linux/fs
Dave Chinner 16fd536737 xfs: track AGs with reclaimable inodes in per-ag radix tree
https://bugzilla.kernel.org/show_bug.cgi?id=16348

When the filesystem grows to a large number of allocation groups,
the summing of recalimable inodes gets expensive. In many cases,
most AGs won't have any reclaimable inodes and so we are wasting CPU
time aggregating over these AGs. This is particularly important for
the inode shrinker that gets called frequently under memory
pressure.

To avoid the overhead, track AGs with reclaimable inodes in the
per-ag radix tree so that we can find all the AGs with reclaimable
inodes via a simple gang tag lookup. This involves setting the tag
when the first reclaimable inode is tracked in the AG, and removing
the tag when the last reclaimable inode is removed from the tree.
Then the summation process becomes a loop walking the radix tree
summing AGs with the reclaim tag set.

This significantly reduces the overhead of scanning - a 6400 AG
filesystea now only uses about 25% of a cpu in kswapd while slab
reclaim progresses instead of being permanently stuck at 100% CPU
and making little progress. Clean filesystems filesystems will see
no overhead and the overhead only increases linearly with the number
of dirty AGs.

Signed-off-by: Dave Chinner <dchinner@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
2010-07-20 09:43:39 +10:00
..
9p drop unused dentry argument to ->fsync 2010-05-27 22:05:02 -04:00
adfs kill spurious reference to vmtruncate 2010-05-27 22:15:42 -04:00
affs drop unused dentry argument to ->fsync 2010-05-27 22:05:02 -04:00
afs writeback: remove writeback_inodes_wbc 2010-07-06 08:54:03 +02:00
autofs fs/: do not fallback to default_llseek() when readdir() uses BKL 2010-05-27 09:12:56 -07:00
autofs4 fs/autofs4: use memdup_user 2010-05-27 09:12:41 -07:00
befs include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h 2010-03-30 22:02:32 +09:00
bfs rename the generic fsync implementations 2010-05-27 22:06:06 -04:00
btrfs writeback: remove writeback_inodes_wbc 2010-07-06 08:54:03 +02:00
cachefiles CacheFiles: Fix error handling in cachefiles_determine_cache_security() 2010-05-12 18:23:58 -07:00
ceph ceph: fix crush device 'out' threshold to 1.0, not 0.1 2010-07-05 09:44:17 -07:00
cifs cifs: remove bogus first_time check in NTLMv2 session setup code 2010-06-16 13:40:18 -04:00
coda drop unused dentry argument to ->fsync 2010-05-27 22:05:02 -04:00
configfs fix setattr error handling in sysfs, configfs 2010-06-04 17:16:29 -04:00
cramfs
debugfs Add x64 support to debugfs 2010-05-19 22:41:57 -04:00
devpts Simplify devpts_get_sb() failure exits 2010-05-21 18:31:12 -04:00
dlm dlm: fix ast ordering for user locks 2010-04-30 14:52:51 -05:00
ecryptfs kill spurious reference to vmtruncate 2010-05-27 22:15:42 -04:00
efs
exofs drop unused dentry argument to ->fsync 2010-05-27 22:05:02 -04:00
exportfs
ext2 ext2: update ctime when changing the file's permission by setfacl 2010-06-25 01:20:37 +02:00
ext3 ext3: update ctime when changing the file's permission by setfacl 2010-06-25 01:20:37 +02:00
ext4 ext4: Fix remaining racy updates of EXT4_I(inode)->i_flags 2010-06-05 11:51:27 -04:00
fat fat: convert to use the new truncate convention. 2010-05-27 22:16:02 -04:00
freevxfs fs/: do not fallback to default_llseek() when readdir() uses BKL 2010-05-27 09:12:56 -07:00
fscache FS-Cache: Remove unneeded null checks 2010-06-01 13:32:11 -07:00
fuse Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse 2010-05-30 09:16:14 -07:00
gfs2 mm: add context argument to shrinker callback 2010-07-19 14:56:17 +10:00
hfs include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h 2010-03-30 22:02:32 +09:00
hfsplus hfsplus: Push down BKL into ioctl function 2010-05-17 05:27:03 +02:00
hostfs drop unused dentry argument to ->fsync 2010-05-27 22:05:02 -04:00
hpfs drop unused dentry argument to ->fsync 2010-05-27 22:05:02 -04:00
