Re: [PATCH] dt-bindings: leds: lacie,ns2-leds: Convert to DT schema

On Wed, Jul 01, 2026 at 06:05:52PM +0000, [email protected] wrote:
> Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
> - [Low] The schema conversion incorrectly restricts LED child node names to `^led(-[0-9a-z]+)?$`, which breaks `dtbs_check` validation for all existing upstream device trees using this binding.
> --
> 
> commit 6c59295f22b17d02a2b26847db54458f820fa4f3
> Author: Ninad Naik <[email protected]>
> 
> dt-bindings: leds: lacie,ns2-leds: Convert to DT schema
> 
> Convert lacie,ns2-leds to DT schema.
> 
> > diff --git a/Documentation/devicetree/bindings/leds/lacie,ns2-leds.yaml b/Documentation/devicetree/bindings/leds/lacie,ns2-leds.yaml
> > new file mode 100644
> > index 0000000000000..679a618906d80
> > --- /dev/null
> > +++ b/Documentation/devicetree/bindings/leds/lacie,ns2-leds.yaml
> > @@ -0,0 +1,87 @@
> [ ... ]
> > +properties:
> > +  compatible:
> > +    const: lacie,ns2-leds
> > +
> > +patternProperties:
> > +  "^led(-[0-9a-z]+)?$":
> 
> [Severity: Low]
> Will this restrictive regex cause existing device trees to fail dtbs_check?
> 
> The old text binding example used a child node named blue-sata, which does
> not match this pattern. Since additionalProperties is set to false below,
> nodes in existing device trees that don't start with led will be rejected
> during validation.
> 
> Should this regex be relaxed to allow existing node names, or should those
> device tree files be updated alongside this conversion?

While yes we prefer what you did, it's perhaps not worth changing these 
old .dts files to match the preference.

Rob