rbd: don't hold spinlock during messenger flush
authorAlex Elder <elder@dreamhost.com>
Wed, 4 Apr 2012 18:35:44 +0000 (13:35 -0500)
committerAlex Elder <elder@dreamhost.com>
Thu, 5 Apr 2012 20:43:58 +0000 (15:43 -0500)
commitcd9d9f5df6098c50726200d4185e9e8da32785b3
tree94e4c7035fc7c7f6596a8f3b59fa27e19a45d4f0
parentc666601a935b94cc0f3310339411b6940de751ba
rbd: don't hold spinlock during messenger flush

A recent change made changes to the rbd_client_list be protected by
a spinlock.  Unfortunately in rbd_put_client(), the lock is taken
before possibly dropping the last reference to an rbd_client, and on
the last reference that eventually calls flush_workqueue() which can
sleep.

The problem was flagged by a debug spinlock warning:
    BUG: spinlock wrong CPU on CPU#3, rbd/27814

The solution is to move the spinlock acquisition and release inside
rbd_client_release(), which is the spot where it's really needed for
protecting the removal of the rbd_client from the client list.

Signed-off-by: Alex Elder <elder@dreamhost.com>
Reviewed-by: Sage Weil <sage@newdream.net>
drivers/block/rbd.c