blk-iocost: don't ignore vrate_min on QD contention

mainline inclusion
from mainline-v5.13-rc1
commit f46ec84b
category: bugfix
bugzilla: 187443, https://gitee.com/openeuler/kernel/issues/I5PFPV
CVE: NA

Reference: https://git.kernel.org/pub/scm/linux/kernel/git/torvalds/linux.git/commit/?id=f46ec84b5acbf8d7067d71a6bbdde213d4b86036

----------------------------------------------------------------------

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: NTejun Heo <tj@kernel.org>
Link: https://lore.kernel.org/r/YIIo1HuyNmhDeiNx@slm.duckdns.orgSigned-off-by: NJens Axboe <axboe@kernel.dk>

confilict:
	block/blk-iocost.c
Signed-off-by: NLi Nan <linan122@huawei.com>
Reviewed-by: NJason Yan <yanaijie@huawei.com>
Signed-off-by: NZheng Zengkai <zhengzengkai@huawei.com>

block/blk-iocost.c

@@ -2341,10 +2341,6 @@ static void ioc_timer_fn(struct timer_list *timer)
 		u64 vrate = ioc->vtime_base_rate;
 		u64 vrate_min = ioc->vrate_min, vrate_max = ioc->vrate_max;
 
-		/* 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