Re: [PATCH v2 1/6] dt-bindings: hwmon: add binding for adi,adt7470

Hi Uwe,

> On Fri, Jul 17, 2026 at 05:59:14PM -0300, Luiz Angelo Daros de Luca wrote:
> > +  "#pwm-cells":
> > +    const: 2
> > +    description: |
> > +      Number of cells in a PWM specifier.
> > +      - The first cell is the PWM channel (0 to 3).
> > +      - The second cell is the PWM period in nanoseconds.
>
> Please use #pwm-cells = <3> and $ref: pwm.yaml#

Thanks for the review. I already switched to #pwm-cells = <3> for
the next version, so that part is settled.

Regarding $ref: /schemas/pwm/pwm.yaml#, it doesn't quite work here.
When I added it and ran
dt_binding_check, it started rejecting the example node:

hwmon@2f (adi,adt7470): $nodename:0: 'hwmon@2f' does not match
'^pwm(@.*|-([0-9]|[1-9][0-9]+))?$'

It turns out pwm.yaml also constrains $nodename as part of the
allOf, so referencing it forces the node name into the pwm@...
pattern. That's not right for this device: the ADT7470 is primarily a
hwmon chip (10 temperature inputs, 4 fan/PWM outputs), and the PWM
channels are just one of the interfaces it exposes, so hwmon@2f is
the correct node name here, not pwm@2f.

So, for the next version, I'm declaring #pwm-cells and
#thermal-sensor-cells directly in the binding instead of pulling in
pwm.yaml via $ref, and pointing to pwm.yaml only in the description
for context:

"#pwm-cells":
const: 3
description:
See Documentation/devicetree/bindings/pwm/pwm.yaml for the
generic PWM binding description.

Let me know if you'd rather see this handled differently (e.g. keeping
the $ref and accepting the pwm@ node name), but I think keeping the
node name aligned with the device's primary function is the more
correct approach.

Regards,

Luiz