[PATCH] dt-bindings: firmware: arm,scmi: Restrict protocol child node properties

[Cristian Marussi]

On Tue, Jan 24, 2023 at 04:20:23PM -0600, Rob Herring wrote:
> The SCMI protocol child nodes are missing any constraints on unknown
> properties. Specifically, either 'unevaluatedProperties' or
> 'additionalProperties' is needed. The current structure with a regex
> match for all child nodes doesn't work for this purpose, so let's move
> the common properties '$defs' entry which each specific protocol node
> can reference and set 'unevaluatedProperties: false'.
> 

Hi Rob,

> Signed-off-by: Rob Herring <robh at kernel.org>
> ---
>  .../bindings/firmware/arm,scmi.yaml           | 43 ++++++++++++++-----
>  1 file changed, 33 insertions(+), 10 deletions(-)
> 
> diff --git a/Documentation/devicetree/bindings/firmware/arm,scmi.yaml b/Documentation/devicetree/bindings/firmware/arm,scmi.yaml
> index 176796931a22..2f7c51c75e85 100644
> --- a/Documentation/devicetree/bindings/firmware/arm,scmi.yaml
> +++ b/Documentation/devicetree/bindings/firmware/arm,scmi.yaml
> @@ -100,7 +100,9 @@ properties:
>        Channel specifier required when using OP-TEE transport.
>  
>    protocol at 11:
> -    type: object
> +    $ref: '#/$defs/protocol-node'
> +    unevaluatedProperties: false
> +
>      properties:
>        reg:
>          const: 0x11
> @@ -112,7 +114,9 @@ properties:
>        - '#power-domain-cells'
>  
>    protocol at 13:
> -    type: object
> +    $ref: '#/$defs/protocol-node'
> +    unevaluatedProperties: false
> +
>      properties:
>        reg:
>          const: 0x13
> @@ -124,7 +128,9 @@ properties:
>        - '#clock-cells'
>  
>    protocol at 14:
> -    type: object
> +    $ref: '#/$defs/protocol-node'
> +    unevaluatedProperties: false
> +
>      properties:
>        reg:
>          const: 0x14
> @@ -136,7 +142,9 @@ properties:
>        - '#clock-cells'
>  
>    protocol at 15:
> -    type: object
> +    $ref: '#/$defs/protocol-node'
> +    unevaluatedProperties: false
> +
>      properties:
>        reg:
>          const: 0x15
> @@ -148,7 +156,9 @@ properties:
>        - '#thermal-sensor-cells'
>  
>    protocol at 16:
> -    type: object
> +    $ref: '#/$defs/protocol-node'
> +    unevaluatedProperties: false
> +
>      properties:
>        reg:
>          const: 0x16
> @@ -160,20 +170,31 @@ properties:
>        - '#reset-cells'
>  
>    protocol at 17:
> -    type: object
> +    $ref: '#/$defs/protocol-node'
> +    unevaluatedProperties: false
> +
>      properties:
>        reg:
>          const: 0x17
>  
>        regulators:
>          type: object
> +        additionalProperties: false
>          description:
>            The list of all regulators provided by this SCMI controller.
>  
> +        properties:
> +          '#address-cells':
> +            const: 1
> +
> +          '#size-cells':
> +            const: 0
> +
>          patternProperties:
> -          '^regulators@[0-9a-f]+$':
> +          '^regulator@[0-9a-f]+$':
>              type: object
>              $ref: "../regulator/regulator.yaml#"
> +            unevaluatedProperties: false
>  
>              properties:
>                reg:
> @@ -184,15 +205,17 @@ properties:
>                - reg
>  
>    protocol at 18:
> -    type: object
> +    $ref: '#/$defs/protocol-node'
> +    unevaluatedProperties: false
> +
>      properties:
>        reg:
>          const: 0x18
>  
>  additionalProperties: false
>  
> -patternProperties:
> -  '^protocol@[0-9a-f]+$':
> +$defs:
> +  protocol-node:
>      type: object
>      description:

so now that the catch-all protocol@ patternProperty is gone in favour
of the 'protocol-node' definition and $refs, does that mean that any
current and future SCMI officially published protocol <N> has to be
added to the above explicit protocol list, even though it does not
have any special additional required property beside reg ?
(like protocol at 18 above...)

As an example SystemPower protocol at 12 is not listed above too and it
has nothing more than a reg=0x12 prop (liek 0x18), but before this patch
was 'covered' by the patternProperty (so Krzysztof shot down, rightly,
my recent attempt to add a distinct protocol at 12 def), but now it does not
seem anymore the case...so will we need to add an explicit protocol node
for any future protocol addition ? (SCMI is extensible up to 255
protos..)

Thanks,
Cristian