md/raid1: close some possible races on write errors during resync
authorNeilBrown <neilb@suse.de>
Thu, 19 Jul 2012 05:59:18 +0000 (15:59 +1000)
committerGreg Kroah-Hartman <gregkh@linuxfoundation.org>
Sun, 29 Jul 2012 15:04:17 +0000 (08:04 -0700)
commitd8ae4bb4a1c12f9cfb373f215e624a9f1fca0767
tree669ac900707bb663af164e972eeef30b519a79c4
parent2ac8a0f58a8782ff8024404a579eee260e8b2010
md/raid1: close some possible races on write errors during resync

commit 58e94ae18478c08229626daece2fc108a4a23261 upstream.

commit 4367af556133723d0f443e14ca8170d9447317cb
   md/raid1: clear bad-block record when write succeeds.

Added a 'reschedule_retry' call possibility at the end of
end_sync_write, but didn't add matching code at the end of
sync_request_write.  So if the writes complete very quickly, or
scheduling makes it seem that way, then we can miss rescheduling
the request and the resync could hang.

Also commit 73d5c38a9536142e062c35997b044e89166e063b
    md: avoid races when stopping resync.

Fix a race condition in this same code in end_sync_write but didn't
make the change in sync_request_write.

This patch updates sync_request_write to fix both of those.
Patch is suitable for 3.1 and later kernels.

Reported-by: Alexander Lyakas <alex.bolshoy@gmail.com>
Original-version-by: Alexander Lyakas <alex.bolshoy@gmail.com>
Signed-off-by: NeilBrown <neilb@suse.de>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>
drivers/md/raid1.c