blk-mq: insert flush request to the front of dispatch queue
PATCH 1/1 blk-mq: insert flush request to the front of dispatch queue
commit 01e99aeca397 ("blk-mq: insert passthrough request into hctx->dispatch directly") may change to add flush request to the tail of dispatch by applying the 'add_head' parameter of blk_mq_sched_insert_request.
Turns out this way causes performance regression on NCQ controller because flush is non-NCQ command, which can't be queued when there is any in-flight NCQ command. When adding flush rq to the front of hctx->dispatch, it is easier to introduce extra time to flush rq's latency compared with adding to the tail of dispatch queue because of S_SCHED_RESTART, then chance of flush merge is increased, and less flush requests may be issued to controller.
So always insert flush request to the front of dispatch queue just like before applying commit 01e99aeca397 ("blk-mq: insert passthrough request into hctx->dispatch directly").
--- block/blk-mq-sched.c | 22 ++++++++++++++++++++++ 1 file changed, 22 insertions(+)
block/blk-mq-sched.c +22 -0
@@ -398,6 +398,28 @@ void blk_mq_sched_insert_request(struct request *rq, bool at_head,WARN_ON (e && (rq->tag !=-1 ));if (blk_mq_sched_bypass_insert (hctx, !!e, rq)) { +/* +* Firstly normal IO request is inserted to scheduler queue or +* sw queue, meantime we add flush request to dispatch queue( +* hctx->dispatch) directly and there is at most one in-flight +* flush request for each hw queue, so it doesn't matter to add +* flush request to tail or front of the dispatch queue. +* +* Secondly in case of NCQ, flush request belongs to non-NCQ +* command, and queueing it will fail when there is any +* in-flight normal IO request(NCQ command). When adding flush +* rq to the front of hctx->dispatch, it is easier to introduce +* extra time to flush rq's latency because of S_SCHED_RESTART +* compared with adding to the tail of dispatch queue, then +* chance of flush merge is increased, and less flush requests +* will be issued to controller. It is observed that ~10% time +* is saved in blktests block/004 on disk attached to AHCI/NCQ +* drive when adding flush rq to the front of hctx->dispatch. +* +* Simply queue flush rq to the front of hctx->dispatch so that +* intensive flush workloads can benefit in case of NCQ HW. +*/ + at_head = (rq->rq_flags &RQF_FLUSH_SEQ ) ? true : at_head;blk_mq_request_bypass_insert (rq, at_head, false); goto run; }
Signature · 1 line
2.20.1
Quoted 13 lines · Ming Lei
commit
01e99aeca397("blk-mq: insert passthrough request intohctx->dispatchdirectly") may change to add flush request to the tail of dispatch by applying the 'add_head' parameter ofblk_mq_sched_insert_request.Turns out this way causes performance regression on NCQ controller because flush is non-NCQ command, which can't be queued when there is any in-flight NCQ command. When adding flush rq to the front of
hctx->dispatch, it is easier to introduce extra time to flush rq's latency compared with adding to the tail of dispatch queue because ofS_SCHED_RESTART, then chance of flush merge is increased, and less flush requests may be issued to controller.
Quoted 3 lines
So always insert flush request to the front of dispatch queue just like before applying commit
01e99aeca397("blk-mq: insert passthrough request intohctx->dispatchdirectly").
Applied, thanks.
Signature · 1 line
Jens Axboe
Quoted 14 lines · Jens Axboe
Quoted 13 lines · Ming Lei
commit
01e99aeca397("blk-mq: insert passthrough request intohctx->dispatchdirectly") may change to add flush request to the tail of dispatch by applying the 'add_head' parameter ofblk_mq_sched_insert_request.Turns out this way causes performance regression on NCQ controller because flush is non-NCQ command, which can't be queued when there is any in-flight NCQ command. When adding flush rq to the front of
hctx->dispatch, it is easier to introduce extra time to flush rq's latency compared with adding to the tail of dispatch queue because ofS_SCHED_RESTART, then chance of flush merge is increased, and less flush requests may be issued to controller.
Quoted 5 lines
Quoted 3 lines
So always insert flush request to the front of dispatch queue just like before applying commit
01e99aeca397("blk-mq: insert passthrough request intohctx->dispatchdirectly").Applied, thanks.
Ming, thank you so much for the patch. Using my SMR SATA drive I confirmed it reduces blktests block/004 runtime as expected. With this patch, the runtime is like before the commit 01e99aeca397. Good.
Signature · 2 lines
Best Regards, Shin'ichiro Kawasaki
Jens Axboe
maintainerQuoted 19 lines · Shinichiro Kawasaki
Quoted 18 lines · Jens Axboe
Quoted 16 lines · Ming Lei
commit
01e99aeca397("blk-mq: insert passthrough request intohctx->dispatchdirectly") may change to add flush request to the tail of dispatch by applying the 'add_head' parameter ofblk_mq_sched_insert_request.Turns out this way causes performance regression on NCQ controller because flush is non-NCQ command, which can't be queued when there is any in-flight NCQ command. When adding flush rq to the front of
hctx->dispatch, it is easier to introduce extra time to flush rq's latency compared with adding to the tail of dispatch queue because ofS_SCHED_RESTART, then chance of flush merge is increased, and less flush requests may be issued to controller.So always insert flush request to the front of dispatch queue just like before applying commit
01e99aeca397("blk-mq: insert passthrough request intohctx->dispatchdirectly").
Quoted 5 lines
Quoted 1 line
Applied, thanks.
Ming, thank you so much for the patch. Using my SMR SATA drive I confirmed it reduces blktests block/004 runtime as expected. With this patch, the runtime is like before the commit
01e99aeca397. Good.
Thanks for testing, it'll go into 5.6.
Signature · 1 line
Jens Axboe