On Wed, Jun 26, 2019 at 11:58:44AM +0200, Thierry Reding wrote:
> On Wed, Jun 26, 2019 at 10:58:27AM +0200, Uwe Kleine-König wrote:
> > On Tue, Jun 25, 2019 at 11:38:39AM +0200, Thierry Reding wrote:
> > > On Mon, Jun 24, 2019 at 12:28:44PM +0100, Daniel Thompson wrote:
> > > > [...] although given pwm-backlight is essentially a wrapper driver
> > > > round a PWM I wondered why the pinctrl was on the backlight node
> > > > (rather than the PWM node).
> > >
> > > I agree with this. We're defining the pin control state for the PWM pin,
> > > so in my opinion it should be the PWM driver that controls it.
> > >
> > > One reason why I think this is important is if we ever end up with a
> > > device that requires pins from two different controllers to be
> > > configured at runtime, then how would we model that? Since pin control
> > > states cannot be aggregated, so you'd have to have multiple "default"
> > > states, each for the pins that they control.
> >
> > I thought you can do:
> >
> > pinctrl-names = "default";
> > pinctrl-0 = <&pinctrl_in_first_pincontroller>, <&pinctrl_in_another_controller>;
> >
> > if two (or more) controllers are involved.
>
> You're right. Both the bindings say that this can be done and the code
> is also there to parse multiple states per pinctrl-* entry.
>
> > > On the other hand if we associate the pin control states with each of
> > > the resources that need those states, then when those resources are
> > > controlled, they will automatically know how to deal with the states.
> > > The top-level device (i.e. backlight) doesn't need to concern itself
> > > with those details.
> >
> > So the options are:
> >
> > a) put "active" and "inactive" pinctrls into the pwm-node, and nothing
> > related to the involved PWM pins in the consumer
> >
> > b) put the PWM pin config in the consumer's "default" pinctrl (and
> > maybe leave it out int "init" if you want smooth taking over).
>
> You can't put it into the "default" state because that state is applied
> before the consumer driver's ->probe().
If you do:
mybacklight {
pinctrl-names = "init", "default";
pinctrl-0 = <&pinctrl_without_pwm>
pinctrl-1 = <&pinctrl_with_pwm>;
...
};
Then nothing is done before probing of the backlight and only when the
probing is done and the pwm is taken over, the PWM-pinctrl is applied.
The only ugly thing here I can identify is that probe() might exit with
the PWM running and then enabling the pinmux for the PWM pin results in
an incomplete period at the beginning. But this happens only in some
corner cases that might not matter. (i.e. if the bootloader enabled the
PWM but didn't setup the pinmux; and if .probe enabled the PWM which we
agreed it probably shouldn't on it's own.)
> > (Or maybe use "enabled" and "disabled" in a) to match the pwm_states
> > .enabled?)
>
> Yeah, I think this is what we'll need to do in order to implement the
> explicit behaviour that we need here.
>
> > The advantages I see in b) over a) are:
> >
> > - "default" and "init" are a known pinctrl concept that most people
> > should have understood.
>
> The problem is that they won't work in this case. The "init" state will
> be applied before the consumer driver's ->probe() if it exists. If it
> doesn't then "default" will be applied instead. Both cases are not
> something that we want if we want to take over the existing
> configuration.
>
> > - You have all pinctrl config for the backlight in a single place.
>
> Depending on your point of view this could be considered a disadvantage.
Yeah, right, this is subjective.
> [...]
> Like I pointed out above, I don't think that's the case. But I don't
> want to overcomplicate things, so if you can prove that it can be done
> with the existing pinctrl helpers, I'd be happy to be proven wrong.
I tried, see above :-)
Best regards
Uwe
--
Pengutronix e.K. | Uwe Kleine-König |
Industrial Linux Solutions | http://www.pengutronix.de/ |
_______________________________________________
dri-devel mailing list
dri-devel@xxxxxxxxxxxxxxxxxxxxx
https://lists.freedesktop.org/mailman/listinfo/dri-devel
