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;
MIME-Version: 1.0
References: <1568878421-12301-1-git-send-email-vincent.guittot@linaro.org>
 <1568878421-12301-5-git-send-email-vincent.guittot@linaro.org> <20191008175537.GC22902@worktop.programming.kicks-ass.net>
In-Reply-To: <20191008175537.GC22902@worktop.programming.kicks-ass.net>
From:   Vincent Guittot <vincent.guittot@linaro.org>
Date:   Tue, 8 Oct 2019 20:47:12 +0200
Message-ID: <CAKfTPtAMEdcpX=h1sNaM-S9R7xB7ZPp5UZGdm_8pBWMYzeB+VA@mail.gmail.com>
Subject: Re: [PATCH v3 04/10] sched/fair: rework load_balance
To:     Peter Zijlstra <peterz@infradead.org>
Cc:     linux-kernel <linux-kernel@vger.kernel.org>,
        Ingo Molnar <mingo@redhat.com>, Phil Auld <pauld@redhat.com>,
        Valentin Schneider <valentin.schneider@arm.com>,
        Srikar Dronamraju <srikar@linux.vnet.ibm.com>,
        Quentin Perret <quentin.perret@arm.com>,
        Dietmar Eggemann <dietmar.eggemann@arm.com>,
        Morten Rasmussen <Morten.Rasmussen@arm.com>,
        Hillf Danton <hdanton@sina.com>
Content-Type: text/plain; charset="UTF-8"
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On Tue, 8 Oct 2019 at 19:55, Peter Zijlstra <peterz@infradead.org> wrote:
>
> On Thu, Sep 19, 2019 at 09:33:35AM +0200, Vincent Guittot wrote:
> > +     if (busiest->group_type == group_asym_packing) {
> > +             /*
> > +              * In case of asym capacity, we will try to migrate all load to
> > +              * the preferred CPU.
> > +              */
> > +             env->balance_type = migrate_load;
> >               env->imbalance = busiest->group_load;
> >               return;
> >       }
>
> I was a bit surprised with this; I sorta expected a migrate_task,1 here.

I have just kept current mechanism

>
> The asym_packing thing has always been a nr_running issue to me. If
> there is something to run, we should run on as many siblings/cores as
> possible, but preferably the 'highest' ranked siblings/cores.

I can probably set the number of running task to migrate instead of the load
>