On Mon, Jul 20, 2020 at 12:53:55PM +0530, Kashyap Desai wrote: > > > > I also noticed nr_hw_queues are now exposed in sysfs - > > > > > > > > > > /sys/devices/pci0000:85/0000:85:00.0/0000:86:00.0/0000:87:04.0/0000:8b > > > > > > > > > :00.0/0000:8c:00.0/0000:8d:00.0/host14/scsi_host/host14/nr_hw_queues:1 > > > > 28 > > > > . > > > > > > That's on my v8 wip branch, so I guess you're picking it up from there. > > > > John - I did more testing on v8 wip branch. CPU hotplug is working as > > expected, but I still see some performance issue on Logical Volumes. > > > > I created 8 Drives Raid-0 VD on MR controller and below is performance > > impact of this RFC. Looks like contention is on single <sdev>. > > > > I used command - "numactl -N 1 fio 1vd.fio --iodepth=128 --bs=4k -- > > rw=randread --cpus_allowed_policy=split --ioscheduler=none -- > > group_reporting --runtime=200 --numjobs=1" > > IOPS without RFC = 300K IOPS with RFC = 230K. > > > > Perf top (shared host tag. IOPS = 230K) > > > > 13.98% [kernel] [k] sbitmap_any_bit_set > > 6.43% [kernel] [k] blk_mq_run_hw_queue > > blk_mq_run_hw_queue function take more CPU which is called from " > scsi_end_request" The problem could be that nr_hw_queues is increased a lot so that sample on blk_mq_run_hw_queue() can be observed now. > It looks like " blk_mq_hctx_has_pending" handles only elevator (scheduler) > case. If queue has ioscheduler=none, we can skip. I case of scheduler=none, > IO will be pushed to hardware queue and it by pass software queue. > Based on above understanding, I added below patch and I can see performance > scale back to expectation. > > Ming mentioned that - we cannot remove blk_mq_run_hw_queues() from IO > completion path otherwise we may see IO hang. So I have just modified > completion path assuming it is only required for IO scheduler case. > https://www.spinics.net/lists/linux-block/msg55049.html > > Please review and let me know if this is good or we have to address with > proper fix. > > diff --git a/block/blk-mq.c b/block/blk-mq.c > index 1be7ac5a4040..b6a5b41b7fc2 100644 > --- a/block/blk-mq.c > +++ b/block/blk-mq.c > @@ -1559,6 +1559,9 @@ void blk_mq_run_hw_queues(struct request_queue *q, > bool async) > struct blk_mq_hw_ctx *hctx; > int i; > > + if (!q->elevator) > + return; > + This way shouldn't be correct, blk_mq_run_hw_queues() is still needed for none because request may not be dispatched successfully by direct issue. Thanks, Ming
