Received-SPF: pass (google.com: domain of linux-kernel+bounces-92623-linux.lists.archive=gmail.com@vger.kernel.org designates 147.75.199.223 as permitted sender) client-ip=147.75.199.223;
Precedence: bulk
MIME-Version: 1.0
References: <20240305095111.777525-1-christian.loehle@arm.com>
 <CAKfTPtC6D4RTP3T8tqEJKA1drGbAANOaGPKqmELd1G8TLocizg@mail.gmail.com> <0e2833ee-0939-44e0-82a2-520a585a0153@arm.com>
In-Reply-To: <0e2833ee-0939-44e0-82a2-520a585a0153@arm.com>
From: Vincent Guittot <vincent.guittot@linaro.org>
Date: Tue, 5 Mar 2024 16:39:39 +0100
Message-ID: <CAKfTPtDtu1-9noPuV0bK=nbkX0XzF8raojb1je=RUPpzKRC85w@mail.gmail.com>
Subject: Re: [PATCH] sched/fair: Remove stale FREQUENCY_UTIL comment
To: Christian Loehle <christian.loehle@arm.com>
Cc: linux-kernel@vger.kernel.org, mingo@redhat.com, peterz@infradead.org, 
	dietmar.eggemann@arm.com
Content-Type: text/plain; charset="UTF-8"

On Tue, 5 Mar 2024 at 16:18, Christian Loehle <christian.loehle@arm.com> wrote:
>
> On 05/03/2024 15:05, Vincent Guittot wrote:
> > On Tue, 5 Mar 2024 at 10:51, Christian Loehle <christian.loehle@arm.com> wrote:
> >>
> >> effective_cpu_util() flags were removed, so remove the comment part
> >> mentioning it.
> >>
> >> commit 9c0b4bb7f6303 ("sched/cpufreq: Rework schedutil governor performance estimation")
> >> reworked effective_cpu_util() removing enum cpu_util_type.
> >>
> >> Signed-off-by: Christian Loehle <christian.loehle@arm.com>
> >> ---
> >>  kernel/sched/fair.c | 2 --
> >>  1 file changed, 2 deletions(-)
> >>
> >> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> >> index 533547e3c90a..6b0c47d00fd3 100644
> >> --- a/kernel/sched/fair.c
> >> +++ b/kernel/sched/fair.c
> >> @@ -7849,8 +7849,6 @@ eenv_pd_max_util(struct energy_env *eenv, struct cpumask *pd_cpus,
> >>                  * Performance domain frequency: utilization clamping
> >>                  * must be considered since it affects the selection
> >>                  * of the performance domain frequency.
> >> -                * NOTE: in case RT tasks are running, by default the
> >> -                * FREQUENCY_UTIL's utilization can be max OPP.
> >
> > Yes, it's no more accurate.
> > Should we update the comment instead of deleting it as the effect is
> > still there ?
> >  "NOTE: in case RT tasks are running, by default the min utilization
> > can be max OPP."
>
> I'm fine with either and that was my first thought here, too, but it did seem like
> the comment was mostly placed there to justify the 'unexpected' high utilization
> when explicitly passing FREQUENCY_UTIL and the need to clamp it then.

The behavior is still there but it has moved in the min parameter .
The problem remains the same when it come to select a compute capacity
with sugov_effective_cpu_perf

> So removing did feel slightly more natural to me anyway.
>
> So alternatively:
>
> From: Christian Loehle <christian.loehle@arm.com>
> Date: Tue, 5 Mar 2024 09:34:41 +0000
> Subject: [PATCH] sched/fair: Remove stale FREQUENCY_UTIL mention
>
> effective_cpu_util() flags were removed, so remove mentioning of the
> flag.
>
> commit 9c0b4bb7f6303 ("sched/cpufreq: Rework schedutil governor performance estimation")
> reworked effective_cpu_util() removing enum cpu_util_type. Modify the
> comment accordingly.
>
> Signed-off-by: Christian Loehle <christian.loehle@arm.com>

Reviewed-by: Vincent Guittot <vincent.guittot@linaro.org>

> ---
>  kernel/sched/fair.c | 4 ++--
>  1 file changed, 2 insertions(+), 2 deletions(-)
>
> diff --git a/kernel/sched/fair.c b/kernel/sched/fair.c
> index 533547e3c90a..f5151cbcd94e 100644
> --- a/kernel/sched/fair.c
> +++ b/kernel/sched/fair.c
> @@ -7849,8 +7849,8 @@ eenv_pd_max_util(struct energy_env *eenv, struct cpumask *pd_cpus,
>                  * Performance domain frequency: utilization clamping
>                  * must be considered since it affects the selection
>                  * of the performance domain frequency.
> -                * NOTE: in case RT tasks are running, by default the
> -                * FREQUENCY_UTIL's utilization can be max OPP.
> +                * NOTE: in case RT tasks are running, by default the min
> +                * utilization can be max OPP.
>                  */
>                 eff_util = effective_cpu_util(cpu, util, &min, &max);
>
> --
> 2.34.1
>