From f46ec84b5acbf8d7067d71a6bbdde213d4b86036 Mon Sep 17 00:00:00 2001 From: Tejun Heo Date: Thu, 22 Apr 2021 21:54:28 -0400 Subject: blk-iocost: don't ignore vrate_min on QD contention ioc_adjust_base_vrate() ignored vrate_min when rq_wait_pct indicates that there is QD contention. The reasoning was that QD depletion always reliably indicates device saturation and thus it's safe to override user specified vrate_min. However, this sometimes leads to unnecessary throttling, especially on really fast devices, because vrate adjustments have delays and inertia. It also confuses users because the behavior violates the explicitly specified configuration. This patch drops the special case handling so that vrate_min is always applied. Signed-off-by: Tejun Heo Link: https://lore.kernel.org/r/YIIo1HuyNmhDeiNx@slm.duckdns.org Signed-off-by: Jens Axboe --- block/blk-iocost.c | 4 ---- 1 file changed, 4 deletions(-) (limited to 'block/blk-iocost.c') diff --git a/block/blk-iocost.c b/block/blk-iocost.c index 98d656bdb42b..e0c4baa01857 100644 --- a/block/blk-iocost.c +++ b/block/blk-iocost.c @@ -987,10 +987,6 @@ static void ioc_adjust_base_vrate(struct ioc *ioc, u32 rq_wait_pct, return; } - /* rq_wait signal is always reliable, ignore user vrate_min */ - if (rq_wait_pct > RQ_WAIT_BUSY_PCT) - vrate_min = VRATE_MIN; - /* * If vrate is out of bounds, apply clamp gradually as the * bounds can change abruptly. Otherwise, apply busy_level -- cgit v1.2.3