iv/linux
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity
c771c14baa
linux/fs
History
Eryu Guan c771c14baa iomap: invalidate page caches should be after iomap_dio_complete() in direct write 
After XFS switching to iomap based DIO (commit acdda3aae1 ("xfs:
use iomap_dio_rw")), I started to notice dio29/dio30 tests failures
from LTP run on ppc64 hosts, and they can be reproduced on x86_64
hosts with 512B/1k block size XFS too.

dio29	diotest3 -b 65536 -n 100 -i 1000 -o 1024000
dio30	diotest6 -b 65536 -n 100 -i 1000 -o 1024000

The failure message is like:
bufcmp: offset 0: Expected: 0x62, got 0x0
diotest03    1  TPASS  :  Read with Direct IO, Write without
diotest03    2  TFAIL  :  diotest3.c:142: comparsion failed; child=98 offset=1425408
diotest03    3  TFAIL  :  diotest3.c:194: Write Direct-child 98 failed

Direct write wrote 0x62 but buffer read got zero. This is because,
when doing direct write to a hole or preallocated file, we
invalidate the page caches before converting the extent from
unwritten state to normal state, which is done by
iomap_dio_complete(), thus leave a window for other buffer reader to
cache the unwritten state extent.

Consider this case, with sub-page blocksize XFS, two processes are
direct writing to different blocksize-aligned regions (say 512B) of
the same preallocated file, and reading the region back via buffered
I/O to compare contents.

process A, region [0,512]		process B, region [512,1024]
xfs_file_write_iter
 xfs_file_aio_dio_write
  iomap_dio_rw
   iomap_apply
   invalidate_inode_pages2_range
   					xfs_file_write_iter
				 	xfs_file_aio_dio_write
					  iomap_dio_rw
					   iomap_apply
					   invalidate_inode_pages2_range
					   iomap_dio_complete
					xfs_file_read_iter
					 xfs_file_buffered_aio_read
					  generic_file_read_iter
					   do_generic_file_read
					    <readahead fills pagecache with 0>
   iomap_dio_complete
xfs_file_read_iter
 <read gets 0 from pagecache>

Process A first invalidates page caches, at this point the
underlying extent is still in unwritten state (iomap_dio_complete
not called yet), and process B finishs direct write and populates
page caches via readahead, which caches zeros in page for region A,
then process A reads zeros from page cache, instead of the actual
data.

Fix it by invalidating page caches after converting unwritten extent
to make sure we read content from disk after extent state changed,
as what we did before switching to iomap based dio.

Also introduce a new 'start' variable to save the original write
offset (iomap_dio_complete() updates iocb->ki_pos), and a 'err'
variable for invalidating caches result, cause we can't reuse 'ret'
anymore.

