Received-SPF: pass (google.com: domain of linux-kernel+bounces-155704-linux.lists.archive=gmail.com@vger.kernel.org designates 139.178.88.99 as permitted sender) client-ip=139.178.88.99;
Precedence: bulk
MIME-Version: 1.0
References: <13515747.uLZWGnKmhe@kreacher> <1913649.CQOukoFCf9@kreacher> <8e26c3cb-1283-4561-95aa-30432f1d13ee@linaro.org>
In-Reply-To: <8e26c3cb-1283-4561-95aa-30432f1d13ee@linaro.org>
From: "Rafael J. Wysocki" <rafael@kernel.org>
Date: Tue, 23 Apr 2024 19:54:49 +0200
Message-ID: <CAJZ5v0h=15LYhukPWmHPK5hvD=2u75rTEiC8oJMVBFziMkB5gQ@mail.gmail.com>
Subject: Re: [PATCH v1 07/16] thermal: gov_power_allocator: Eliminate a
 redundant variable
To: Daniel Lezcano <daniel.lezcano@linaro.org>
Cc: "Rafael J. Wysocki" <rjw@rjwysocki.net>, Linux PM <linux-pm@vger.kernel.org>, 
	LKML <linux-kernel@vger.kernel.org>, Lukasz Luba <lukasz.luba@arm.com>, 
	Srinivas Pandruvada <srinivas.pandruvada@linux.intel.com>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

On Tue, Apr 23, 2024 at 7:35=E2=80=AFPM Daniel Lezcano
<daniel.lezcano@linaro.org> wrote:
>
> On 10/04/2024 18:12, Rafael J. Wysocki wrote:
> > From: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> >
> > Notice that the passive field in struct thermal_zone_device is not
> > used by the Power Allocator governor itself and so the ordering of
> > its updates with respect to allow_maximum_power() or allocate_power()
> > does not matter.
> >
> > Accordingly, make power_allocator_manage() update that field right
> > before returning, which allows the current value of it to be passed
> > directly to allow_maximum_power() without using the additional update
> > variable that can be dropped.
> >
> > Signed-off-by: Rafael J. Wysocki <rafael.j.wysocki@intel.com>
> > ---
>
> The step_wise and the power allocator are changing the tz->passive
> values, so telling the core to start and stop the passive mitigation time=
r.
>
> It looks strange that a plugin controls the core internal and not the
> opposite.
>
> I'm wondering if it would not make sense to have the following ops:
>
>         .start
>         .stop
>
> .start is called when the first trip point is crossed the way up
> .stop is called when the first trip point is crossed the way down
>
>   - The core is responsible to start and stop the passive mitigation time=
r.
>
>   - the governors do no longer us tz->passive
>
> The reset of the governor can happen at start or stop, as well as the
> device cooling states.

I have a patch that simply increments tz->passive when a passive trip
point is passed on the way up and decrements it when a passive trip
point is crossed on the way down.  It appears to work reasonably well.