From: Ming Lei Date: Mon, 3 Jul 2017 12:37:14 +0000 (+0800) Subject: blk-mq-sched: fix performance regression of mq-deadline X-Git-Url: https://git.stricted.de/?a=commitdiff_plain;h=32825c45ff8f4cce937ab85b030dc693ceb1aa0a;p=GitHub%2FLineageOS%2Fandroid_kernel_motorola_exynos9610.git blk-mq-sched: fix performance regression of mq-deadline When mq-deadline is taken, IOPS of sequential read and seqential write is observed more than 20% drop on sata(scsi-mq) devices, compared with using 'none' scheduler. The reason is that the default nr_requests for scheduler is too big for small queuedepth devices, and latency is increased much. Since the principle of taking 256 requests for mq scheduler is based on 128 queue depth, this patch changes into double size of min(hw queue_depth, 128). Signed-off-by: Ming Lei Signed-off-by: Jens Axboe --- diff --git a/block/blk-mq-sched.c b/block/blk-mq-sched.c index 7f0dc48ffb40..4ab69435708c 100644 --- a/block/blk-mq-sched.c +++ b/block/blk-mq-sched.c @@ -515,10 +515,12 @@ int blk_mq_init_sched(struct request_queue *q, struct elevator_type *e) } /* - * Default to 256, since we don't split into sync/async like the - * old code did. Additionally, this is a per-hw queue depth. + * Default to double of smaller one between hw queue_depth and 128, + * since we don't split into sync/async like the old code did. + * Additionally, this is a per-hw queue depth. */ - q->nr_requests = 2 * BLKDEV_MAX_RQ; + q->nr_requests = 2 * min_t(unsigned int, q->tag_set->queue_depth, + BLKDEV_MAX_RQ); queue_for_each_hw_ctx(q, hctx, i) { ret = blk_mq_sched_alloc_tags(q, hctx, i);