From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <stable-owner@vger.kernel.org>
Received: from mout.gmx.net ([212.227.17.20]:51985 "EHLO mout.gmx.net"
        rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP
        id S967713AbeBNNZq (ORCPT <rfc822;stable@vger.kernel.org>);
        Wed, 14 Feb 2018 08:25:46 -0500
Message-ID: <1518614741.8220.19.camel@gmx.de>
Subject: Re: [PATCH BUGFIX V3] block, bfq: add requeue-request hook
From: Mike Galbraith <efault@gmx.de>
To: Oleksandr Natalenko <oleksandr@natalenko.name>,
        Paolo Valente <paolo.valente@linaro.org>
Cc: Jens Axboe <axboe@kernel.dk>, stable <stable@vger.kernel.org>
Date: Wed, 14 Feb 2018 14:25:41 +0100
In-Reply-To: <30107c3d7cc4dadb0d6949b7dd30d565@natalenko.name>
References: <20180207211920.6343-1-paolo.valente@linaro.org>
         <b5d4a3bc8946db0a696e8398b68adba0@natalenko.name>
         <1518197379.26824.31.camel@gmx.de> <6394471.U0O273vb9H@natalenko.name>
         <9E24F648-C93A-4CEA-A1B6-B041540CEAAE@linaro.org>
         <1518434553.13087.7.camel@gmx.de>
         <805e9d5af9aed3fbc2697fdb0ce51e88@natalenko.name>
         <1518498175.6944.48.camel@gmx.de>
         <816E0B1B-B2D9-4604-A8CB-1E32AFBF6C22@linaro.org>
         <1518504174.6944.71.camel@gmx.de> <1518504640.6944.73.camel@gmx.de>
         <E25879D9-0AF9-4924-A5AF-91FB69DD324C@linaro.org>
         <1518532236.15792.25.camel@gmx.de> <1518587910.5647.14.camel@gmx.de>
         <1518591888.6752.12.camel@gmx.de> <1518592512.6752.14.camel@gmx.de>
         <07DA441B-C2E8-4F65-B674-11C87D2F084B@linaro.org>
         <30107c3d7cc4dadb0d6949b7dd30d565@natalenko.name>
Content-Type: text/plain; charset="ISO-8859-15"
Mime-Version: 1.0
Content-Transfer-Encoding: 7bit
Sender: stable-owner@vger.kernel.org
List-ID: <stable.vger.kernel.org>

On Wed, 2018-02-14 at 10:55 +0100, Oleksandr Natalenko wrote:
> Hi.
> 
> 14.02.2018 09:56, Paolo Valente wrote:
> > This has evidently to do with tag pressure.  I've looked for a way to
> > easily reduce the number of tags online, so as to put your system in
> > the bad spot deterministically.  But at no avail.  Does anyone know a
> > way to do it?
> 
> Isn't this a case for block device queue_depth? IIRC, the amount of tags 
> is derived from it.

/sys/devices/pci0000:00/0000:00:1f.2/ata1/host0/target0:0:0/0:0:0:0/queue_depth
/sys/devices/pci0000:00/0000:00:1f.2/ata6/host5/target5:0:0/5:0:0:0/queue_depth
/sys/devices/pci0000:00/0000:00:1f.2/ata2/host1/target1:0:0/1:0:0:0/queue_depth
/sys/devices/pci0000:00/0000:00:1f.2/ata5/host4/target4:0:0/4:0:0:0/queue_depth

FWIW, these guys are all 31.