out-of-range-compare]
if (remote_arg->area_size > LONG_MAX / sizeof(unsigned long))
~~~~~~~~~~~~~~~~~~~~~ ^ ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
The compiler can correctly optimize the check away and the code appears
correct to me, so just add a cast to avoid the warning.
Link: https://lkml.kernel.org/r/20240328143051.1069575-5-arnd@kernel.org
Signed-off-by: Arnd Bergmann <arnd@arndb.de>
Reviewed-by: Justin Stitt <justinstitt@google.com>
Cc: Andrey Konovalov <andreyknvl@gmail.com>
Cc: Bill Wendling <morbo@google.com>
Cc: Dmitry Vyukov <dvyukov@google.com>
Cc: Nathan Chancellor <nathan@kernel.org>
Cc: Nick Desaulniers <ndesaulniers@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
After commit 443cbaf9e2 ("crash: split vmcoreinfo exporting code out
from crash_core.c"), Kconfig item CRASH_CORE has gone away in kernel.
Items VMCORE_INFO and CRASH_RESERVE are used instead.
So clean up the outdated description about CRASH_CORE and update it
accordingly.
Link: https://lkml.kernel.org/r/20240329132825.1102459-3-bhe@redhat.com
Signed-off-by: Baoquan He <bhe@redhat.com>
Cc: Jonathan Corbet <corbet@lwn.net>
Cc: Geert Uytterhoeven <geert@linux-m68k.org>
Cc: Huacai Chen <chenhuacai@kernel.org>
Cc: WANG Xuerui <kernel@xen0n.name>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Add short-circuit code to speed up searching
Link: https://lkml.kernel.org/r/20240328125203.20892-4-heming.zhao@suse.com
Signed-off-by: Heming Zhao <heming.zhao@suse.com>
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: Changwei Ge <gechangwei@live.cn>
Cc: Gang He <ghe@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Jun Piao <piaojun@huawei.com>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Mark Fasheh <mark@fasheh.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Patch series "improve write IO performance when fragmentation is high",
v6.
This patch (of 4):
After introducing gd->bg_contig_free_bits, the code path
'ocfs2_cluster_group_search() => ocfs2_local_alloc_seen_free_bits()'
becomes death when all the gd->bg_contig_free_bits are set to the correct
value. This patch relocates ocfs2_local_alloc_seen_free_bits() to a more
appropriate location. (The new place being ocfs2_block_group_set_bits().)
In ocfs2_local_alloc_seen_free_bits(), the scope of the spin-lock has been
adjusted to reduce meaningless lock races. e.g: when userspace creates &
deletes 1 cluster_size files in parallel, acquiring the spin-lock in
ocfs2_local_alloc_seen_free_bits() is totally pointless and impedes IO
performance.
Link: https://lkml.kernel.org/r/20240328125203.20892-3-heming.zhao@suse.com
Signed-off-by: Heming Zhao <heming.zhao@suse.com>
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: Mark Fasheh <mark@fasheh.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Changwei Ge <gechangwei@live.cn>
Cc: Gang He <ghe@suse.com>
Cc: Jun Piao <piaojun@huawei.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
The group_search function ocfs2_cluster_group_search() should
bypass groups with insufficient space to avoid unnecessary
searches.
This patch is particularly useful when ocfs2 is handling huge
number small files, and volume fragmentation is very high.
In this case, ocfs2 is busy with looking up available la window
from //global_bitmap.
This patch introduces a new member in the Group Description (gd)
struct called 'bg_contig_free_bits', representing the max
contigous free bits in this gd. When ocfs2 allocates a new
la window from //global_bitmap, 'bg_contig_free_bits' helps
expedite the search process.
Let's image below path.
1. la state (->local_alloc_state) is set THROTTLED or DISABLED.
2. when user delete a large file and trigger
ocfs2_local_alloc_seen_free_bits set osb->local_alloc_state
unconditionally.
3. a write IOs thread run and trigger the worst performance path
```
ocfs2_reserve_clusters_with_limit
ocfs2_reserve_local_alloc_bits
ocfs2_local_alloc_slide_window //[1]
+ ocfs2_local_alloc_reserve_for_window //[2]
+ ocfs2_local_alloc_new_window //[3]
ocfs2_recalc_la_window
```
[1]:
will be called when la window bits used up.
[2]:
under la state is ENABLED, and this func only check global_bitmap
free bits, it will succeed in general.
[3]:
will use the default la window size to search clusters then fail.
ocfs2_recalc_la_window attempts other la window sizes.
the timing complexity is O(n^4), resulting in a significant time
cost for scanning global bitmap. This leads to a dramatic slowdown
in write I/Os (e.g., user space 'dd').
i.e.
an ocfs2 partition size: 1.45TB, cluster size: 4KB,
la window default size: 106MB.
