Date: Mon, 10 Apr 2017 10:40:38 +0200
From: Peter Zijlstra <peterz@infradead.org>
To: Dietmar Eggemann <dietmar.eggemann@arm.com>
Cc: Paul Turner <pjt@google.com>, Yuyang Du <yuyang.du@intel.com>,
        Ingo Molnar <mingo@kernel.org>, LKML <linux-kernel@vger.kernel.org>,
        Benjamin Segall <bsegall@google.com>,
        Morten Rasmussen <morten.rasmussen@arm.com>,
        Vincent Guittot <vincent.guittot@linaro.org>,
        Matt Fleming <matt@codeblueprint.co.uk>,
        Mike Galbraith <umgwanakikbuti@gmail.com>
Subject: Re: [RESEND PATCH 2/2] sched/fair: Optimize __update_sched_avg()
Message-ID: <20170410084038.eg6nyyle6jwt5rmb@hirez.programming.kicks-ass.net>
References: <20170329000442.GC2459@ydu19desktop>
 <20170329104126.lg6ismevfbqywpcj@hirez.programming.kicks-ass.net>
 <CAPM31RLHHKVBvVn04ZmvNQ=69f3M9WyFBUNPOoO-8=Q207ioxA@mail.gmail.com>
 <20170330121658.6mo7datma4ssw7st@hirez.programming.kicks-ass.net>
 <20170330141428.deiduft5btwid6ov@hirez.programming.kicks-ass.net>
 <CAPM31RJL_-Ty+_5r4gwvPvF+WbsEQ_0YETpBVHWbQmtMM=SyQg@mail.gmail.com>
 <20170331070141.ygghjaamv6bswbyl@hirez.programming.kicks-ass.net>
 <CAPM31RKyCe95r-y-J3DC7rTcY0cP0NFfXG1A6WNV9b2OqC2c2A@mail.gmail.com>
 <20170331112358.422kb5aosbq3k5tn@hirez.programming.kicks-ass.net>
 <2f4c2d4e-f56e-8681-1a32-621eb2ad8c35@arm.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <2f4c2d4e-f56e-8681-1a32-621eb2ad8c35@arm.com>
User-Agent: NeoMutt/20170113 (1.7.2)
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 447
Lines: 11

On Mon, Apr 10, 2017 at 08:39:10AM +0100, Dietmar Eggemann wrote:

> Are you planning to get rid of scale_load_down() for weight because of
> the correctness problems of the signal mentioned by Paul in this thread?

Yes; it would be good to (again) use all 20 bits of weight for the
load_avg things. It was one of the reasons we went to 20 in the first
place after all.

The patch was a quick draft and known broken, but thanks for having a
look.