Re: [PATCH v3 1/2] regulator: dt-bindings: document the SGM3804 Dual Output regulator

[Krzysztof Kozlowski]

On Mon, May 04, 2026 at 03:22:49PM +0200, Neil Armstrong wrote:
> On 5/4/26 14:44, Neil Armstrong wrote:
> > +        regulator@3e {
> > +            compatible = "sgmicro,sgm3804";
> > +            reg = <0x3e>;
> > +
> > +            vin-supply = <&vin_reg>;
> > +
> > +            enable-gpios = <&gpio 17 GPIO_ACTIVE_HIGH>,
> > +                           <&gpio 18 GPIO_ACTIVE_HIGH>;
> 
> So this fails with Rob's bot and with DTBS_CHECK with board DT but I can't make
> it fail with dt_binding_check on my side.

Drop DT_SCHEMA_FILES and test against all bindings to reproduce this.

> 
> enable-gpios: [[88, 59, 0], [88, 58, 0]] is too long
> 
> But I found out in Documentation/devicetree/bindings/gpio/gpio-consumer-common.yaml :
>   enable-gpios:
>     maxItems: 1
>     description:
>       GPIO connected to the enable control pin.
> 
> This was rather unexpected...
> 
> So I'll probably need update the bindings to mode the enable gpios in the subnodes instead.
> 

Yes, if these GPIOs control specific regulators, convention is to put
them inside the regulator subnodes. Many other devices do it. This is a
bit different approach than with input supplies, which are supposed to
be put in top-level, even if they are dedicated to specific regulators.


Best regards,
Krzysztof