Received-SPF: pass (google.com: domain of linux-kernel+bounces-9586-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: <20231221211222.1380658-1-martin.blumenstingl@googlemail.com>
 <tek6c6symqgm6x6ujh4m67q32en24pzrkjbchffir7qljo4gor@7qpu4zmgyzpq>
 <CAFBinCAxh0xU2mDRX3t42j6oJ534p9RPUV+dYoRe0oacTw_7iA@mail.gmail.com> <2f2bc3xvemk2x3sno65so6vglmpavjtyeiqzy6yyzwvx5hqtmi@tsfx2hr7rmqp>
In-Reply-To: <2f2bc3xvemk2x3sno65so6vglmpavjtyeiqzy6yyzwvx5hqtmi@tsfx2hr7rmqp>
From: Martin Blumenstingl <martin.blumenstingl@googlemail.com>
Date: Fri, 22 Dec 2023 11:12:30 +0100
Message-ID: <CAFBinCCLorBkGmpeUiep6gT7N__2641ec+f=hJyUgVEv1x6EdA@mail.gmail.com>
Subject: Re: [RFC PATCH v1] regulator: pwm-regulator: Fix continuous
 get_voltage for disabled PWM
To: =?UTF-8?Q?Uwe_Kleine=2DK=C3=B6nig?= <u.kleine-koenig@pengutronix.de>
Cc: lgirdwood@gmail.com, broonie@kernel.org, linux-pwm@vger.kernel.org, 
	linux-amlogic@lists.infradead.org, linux-kernel@vger.kernel.org, 
	Thierry Reding <thierry.reding@gmail.com>, Heiner Kallweit <hkallweit1@gmail.com>, 
	Dmitry Rokosov <ddrokosov@sberdevices.ru>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

Hello Uwe,

On Fri, Dec 22, 2023 at 8:10=E2=80=AFAM Uwe Kleine-K=C3=B6nig
<u.kleine-koenig@pengutronix.de> wrote:
[...]
> Also the calculation is wrong: If a relative duty-cyle in the interval
> [91%; 0%] maps lineary to [860 mV; 1140 mV] you get 1100 mV at
>
>              1100 mV - 860 mV
>         91 + ---------------- * (0 - 91) =3D 13
>              1140 mV - 860 mV
>
> (If the calculations in the driver used signed multiplication and
> division, all the checks for max_uV_duty < min_uV_duty could just go
> away.)
>
> So you want
>
> +               pwm-dutycycle-range =3D <13 0>;
Thank you!

> (if this restriction is really necessary).
I could not find a way around this.
Without this change pwm_regulator_set_voltage() is called with req_min
860mV and req_max 1140mV.
pwm_regulator_set_voltage() will then pick the lowest possible
voltage, which then results in 860mV (exactly what I get without any
patches).

To be able to keep the original minimum voltage in .dts would be to
work on what Mark suggested where he said:
"I'd expect a change in the init_state() function, possibly one that
programs the PWM to reflect the actual hardware state"

[...]
> > -     if (pstate.enabled)
> > -             voltage =3D pwm_get_relative_duty_cycle(&pstate, duty_uni=
t);
> > -     else if (max_uV_duty < min_uV_duty)
> > -             voltage =3D max_uV_duty;
> > -     else
> > -             voltage =3D min_uV_duty;
> > +     voltage =3D pwm_get_relative_duty_cycle(&pstate, duty_unit);
>
> I'd add here:
>
>         if (voltage < min(max_uV_duty, min_uV_duty) ||
>             voltage > max(max_uV_duty, min_uV_duty))
>                 return -ENOTRECOVERABLE;
I can do that - although I think it should be a separate change.


Best regards,
Martin