dm rq: fix a race condition in rq_completed()

It is required to hold the queue lock when calling blk_run_queue_async()
to avoid that a race between blk_run_queue_async() and
blk_cleanup_queue() is triggered.

Cc: stable@vger.kernel.org
Signed-off-by: Bart Van Assche <bart.vanassche@sandisk.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>

drivers/md/dm-rq.c

@@ -226,6 +226,9 @@ static void rq_end_stats(struct mapped_device *md, struct request *orig)
  */
 static void rq_completed(struct mapped_device *md, int rw, bool run_queue)
 {
+	struct request_queue *q = md->queue;
+	unsigned long flags;
+
 	atomic_dec(&md->pending[rw]);
 
 	/* nudge anyone waiting on suspend queue */
@@ -238,8 +241,11 @@ static void rq_completed(struct mapped_device *md, int rw, bool run_queue)
 	 * back into ->request_fn() could deadlock attempting to grab the
 	 * queue lock again.
 	 */
-	if (!md->queue->mq_ops && run_queue)
-		blk_run_queue_async(md->queue);
+	if (!q->mq_ops && run_queue) {
+		spin_lock_irqsave(q->queue_lock, flags);
+		blk_run_queue_async(q);
+		spin_unlock_irqrestore(q->queue_lock, flags);
+	}
 
 	/*
 	 * dm_put() must be at the end of this function. See the comment above