davidlt/kernel-ark
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity
74614de17d
kernel-ark/fs
History
Tim Chen d23da150a3 fs/superblock: avoid locking counting inodes and dentries before reclaiming them 
We remove the call to grab_super_passive in call to super_cache_count.
This becomes a scalability bottleneck as multiple threads are trying to do
memory reclamation, e.g.  when we are doing large amount of file read and
page cache is under pressure.  The cached objects quickly got reclaimed
down to 0 and we are aborting the cache_scan() reclaim.  But counting
creates a log jam acquiring the sb_lock.

We are holding the shrinker_rwsem which ensures the safety of call to
list_lru_count_node() and s_op->nr_cached_objects.  The shrinker is
unregistered now before ->kill_sb() so the operation is safe when we are
doing unmount.

The impact will depend heavily on the machine and the workload but for a
small machine using postmark tuned to use 4xRAM size the results were

                                  3.15.0-rc5            3.15.0-rc5
                                     vanilla         shrinker-v1r1
Ops/sec Transactions         21.00 (  0.00%)       24.00 ( 14.29%)
Ops/sec FilesCreate          39.00 (  0.00%)       44.00 ( 12.82%)
Ops/sec CreateTransact       10.00 (  0.00%)       12.00 ( 20.00%)
Ops/sec FilesDeleted       6202.00 (  0.00%)     6202.00 (  0.00%)
Ops/sec DeleteTransact       11.00 (  0.00%)       12.00 (  9.09%)
Ops/sec DataRead/MB          25.97 (  0.00%)       29.10 ( 12.05%)
Ops/sec DataWrite/MB         49.99 (  0.00%)       56.02 ( 12.06%)

ffsb running in a configuration that is meant to simulate a mail server showed

                                 3.15.0-rc5             3.15.0-rc5
                                    vanilla          shrinker-v1r1
Ops/sec readall           9402.63 (  0.00%)      9567.97 (  1.76%)
Ops/sec create            4695.45 (  0.00%)      4735.00 (  0.84%)
Ops/sec delete             173.72 (  0.00%)       179.83 (  3.52%)
Ops/sec Transactions     14271.80 (  0.00%)     14482.81 (  1.48%)
Ops/sec Read                37.00 (  0.00%)        37.60 (  1.62%)
Ops/sec Write               18.20 (  0.00%)        18.30 (  0.55%)

