Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 23.128.96.35 as permitted sender) client-ip=23.128.96.35;
MIME-Version: 1.0
References: <CABCjUKBXPLNaBjELHqz=Oke_JsxXLQauTWrsNmTp3KToW3-9MQ@mail.gmail.com>
 <CABCjUKCRXeT2eExm6+y7cZCX4yEbj5MbjFnCHv4xvRPL5ce13g@mail.gmail.com> <CAJWu+opvq3dTk4dQmmR+7fs2KcrMELvt0iJcDeRQ9o9cnAVHWA@mail.gmail.com>
In-Reply-To: <CAJWu+opvq3dTk4dQmmR+7fs2KcrMELvt0iJcDeRQ9o9cnAVHWA@mail.gmail.com>
From:   Joel Fernandes <joelaf@google.com>
Date:   Wed, 4 Oct 2023 12:16:50 -0400
Message-ID: <CAJWu+or4yNoX3Dd=hCKbgM=JbVjzjV4M3o8Y6ihhJ_JLThEbFg@mail.gmail.com>
Subject: Re: NOHZ interaction between IPI-less kick_ilb() and nohz_csd_func().
To:     Suleiman Souhlal <suleiman@google.com>,
        Frederic Weisbecker <frederic@kernel.org>
Cc:     Vincent Guittot <vincent.guittot@linaro.org>,
        Ingo Molnar <mingo@redhat.com>,
        Juri Lelli <juri.lelli@redhat.com>,
        Linux Kernel <linux-kernel@vger.kernel.org>,
        Steven Rostedt <rostedt@google.com>,
        Vineeth Pillai <vineethrp@google.com>,
        Youssef Esmat <youssefesmat@google.com>,
        Peter Zijlstra <peterz@infradead.org>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable
Precedence: bulk

On Wed, Oct 4, 2023 at 12:09=E2=80=AFPM Joel Fernandes <joelaf@google.com> =
wrote:
>
> +Frederic Weisbecker
>
> On Wed, Sep 13, 2023 at 10:32=E2=80=AFAM Suleiman Souhlal <suleiman@googl=
e.com> wrote:
> >
> > (I forgot to also add Vincent...)
> >
> > On Wed, Sep 13, 2023 at 9:49=E2=80=AFPM Suleiman Souhlal <suleiman@goog=
le.com> wrote:
> > >
> > > Hello,
> > >
> > > I noticed that on x86 machines that have MWAIT, with NOHZ, when the
> > > kernel decides to kick the idle load balance on another CPU in
> > > kick_ilb(), there's an optimization that makes it avoid using an IPI
> > > and instead exploit the fact that the remote CPU is MWAITing on the
> > > thread_info flags, by just setting TIF_NEED_RESCHED, in
> > > call_function_single_prep_ipi().
> > > However, on the remote CPU, in nohz_csd_func(), we end up not raising
> > > the sched softirq due to NEED_RESCHED being set, so the ILB doesn't
> > > end up getting done.
> > >
> > > Is this intended?

Just thinking out loud I was wondering how nohz-ILB really matters if
based on what Suleiman is saying - it is not even triggering on x86
due to the mwait optimization. And if it does matter, how much
improvement will fixing this bug give. I think at least on ARM, I
remember it matters.

I am meanwhile looking at it more closely...

thanks,

 - Joel