Re: [RFC PATCH] dt-bindings: power: Add power-limit-controller schema

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 3 potential issue(s) to consider:
- [Medium] Node name regexes for domains and limits incorrectly restrict unit addresses to decimal digits, breaking on standard hex addresses.
- [Medium] Missing `#address-cells` and `#size-cells` properties for nodes with `reg`-bearing children, causing dtc warnings and schema validation failures.
- [Medium] Use of non-standard `-microsecond` suffix prevents automatic schema typing.
--

--- Patch [1]: [RFC PATCH] dt-bindings: power: Add power-limit-controller schema ---
commit 04e63fa5b3b701b0306a74eba5521673337890b4
Author: Manaf Meethalavalappu Pallikunhi <manaf.pallikunhi@oss.qualcomm.com>

dt-bindings: power: Add power-limit-controller schema
    
This RFC proposes a new device tree binding schema for power limit
controllers that manage SoC power domains with hardware-enforced power
capping capabilities.

> diff --git a/Documentation/devicetree/bindings/power/limits/power-limit-controller.yaml b/Documentation/devicetree/bindings/power/limits/power-limit-controller.yaml
> new file mode 100644
> index 0000000000000..9cd4d9d6414d9
> --- /dev/null
> +++ b/Documentation/devicetree/bindings/power/limits/power-limit-controller.yaml
> @@ -0,0 +1,238 @@
> +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
> +%YAML 1.2
> +---
> +$id: http://devicetree.org/schemas/power/limits/power-limit-controller.yaml#
> +$schema: http://devicetree.org/meta-schemas/core.yaml#

[ ... ]

> +properties:
> +  $nodename:
> +    pattern: "^power-limits(@.*)?$"
> +
> +  '#power-limit-domain-cells':
> +    description: |
> +      Number of cells in a power limit domain specifier for child domains.
> +      Typically 1, representing the domain index.
> +    const: 1

[Severity: Medium]
Will this schema cause dtc warnings for missing #address-cells and #size-cells?

Since the child nodes like power-limit-domain@0 require a reg property,
wouldn't the root node need to define #address-cells and #size-cells
properties here in the schema?

> +
> +patternProperties:
> +  "^power-limit-domain@[0-9]+$":

[Severity: Medium]
Does this regular expression restrict unit addresses to only decimal digits?

Standard device trees use hexadecimal for unit addresses, so a node like
power-limit-domain@a would fail to match this pattern. Could this be
updated to [0-9a-f]+ instead?

> +    type: object
> +    description: |
> +      Individual power limit domain/zone under this controller.
> +      Each domain can either:
> +        - Have power limit constraints (with power-limits child node)
> +        - Be monitoring-only (without power-limits child node)

[ ... ]

> +      power-limits:
> +        type: object
> +        description: |
> +          Container node for power limit constraints within this domain.
> +          Each child node represents a power limit constraint index.
> +
> +          This node is optional. If omitted (or if monitoring-only is set),
> +          the domain provides only power/energy measurement without limits.
> +
> +        patternProperties:
> +          "^power-limit@[0-9]+$":

[Severity: Medium]
Similarly, does this constraint node pattern also improperly restrict unit
addresses to decimal digits instead of hexadecimal?

> +            type: object
> +            description: |
> +              Individual power limit constraint configuration.

[ ... ]

> +              power-limit-microwatt:
> +                description: |
> +                  Default power limit value for this constraint at boot/reset.
> +                  This is the initial value that will be programmed.
> +
> +              time-window-min-microsecond:

[Severity: Medium]
Does the -microsecond suffix prevent dtschema from automatically inferring
property types?

The standard suffix recognized by the core schema for microseconds is -us.
Without the standard suffix, the schema validator cannot automatically infer
the property types (such as uint32), which could cause missing type definition
errors during dt_binding_check.

> +                description: |
> +                  Minimum time window for power averaging.
> +                  Shorter windows allow faster response to power excursions.

[ ... ]

> +            required:
> +              - reg
> +
> +            additionalProperties: true
> +
> +        additionalProperties: false

[Severity: Medium]
Will this strict additionalProperties: false prevent adding #address-cells
and #size-cells to the power-limits container?

Because the power-limit@N child nodes require reg properties, this container
node will need these cell-count properties, but they aren't explicitly defined
in this schema block.

> +
> +    required:
> +      - reg
> +
> +    additionalProperties: true
> +
> +additionalProperties: true
> +
> +examples:
> +  - |
> +    // Multi-domain power limit controller with mixed capabilities
> +    // Demonstrates multiple domains with and without power limit constraints
> +    power-limits@ef3b000 {
> +        compatible = "qcom,glymur-spel";
> +        reg = <0x0ef3b000 0x1000>;
> +        #power-limit-domain-cells = <1>;

[Severity: Medium]
Could this example fail dt_binding_check due to the missing #address-cells
and #size-cells properties in the power-limits@ef3b000 node?

> +
> +        // Domain 0: System domain with full power limit control (PL1/PL2)
> +        sys_domain: power-limit-domain@0 {
> +            reg = <0>;

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260709180727.4015267-1-manaf.pallikunhi@oss.qualcomm.com?part=1