linux-kernel - Re: [RFC PATCH v3 05/10] sched/topology: Reference the Energy Model of CPUs when available

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] [day] [month] [year] [list]

Message-ID: <20180620075809.GA23168@e108498-lin.cambridge.arm.com>
Date:   Wed, 20 Jun 2018 08:58:09 +0100
From:   Quentin Perret <quentin.perret@....com>
To:     Peter Zijlstra <peterz@...radead.org>
Cc:     rjw@...ysocki.net, gregkh@...uxfoundation.org,
        linux-kernel@...r.kernel.org, linux-pm@...r.kernel.org,
        mingo@...hat.com, dietmar.eggemann@....com,
        morten.rasmussen@....com, chris.redpath@....com,
        patrick.bellasi@....com, valentin.schneider@....com,
        vincent.guittot@...aro.org, thara.gopinath@...aro.org,
        viresh.kumar@...aro.org, tkjos@...gle.com, joelaf@...gle.com,
        smuckle@...gle.com, adharmap@...cinc.com, skannan@...cinc.com,
        pkondeti@...eaurora.org, juri.lelli@...hat.com,
        edubezval@...il.com, srinivas.pandruvada@...ux.intel.com,
        currojerez@...eup.net, javi.merino@...nel.org
Subject: Re: [RFC PATCH v3 05/10] sched/topology: Reference the Energy Model
 of CPUs when available

On Tuesday 19 Jun 2018 at 20:42:50 (+0200), Peter Zijlstra wrote:
> On Tue, Jun 19, 2018 at 06:13:17PM +0100, Quentin Perret wrote:
> > Would replacing the static_key by a flag attached to the root_domain be
> > reasonable ? 
> 
> Keep the static key as is, enable if any root domain needs it.

OK. So basically the semantics of the static key becomes: it is set if
at least one root domain has a non-empty list of freq domains. And a
root_domain has a non-empty list of frequency domains if it meets all
conditions for EAS (SD_ASYM_CPUCAPACITY flag set, low complexity for
the "portion" of the EM covering it, ...)

And then, checking the status of the list (empty or not) for a specific
root domain can replace my flag I guess. I'll need to check that somewhere
around select_task_rq_fair() in addition to the check on the static key to
decide if a task can go in find_energy_efficient_cpu().

> > Sorry but I didn't understand that ...
> 
> If you have only a single freq-domain, there is nothing to do, right?

Ah, right. Since we already agreed that all CPUs in a freq domain must
have the same micro-arch, that's correct. Checking the presence of the
SD_ASYM_CPUCAPACITY flag should cover this case I guess.

Thanks,
Quentin