Re: [PATCH v4 2/3] dt-bindings: display: Fix BCM2835 HVS bindings for BCM2712

[Krzysztof Kozlowski <krzk@kernel.org>]

On 19/12/2024 12:54, Dave Stevenson wrote:
> Hi Krzysztof
> 
> On Thu, 19 Dec 2024 at 08:42, Krzysztof Kozlowski <krzk@kernel.org> wrote:
>>
>> On Wed, Dec 18, 2024 at 02:48:33PM +0000, Dave Stevenson wrote:
>>> Commit 6cfcbe548a3a ("dt-bindings: display: Add BCM2712 HVS bindings")
>>> added the compatible string for BCM2712, but missed out that
>>> the number of interrupts and clocks changed too, and both need to be
>>> named.
>>>
>>> Update to validate clock, interrupts, and their names for the variants.
>>>
>>> Fixes: 6cfcbe548a3a ("dt-bindings: display: Add BCM2712 HVS bindings")
>>> Signed-off-by: Dave Stevenson <dave.stevenson@raspberrypi.com>
>>> ---
>>>  .../bindings/display/brcm,bcm2835-hvs.yaml         | 84 ++++++++++++++++++----
>>>  1 file changed, 70 insertions(+), 14 deletions(-)
>>>
>>> diff --git a/Documentation/devicetree/bindings/display/brcm,bcm2835-hvs.yaml b/Documentation/devicetree/bindings/display/brcm,bcm2835-hvs.yaml
>>> index f91c9dce2a44..fd25ee5ce301 100644
>>> --- a/Documentation/devicetree/bindings/display/brcm,bcm2835-hvs.yaml
>>> +++ b/Documentation/devicetree/bindings/display/brcm,bcm2835-hvs.yaml
>>> @@ -20,11 +20,20 @@ properties:
>>>      maxItems: 1
>>>
>>>    interrupts:
>>> -    maxItems: 1
>>> +    minItems: 1
>>> +    maxItems: 3
>>> +
>>> +  interrupt-names:
>>> +    minItems: 1
>>> +    maxItems: 3
>>>
>>>    clocks:
>>> -    maxItems: 1
>>> -    description: Core Clock
>>> +    minItems: 1
>>> +    maxItems: 2
>>> +
>>> +  clock-names:
>>> +    minItems: 1
>>> +    maxItems: 2
>>>
>>>  required:
>>>    - compatible
>>> @@ -33,17 +42,64 @@ required:
>>>
>>>  additionalProperties: false
>>>
>>> -if:
>>> -  properties:
>>> -    compatible:
>>> -      contains:
>>> -        enum:
>>> -          - brcm,bcm2711-hvs
>>> -          - brcm,bcm2712-hvs
>>> -
>>> -then:
>>> -  required:
>>> -    - clocks
>>> +allOf:
>>> +  - if:
>>> +      properties:
>>> +        compatible:
>>> +          contains:
>>> +            const: brcm,bcm2711-hvs
>>> +
>>> +    then:
>>> +      properties:
>>> +        clocks:
>>> +          items:
>>> +            - description: Core Clock
>>> +        interrupts:
>>> +          maxItems: 1
>>
>>
>> clock-names and interrupt-names: false, unless driver needs them but all
>> this should be explained in the commit msg because it would be a change
>> to the binding.
> 
> False it is then.
> 
> Is there actually a full guide to binding requirements?
> https://www.kernel.org/doc/html/latest/devicetree/bindings/writing-schema.html
> is the closest I've found, but it doesn't obviously cover these types
> of things.
> 
>>> +
>>> +      required:
>>> +        - clocks
>>> +
>>> +  - if:
>>> +      properties:
>>> +        compatible:
>>> +          contains:
>>> +            const: brcm,bcm2712-hvs
>>> +
>>> +    then:
>>> +      properties:
>>> +        clocks:
>>> +          minItems: 2
>>> +          maxItems: 2
>>> +        clock-names:
>>> +          items:
>>> +            - const: core
>>> +            - const: disp
>>> +        interrupts:
>>> +          items:
>>> +            - description: Channel 0 End of frame
>>> +            - description: Channel 1 End of frame
>>> +            - description: Channel 2 End of frame
>>> +        interrupt-names:
>>> +          items:
>>> +            - const: ch0-eof
>>> +            - const: ch1-eof
>>> +            - const: ch2-eof
>>> +      required:
>>> +        - clocks
>>> +        - clock-names
>>> +        - interrupt-names
>>
>> My previous comment still stands. Reply to people's feedback instead of
>> ignoring it.
> 
> Your previous comment was
> "Why requiring last two names? Commit msg does not explain that."
> 
> I didn't ignore it. The driver needs them, and the commit msg states

Uh, so someone added undocumented ABI. How this passed any checks? Or no
one cared to run validation?


> " but missed out that the number of interrupts and clocks changed too,
> and *both need to be
> named*.

Ah, I really did not get this.

> Update to validate clock, interrupts, and *their names* for the variants."

Please say explicitly that since some commit driver needs this. There
should be a clear reason in the commit msg why you are adding this ABI.


Best regards,
Krzysztof