Btrfs: fix qgroup rescan worker initialization
authorFilipe Manana <fdmanana@suse.com>
Thu, 24 Nov 2016 02:09:04 +0000 (02:09 +0000)
committerFilipe Manana <fdmanana@suse.com>
Fri, 25 Nov 2016 18:06:50 +0000 (18:06 +0000)
commit8d9eddad19467b008e0c881bc3133d7da94b7ec1
tree59713176da0ae96b4f0b53afe000201657382907
parentf177d73949bf758542ca15a1c1945bd2e802cc65
Btrfs: fix qgroup rescan worker initialization

We were setting the qgroup_rescan_running flag to true only after the
rescan worker started (which is a task run by a queue). So if a user
space task starts a rescan and immediately after asks to wait for the
rescan worker to finish, this second call might happen before the rescan
worker task starts running, in which case the rescan wait ioctl returns
immediatley, not waiting for the rescan worker to finish.

This was making the fstest btrfs/022 fail very often.

Fixes: d2c609b834d6 (btrfs: properly track when rescan worker is running)
Cc: stable@vger.kernel.org # 4.4+
Signed-off-by: Filipe Manana <fdmanana@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
fs/btrfs/qgroup.c