hppfs drop unused dentry argument to ->fsync 2010-05-27 22:05:02 -04:00
hugetlbfs rename the generic fsync implementations 2010-05-27 22:06:06 -04:00
isofs fs/: do not fallback to default_llseek() when readdir() uses BKL 2010-05-27 09:12:56 -07:00
jbd ext3: Fix waiting on transaction during fsync 2010-05-21 19:30:41 +02:00
jbd2 jbd2/ocfs2: Fix block checksumming when a buffer is used in several transactions 2010-07-15 15:17:47 -07:00
jffs2 Merge git://git.infradead.org/~dwmw2/mtd-2.6.35 2010-06-07 17:10:06 -07:00
jfs Merge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs-2.6 2010-05-30 09:11:11 -07:00
lockd include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h 2010-03-30 22:02:32 +09:00
logfs drop unused dentry argument to ->fsync 2010-05-27 22:05:02 -04:00
minix Minix: Clean up left over label 2010-06-04 17:16:30 -04:00
ncpfs drop unused dentry argument to ->fsync 2010-05-27 22:05:02 -04:00
nfs mm: add context argument to shrinker callback 2010-07-19 14:56:17 +10:00
nfs_common include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h 2010-03-30 22:02:32 +09:00
nfsd Merge branch 'for-2.6.35' of git://linux-nfs.org/~bfields/linux 2010-06-09 12:43:04 -07:00
nilfs2 nilfs2: remove obsolete declarations of cache constructor and destructor 2010-05-31 20:50:29 +09:00
nls
notify Saner locking around deactivate_super() 2010-05-21 18:31:14 -04:00
ntfs drop unused dentry argument to ->fsync 2010-05-27 22:05:02 -04:00
ocfs2 Merge branch 'upstream-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jlbec/ocfs2 2010-07-18 10:09:25 -07:00
omfs rename the generic fsync implementations 2010-05-27 22:06:06 -04:00
openpromfs
partitions [S390] dasd: use correct label location for diag fba disks 2010-07-19 09:22:50 +02:00
proc nommu: add '[stack]' label to /proc/pid/maps output 2010-06-29 15:29:30 -07:00
qnx4 rename the generic fsync implementations 2010-05-27 22:06:06 -04:00
quota mm: add context argument to shrinker callback 2010-07-19 14:56:17 +10:00
ramfs fs: convert simple fs to new truncate 2010-05-27 22:15:47 -04:00
reiserfs Merge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs-2.6 2010-05-30 09:11:11 -07:00
romfs
smbfs kill spurious reference to vmtruncate 2010-05-27 22:15:42 -04:00
squashfs squashfs: fix name reading in squashfs_xattr_get 2010-05-23 08:27:42 +01:00
sysfs fix setattr error handling in sysfs, configfs 2010-06-04 17:16:29 -04:00
sysv sysvfs: fix NULL deref. when allocating new inode 2010-06-29 15:29:32 -07:00
ubifs mm: add context argument to shrinker callback 2010-07-19 14:56:17 +10:00
udf Merge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs-2.6 2010-05-30 09:11:11 -07:00
ufs Merge branch 'for_linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jack/linux-fs-2.6 2010-05-30 09:11:11 -07:00
xfs xfs: track AGs with reclaimable inodes in per-ag radix tree 2010-07-20 09:43:39 +10:00
aio.c get rid of the magic around f_count in aio 2010-05-27 22:03:07 -04:00
anon_inodes.c Revert "anon_inode: set S_IFREG on the anon_inode" 2010-05-27 22:03:05 -04:00
attr.c fs: introduce new truncate sequence 2010-05-27 22:15:33 -04:00
bad_inode.c drop unused dentry argument to ->fsync 2010-05-27 22:05:02 -04:00
binfmt_aout.c include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h 2010-03-30 22:02:32 +09:00
binfmt_elf_fdpic.c binfmt_elf_fdpic: Fix clear_user() error handling 2010-06-01 08:11:06 -07:00
binfmt_elf.c coredump: pass mm->flags as a coredump parameter for consistency 2010-03-06 11:26:46 -08:00
binfmt_em86.c include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h 2010-03-30 22:02:32 +09:00
binfmt_flat.c flat: tweak default stack alignment 2010-06-29 15:29:31 -07:00
binfmt_misc.c
binfmt_script.c include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h 2010-03-30 22:02:32 +09:00
binfmt_som.c
bio-integrity.c include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h 2010-03-30 22:02:32 +09:00
bio.c Merge branch 'master' into for-linus 2010-03-19 08:05:10 +01:00
block_dev.c block: remove duplicate BUG_ON() in bd_finish_claiming() 2010-06-10 19:08:34 +02:00
buffer.c fs: introduce new truncate sequence 2010-05-27 22:15:33 -04:00
char_dev.c
compat_binfmt_elf.c elf coredump: replace ELF_CORE_EXTRA_* macros by functions 2010-03-06 11:26:45 -08:00
compat_ioctl.c pktcdvd: improve BKL and compat_ioctl.c usage 2010-04-29 08:44:37 -07:00
compat.c fs/compat_rw_copy_check_uvector: add missing compat_ptr call 2010-06-04 15:21:44 -07:00
dcache.c mm: add context argument to shrinker callback 2010-07-19 14:56:17 +10:00
dcookies.c
direct-io.c fs: introduce new truncate sequence 2010-05-27 22:15:33 -04:00
drop_caches.c new helper: iterate_supers() 2010-05-21 18:31:16 -04:00
eventfd.c include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h 2010-03-30 22:02:32 +09:00
eventpoll.c sched, wait: Use wrapper functions 2010-05-11 17:43:58 +02:00
exec.c exit: avoid sig->count in de_thread/__exit_signal synchronization 2010-05-27 09:12:46 -07:00
fcntl.c fs/fcntl.c:kill_fasync_rcu() fa_lock must be IRQ-safe 2010-06-29 15:29:32 -07:00
fifo.c include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h 2010-03-30 22:02:32 +09:00
file_table.c get rid of the magic around f_count in aio 2010-05-27 22:03:07 -04:00
file.c fs: use rlimit helpers 2010-03-06 11:26:29 -08:00
filesystems.c include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h 2010-03-30 22:02:32 +09:00
fs_struct.c
fs-writeback.c writeback: simplify the write back thread queue 2010-07-06 08:59:53 +02:00
generic_acl.c fs: xattr_handler table should be const 2010-05-21 18:31:18 -04:00
inode.c mm: add context argument to shrinker callback 2010-07-19 14:56:17 +10:00
internal.h Bury __put_super_and_need_restart() 2010-05-21 18:31:16 -04:00
ioctl.c Introduce freeze_super and thaw_super for the fsfreeze ioctl 2010-05-21 18:31:18 -04:00
ioprio.c include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h 2010-03-30 22:02:32 +09:00
Kconfig Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/sage/ceph-client 2010-03-19 09:43:06 -07:00
Kconfig.binfmt
libfs.c wrong type for 'magic' argument in simple_fill_super() 2010-06-04 17:16:28 -04:00
locks.c Merge branch 'for-next' into for-linus 2010-03-08 16:55:37 +01:00
Makefile Take statfs variants to fs/statfs.c 2010-05-21 18:31:17 -04:00
mbcache.c mm: add context argument to shrinker callback 2010-07-19 14:56:17 +10:00
mpage.c include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h 2010-03-30 22:02:32 +09:00
namei.c VFS: fix recent breakage of FS_REVAL_DOT 2010-05-27 22:03:06 -04:00
namespace.c Merge branch 'next' into for-linus 2010-05-18 08:57:00 +10:00
nfsctl.c
no-block.c
open.c Take statfs variants to fs/statfs.c 2010-05-21 18:31:17 -04:00
pipe.c pipe: fix check in "set size" fcntl 2010-06-10 19:08:34 +02:00
pnode.c
pnode.h
posix_acl.c
read_write.c vfs: introduce noop_llseek() 2010-05-27 09:12:56 -07:00
read_write.h
readdir.c
select.c Add generic sys_old_select() 2010-03-12 15:52:32 -08:00
seq_file.c seq_file: fix new kernel-doc warnings 2010-03-07 15:48:26 -08:00
signalfd.c include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h 2010-03-30 22:02:32 +09:00
splice.c splice: check f_mode for seekable file 2010-06-30 08:12:37 +02:00
stack.c
stat.c
statfs.c Take statfs variants to fs/statfs.c 2010-05-21 18:31:17 -04:00
super.c fs: fix superblock iteration race 2010-06-29 10:38:22 -07:00
sync.c Merge branch 'master' into for-linus 2010-06-01 12:42:12 +02:00
timerfd.c fs/timerfd.c: make use of wait_event_interruptible_locked_irq() 2010-05-20 13:21:42 -07:00
utimes.c
xattr_acl.c include cleanup: Update gfp.h and slab.h includes to prepare for breaking implicit slab.h inclusion from percpu.h 2010-03-30 22:02:32 +09:00
xattr.c fs: xattr_handler table should be const 2010-05-21 18:31:18 -04:00