Re: [PATCH v2 1/4] dt-bindings: reset: nuvoton,ma35d1-reset: add simple-mfd and child node support

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] The child node pattern forces a unit address and `reg` property, but the intended child device's binding strictly prohibits `reg`, creating a validation Catch-22.
--

commit 64487b3a088a3cb8b25dc58df4266bfb284945b3
Author: Joey Lu <a0987203069@gmail.com>

dt-bindings: reset: nuvoton,ma35d1-reset: add simple-mfd and child node support

This commit updates the MA35D1 reset binding to allow simple-mfd and child
nodes. It adds #address-cells, #size-cells, and a pattern property to support
instantiating children like the USB PHY.

> diff --git a/Documentation/devicetree/bindings/reset/nuvoton,ma35d1-reset.yaml b/Documentation/devicetree/bindings/reset/nuvoton,ma35d1-reset.yaml
> index 3ce7dcecd87ae..1fda7e8f4b5d1 100644
> --- a/Documentation/devicetree/bindings/reset/nuvoton,ma35d1-reset.yaml
> +++ b/Documentation/devicetree/bindings/reset/nuvoton,ma35d1-reset.yaml
[ ... ]
> @@ -26,6 +28,16 @@ properties:
>    '#reset-cells':
>      const: 1
>  
> +  '#address-cells':
> +    const: 1
> +
> +  '#size-cells':
> +    const: 1
> +
> +patternProperties:
> +  "^.*@[0-9a-f]+$":
> +    type: object

[Severity: Medium]
This pattern property forces child nodes to have a unit address, which
inherently requires a reg property to avoid a dtc warning ("node has a unit 
name, but no reg or ranges property").

However, looking at the child PHY binding 
(Documentation/devicetree/bindings/phy/nuvoton,ma35d1-usb2-phy.yaml), it 
strictly prohibits reg via additionalProperties: false.

If we provide the reg property required by the parent's unit address, it will
fail the PHY schema validation.

Additionally, the PHY driver doesn't seem to act as a true MFD child. Looking
at drivers/phy/nuvoton/phy-ma35d1-usb2.c:ma35_usb_phy_probe():

    p_phy->sysreg = syscon_regmap_lookup_by_phandle(pdev->dev.of_node, "nuvoton,sys");

The driver hardcodes its offset and uses a phandle instead of expecting a reg 
property.

Could this create a validation failure for the child PHY node, and are the
bindings and the driver structurally out of sync here?

> +
>  required:
>    - compatible
>    - reg

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260625023958.569299-1-a0987203069@gmail.com?part=1

-- 
linux-phy mailing list
linux-phy@lists.infradead.org
https://lists.infradead.org/mailman/listinfo/linux-phy