[<prev] [next>] [<thread-prev] [day] [month] [year] [list]
Message-ID: <aed2c239-8169-d30f-1410-ac0b5ae10cf1@kernel.dk>
Date: Fri, 23 Apr 2021 08:55:15 -0600
From: Jens Axboe <axboe@...nel.dk>
To: Tejun Heo <tj@...nel.org>
Cc: linux-block@...r.kernel.org, kernel-team@...com, dskarlat@...com,
dschatzberg@...com, linux-kernel@...r.kernel.org
Subject: Re: [PATCH for-5.13/block] blk-iocost: don't ignore vrate_min on QD
contention
On 4/22/21 7:54 PM, Tejun Heo wrote:
> 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.
Applied, thanks.
--
Jens Axboe
Powered by blists - more mailing lists