linux-kernel - Re: [PATCH 1/2] cpufreq: schedutil: Fix superfluous updates caused by need_freq

lists.openwall.net		lists / announce owl-users owl-dev john-users john-dev passwdqc-users yescrypt popa3d-users / oss-security kernel-hardening musl sabotage tlsify passwords / crypt-dev xvendor / Bugtraq Full-Disclosure linux-kernel linux-netdev linux-ext4 linux-hardening linux-cve-announce PHC
Open Source and information security mailing list archives

Hash Suite: Windows password security audit tool. GUI, reports in PDF.

[<prev] [next>] [<thread-prev] [thread-next>] [day] [month] [year] [list]

Message-ID: <CAB8ipk8AAFFtV3OA4S=g9X9AXsC4Ntr911DLLRkhgQtvMvXAvg@mail.gmail.com>
Date: Thu, 10 Apr 2025 10:13:04 +0800
From: Xuewen Yan <xuewen.yan94@...il.com>
To: Sultan Alsawaf <sultan@...neltoast.com>
Cc: Stephan Gerhold <stephan.gerhold@...aro.org>, "Rafael J. Wysocki" <rafael@...nel.org>, 
	Viresh Kumar <viresh.kumar@...aro.org>, Ingo Molnar <mingo@...hat.com>, 
	Peter Zijlstra <peterz@...radead.org>, Juri Lelli <juri.lelli@...hat.com>, 
	Vincent Guittot <vincent.guittot@...aro.org>, Dietmar Eggemann <dietmar.eggemann@....com>, 
	Steven Rostedt <rostedt@...dmis.org>, Ben Segall <bsegall@...gle.com>, Mel Gorman <mgorman@...e.de>, 
	Valentin Schneider <vschneid@...hat.com>, linux-pm@...r.kernel.org, 
	linux-kernel@...r.kernel.org, regressions@...ts.linux.dev, 
	Johan Hovold <johan@...nel.org>
Subject: Re: [PATCH 1/2] cpufreq: schedutil: Fix superfluous updates caused by need_freq_update

On Thu, Apr 10, 2025 at 10:09 AM Sultan Alsawaf <sultan@...neltoast.com> wrote:
>
> On Thu, Apr 10, 2025 at 10:06:41AM +0800, Xuewen Yan wrote:
> > On Thu, Apr 10, 2025 at 9:49 AM Sultan Alsawaf <sultan@...neltoast.com> wrote:
> > >
> > > On Wed, Apr 09, 2025 at 07:48:05PM +0800, Xuewen Yan wrote:
> > > > Or can we modify it as follows?
> > > >
> > > > -->8--
> > > >
> > > > diff --git a/kernel/sched/cpufreq_schedutil.c b/kernel/sched/cpufreq_schedutil.c
> > > > index 1a19d69b91ed..0e8d3b92ffe7 100644
> > > > --- a/kernel/sched/cpufreq_schedutil.c
> > > > +++ b/kernel/sched/cpufreq_schedutil.c
> > > > @@ -83,7 +83,7 @@ static bool sugov_should_update_freq(struct
> > > > sugov_policy *sg_policy, u64 time)
> > > >
> > > >         if (unlikely(sg_policy->limits_changed)) {
> > > >                 sg_policy->limits_changed = false;
> > > > -               sg_policy->need_freq_update =
> > > > cpufreq_driver_test_flags(CPUFREQ_NEED_UPDATE_LIMITS);
> > > > +               sg_policy->need_freq_update = true;
> > > >                 return true;
> > > >         }
> > > >
> > > > @@ -95,11 +95,15 @@ static bool sugov_should_update_freq(struct
> > > > sugov_policy *sg_policy, u64 time)
> > > >  static bool sugov_update_next_freq(struct sugov_policy *sg_policy, u64 time,
> > > >                                    unsigned int next_freq)
> > > >  {
> > > > -       if (sg_policy->need_freq_update)
> > > > +       if (sg_policy->need_freq_update) {
> > > >                 sg_policy->need_freq_update = false;
> > > > -       else if (sg_policy->next_freq == next_freq)
> > > > -               return false;
> > > > +               if (cpufreq_driver_test_flags(CPUFREQ_NEED_UPDATE_LIMITS))
> > > > +                       goto change;
> > > > +       }
> > > >
> > > > +       if (sg_policy->next_freq == next_freq)
> > > > +               return false;
> > > > +change:
> > > >         sg_policy->next_freq = next_freq;
> > > >         sg_policy->last_freq_update_time = time;
> > >
> > > If CPUFREQ_NEED_UPDATE_LIMITS isn't specified, then there's no need to request a
> > > frequency switch from the driver when the current frequency is exactly the same
> > > as the next frequency.
> >
> > Yes, the following check would return false:
> >
> >  +       if (sg_policy->next_freq == next_freq)
> >  +               return false;
>
> But what does that change fix? In fact, that change causes a limits update to
> trigger a frequency switch request to the driver even when the new frequency is
> the same as the current one.

We set the sg_policy->need_freq_update = false instead of
sg_policy->need_freq_update =
cpufreq_driver_test_flags(CPUFREQ_NEED_UPDATE_LIMITS),
to fix the original issue, and then add the
+               if (cpufreq_driver_test_flags(CPUFREQ_NEED_UPDATE_LIMITS))
+                       goto change;

to allow cpufreq to update when CPUFREQ_NEED_UPDATE_LIMITS is set.

>
> Sultan