linux-kernel - Re: [RFC] Documentation/scheduler/schedutil.txt

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: <20201120092724.GB2653684@google.com>
Date:   Fri, 20 Nov 2020 09:27:24 +0000
From:   Quentin Perret <qperret@...gle.com>
To:     Viresh Kumar <viresh.kumar@...aro.org>
Cc:     Peter Zijlstra <peterz@...radead.org>,
        "Rafael J. Wysocki" <rjw@...ysocki.net>,
        Ingo Molnar <mingo@...nel.org>,
        Thomas Gleixner <tglx@...utronix.de>,
        Vincent Guittot <vincent.guittot@...aro.org>,
        Morten Rasmussen <morten.rasmussen@....com>,
        dietmar.eggemann@....com, patrick.bellasi@...bug.net,
        lenb@...nel.org, linux-kernel@...r.kernel.org,
        valentin.schneider@....com, ionela.voinescu@....com
Subject: Re: [RFC] Documentation/scheduler/schedutil.txt

On Friday 20 Nov 2020 at 14:49:04 (+0530), Viresh Kumar wrote:
>     This is unlikely to be an issue on systems where cpufreq policies are
>     shared between multiple CPUs, because in those cases the policy
>     utilization is computed as the maximum of the CPU utilization values
>     over the whole policy and if that turns out to be low, reducing the
>     frequency for the policy most likely is a good idea anyway.

Hmm, I'm not sure I agree with this actually. We may be migrating the
task to a different policy altogether. And even if we migrate to another
CPU in the current policy, the task util_avg may be small just because
it was packed with other tasks on a rq, which means it may not increase
the util of the destination rq by much.

ISTR Douglas' EM-based schedutil boosting series was addressing that at
some point, I'll go have a look back at that discussion...