Received-SPF: pass (google.com: best guess record for domain of linux-kernel-owner@vger.kernel.org designates 209.132.180.67 as permitted sender) client-ip=209.132.180.67;
Date:   Thu, 18 Jan 2018 10:32:34 +0000
From:   Morten Rasmussen <morten.rasmussen@foss.arm.com>
To:     Peter Zijlstra <peterz@infradead.org>
Cc:     Vincent Guittot <vincent.guittot@linaro.org>,
        Ingo Molnar <mingo@kernel.org>,
        linux-kernel <linux-kernel@vger.kernel.org>,
        Brendan Jackman <brendan.jackman@arm.com>,
        Dietmar Eggemann <dietmar.eggemann@arm.com>,
        Morten Rasmussen <morten.rasmussen@arm.com>
Subject: Re: [RFC PATCH 2/5] sched: Add NOHZ_STATS_KICK
Message-ID: <20180118103234.GC28799@e105550-lin.cambridge.arm.com>
References: <CAKfTPtCVgnO1zcD_X4Ad-T0auWmN-TJ0dezv0P5wEtubkbgHyQ@mail.gmail.com>
 <20171222075934.f6yenvcb2zkf2ysd@hirez.programming.kicks-ass.net>
 <CAKfTPtC0+5zjvgNXMyVcvnnyJh0giQgJHtx+5p-GEpjUW1x8eg@mail.gmail.com>
 <20171222082915.4lcb7xyyooqyjpia@hirez.programming.kicks-ass.net>
 <20171222091221.ow5vn3ydx3hj4nht@hirez.programming.kicks-ass.net>
 <CAKfTPtAzutuyA7D2qOErKPwdFaRK082YAtVn4ci8hAOUvL3zMQ@mail.gmail.com>
 <20171222185629.lysjebfifgdwvvhu@hirez.programming.kicks-ass.net>
 <20171222204247.kyc6ugyyu3ei7zhs@hirez.programming.kicks-ass.net>
 <20180102154457.GB28799@e105550-lin.cambridge.arm.com>
 <20180115094317.GN6176@hirez.programming.kicks-ass.net>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <20180115094317.GN6176@hirez.programming.kicks-ass.net>
User-Agent: Mutt/1.5.24 (2015-08-30)
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Mon, Jan 15, 2018 at 10:43:18AM +0100, Peter Zijlstra wrote:
> On Tue, Jan 02, 2018 at 03:44:57PM +0000, Morten Rasmussen wrote:
> 
> > Vincent already proposed, why can't we just modify Brendan's
> > CPU_NEWLY_IDLE proposal to do a stats update from idle_balance() every
> > 32ms regardless of whether we need to load-balance?
> 
> I think that code is there, no?
> 
> Subject: sched: Update blocked load from NEWIDLE

The mechanics are there, but I think the problem is that the
idle_balance() bails out before we get to it in some cases. If we only
have a few small periodic tasks running rd->overload won't be set and
idle_balance() returns before doing anything.

We would need some sort of check to see if a PELT update is due and make
sure it happens, even if idle_balance() has nothing to do.