linux/fs
Chris Mason 4f003fd32b ext3: Add locking to ext3_do_update_inode
I've been struggling with this off and on while I've been testing the
data=guarded work.  The symptom is corrupted orphan lists and inodes
with the wrong i_size stored on disk.  I was convinced the
data=guarded code was just missing a call to ext3_mark_inode_dirty, but
tracing showed the i_disksize I was sending to ext3_mark_inode_dirty
wasn't actually making it to the drive.

ext3_mark_inode_dirty can be called without locks held (atime updates
and a few others), so the data=guarded code uses locks while updating
the in-memory inode, and then calls ext3_mark_inode_dirty
without any locks held.

But, ext3_mark_inode_dirty has no internal locking to make sure that
only one CPU is updating the buffer head at a time.  Generally this
works out ok because everyone that changes the inode then calls
ext3_mark_inode_dirty themselves.  Even though it races, eventually
someone updates the buffer heads and things move on.

But there is still a risk of the wrong values getting in, and the
data=guarded code seems to hit the race very often.

Since everyone that changes the inode also logs it, it should be
possible to fix this with some memory barriers.  I'll leave that as an
exercise to the reader and lock the buffer head instead.

It it probably a good idea to have a different patch series for lockless
bit flipping on the ext3 i_state field.  ext3_do_update_inode &= clears
EXT3_STATE_NEW without any locks held.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
Signed-off-by: Jan Kara <jack@suse.cz>
2009-09-16 17:44:11 +02:00
..
9p 9p: remove unnecessary v9fses->options which duplicates the mount string 2009-08-17 16:42:28 -05:00
adfs headers: smp_lock.h redux 2009-07-12 12:22:34 -07:00
affs affs: add ->sync_fs 2009-06-11 21:36:14 -04:00
afs writeback: get rid of wbc->for_writepages 2009-09-16 15:16:18 +02:00
autofs switch follow_down() 2009-06-11 21:36:01 -04:00
autofs4 autofs4 - fix missed case when changing to use struct path 2009-08-31 17:44:05 -10:00
befs Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs-2.6 2009-06-17 08:46:57 -07:00
bfs headers: smp_lock.h redux 2009-07-12 12:22:34 -07:00
btrfs fs: Assign bdi in super_block 2009-09-16 15:18:51 +02:00
cachefiles enforce ->sync_fs is only called for rw superblock 2009-06-11 21:36:06 -04:00
cifs cifs: consolidate reconnect logic in smb_init routines 2009-09-03 15:30:48 +00:00
coda splice: implement default splice_read method 2009-05-11 14:13:10 +02:00
configfs writeback: add name to backing_dev_info 2009-09-11 09:20:26 +02:00
cramfs
debugfs debugfs: use specified mode to possibly mark files read/write only 2009-06-15 21:30:28 -07:00
devpts devpts: remove module-related code 2009-06-24 08:15:24 -04:00
dlm Merge branch 'master' of master.kernel.org:/pub/scm/linux/kernel/git/davem/net-2.6 2009-08-12 17:44:53 -07:00
ecryptfs eCryptfs: parse_tag_3_packet check tag 3 packet encrypted key size 2009-07-28 14:26:06 -07:00
efs get rid of BKL in fs/efs 2009-06-17 00:36:36 -04:00
exofs headers: smp_lock.h redux 2009-07-12 12:22:34 -07:00
exportfs
ext2 ext2: Update comment about generic_osync_inode 2009-09-14 17:08:16 +02:00
ext3 ext3: Add locking to ext3_do_update_inode 2009-09-16 17:44:11 +02:00
ext4 ext4: Remove syncing logic from ext4_file_write 2009-09-14 17:08:16 +02:00
fat fat: Opencode sync_page_range_nolock() 2009-09-14 17:08:17 +02:00
freevxfs headers: smp_lock.h redux 2009-07-12 12:22:34 -07:00
