Date: Sat, 8 Oct 2016 12:46:38 +0200
From: Heinz Diehl <htd+ml@fritha.org>
To: linux-kernel@vger.kernel.org
Subject: Re: [PATCH V3 00/11] block-throttle: add .high limit
Message-ID: <20161008104638.GA3727@fritha.org>
References: <20161004155616.GB4205@htj.duckdns.org>
 <A5525664-DF90-4604-B64A-E793BBC0CB6A@unimore.it>
 <20161004162759.GD4205@htj.duckdns.org>
 <278BCC7B-ED58-4FDF-9243-FAFC3F862E4D@unimore.it>
 <20161004172852.GB73678@anikkar-mbp.local.dhcp.thefacebook.com>
 <EE90477C-FD56-48C1-8A45-97AF39A8BB0C@unimore.it>
 <20161004185413.GF4205@htj.duckdns.org>
 <F9A2F063-F5A2-49A6-8DFE-91ACDE380645@unimore.it>
 <20161004191427.GG4205@htj.duckdns.org>
 <CACRpkdavM0ZGdDjXAsEG1MMy4WuAnudNw9KL8eFJFsfm_ptEBg@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CACRpkdavM0ZGdDjXAsEG1MMy4WuAnudNw9KL8eFJFsfm_ptEBg@mail.gmail.com>
Organization: private site
OpenPGP: id=60F4A92C; url=http://www.fritha.org/htd.asc;
 preference=signencrypt
User-Agent: Mutt/1.7+28 (8963e77577ad) (2016-08-17)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 625
Lines: 17

On 06.10.2016, Linus Walleij wrote: 

> Not just desktops, also Android phones.
 
> So why not have BFQ as a separate scheduling policy upstream,
> alongside CFQ, deadline and noop?

Please, just do it! CFQ and deadline perform horribly on Android, and
the same is true for desktop systems based on SSD drives.

What e.g. this test shows is true for all of my machines:
https://www.youtube.com/watch?v=1cjZeaCXIyM

I've been using BFQ nearly right from its inception, and it improves
all of my systems noticeably, in fact in such a positive way that I
don't update to a newer kernel unless there is a BFQ patch ready for
it.