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:   Wed, 9 May 2018 14:10:01 +0530
From:   Viresh Kumar <viresh.kumar@linaro.org>
To:     "Rafael J. Wysocki" <rafael@kernel.org>
Cc:     Joel Fernandes <joel@joelfernandes.org>,
        Juri Lelli <juri.lelli@redhat.com>,
        Claudio Scordino <claudio@evidence.eu.com>,
        Linux Kernel Mailing List <linux-kernel@vger.kernel.org>,
        "Rafael J . Wysocki" <rafael.j.wysocki@intel.com>,
        Peter Zijlstra <peterz@infradead.org>,
        Ingo Molnar <mingo@redhat.com>,
        Patrick Bellasi <patrick.bellasi@arm.com>,
        Luca Abeni <luca.abeni@santannapisa.it>,
        Joel Fernandes <joelaf@google.com>,
        Linux PM <linux-pm@vger.kernel.org>
Subject: Re: [RFC PATCH] sched/cpufreq/schedutil: handling urgent frequency
 requests
Message-ID: <20180509084001.bghnwpv3a3xnuxce@vireshk-i7>
References: <1525704215-8683-1-git-send-email-claudio@evidence.eu.com>
 <20180508065435.bcht6dyb3rpp6gk5@vireshk-i7>
 <20180509045425.GA158882@joelaf.mtv.corp.google.com>
 <20180509064530.GA1681@localhost.localdomain>
 <20180509080644.GA76874@joelaf.mtv.corp.google.com>
 <CAJZ5v0iFZ5+6BQCism-o8bnOxD8KveueJ5CT58F9uH_7_pYOXQ@mail.gmail.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Disposition: inline
In-Reply-To: <CAJZ5v0iFZ5+6BQCism-o8bnOxD8KveueJ5CT58F9uH_7_pYOXQ@mail.gmail.com>
User-Agent: NeoMutt/20180323-120-3dd1ac
Sender: linux-kernel-owner@vger.kernel.org
Precedence: bulk

On 09-05-18, 10:30, Rafael J. Wysocki wrote:
> On Wed, May 9, 2018 at 10:06 AM, Joel Fernandes <joel@joelfernandes.org> wrote:
> > How about this? Will use the latest request, and also doesn't do unnecessary
> > irq_work_queue:

I almost wrote the same stuff before I went for lunch :)

> > (untested)
> > -----8<--------
> > diff --git a/kernel/sched/cpufreq_schedutil.c b/kernel/sched/cpufreq_schedutil.c
> > index d2c6083304b4..6a3e42b01f52 100644
> > --- a/kernel/sched/cpufreq_schedutil.c
> > +++ b/kernel/sched/cpufreq_schedutil.c
> > @@ -38,7 +38,7 @@ struct sugov_policy {
> >         struct                  mutex work_lock;
> >         struct                  kthread_worker worker;
> >         struct task_struct      *thread;
> > -       bool                    work_in_progress;
> > +       bool                    work_in_progress; /* Has kthread been kicked */
> >
> >         bool                    need_freq_update;
> >  };
> > @@ -92,9 +92,6 @@ static bool sugov_should_update_freq(struct sugov_policy *sg_policy, u64 time)
> >             !cpufreq_can_do_remote_dvfs(sg_policy->policy))
> >                 return false;
> >
> > -       if (sg_policy->work_in_progress)
> > -               return false;
> > -
> 
> Why this change?
> 
> Doing the below is rather pointless if work_in_progress is set, isn't it?
> 
> You'll drop the results of it on the floor going forward anyway then AFAICS.
> 
> >         if (unlikely(sg_policy->need_freq_update)) {
> >                 sg_policy->need_freq_update = false;
> >                 /*
> > @@ -129,8 +126,11 @@ static void sugov_update_commit(struct sugov_policy *sg_policy, u64 time,
> >                 policy->cur = next_freq;
> >                 trace_cpu_frequency(next_freq, smp_processor_id());
> >         } else {
> > -               sg_policy->work_in_progress = true;
> > -               irq_work_queue(&sg_policy->irq_work);
> > +               /* work_in_progress helps us not queue unnecessarily */
> > +               if (!sg_policy->work_in_progress) {
> > +                       sg_policy->work_in_progress = true;
> > +                       irq_work_queue(&sg_policy->irq_work);
> > +               }
> >         }
> >  }

Right, none of the above changes are required now.

> > @@ -381,13 +381,23 @@ sugov_update_shared(struct update_util_data *hook, u64 time, unsigned int flags)
> >  static void sugov_work(struct kthread_work *work)
> >  {
> >         struct sugov_policy *sg_policy = container_of(work, struct sugov_policy, work);
> > +       unsigned int freq;
> > +
> > +       /*
> > +        * Hold sg_policy->update_lock just enough to handle the case where:
> > +        * if sg_policy->next_freq is updated before work_in_progress is set to
> > +        * false, we may miss queueing the new update request since
> > +        * work_in_progress would appear to be true.
> > +        */
> > +       raw_spin_lock(&sg_policy->update_lock);
> > +       freq = sg_policy->next_freq;
> > +       sg_policy->work_in_progress = false;
> > +       raw_spin_unlock(&sg_policy->update_lock);

One problem we still have is that sg_policy->update_lock is only used
in the shared policy case and not in the single CPU per policy case,
so the race isn't solved there yet.

> >         mutex_lock(&sg_policy->work_lock);
> > -       __cpufreq_driver_target(sg_policy->policy, sg_policy->next_freq,
> > +       __cpufreq_driver_target(sg_policy->policy, freq,
> >                                 CPUFREQ_RELATION_L);
> >         mutex_unlock(&sg_policy->work_lock);
> > -
> > -       sg_policy->work_in_progress = false;
> >  }
> >
> >  static void sugov_irq_work(struct irq_work *irq_work)

-- 
viresh