Signed-off-by: Tim Chen <tim.c.chen@linux.intel.com>
Signed-off-by: Mel Gorman <mgorman@suse.de>
Cc: Johannes Weiner <hannes@cmpxchg.org>
Cc: Hugh Dickins <hughd@google.com>
Cc: Dave Chinner <david@fromorbit.com>
Tested-by: Yuanhan Liu <yuanhan.liu@linux.intel.com>
Cc: Bob Liu <bob.liu@oracle.com>
Cc: Jan Kara <jack@suse.cz>
Acked-by: Rik van Riel <riel@redhat.com>
Cc: Al Viro <viro@zeniv.linux.org.uk>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>
2014-06-04 16:54:11 -07:00
..
9p fs/9p: kerneldoc fixes 2014-06-04 16:53:55 -07:00
adfs fs/adfs/super.c: add __init to init_inodecache() 2014-04-07 16:36:08 -07:00
affs fs/affs/super.c: bugfix / double free 2014-05-06 13:05:00 -07:00
afs File locking related changes for v3.16 2014-06-04 08:12:50 -07:00
autofs4 autofs: fix lockref lookup 2014-05-06 13:04:59 -07:00
befs Major changes for 3.14 include support for the newly added ZERO_RANGE 2014-04-04 15:39:39 -07:00
bfs fs/bfs/inode.c: add __init to init_inodecache() 2014-04-07 16:36:08 -07:00
btrfs mm: non-atomically mark page accessed during page cache allocation where possible 2014-06-04 16:54:10 -07:00
cachefiles Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2014-04-12 14:49:50 -07:00
ceph locks: ensure that fl_owner is always initialized properly in flock and lease codepaths 2014-06-02 08:09:29 -04:00
cifs cifs: fix actimeo=0 corner case when cifs_i->time == jiffies 2014-04-24 22:37:03 -05:00
coda Major changes for 3.14 include support for the newly added ZERO_RANGE 2014-04-04 15:39:39 -07:00
configfs fs/configfs: use pr_fmt 2014-06-04 16:53:53 -07:00
cramfs Major changes for 3.14 include support for the newly added ZERO_RANGE 2014-04-04 15:39:39 -07:00
debugfs Major changes for 3.14 include support for the newly added ZERO_RANGE 2014-04-04 15:39:39 -07:00
devpts fs: push sync_filesystem() down to the file system's remount_fs() 2014-03-13 10:14:33 -04:00
dlm net: Fix use after free by removing length arg from sk_data_ready callbacks. 2014-04-11 16:15:36 -04:00
ecryptfs Merge branch 'cross-rename' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs 2014-04-04 14:03:05 -07:00
efivarfs
efs Major changes for 3.14 include support for the newly added ZERO_RANGE 2014-04-04 15:39:39 -07:00
exofs Merge branch 'for-linus' of git://git.open-osd.org/linux-open-osd 2014-04-10 14:33:02 -07:00
exportfs
ext2 Merge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs 2014-04-07 17:59:17 -07:00
ext3 Merge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs 2014-04-07 17:59:17 -07:00
ext4 mm: non-atomically mark page accessed during page cache allocation where possible 2014-06-04 16:54:10 -07:00
f2fs mm: non-atomically mark page accessed during page cache allocation where possible 2014-06-04 16:54:10 -07:00
fat Major changes for 3.14 include support for the newly added ZERO_RANGE 2014-04-04 15:39:39 -07:00
freevxfs Major changes for 3.14 include support for the newly added ZERO_RANGE 2014-04-04 15:39:39 -07:00
fscache fs/fscache: replace seq_printf by seq_puts 2014-06-04 16:53:52 -07:00
fuse mm: non-atomically mark page accessed during page cache allocation where possible 2014-06-04 16:54:10 -07:00
gfs2 mm: non-atomically mark page accessed during page cache allocation where possible 2014-06-04 16:54:10 -07:00
hfs Major changes for 3.14 include support for the newly added ZERO_RANGE 2014-04-04 15:39:39 -07:00
hfsplus Major changes for 3.14 include support for the newly added ZERO_RANGE 2014-04-04 15:39:39 -07:00
hostfs mm + fs: store shadow entries in page cache 2014-04-03 16:21:01 -07:00
hpfs Major changes for 3.14 include support for the newly added ZERO_RANGE 2014-04-04 15:39:39 -07:00
hppfs
hugetlbfs fs/hugetlbfs/inode.c: remove null test before kfree 2014-06-04 16:54:11 -07:00
isofs Merge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs 2014-04-07 17:59:17 -07:00
jbd
jbd2 arch: Mass conversion of smp_mb__*() 2014-04-18 14:20:48 +02:00
jffs2 MTD updates for 3.15: 2014-04-07 10:17:30 -07:00
jfs fs/jfs/super.c: convert simple_str to kstr 2014-06-03 14:14:00 -05:00
kernfs kernfs: move the last knowledge of sysfs out from kernfs 2014-05-27 14:33:17 -07:00
lockd lockd: ensure we tear down any live sockets when socket creation fails during lockd_up 2014-03-28 10:43:08 -04:00
logfs mm + fs: store shadow entries in page cache 2014-04-03 16:21:01 -07:00
minix Major changes for 3.14 include support for the newly added ZERO_RANGE 2014-04-04 15:39:39 -07:00
ncpfs Merge git://git.kernel.org/pub/scm/linux/kernel/git/davem/net 2014-04-12 17:31:22 -07:00
nfs File locking related changes for v3.16 2014-06-04 08:12:50 -07:00
nfs_common
nfsd nfsd4: warn on finding lockowner without stateid's 2014-05-21 11:11:21 -04:00
nilfs2 mm: implement ->map_pages for page cache 2014-04-07 16:35:53 -07:00
nls
notify fanotify: check file flags passed in fanotify_init 2014-06-04 16:53:52 -07:00
ntfs mm: non-atomically mark page accessed during page cache allocation where possible 2014-06-04 16:54:10 -07:00
ocfs2 fs/buffer.c: remove block_write_full_page_endio() 2014-06-04 16:54:02 -07:00