Signed-off-by: Eryu Guan <eguan@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
Reviewed-by: Darrick J. Wong <darrick.wong@oracle.com>
Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
2017-03-06 09:50:01 -08:00
..
9p Merge branch 'work.misc' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2017-03-03 21:44:35 -08:00
adfs
affs sched/headers: Prepare to remove <linux/cred.h> inclusion from <linux/sched.h> 2017-03-02 08:42:31 +01:00
afs Merge branch 'rebased-statx' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2017-03-03 11:38:56 -08:00
autofs4 sched/headers: Prepare to remove <linux/cred.h> inclusion from <linux/sched.h> 2017-03-02 08:42:31 +01:00
befs sched/headers: Prepare to remove <linux/cred.h> inclusion from <linux/sched.h> 2017-03-02 08:42:31 +01:00
bfs Replace <asm/uaccess.h> with <linux/uaccess.h> globally 2016-12-24 11:46:01 -08:00
btrfs Merge branch 'rebased-statx' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2017-03-03 11:38:56 -08:00
cachefiles sched/headers: Prepare to remove <linux/cred.h> inclusion from <linux/sched.h> 2017-03-02 08:42:31 +01:00
ceph Merge branch 'rebased-statx' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2017-03-03 11:38:56 -08:00
cifs Merge branch 'for-next' of git://git.samba.org/sfrench/cifs-2.6 2017-03-03 16:00:59 -08:00
coda Merge branch 'rebased-statx' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2017-03-03 11:38:56 -08:00
configfs Replace <asm/uaccess.h> with <linux/uaccess.h> globally 2016-12-24 11:46:01 -08:00
cramfs
crypto KEYS: Differentiate uses of rcu_dereference_key() and user_key_payload() 2017-03-02 10:09:00 +11:00
debugfs Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace 2017-02-23 20:33:51 -08:00
devpts
dlm sched/headers: Prepare to move signal wakeup & sigpending methods from <linux/sched.h> into <linux/sched/signal.h> 2017-03-02 08:42:32 +01:00
ecryptfs Merge branch 'rebased-statx' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2017-03-03 11:38:56 -08:00
efivarfs
efs Replace <asm/uaccess.h> with <linux/uaccess.h> globally 2016-12-24 11:46:01 -08:00
exofs locking/atomic, kref: Add kref_read() 2017-01-14 11:37:18 +01:00
exportfs Merge branch 'rebased-statx' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2017-03-03 11:38:56 -08:00
ext2 sched/headers: Prepare to remove <linux/cred.h> inclusion from <linux/sched.h> 2017-03-02 08:42:31 +01:00
ext4 Merge branch 'rebased-statx' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2017-03-03 11:38:56 -08:00
f2fs Merge branch 'rebased-statx' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2017-03-03 11:38:56 -08:00
fat statx: Add a system call to make enhanced file info available 2017-03-02 20:51:15 -05:00
freevxfs
fscache KEYS: Differentiate uses of rcu_dereference_key() and user_key_payload() 2017-03-02 10:09:00 +11:00
fuse Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/fuse 2017-03-03 12:14:13 -08:00
gfs2 Merge branch 'rebased-statx' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2017-03-03 11:38:56 -08:00
hfs Merge branch 'work.misc' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2017-03-03 21:44:35 -08:00
hfsplus sched/headers: Prepare to remove <linux/cred.h> inclusion from <linux/sched.h> 2017-03-02 08:42:31 +01:00
hostfs
hpfs sched/headers: Prepare to move signal wakeup & sigpending methods from <linux/sched.h> into <linux/sched/signal.h> 2017-03-02 08:42:32 +01:00
hugetlbfs sched/headers: Prepare to move signal wakeup & sigpending methods from <linux/sched.h> into <linux/sched/signal.h> 2017-03-02 08:42:32 +01:00
isofs sched/headers: Prepare to remove <linux/cred.h> inclusion from <linux/sched.h> 2017-03-02 08:42:31 +01:00
jbd2 For this cycle we add support for the shutdown ioctl, which is 2017-02-20 18:24:39 -08:00
jffs2 sched/headers: Prepare to move signal wakeup & sigpending methods from <linux/sched.h> into <linux/sched/signal.h> 2017-03-02 08:42:32 +01:00
jfs fs: add i_blocksize() 2017-02-27 18:43:46 -08:00
kernfs Merge branch 'rebased-statx' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2017-03-03 11:38:56 -08:00
lockd sched/headers: Prepare for new header dependencies before moving code to <linux/sched/signal.h> 2017-03-02 08:42:29 +01:00
minix statx: Add a system call to make enhanced file info available 2017-03-02 20:51:15 -05:00
ncpfs Merge branch 'WIP.sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2017-03-03 10:16:38 -08:00
