Re: [PATCH 1/2] dt-bindings: hwmon: pmbus: Add Analog Devices MAX16545/MAX16550 and Volterra VT7505

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] The `adi,ocp-severe-filter` property exposes raw register bit values instead of standard physical units.
--

--- Patch [1]: [PATCH 1/2] dt-bindings: hwmon: pmbus: Add Analog Devices MAX16545/MAX16550 and Volterra VT7505 ---
commit e93633b9753ab840614d58e40b8f32b35288cabc
Author: Sanman Pradhan <psanman@juniper.net>

dt-bindings: hwmon: pmbus: Add Analog Devices MAX16545/MAX16550 and Volterra VT7505

This commit adds device tree binding documentation for the Analog Devices
MAX16545, MAX16550, and the Volterra VT7505 PMBus controllers. These
controllers provide voltage, current, power, and temperature monitoring.

> diff --git a/Documentation/devicetree/bindings/hwmon/pmbus/adi,vt7505.yaml b/Documentation/devicetree/bindings/hwmon/pmbus/adi,vt7505.yaml
> new file mode 100644
> index 0000000000000..286c88eb7be8f
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/hwmon/pmbus/adi,vt7505.yaml

[ ... ]

> +  adi,ocp-severe-filter:
> +    $ref: /schemas/types.yaml#/definitions/uint32
> +    description:
> +      Deglitch time applied to the severe overcurrent-protection comparator,
> +      programmed into the MFR_CONFIG[15:14] field. The MAX16550 and VT7505
> +      expose this field; it decodes as 0 = 0 us, 1 = 1 us, 2 = 2 us and
> +      3 = 10 us. On the MAX16545 the severe overcurrent delay is fixed and
> +      these bits are reserved.
> +    enum: [0, 1, 2, 3]

[Severity: Medium]
Should this property use physical units instead of raw register values?

Device tree properties typically describe hardware configurations abstractly
using standard units like microseconds. Exposing the raw register bit values
(0, 1, 2, 3) rather than the actual time durations (0, 1, 2, 10) creates a
non-standard ABI that leaks internal hardware encoding.

Once merged, this raw-value ABI must be supported indefinitely. Would it be
better to define the accepted values as 0, 1, 2, and 10 to match the physical
time delays?

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260722150324.164799-1-sanman.pradhan@hpe.com?part=1