The partition is fragmentation by creating & deleting huge mount of
small files.
before this patch, the timing of [3] should be
(the number got from real world):
- la window size change order (size: MB):
106, 53, 26.5, 13, 6.5, 3.25, 1.6, 0.8
only 0.8MB succeed, 0.8MB also triggers la window to disable.
ocfs2_local_alloc_new_window retries 8 times, first 7 times totally
runs in worst case.
- group chain number: 242
ocfs2_claim_suballoc_bits calls for-loop 242 times
- each chain has 49 block group
ocfs2_search_chain calls while-loop 49 times
- each bg has 32256 blocks
ocfs2_block_group_find_clear_bits calls while-loop for 32256 bits.
for ocfs2_find_next_zero_bit uses ffz() to find zero bit, let's use
(32256/64) (this is not worst value) for timing calucation.
the loop times: 7*242*49*(32256/64) = 41835024 (~42 million times)
In the worst case, user space writes 1MB data will trigger 42M scanning
times.
under this patch, the timing is '7*242*49 = 83006', reduced by three
orders of magnitude.
Link: https://lkml.kernel.org/r/20240328125203.20892-2-heming.zhao@suse.com
Signed-off-by: Heming Zhao <heming.zhao@suse.com>
Reviewed-by: Joseph Qi <joseph.qi@linux.alibaba.com>
Cc: Changwei Ge <gechangwei@live.cn>
Cc: Gang He <ghe@suse.com>
Cc: Joel Becker <jlbec@evilplan.org>
Cc: Jun Piao <piaojun@huawei.com>
Cc: Junxiao Bi <junxiao.bi@oracle.com>
Cc: Mark Fasheh <mark@fasheh.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
While browsing through ChromeOS crash reports, I found one with an
allocation failure that looked like this:
chrome: page allocation failure: order:7,
mode:0x40dc0(GFP_KERNEL|__GFP_COMP|__GFP_ZERO),
nodemask=(null),cpuset=urgent,mems_allowed=0
CPU: 7 PID: 3295 Comm: chrome Not tainted
5.15.133-20574-g8044615ac35c #1 (HASH:1162 1)
Hardware name: Google Lazor (rev3 - 8) with KB Backlight (DT)
Call trace:
...
warn_alloc+0x104/0x174
__alloc_pages+0x5f0/0x6e4
kmalloc_order+0x44/0x98
kmalloc_order_trace+0x34/0x124
__kmalloc+0x228/0x36c
__regset_get+0x68/0xcc
regset_get_alloc+0x1c/0x28
elf_core_dump+0x3d8/0xd8c
do_coredump+0xeb8/0x1378
get_signal+0x14c/0x804
...
An order 7 allocation is (1 << 7) contiguous pages, or 512K. It's not
a surprise that this allocation failed on a system that's been running
for a while.
More digging showed that it was fairly easy to see the order 7
allocation by just sending a SIGQUIT to chrome (or other processes) to
generate a core dump. The actual amount being allocated was 279,584
bytes and it was for "core_note_type" NT_ARM_SVE.
There was quite a bit of discussion [1] on the mailing lists in
response to my v1 patch attempting to switch to vmalloc. The overall
conclusion was that we could likely reduce the 279,584 byte allocation
by quite a bit and Mark Brown has sent a patch to that effect [2].
However even with the 279,584 byte allocation gone there are still
65,552 byte allocations. These are just barely more than the 65,536
bytes and thus would require an order 5 allocation.
An order 5 allocation is still something to avoid unless necessary and
nothing needs the memory here to be contiguous. Change the allocation
to kvzalloc() which should still be efficient for small allocations
but doesn't force the memory subsystem to work hard (and maybe fail)
at getting a large contiguous chunk.
[1] https://lore.kernel.org/r/20240201171159.1.Id9ad163b60d21c9e56c2d686b0cc9083a8ba7924@changeid
[2] https://lore.kernel.org/r/20240203-arm64-sve-ptrace-regset-size-v1-1-2c3ba1386b9e@kernel.org
Link: https://lkml.kernel.org/r/20240205092626.v2.1.Id9ad163b60d21c9e56c2d686b0cc9083a8ba7924@changeid
Signed-off-by: Douglas Anderson <dianders@chromium.org>
Reviewed-by: Catalin Marinas <catalin.marinas@arm.com>
Cc: Al Viro <viro@ZenIV.linux.org.uk>
Cc: Christian Brauner <brauner@kernel.org>
Cc: Dave Martin <Dave.Martin@arm.com>
Cc: Eric Biederman <ebiederm@xmission.com>
Cc: Jan Kara <jack@suse.cz>
Cc: Kees Cook <keescook@chromium.org>
Cc: Mark Brown <broonie@kernel.org>
Cc: Matthew Wilcox (Oracle) <willy@infradead.org>
Cc: Oleg Nesterov <oleg@redhat.com>
Cc: Will Deacon <will@kernel.org>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
