Re: [PATCH v1 4/7] dt-bindings: crypto: fsl-dcp: Add i.MX6SL, i.MX6SLL, and i.MX6ULL support

[Conor Dooley]

On Thu, Jun 01, 2023 at 12:14:48PM +0200, Oleksij Rempel wrote:
> Add support for i.MX6SL, i.MX6SLL, and i.MX6ULL to the 'fsl-dcp' binding
> to address the following dtbs_check warning:
>   imx6ull-jozacp.dtb: crypto@2280000: compatible:0: 'fsl,imx6ull-dcp' is
>     not one of ['fsl,imx23-dcp', 'fsl,imx28-dcp']
>   From schema: Documentation/devicetree/bindings/crypto/fsl-dcp.yaml
> 
>   imx6ull-jozacp.dtb: crypto@2280000: compatible: ['fsl,imx6ull-dcp',
>     'fsl,imx28-dcp'] is too long
>   From schema: Documentation/devicetree/bindings/crypto/fsl-dcp.yaml
> 
> Signed-off-by: Oleksij Rempel <o.rempel@xxxxxxxxxxxxxx>
> ---
>  .../devicetree/bindings/crypto/fsl-dcp.yaml   | 19 ++++++++++++++++---
>  1 file changed, 16 insertions(+), 3 deletions(-)
> 
> diff --git a/Documentation/devicetree/bindings/crypto/fsl-dcp.yaml b/Documentation/devicetree/bindings/crypto/fsl-dcp.yaml
> index 99be01539fcd..8af393b9f3ca 100644
> --- a/Documentation/devicetree/bindings/crypto/fsl-dcp.yaml
> +++ b/Documentation/devicetree/bindings/crypto/fsl-dcp.yaml
> @@ -11,9 +11,22 @@ maintainers:
>  
>  properties:
>    compatible:
> -    enum:
> -      - fsl,imx23-dcp
> -      - fsl,imx28-dcp
> +    oneOf:
> +      - const: fsl,imx23-dcp
> +      - const: fsl,imx28-dcp
> +      - const: fsl,imx6sl-dcp
> +      - const: fsl,imx6sll-dcp
> +      - const: fsl,imx6ull-dcp

Confused again here chief, why allow these to appear on their own if
their are all compatible with the imx28-dcp?

> +      - items:
> +          - enum:
> +              - fsl,imx6sl-dcp
> +              - fsl,imx6sll-dcp
> +              - fsl,imx6ull-dcp
> +          - const: fsl,imx28-dcp
> +      - items:
> +          - enum:
> +              - fsl,imx28-dcp
> +          - const: fsl,imx23-dcp

I don't get this either. Why set the imx23-dcp as the fallback for the
imx28-dcp, when the imx28-dcp is being used as the fallback for the imx6
stuff?

I get the impression that some of the devicetrees should be fixed up,
rather than adding these sorta odd conditions to the bindings.

To me it'd make sense to swap everything that uses imx28-dcp as a
fallback to use imx23-dcp instead, since that is the most generic one?

What am I missing?

Cheers,
Conor.

>  
>    reg:
>      maxItems: 1
> -- 
> 2.39.2
>

Attachment: signature.asc
Description: PGP signature