在 2023-06-21星期三的 00:50 +0100，Andre Przywara写道：
> On Tue, 20 Jun 2023 15:00:20 -0500
> Chris Morgan <macroalpha82@xxxxxxxxx> wrote:
>
> Hi Chris,
>
> thanks for taking care of upstreaming, cute little device.
>
> > From: Chris Morgan <macromorgan@xxxxxxxxxxx>
> >
> > Add a default pinctrl for the pwm function.
> >
> > Signed-off-by: Chris Morgan <macromorgan@xxxxxxxxxxx>
>
>
> > ---
> > arch/arm/boot/dts/sun8i-v3s.dtsi | 7 +++++++
> > 1 file changed, 7 insertions(+)
> >
> > diff --git a/arch/arm/boot/dts/sun8i-v3s.dtsi
> > b/arch/arm/boot/dts/sun8i-v3s.dtsi
> > index b001251644f7..e5977524abe2 100644
> > --- a/arch/arm/boot/dts/sun8i-v3s.dtsi
> > +++ b/arch/arm/boot/dts/sun8i-v3s.dtsi
> > @@ -414,6 +414,11 @@ mmc1_pins: mmc1-pins {
> > bias-pull-up;
> > };
> >
> > + pwm_pins: pwm-pins {
> > + pins = "PB4";
> > + function = "pwm0";
> > + };
> > +
> > spi0_pins: spi0-pins {
> > pins = "PC0", "PC1", "PC2", "PC3";
> > function = "spi0";
> > @@ -441,6 +446,8 @@ pwm: pwm@1c21400 {
> > "allwinner,sun7i-a20-pwm";
> > reg = <0x01c21400 0xc>;
> > clocks = <&osc24M>;
> > + pinctrl-0 = <&pwm_pins>;
> > + pinctrl-names = "default";
>
> How is this supposed to work with multiple channels? There is PWM1 on
> PB5. If one wants to potentially use that, we would need to add a
> reference to those pins here as well, and they would all be muxed to
> PWM upon the PWM controller probing?
>
> So while I see that it's the only output pin for PWM0, this might
> still
> need to go into the board .dts, alongside the status = "okay"; line.
> So
> each board would specify exactly the pins it needs (PWM0 only, PWM1
> only, both or none).
Maybe we should have two PWM pinctrl nodes, pwm0_pins and pwm1_pins,
and then enable the controller (with the pinctrl property) in
individual board DTs.
>
> Otherwise I compared this against the manual and Linux pinctrl
> driver,
> it all matches up.
>
> Cheers,
> Andre
>
> > #pwm-cells = <3>;
> > status = "disabled";
> > };
>