fscache FS-Cache: Fixup renamed filenames in comments in internal.h 2009-05-27 10:20:13 -07:00
fuse fs: Assign bdi in super_block 2009-09-16 15:18:51 +02:00
gfs2 Merge branch 'for-2.6.32' of git://git.kernel.dk/linux-2.6-block 2009-09-14 17:55:15 -07:00
hfs headers: smp_lock.h redux 2009-07-12 12:22:34 -07:00
hfsplus headers: smp_lock.h redux 2009-07-12 12:22:34 -07:00
hostfs hostfs: set maximum filesize in superblock for proper LFS support 2009-06-30 18:56:03 -07:00
hpfs headers: smp_lock.h redux 2009-07-12 12:22:34 -07:00
hppfs
hugetlbfs writeback: add name to backing_dev_info 2009-09-11 09:20:26 +02:00
isofs isofs: fix Joliet regression 2009-07-10 19:18:59 -07:00
jbd jbd: Annotate transaction start also for journal_restart() 2009-09-16 17:44:10 +02:00
jbd2 writeback: get rid of wbc->for_writepages 2009-09-16 15:16:18 +02:00
jffs2 jffs2/jfs/xfs: switch over to 'check_acl' rather than 'permission()' 2009-09-08 11:09:04 -07:00
jfs jffs2/jfs/xfs: switch over to 'check_acl' rather than 'permission()' 2009-09-08 11:09:04 -07:00
lockd lockd: Replace nsm_display_address() with rpc_ntop() 2009-08-09 15:09:39 -04:00
minix Making fs/minix/minix.h double including safe 2009-06-22 11:34:42 -07:00
ncpfs NLS: update handling of Unicode 2009-06-15 21:44:43 -07:00
nfs fs: Assign bdi in super_block 2009-09-16 15:18:51 +02:00
nfs_common
nfsd Merge branch 'nfs-for-2.6.32' 2009-09-11 14:59:37 -04:00
nilfs2 fs: remove bdev->bd_inode_backing_dev_info 2009-09-16 15:16:18 +02:00
nls NLS: update handling of Unicode 2009-06-15 21:44:43 -07:00
notify inotify: update the group mask on mark addition 2009-08-28 12:51:14 -04:00
ntfs ntfs: Use new syncing helpers and update comments 2009-09-14 17:08:16 +02:00
ocfs2 ocfs2: Update syncing after splicing to match generic version 2009-09-14 17:08:16 +02:00
omfs switch omfs to simple_fsync() 2009-06-11 21:36:13 -04:00
openpromfs
partitions driver model: constify attribute groups 2009-09-15 09:50:47 -07:00
proc mm: revert "oom: move oom_adj value" 2009-08-18 16:31:13 -07:00
qnx4 fs/qnx4: sanitize includes 2009-06-11 21:36:12 -04:00
quota quota: Silence lockdep on quota_on 2009-07-30 17:31:23 +02:00
ramfs writeback: add name to backing_dev_info 2009-09-11 09:20:26 +02:00
reiserfs headers: smp_lock.h redux 2009-07-12 12:22:34 -07:00
romfs ROMFS: romfs_dev_read() error ignored 2009-05-09 10:49:41 -04:00
smbfs push BKL down into ->put_super 2009-06-11 21:36:07 -04:00
squashfs headers: smp_lock.h redux 2009-07-12 12:22:34 -07:00
sysfs Merge branch 'writeback' of git://git.kernel.dk/linux-2.6-block 2009-09-11 09:17:05 -07:00
sysv get rid of BKL in fs/sysv 2009-06-17 00:36:37 -04:00
ubifs writeback: separate starting of sync vs opportunistic writeback 2009-09-16 15:18:52 +02:00
udf udf: Fix possible corruption when close races with write 2009-09-14 19:13:01 +02:00
ufs ufs: sector_t cannot be negative 2009-06-18 13:03:46 -07:00
xfs xfs: Convert sync_page_range() to simple filemap_write_and_wait_range() 2009-09-14 17:08:17 +02:00
aio.c eventfd: revised interface and cleanups 2009-06-30 18:55:58 -07:00
anon_inodes.c fs: Provide empty .set_page_dirty() aop for anon inodes 2009-06-18 14:46:10 +02:00
attr.c
bad_inode.c
binfmt_aout.c
binfmt_elf_fdpic.c elf_core_dump: use rcu_read_lock() to access ->real_parent 2009-06-18 13:03:52 -07:00
binfmt_elf.c binfmt_elf: fix PT_INTERP bss handling 2009-09-10 20:11:12 +10:00
binfmt_em86.c
binfmt_flat.c flat: fix uninitialized ptr with shared libs 2009-08-07 10:39:57 -07:00
binfmt_misc.c
binfmt_script.c
binfmt_som.c
