dm thin: fix race condition when destroying thin pool workqueue
authorNikolay Borisov <kernel@kyup.com>
Thu, 17 Dec 2015 16:03:35 +0000 (18:03 +0200)
committerJiri Slaby <jslaby@suse.cz>
Fri, 12 Feb 2016 07:58:10 +0000 (08:58 +0100)
commit021cc4c804dd34b3a4fca8175506bb8be380242d
tree92758d3c386acb895e05c22238a37a83da8f8a70
parentd43afc4a7e793dbe84b16e2b35e2522150adc260
dm thin: fix race condition when destroying thin pool workqueue

commit 18d03e8c25f173f4107a40d0b8c24defb6ed69f3 upstream.

When a thin pool is being destroyed delayed work items are
cancelled using cancel_delayed_work(), which doesn't guarantee that on
return the delayed item isn't running.  This can cause the work item to
requeue itself on an already destroyed workqueue.  Fix this by using
cancel_delayed_work_sync() which guarantees that on return the work item
is not running anymore.

Fixes: 905e51b39a555 ("dm thin: commit outstanding data every second")
Fixes: 85ad643b7e7e5 ("dm thin: add timeout to stop out-of-data-space mode holding IO forever")
Signed-off-by: Nikolay Borisov <kernel@kyup.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Cc: Nikolay Borisov <kernel@kyup.com>
Signed-off-by: Jiri Slaby <jslaby@suse.cz>
drivers/md/dm-thin.c