nfs Merge branch 'rebased-statx' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2017-03-03 11:38:56 -08:00
nfs_common
nfsd Merge branch 'rebased-statx' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2017-03-03 11:38:56 -08:00
nilfs2 sched/headers: Prepare to move signal wakeup & sigpending methods from <linux/sched.h> into <linux/sched/signal.h> 2017-03-02 08:42:32 +01:00
nls
notify sched/headers: Prepare to move signal wakeup & sigpending methods from <linux/sched.h> into <linux/sched/signal.h> 2017-03-02 08:42:32 +01:00
ntfs sched/headers: Prepare to move signal wakeup & sigpending methods from <linux/sched.h> into <linux/sched/signal.h> 2017-03-02 08:42:32 +01:00
ocfs2 Merge branch 'rebased-statx' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2017-03-03 11:38:56 -08:00
omfs sched/headers: Prepare to remove <linux/cred.h> inclusion from <linux/sched.h> 2017-03-02 08:42:31 +01:00
openpromfs Replace <asm/uaccess.h> with <linux/uaccess.h> globally 2016-12-24 11:46:01 -08:00
orangefs Merge branch 'rebased-statx' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2017-03-03 11:38:56 -08:00
overlayfs Merge branch 'overlayfs-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/mszeredi/vfs 2017-03-03 12:02:42 -08:00
proc Merge branch 'rebased-statx' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2017-03-03 11:38:56 -08:00
pstore fs/pstore: fs/squashfs: change usage of LZ4 to work with new LZ4 version 2017-02-24 17:46:57 -08:00
qnx4
qnx6
quota sched/headers: Prepare to remove <linux/cred.h> inclusion from <linux/sched.h> 2017-03-02 08:42:31 +01:00
ramfs Replace <asm/uaccess.h> with <linux/uaccess.h> globally 2016-12-24 11:46:01 -08:00
reiserfs fs: add i_blocksize() 2017-02-27 18:43:46 -08:00
romfs romfs: use different way to generate fsid for BLOCK or MTD 2017-01-24 16:26:14 -08:00
squashfs fs/pstore: fs/squashfs: change usage of LZ4 to work with new LZ4 version 2017-02-24 17:46:57 -08:00
sysfs
sysv statx: Add a system call to make enhanced file info available 2017-03-02 20:51:15 -05:00
tracefs
ubifs statx: Add a system call to make enhanced file info available 2017-03-02 20:51:15 -05:00
udf statx: Add a system call to make enhanced file info available 2017-03-02 20:51:15 -05:00
ufs Replace <asm/uaccess.h> with <linux/uaccess.h> globally 2016-12-24 11:46:01 -08:00
xfs Merge branch 'rebased-statx' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2017-03-03 11:38:56 -08:00
aio.c Merge branch 'WIP.sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2017-03-03 10:16:38 -08:00
anon_inodes.c Replace <asm/uaccess.h> with <linux/uaccess.h> globally 2016-12-24 11:46:01 -08:00
attr.c sched/headers: Prepare for new header dependencies before moving code to <linux/sched/signal.h> 2017-03-02 08:42:29 +01:00
bad_inode.c statx: Add a system call to make enhanced file info available 2017-03-02 20:51:15 -05:00
binfmt_aout.c sched/headers: Prepare for new header dependencies before moving code to <linux/sched/task_stack.h> 2017-03-02 08:42:36 +01:00
binfmt_elf_fdpic.c sched/headers: Prepare to move cputime functionality from <linux/sched.h> into <linux/sched/cputime.h> 2017-03-02 08:42:39 +01:00
binfmt_elf.c sched/headers: Prepare to move cputime functionality from <linux/sched.h> into <linux/sched/cputime.h> 2017-03-02 08:42:39 +01:00
binfmt_em86.c
binfmt_flat.c sched/headers: Prepare for new header dependencies before moving code to <linux/sched/task_stack.h> 2017-03-02 08:42:36 +01:00
binfmt_misc.c sched/headers: Prepare to remove the <linux/mm_types.h> dependency from <linux/sched.h> 2017-03-02 08:42:37 +01:00
binfmt_script.c
block_dev.c block: Initialize bd_bdi on inode initialization 2017-03-02 08:56:59 -07:00
buffer.c sched/headers: Prepare for the reduction of <linux/sched.h>'s signal API dependency 2017-03-02 08:42:37 +01:00
char_dev.c
compat_binfmt_elf.c fs/binfmt: Convert obsolete cputime type to nsecs 2017-02-01 09:13:51 +01:00
compat_ioctl.c Replace <asm/uaccess.h> with <linux/uaccess.h> globally 2016-12-24 11:46:01 -08:00
compat.c sched/headers: Prepare to remove <linux/cred.h> inclusion from <linux/sched.h> 2017-03-02 08:42:31 +01:00
coredump.c sched/headers: Prepare for new header dependencies before moving code to <linux/sched/task_stack.h> 2017-03-02 08:42:36 +01:00
dax.c sched/headers: Prepare for the reduction of <linux/sched.h>'s signal API dependency 2017-03-02 08:42:37 +01:00
dcache.c mnt: Protect the mountpoint hashtable with mount_lock 2017-01-10 13:34:43 +13:00
