Re: [PATCH 1/2] dt-bindings: regulator: qcom,usb-vbus-regulator: add qcom,pm4125-vbus-reg

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 2 potential issue(s) to consider:
- [High] Removing the fallback compatible string for `qcom,pm4125-vbus-reg` breaks backwards compatibility and schema validation for existing device trees.
- [Low] Unit-address mismatch in the newly added pm4125 device tree example causes a dtc compiler warning/error.
--

commit d88778938ce0889e7c6904b04eca901125a815cb
Author: Rakesh Kota <rakesh.kota@oss.qualcomm.com>

dt-bindings: regulator: qcom,usb-vbus-regulator: add qcom,pm4125-vbus-reg

This commit moves qcom,pm4125-vbus-reg from the fallback items list into
a standalone enum since it has a different register layout. It also updates
the required properties for this variant to use microvolt instead of
microamp, and adds a new example.

> diff --git a/Documentation/devicetree/bindings/regulator/qcom,usb-vbus-regulator.yaml b/Documentation/devicetree/bindings/regulator/qcom,usb-vbus-regulator.yaml
> --- a/Documentation/devicetree/bindings/regulator/qcom,usb-vbus-regulator.yaml
> +++ b/Documentation/devicetree/bindings/regulator/qcom,usb-vbus-regulator.yaml

[ ... ]

>  properties:
>    compatible:
>      oneOf:
>        - enum:
>            - qcom,pm8150b-vbus-reg
> +          - qcom,pm4125-vbus-reg
>        - items:
>            - enum:
> -              - qcom,pm4125-vbus-reg
>                - qcom,pm6150-vbus-reg
>                - qcom,pm7250b-vbus-reg

[Severity: High]
Does removing this fallback compatible string break backwards compatibility
for existing device trees?

Previously, the binding allowed the compatible string list to be
qcom,pm4125-vbus-reg, qcom,pm8150b-vbus-reg, which is currently used in
arch/arm64/boot/dts/qcom/pm4125.dtsi.

Additionally, the required properties for qcom,pm4125-vbus-reg are changed
from regulator-min/max-microamp to regulator-min/max-microvolt. Since
existing device trees like qrb2210-rb1.dts rely on the microamp properties,
will this cause make dtbs_check to fail for existing in-tree device trees?

[ ... ]

> +  - |
> +     pmic {
> +        #address-cells = <1>;
> +        #size-cells = <0>;
> +
> +        usb-vbus-regulator@5000 {
> +            compatible = "qcom,pm4125-vbus-reg";
> +            reg = <0x1100>;

[Severity: Low]
Will this unit-address mismatch cause a compiler warning during validation?

The node name specifies an address of 5000, but the reg property specifies
0x1100. This usually triggers a unit_address_vs_reg warning from the dtc
compiler during make dt_binding_check.

Should the node name be updated to match the reg property address?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260520-add_pm4125-vbus-reg-v1-0-f6d15d7dbbe8@oss.qualcomm.com?part=1