Re: [PATCH v3 12/16] pinctrl: starfive: Add pinctrl driver for StarFive SoCs

[Emil Renner Berthing]

On Tue, 9 Nov 2021 at 21:29, Linus Walleij <linus.walleij@xxxxxxxxxx> wrote:
> On Tue, Nov 9, 2021 at 10:40 AM Emil Renner Berthing <kernel@xxxxxxxx> wrote:
> > On Tue, 9 Nov 2021 at 10:34, Andy Shevchenko <andy.shevchenko@xxxxxxxxx> wrote:
>
> > > > The order the different states are blinked depends entirely on
> > > > how the pinctrl framework parses the device tree. I still think it
> > > > would be more natural to cleanly go to the end result without this
> > > > blinking.
> >
> > Hmm.. but if going through the different states is what you want, then
> > wouldn't you need the device tree to have an ordered list of the
> > states rather than just a single node and also a way to tune how long
> > time the different states are blinked?
>
> In a way you are correct that the DT is a functional language and it's
> a bit lite a style sheet or prolog or something in that the end reduction
> is what counts.
>
> In this case, I would say something is weird if there are interim states,
> the yaml validation should not allow you to set the same thing back
> and forth in your DTS file.

Yes, exactly.

> Alas we are not perfect as in yaml validation isn't perfect either.
> I can't see what the problem is really, just write proper DTS files
> and there will not be any interim states, right?

No, I agree. I think it's only that Andy wasn't sure if these interim
states might be meaningful/useful.

> And if it is possible
> to write DTS files that have states and sequence requirements,
> these should be caught in validation. Should be.
>
> Yours,
> Linus Walleij