linux-kernel - Re: [PATCH v2 0/3] EM / PM: Inefficient OPPs

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 for Android: free password hash cracker in your pocket

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

Message-ID: <4a3db514-bec5-394f-ec3f-15c23b44b8f6@arm.com>
Date:   Wed, 26 May 2021 12:50:45 +0100
From:   Lukasz Luba <lukasz.luba@....com>
To:     Viresh Kumar <viresh.kumar@...aro.org>
Cc:     Vincent Donnefort <vincent.donnefort@....com>,
        peterz@...radead.org, rjw@...ysocki.net,
        vincent.guittot@...aro.org, qperret@...gle.com,
        linux-kernel@...r.kernel.org, ionela.voinescu@....com,
        dietmar.eggemann@....com
Subject: Re: [PATCH v2 0/3] EM / PM: Inefficient OPPs



On 5/26/21 11:39 AM, Lukasz Luba wrote:
> 

[snip]

To summarize:
- we don't want to disable some OPPs
- we want to give a 'hint' from energy perspective
- we rely on SchedUtil 2nd stage which clamps this
freq hint value to allowed OPPs which might set
actually not the one what we see as 'efficient'
-- we don't harm some existing platform which might
    needs these 'inefficient' OPPs in some use cases
- we pay some extra cost in this SchedUtil freq
switch path, which shouldn't harm too much.
-- we pay this cost only for arm/arm64 platforms
    which use EM
-- this cost is balanced by the benefit that we see in
    benchmarks and measured energy
-- the LUT might limit the impact

I hope this would help to better understand the scope and
impact of this patch set.

Regards,
Lukasz