From: Ming Lei <ming.lei@xxxxxxxxxx> commit c48dac137a62a5d6fa1ef3fa445cbd9c43655a76 upstream. The original comment says: q->sysfs_lock must be held to provide mutual exclusion between elevator_switch() and here. Which is simply wrong. elevator_init_mq() is only called from blk_mq_init_allocated_queue, which is always called before the request queue is registered via blk_register_queue(), for dm-rq or normal rq based driver. However, queue's kobject is only exposed and added to sysfs in blk_register_queue(). So there isn't such race between elevator_switch() and elevator_init_mq(). So avoid to hold q->sysfs_lock in elevator_init_mq(). Cc: Christoph Hellwig <hch@xxxxxxxxxxxxx> Cc: Hannes Reinecke <hare@xxxxxxxx> Cc: Greg KH <gregkh@xxxxxxxxxxxxxxxxxxx> Cc: Mike Snitzer <snitzer@xxxxxxxxxx> Cc: Bart Van Assche <bvanassche@xxxxxxx> Cc: Damien Le Moal <damien.lemoal@xxxxxxx> Reviewed-by: Bart Van Assche <bvanassche@xxxxxxx> Signed-off-by: Ming Lei <ming.lei@xxxxxxxxxx> Signed-off-by: Jens Axboe <axboe@xxxxxxxxx> Signed-off-by: Jack Wang <jinpu.wang@xxxxxxxxxxxxxxx> Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx> --- block/elevator.c | 14 +++++--------- 1 file changed, 5 insertions(+), 9 deletions(-) --- a/block/elevator.c +++ b/block/elevator.c @@ -980,23 +980,19 @@ int elevator_init_mq(struct request_queu if (q->nr_hw_queues != 1) return 0; - /* - * q->sysfs_lock must be held to provide mutual exclusion between - * elevator_switch() and here. - */ - mutex_lock(&q->sysfs_lock); + WARN_ON_ONCE(test_bit(QUEUE_FLAG_REGISTERED, &q->queue_flags)); + if (unlikely(q->elevator)) - goto out_unlock; + goto out; e = elevator_get(q, "mq-deadline", false); if (!e) - goto out_unlock; + goto out; err = blk_mq_init_sched(q, e); if (err) elevator_put(e); -out_unlock: - mutex_unlock(&q->sysfs_lock); +out: return err; }