dcookies.c Replace <asm/uaccess.h> with <linux/uaccess.h> globally 2016-12-24 11:46:01 -08:00
direct-io.c fs: add i_blocksize() 2017-02-27 18:43:46 -08:00
drop_caches.c
eventfd.c sched/headers: Prepare to move signal wakeup & sigpending methods from <linux/sched.h> into <linux/sched/signal.h> 2017-03-02 08:42:32 +01:00
eventpoll.c sched/headers: Prepare to move signal wakeup & sigpending methods from <linux/sched.h> into <linux/sched/signal.h> 2017-03-02 08:42:32 +01:00
exec.c sched/headers: Prepare for new header dependencies before moving code to <linux/sched/task.h> 2017-03-02 08:42:35 +01:00
fcntl.c sched/headers: Prepare for new header dependencies before moving code to <linux/sched/task.h> 2017-03-02 08:42:35 +01:00
fhandle.c Replace <asm/uaccess.h> with <linux/uaccess.h> globally 2016-12-24 11:46:01 -08:00
file_table.c sched/headers: Prepare to remove <linux/cred.h> inclusion from <linux/sched.h> 2017-03-02 08:42:31 +01:00
file.c sched/headers: Prepare for new header dependencies before moving code to <linux/sched/signal.h> 2017-03-02 08:42:29 +01:00
filesystems.c Replace <asm/uaccess.h> with <linux/uaccess.h> globally 2016-12-24 11:46:01 -08:00
fs_pin.c
fs_struct.c sched/headers: Prepare for new header dependencies before moving code to <linux/sched/task.h> 2017-03-02 08:42:35 +01:00
fs-writeback.c
inode.c
internal.h iomap: constify struct iomap_ops 2017-01-30 16:32:25 -08:00
ioctl.c sched/headers: Prepare for the reduction of <linux/sched.h>'s signal API dependency 2017-03-02 08:42:37 +01:00
iomap.c iomap: invalidate page caches should be after iomap_dio_complete() in direct write 2017-03-06 09:50:01 -08:00
Kconfig dax: fix build warnings with FS_DAX and !FS_IOMAP 2017-01-24 16:26:14 -08:00
Kconfig.binfmt
libfs.c Merge branch 'rebased-statx' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2017-03-03 11:38:56 -08:00
locks.c Replace <asm/uaccess.h> with <linux/uaccess.h> globally 2016-12-24 11:46:01 -08:00
Makefile logfs: remove from tree 2016-12-14 23:48:11 -05:00
mbcache.c
mount.h mnt: Tuck mounts under others instead of creating shadow/side mounts. 2017-02-04 00:01:06 +13:00
mpage.c fs: add i_blocksize() 2017-02-27 18:43:46 -08:00
namei.c Merge remote-tracking branch 'ovl/for-viro' into for-linus 2017-03-02 06:41:22 -05:00
namespace.c sched/headers: Prepare to move 'init_task' and 'init_thread_union' from <linux/sched.h> to <linux/sched/task.h> 2017-03-02 08:42:38 +01:00
no-block.c
nsfs.c nsfs: Add an ioctl() to return owner UID of a userns 2017-02-03 14:35:43 +13:00
open.c vfs: wrap write f_ops with file_{start,end}_write() 2017-02-07 15:05:04 +01:00
pipe.c Replace <asm/uaccess.h> with <linux/uaccess.h> globally 2016-12-24 11:46:01 -08:00
pnode.c mnt: Tuck mounts under others instead of creating shadow/side mounts. 2017-02-04 00:01:06 +13:00
pnode.h mnt: Tuck mounts under others instead of creating shadow/side mounts. 2017-02-04 00:01:06 +13:00
posix_acl.c sched/headers: Prepare to remove <linux/cred.h> inclusion from <linux/sched.h> 2017-03-02 08:42:31 +01:00
proc_namespace.c sched/headers: Prepare to move the task_lock()/unlock() APIs to <linux/sched/task.h> 2017-03-02 08:42:38 +01:00
read_write.c Merge branch 'WIP.sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2017-03-03 10:16:38 -08:00
readdir.c Replace <asm/uaccess.h> with <linux/uaccess.h> globally 2016-12-24 11:46:01 -08:00
select.c sched/headers: Prepare for new header dependencies before moving code to <linux/sched/signal.h> 2017-03-02 08:42:29 +01:00
seq_file.c Replace <asm/uaccess.h> with <linux/uaccess.h> globally 2016-12-24 11:46:01 -08:00
signalfd.c
splice.c Merge branch 'WIP.sched-core-for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/tip/tip 2017-03-03 10:16:38 -08:00
stack.c
stat.c Merge branch 'rebased-statx' of git://git.kernel.org/pub/scm/linux/kernel/git/viro/vfs 2017-03-03 11:38:56 -08:00
statfs.c
super.c Merge branch 'for-linus' of git://git.kernel.org/pub/scm/linux/kernel/git/ebiederm/user-namespace 2017-02-23 20:33:51 -08:00
sync.c vfs: use helper for calling f_op->fsync() 2017-02-20 16:51:23 +01:00
timerfd.c timerfd: Protect the might cancel mechanism proper 2017-02-10 11:15:09 +01:00
userfaultfd.c sched/headers: Prepare to move signal wakeup & sigpending methods from <linux/sched.h> into <linux/sched/signal.h> 2017-03-02 08:42:32 +01:00
utimes.c Replace <asm/uaccess.h> with <linux/uaccess.h> globally 2016-12-24 11:46:01 -08:00
xattr.c Replace <asm/uaccess.h> with <linux/uaccess.h> globally 2016-12-24 11:46:01 -08:00