Received-SPF: pass (google.com: domain of linux-kernel-owner@vger.kernel.org designates 2620:137:e000::1:20 as permitted sender) client-ip=2620:137:e000::1:20;
MIME-Version: 1.0
References: <5708364.DvuYhMxLoT@kreacher> <CAJZ5v0hNZLvt+bJLvvibmk3dw_u5XnYNS020r7QLWLiiWLZSgQ@mail.gmail.com>
 <CAAYoRsVHFLpQ91DqUZygZETsYC31rhmgFru=19C-veXSH+dy5w@mail.gmail.com>
In-Reply-To: <CAAYoRsVHFLpQ91DqUZygZETsYC31rhmgFru=19C-veXSH+dy5w@mail.gmail.com>
From:   "Rafael J. Wysocki" <rafael@kernel.org>
Date:   Thu, 10 Aug 2023 19:20:32 +0200
Message-ID: <CAJZ5v0iYCwAaMgpD=_M+Ynwx50UpQDV5GU4Lj1tkTKKAEt5xoA@mail.gmail.com>
Subject: Re: [RFT] [PATCH v1] cpuidle: menu: Skip tick_nohz_get_sleep_length()
 call in some cases
To:     Doug Smythies <dsmythies@telus.net>
Cc:     "Rafael J. Wysocki" <rafael@kernel.org>,
        Linux PM <linux-pm@vger.kernel.org>,
        Anna-Maria Behnsen <anna-maria@linutronix.de>,
        Peter Zijlstra <peterz@infradead.org>,
        LKML <linux-kernel@vger.kernel.org>,
        Frederic Weisbecker <frederic@kernel.org>,
        Kajetan Puchalski <kajetan.puchalski@arm.com>,
        Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>,
        "Rafael J. Wysocki" <rjw@rjwysocki.net>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 8BIT
Precedence: bulk

Hi Doug,

On Thu, Aug 10, 2023 at 6:28 PM Doug Smythies <dsmythies@telus.net> wrote:
>
> On Wed, Aug 9, 2023 at 11:16 AM Rafael J. Wysocki <rafael@kernel.org> wrote:
> > On Wed, Aug 9, 2023 at 4:53 PM Rafael J. Wysocki <rjw@rjwysocki.net> wrote:
> > >
> > > From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> > >
> > > Because the cost of calling tick_nohz_get_sleep_length() may increase
> > > in the future, reorder the code in menu_select() so it first uses the
> > > statistics to determine the expected idle duration.  If that value is
> > > higher than RESIDENCY_THRESHOLD_NS, tick_nohz_get_sleep_length() will
> > > be called to obtain the time till the closest timer and refine the
> > > idle duration prediction if necessary.
> > >
> > > This causes the governor to always take the full overhead of
> > > get_typical_interval() with the assumption that the cost will be
> > > amortized by skipping the tick_nohz_get_sleep_length() call in the
> > > cases when the predicted idle duration is relatively very small.
> > >
> > > Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> > > ---
> ... deleted...
> >
> > This patch is now present in the git branch at
> >
> > git://git.kernel.org/pub/scm/linux/kernel/git/rafael/linux-pm.git \
> >  pm-cpuidle-gov
> >
> > along with the previous teo governor changes.
>
> Hi Rafael,
>
> Thanks for the branch and adding it to the previous 6.5-rc4 code,
> as now I can re-use the menu baseline tests already done.
>
> My test computer boots by default to use the teo idle governor.
> When I change to the menu governor, my system becomes
> unresponsive and I have to re-boot.
>
> Is anyone else having this difficulty?

There is a missing check it get_typical_interval(), my bad.

I'll send a v2 of the menu governor patch shortly and I'll update the
pm-cpuidle-gov branch accordingly.

Thanks!