linux-kernel - Re: [PATCH] cpufreq: acpi: Don't enable boost on policy exit

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: <CAJZ5v0iO4=nHcATzPyiKiWumdETFRR32C97K_RH=yhD--Tai=g@mail.gmail.com>
Date: Thu, 17 Apr 2025 14:39:43 +0200
From: "Rafael J. Wysocki" <rafael@...nel.org>
To: Viresh Kumar <viresh.kumar@...aro.org>
Cc: Nicholas Chin <nic.c3.14@...il.com>, linux-kernel@...r.kernel.org, 
	linux-pm@...r.kernel.org, rafael.j.wysocki@...el.com, rafael@...nel.org, 
	vincent.guittot@...aro.org, zhenglifeng1@...wei.com
Subject: Re: [PATCH] cpufreq: acpi: Don't enable boost on policy exit

On Thu, Apr 17, 2025 at 7:09 AM Viresh Kumar <viresh.kumar@...aro.org> wrote:
>
> Copying more information from Bugzilla here (Nicholas, it would be
> faster if you can put all your observations here first, more people
> are looking at emails than bugzilla).
>
> > Nicholas Chin wrote:
> > I did some more testing and debugging and it seems like when
> > cpufreq_online() runs after waking the system, policy->boost_enabled
> > and cpufreq_boost_enabled() are both 0, so the set_boost() at the end
> > of that function is never run.
>
> Right, this is what I wanted to do with the $Subject patch. Don't
> update boost anymore in suspend/resume

This is going to work for suspend-to-idle, but not necessarily for S3.

BTW, the patch is correct IMV, so I'm not going to drop it, but it
looks like something more is needed on top of it.

> > cpufreq_boost_enabled() being 0 indicates that the MSR has boosting
> > disabled, but when I read out that MSR using rdmsr the bit seems to
> > indicate that it is actually enabled (I am aware of the inverted logic
> > of that bit). set_boost() seems to be the only place in the kernel
> > that causes that MSR to be modified, and I didn't see any extra calls
> > to it in my debug logs, so it seems like something else (outside the
> > kernel?) is setting that MSR.
>
> And this is what I feel too, something else in kernel or outside of it
> is doing something tricky.

On a resume from S3, you actually don't know if the platform firmware
has preserved the configuration from before the suspend transition.
It may not.