From: "Rafael J. Wysocki" <rafael@kernel.org>
To: Viresh Kumar <viresh.kumar@linaro.org>
Cc: Doug Smythies <dsmythies@telus.net>,
Rafael Wysocki <rjw@rjwysocki.net>,
Ingo Molnar <mingo@redhat.com>,
Peter Zijlstra <peterz@infradead.org>,
Linux PM <linux-pm@vger.kernel.org>,
Vincent Guittot <vincent.guittot@linaro.org>,
Joel Fernandes <joel@joelfernandes.org>,
"v4 . 18+" <stable@vger.kernel.org>,
Linux Kernel Mailing List <linux-kernel@vger.kernel.org>
Subject: Re: [PATCH] cpufreq: schedutil: Don't skip freq update when limits change
Date: Tue, 23 Jul 2019 12:27:32 +0200 [thread overview]
Message-ID: <CAJZ5v0ji+ksapJ4kc2m5UM_O+AShAvJWmYhTQHiXiHnpTq+xRg@mail.gmail.com> (raw)
In-Reply-To: <20190723091551.nchopfpqlmdmzvge@vireshk-i7>
On Tue, Jul 23, 2019 at 11:15 AM Viresh Kumar <viresh.kumar@linaro.org> wrote:
>
> On 23-07-19, 00:10, Doug Smythies wrote:
> > On 2019.07.21 23:52 Viresh Kumar wrote:
> >
> > > To avoid reducing the frequency of a CPU prematurely, we skip reducing
> > > the frequency if the CPU had been busy recently.
> > >
> > > This should not be done when the limits of the policy are changed, for
> > > example due to thermal throttling. We should always get the frequency
> > > within limits as soon as possible.
> > >
> > > Fixes: ecd288429126 ("cpufreq: schedutil: Don't set next_freq to UINT_MAX")
> > > Cc: v4.18+ <stable@vger.kernel.org> # v4.18+
> > > Reported-by: Doug Smythies <doug.smythies@gmail.com>
> > > Signed-off-by: Viresh Kumar <viresh.kumar@linaro.org>
> > > ---
> > > @Doug: Please try this patch, it must fix the issue you reported.
> >
> > It fixes the driver = acpi-cpufreq ; governor = schedutil test case
> > It does not fix the driver = intel_cpufreq ; governor = schedutil test case
> >
> > I have checked my results twice, but will check again in the day or two.
>
> The patch you tried to revert wasn't doing any driver specific stuff
> but only schedutil. If that revert fixes your issue with both the
> drivers, then this patch should do it as well.
>
> I am clueless now on what can go wrong with intel_cpufreq driver with
> schedutil now.
>
> Though there is one difference between intel_cpufreq and acpi_cpufreq,
> intel_cpufreq has fast_switch_possible=true and so it uses slightly
> different path in schedutil. I tried to look from that perspective as
> well but couldn't find anything wrong.
acpi-cpufreq should use fast switching on the Doug's system too.
> If you still find intel_cpufreq to be broken, even with this patch,
> please set fast_switch_possible=false instead of true in
> __intel_pstate_cpu_init() and try tests again. That shall make it very
> much similar to acpi-cpufreq driver.
I wonder if this helps. Even so, we want fast switching to be used by
intel_cpufreq.
Anyway, it looks like the change reverted by the Doug's patch
introduced a race condition that had not been present before. Namely,
need_freq_update is cleared in get_next_freq() when it is set _or_
when the new freq is different from the cached one, so in the latter
case if it happens to be set by sugov_limits() after evaluating
sugov_should_update_freq() (which returned 'true' for timing reasons),
that update will be lost now. [Previously the update would not be
lost, because the clearing of need_freq_update depended only on its
current value.] Where it matters is that in the "need_freq_update set"
case, the "premature frequency reduction avoidance" should not be
applied (as you noticed and hence the $subject patch).
However, even with the $subject patch, need_freq_update may still be
set by sugov_limits() after the check added by it and then cleared by
get_next_freq(), so it doesn't really eliminate the problem.
IMO eliminating would require invalidating next_freq this way or
another when need_freq_update is set in sugov_should_update_freq(),
which was done before commit ecd2884291261e3fddbc7651ee11a20d596bb514.
next prev parent reply other threads:[~2019-07-23 10:27 UTC|newest]
Thread overview: 27+ messages / expand[flat|nested] mbox.gz Atom feed top
2019-07-18 6:26 [PATCH] Revert "cpufreq: schedutil: Don't set next_freq to UINT_MAX" Doug Smythies
2019-07-18 10:28 ` Viresh Kumar
2019-07-18 15:46 ` Doug Smythies
2019-07-22 6:49 ` Viresh Kumar
2019-07-22 6:51 ` [PATCH] cpufreq: schedutil: Don't skip freq update when limits change Viresh Kumar
2019-07-23 7:10 ` Doug Smythies
2019-07-23 9:13 ` Rafael J. Wysocki
2019-07-23 9:15 ` Viresh Kumar
2019-07-23 10:27 ` Rafael J. Wysocki [this message]
2019-07-24 11:43 ` Viresh Kumar
2019-07-25 15:20 ` Doug Smythies
2019-07-26 3:26 ` Viresh Kumar
2019-07-26 6:57 ` Viresh Kumar
2019-07-29 7:55 ` Doug Smythies
2019-07-29 8:32 ` Viresh Kumar
2019-07-29 8:37 ` Rafael J. Wysocki
2019-08-01 0:20 ` Doug Smythies
2019-08-01 6:17 ` Viresh Kumar
2019-08-01 7:47 ` Rafael J. Wysocki
2019-08-01 7:55 ` Viresh Kumar
2019-08-01 17:57 ` Doug Smythies
2019-08-02 3:48 ` Viresh Kumar
2019-08-02 9:11 ` Rafael J. Wysocki
2019-08-02 9:19 ` Rafael J. Wysocki
2019-08-06 4:00 ` Viresh Kumar
2019-07-31 2:58 Viresh Kumar
2019-07-31 23:19 ` Doug Smythies
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=CAJZ5v0ji+ksapJ4kc2m5UM_O+AShAvJWmYhTQHiXiHnpTq+xRg@mail.gmail.com \
--to=rafael@kernel.org \
--cc=dsmythies@telus.net \
--cc=joel@joelfernandes.org \
--cc=linux-kernel@vger.kernel.org \
--cc=linux-pm@vger.kernel.org \
--cc=mingo@redhat.com \
--cc=peterz@infradead.org \
--cc=rjw@rjwysocki.net \
--cc=stable@vger.kernel.org \
--cc=vincent.guittot@linaro.org \
--cc=viresh.kumar@linaro.org \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).