On 17/08/2020 11:08, Ming Lei wrote:
Now the request queue is run in scsi_end_request() unconditionally if both target queue and host queue is ready. We should have re-run request queue only after this device queue becomes busy for restarting this LUN only. Recently Long Li reported that cost of run queue may be very heavy in case of high queue depth. So improve this situation by only running the request queue when this LUN is busy. Cc: Ewan D. Milne <emilne@xxxxxxxxxx> Cc: Kashyap Desai <kashyap.desai@xxxxxxxxxxxx> Cc: Hannes Reinecke <hare@xxxxxxx> Cc: Bart Van Assche <bvanassche@xxxxxxx> Cc: Long Li <longli@xxxxxxxxxxxxx> Cc: John Garry <john.garry@xxxxxxxxxx> Cc: linux-block@xxxxxxxxxxxxxxx Reported-by: Long Li <longli@xxxxxxxxxxxxx> Tested-by: Kashyap Desai <kashyap.desai@xxxxxxxxxxxx> Signed-off-by: Ming Lei <ming.lei@xxxxxxxxxx>
FWIW, this looks ok to me, and tested on scsi_debug showed less time in blk_mq_run_hw_queues() [for reduced queue depth]:
Reviewed-by: John Garry <john.garry@xxxxxxxxxx> thanks