omfs mm + fs: store shadow entries in page cache 2014-04-03 16:21:01 -07:00
openpromfs fs: push sync_filesystem() down to the file system's remount_fs() 2014-03-13 10:14:33 -04:00
proc mm: softdirty: clear VM_SOFTDIRTY flag inside clear_refs_write() instead of clear_soft_dirty() 2014-06-04 16:53:56 -07:00
pstore Major changes for 3.14 include support for the newly added ZERO_RANGE 2014-04-04 15:39:39 -07:00
qnx4 fs: push sync_filesystem() down to the file system's remount_fs() 2014-03-13 10:14:33 -04:00
qnx6 fs: push sync_filesystem() down to the file system's remount_fs() 2014-03-13 10:14:33 -04:00
quota Merge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs 2014-04-07 17:59:17 -07:00
ramfs
reiserfs Merge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs 2014-04-07 17:59:17 -07:00
romfs fs: push sync_filesystem() down to the file system's remount_fs() 2014-03-13 10:14:33 -04:00
squashfs fs/squashfs/squashfs.h: replace pr_warning by pr_warn 2014-06-04 16:53:52 -07:00
sysfs kernfs: move the last knowledge of sysfs out from kernfs 2014-05-27 14:33:17 -07:00
sysv Major changes for 3.14 include support for the newly added ZERO_RANGE 2014-04-04 15:39:39 -07:00
ubifs Merge branch 'locking-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip into next 2014-06-03 12:57:53 -07:00
udf Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2014-04-12 14:49:50 -07:00
ufs fs/ufs: remove unused ufs_super_block_third pointer 2014-04-07 16:36:16 -07:00
xfs xfs: list_lru_init returns a negative error 2014-05-15 09:23:24 +10:00
aio.c aio: fix potential leak in aio_run_iocb(). 2014-05-01 08:37:43 -04:00
anon_inodes.c vfs: Allocate anon_inode_inode in anon_inode_init() 2014-03-27 09:52:54 -07:00
attr.c
bad_inode.c
binfmt_aout.c
binfmt_elf_fdpic.c
binfmt_elf.c exec: kill the unnecessary mm->def_flags setting in load_elf_binary() 2014-04-07 16:35:52 -07:00
binfmt_em86.c
binfmt_flat.c
binfmt_misc.c binfmt_misc: add missing 'break' statement 2014-04-03 16:21:16 -07:00
binfmt_script.c
binfmt_som.c
block_dev.c fs/block_dev.c: add bdev_read_page() and bdev_write_page() 2014-06-04 16:54:02 -07:00
buffer.c mm: non-atomically mark page accessed during page cache allocation where possible 2014-06-04 16:54:10 -07:00
char_dev.c
compat_binfmt_elf.c
compat_ioctl.c
compat.c locks: rename file-private locks to "open file description locks" 2014-04-22 08:23:58 -04:00
coredump.c coredump: fix va_list corruption 2014-04-19 13:23:31 -07:00
dcache.c dcache: add missing lockdep annotation 2014-05-31 09:13:21 -07:00
dcookies.c
direct-io.c xfs: update for 3.15-rc1 2014-04-04 15:50:08 -07:00
drop_caches.c drop_caches: add some documentation and info message 2014-04-03 16:21:04 -07:00
eventfd.c
eventpoll.c
exec.c metag: Reduce maximum stack size to 256MB 2014-05-15 00:00:35 +01:00
fcntl.c locks: rename file-private locks to "open file description locks" 2014-04-22 08:23:58 -04:00
fhandle.c
file_table.c Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2014-04-12 14:49:50 -07:00
file.c Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2014-04-12 14:49:50 -07:00
filesystems.c sys_sysfs: Add CONFIG_SYSFS_SYSCALL 2014-04-03 16:21:05 -07:00
fs_struct.c
fs-writeback.c One of the main highlights this time, is not the patches themselves 2014-04-04 14:49:16 -07:00
inode.c Major changes for 3.14 include support for the newly added ZERO_RANGE 2014-04-04 15:39:39 -07:00
internal.h
ioctl.c
Kconfig
Kconfig.binfmt
libfs.c fs/libfs.c: add generic data flush to fsync 2014-06-04 16:53:55 -07:00
locks.c locks: add some tracepoints in the lease handling code 2014-06-02 08:09:30 -04:00
Makefile block: move ioprio.c from fs/ to block/ 2014-05-19 11:02:18 -06:00
mbcache.c ext4: each filesystem creates and uses its own mb_cache 2014-03-18 19:24:49 -04:00
mount.h reduce m_start() cost... 2014-04-01 23:19:09 -04:00
mpage.c fs/block_dev.c: add bdev_read_page() and bdev_write_page() 2014-06-04 16:54:02 -07:00
namei.c fix races between __d_instantiate() and checks of dentry flags 2014-04-19 12:30:58 -04:00
namespace.c VFS: Make delayed_free() call free_vfsmnt() 2014-04-01 23:19:18 -04:00
no-block.c
open.c These are regression and bug fixes for ext4. 2014-04-20 20:43:47 -07:00
pipe.c switch pipe_read() to copy_page_to_iter() 2014-04-01 23:19:22 -04:00
pnode.c smarter propagate_mnt() 2014-04-01 23:19:08 -04:00
pnode.h smarter propagate_mnt() 2014-04-01 23:19:08 -04:00
posix_acl.c posix_acl: handle NULL ACL in posix_acl_equiv_mode 2014-05-06 13:58:42 -04:00
proc_namespace.c reduce m_start() cost... 2014-04-01 23:19:09 -04:00
read_write.c Merge branch 'compat' of git://git.kernel.org/pub/scm/linux/kernel/git/s390/linux 2014-03-31 14:32:17 -07:00
readdir.c fanotify: create FAN_ACCESS event for readdir 2014-06-04 16:53:52 -07:00
select.c
seq_file.c
signalfd.c
splice.c vfs: fix vmplice_to_user() 2014-05-28 01:54:52 -04:00
stack.c
stat.c
statfs.c
super.c fs/superblock: avoid locking counting inodes and dentries before reclaiming them 2014-06-04 16:54:11 -07:00
sync.c
timerfd.c
utimes.c
xattr.c