dm mpath: don't call blk_mq_delay_run_hw_queue() in case of BLK_STS_RESOURCENotice: this object is not reachable from any branch.

If .queue_rq() returns BLK_STS_RESOURCE, blk-mq will rerun the queue in
these three situations:

1) if BLK_MQ_S_SCHED_RESTART is set
- queue is rerun after one rq is completed, see blk_mq_sched_restart()
which is run from blk_mq_free_request()

2) run out of driver tag
- queue is rerun after one tag is freed

3) otherwise
- queue is run immediately in blk_mq_dispatch_rq_list()

dm-mpath's random (100ms) delayed running of hw queue was introduced by
commit 6077c2d706097c0 ("dm rq: Avoid that request processing stalls
sporadically"), but it never explained the reasoning behind it.  At best
it was a workaround.  And calling blk_mq_delay_run_hw_queue() inside
.queue_rq() is a horrible hack because it breaks BLK_MQ_S_SCHED_RESTART,
which degrades I/O peformance a lot.  So remove it.

Signed-off-by: Ming Lei <ming.lei@redhat.com>
Signed-off-by: Mike Snitzer <snitzer@redhat.com>
Notice: this object is not reachable from any branch.

1 files changed, 0 insertions, 1 deletions

diff --git a/drivers/md/dm-rq.c b/drivers/md/dm-rq.c
index 9d32f25489c27..cbe8a06ef8b01 100644
--- a/drivers/md/dm-rq.c
+++ b/drivers/md/dm-rq.c
@@ -758,7 +758,6 @@ static blk_status_t dm_mq_queue_rq(struct blk_mq_hw_ctx *hctx,
 		/* Undo dm_start_request() before requeuing */
 		rq_end_stats(md, rq);
 		rq_completed(md, rq_data_dir(rq), false);
-		blk_mq_delay_run_hw_queue(hctx, 100/*ms*/);
 		return BLK_STS_RESOURCE;
 	}