Messages in this thread | | | Date | Sat, 26 Sep 2015 13:32:56 -0400 | From | Tejun Heo <> | Subject | Re: [PATCH v4 6/7] blk-mq: fix freeze queue race |
| |
Hello,
On Sun, Sep 27, 2015 at 02:09:24AM +0900, Akinobu Mita wrote: > @@ -420,7 +420,9 @@ static void blk_mq_sysfs_init(struct request_queue *q) > /* see blk_register_queue() */ > void blk_mq_finish_init(struct request_queue *q) > { > + mutex_lock(&q->mq_freeze_lock); > percpu_ref_switch_to_percpu(&q->mq_usage_counter); > + mutex_unlock(&q->mq_freeze_lock);
This looks weird to me. What can it race against at this point?
> @@ -115,11 +115,15 @@ void blk_mq_freeze_queue_start(struct request_queue *q) > { > int freeze_depth; > > + mutex_lock(&q->mq_freeze_lock); > + > freeze_depth = atomic_inc_return(&q->mq_freeze_depth);
It doesn't have to be an atomic anymore, right?
Thanks.
-- tejun
| |