bio-integrity.c block: Create bip slabs with embedded integrity vectors 2009-07-01 10:56:25 +02:00
bio.c block: fix sg SG_DXFER_TO_FROM_DEV regression 2009-07-10 20:31:53 +02:00
block_dev.c fs: remove bdev->bd_inode_backing_dev_info 2009-09-16 15:16:18 +02:00
buffer.c writeback: switch to per-bdi threads for flushing data 2009-09-11 09:20:25 +02:00
char_dev.c Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tiwai/sound-2.6 2009-09-11 09:19:35 -07:00
compat_binfmt_elf.c
compat_ioctl.c compat_ioctl: hook up compat handler for FIEMAP ioctl 2009-08-07 10:39:56 -07:00
compat.c exec: do not sleep in TASK_TRACED under ->cred_guard_mutex 2009-09-05 11:30:42 -07:00
dcache.c sched: Pull up the might_sleep() check into cond_resched() 2009-07-18 15:51:44 +02:00
dcookies.c
direct-io.c block: Do away with the notion of hardsect_size 2009-05-22 23:22:54 +02:00
drop_caches.c mm: remove __invalidate_mapping_pages variant 2009-06-16 19:47:43 -07:00
eventfd.c eventfd: revised interface and cleanups 2009-06-30 18:55:58 -07:00
eventpoll.c epoll: fix nested calls support 2009-06-18 13:03:41 -07:00
exec.c exec: do not sleep in TASK_TRACED under ->cred_guard_mutex 2009-09-05 11:30:42 -07:00
fcntl.c headers: smp_lock.h redux 2009-07-12 12:22:34 -07:00
fifo.c
file_table.c fs: move mark_files_ro into file_table.c 2009-06-11 21:36:02 -04:00
file.c
filesystems.c fs: Mark get_filesystem_list() as __init function. 2009-04-20 23:02:52 -04:00
fs_struct.c
fs-writeback.c writeback: fix possible bdi writeback refcounting problem 2009-09-16 15:18:53 +02:00
generic_acl.c
inode.c fs: remove bdev->bd_inode_backing_dev_info 2009-09-16 15:16:18 +02:00
internal.h Trim a bit of crap from fs.h 2009-06-11 21:36:07 -04:00
ioctl.c fs: Add new pre-allocation ioctls to vfs for compatibility with legacy xfs ioctls 2009-06-24 08:15:27 -04:00
ioprio.c
Kconfig fs/Kconfig: move nilfs2 outside misc filesystems 2009-09-14 18:27:16 +09:00
Kconfig.binfmt
libfs.c vfs: make get_sb_pseudo set s_maxbytes to value that can be cast to signed 2009-08-18 16:31:12 -07:00
locks.c Merge branch 'sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/linux-2.6-tip 2009-09-11 13:23:18 -07:00
Makefile nilfs2: update makefile and Kconfig 2009-04-07 08:31:16 -07:00
mbcache.c
mpage.c ext4: Properly initialize the buffer_head state 2009-05-13 15:13:42 -04:00
namei.c Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/jmorris/security-testing-2.6 2009-09-11 08:55:49 -07:00
namespace.c vfs: mnt_want_write_file(): fix special file handling 2009-08-07 10:39:56 -07:00
nfsctl.c
no-block.c
open.c CRED: Add some configurable debugging [try #6] 2009-09-02 21:29:01 +10:00
pipe.c lockdep: Fix lockdep annotation for pipe_double_lock() 2009-07-22 21:14:14 +02:00
pnode.c
pnode.h
posix_acl.c
read_write.c splice: implement default splice_read method 2009-05-11 14:13:10 +02:00
read_write.h
readdir.c
select.c poll/select: initialize triggered field of struct poll_wqueues 2009-08-15 18:40:11 -07:00
seq_file.c seq_file: add function to write binary data 2009-06-18 13:03:57 -07:00
signalfd.c
splice.c Merge branch 'for-2.6.32' of git://git.kernel.dk/linux-2.6-block 2009-09-14 17:55:15 -07:00
stack.c
stat.c kill vfs_stat_fd / vfs_lstat_fd 2009-04-20 23:02:52 -04:00
super.c fs: Assign bdi in super_block 2009-09-16 15:18:51 +02:00
sync.c fs: Assign bdi in super_block 2009-09-16 15:18:51 +02:00
timerfd.c
utimes.c
xattr_acl.c
xattr.c VFS: Factor out part of vfs_setxattr so it can be called from the SELinux hook for inode_setsecctx. 2009-09-10 10:11:22 +10:00