linux-kernel - Re: [PATCH v1 1/2] cpuidle: governors: teo: Fix tick_intercepts handling in teo

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: <CAJZ5v0iyqW2hC1PJT+XqaSc9i2pvKOTjb+1w+jkCLpJSmA=9Lg@mail.gmail.com>
Date: Mon, 17 Nov 2025 17:15:55 +0100
From: "Rafael J. Wysocki" <rafael@...nel.org>
To: Christian Loehle <christian.loehle@....com>
Cc: "Rafael J. Wysocki" <rafael@...nel.org>, Linux PM <linux-pm@...r.kernel.org>, 
	LKML <linux-kernel@...r.kernel.org>, Aboorva Devarajan <aboorvad@...ux.ibm.com>
Subject: Re: [PATCH v1 1/2] cpuidle: governors: teo: Fix tick_intercepts
 handling in teo_update()

On Mon, Nov 17, 2025 at 10:06 AM Christian Loehle
<christian.loehle@....com> wrote:
>
> On 11/16/25 12:34, Rafael J. Wysocki wrote:
> > From: Rafael J. Wysocki <rafael.j.wysocki@...el.com>
> >
> > The condition deciding whether or not to increase cpu_data->tick_intercepts
> > in teo_update() is reverse, so fix it.
> >
> > Fixes: d619b5cc6780 ("cpuidle: teo: Simplify counting events used for tick management")
> > Cc: All applicable <stable@...r.kernel.org>
> > Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@...el.com>
> > ---
> >
> > I'm planning to apply this for 6.19 on top of
> >
> > https://lore.kernel.org/linux-pm/6228387.lOV4Wx5bFT@rafael.j.wysocki/
> >
> > because that patch (indirectly) depends on commit d619b5cc6780.
> >
> > ---
> >  drivers/cpuidle/governors/teo.c |    2 +-
> >  1 file changed, 1 insertion(+), 1 deletion(-)
> >
> > --- a/drivers/cpuidle/governors/teo.c
> > +++ b/drivers/cpuidle/governors/teo.c
> > @@ -251,7 +251,7 @@ static void teo_update(struct cpuidle_dr
> >               cpu_data->state_bins[idx_timer].hits += PULSE;
> >       } else {
> >               cpu_data->state_bins[idx_duration].intercepts += PULSE;
> > -             if (TICK_NSEC <= measured_ns)
> > +             if (measured_ns <= TICK_NSEC)
>
> nit: Why <= instead of <?

Because it was <= before.

> I guess it really doesn't matter with measured_ns only being a rough approximation
> with an error in the order of wakeup-latency.

Right and moreover, TICK_NSEC is an upper bound for tick wakeups, they
occur earlier as a rule.

> Reviewed-by:
> Christian Loehle <christian.loehle@....com>
>
> Let me go write some tests for all these edge cases :/
>
> IIRC Aboorva's power systems have no idle state deeper than TICK_NSEC, so
> this might make a big difference here, hence CCed.

Like x86 systems with HZ < 1000 which are the majority nowadays AFAICS.

Thanks!