cfq-iosched: fix race in cfq_set_request()

We need to hold the queue lock over the reference increment, it's not atomic anymore. Signed-off-by: Jens Axboe <jaxboe@fusionio.com>
git-mirror · Mar 7, 2011 · 93803e0 · 93803e0
1 parent da52777
commit 93803e0
Showing 1 changed file with 1 addition and 2 deletions.
diff --git a/block/cfq-iosched.c b/block/cfq-iosched.c
@@ -3675,12 +3675,11 @@ cfq_set_request(struct request_queue *q, struct request *rq, gfp_t gfp_mask)
 
 	cfqq->allocated[rw]++;
 
-	spin_unlock_irqrestore(q->queue_lock, flags);
-
 	cfqq->ref++;
 	rq->elevator_private[0] = cic;
 	rq->elevator_private[1] = cfqq;
 	rq->elevator_private[2] = cfq_ref_get_cfqg(cfqq->cfqg);
+	spin_unlock_irqrestore(q->queue_lock, flags);
 	return 0;
 
 queue_fail: