Re: [PATCH 8/9] dt-bindings: phy: sparx5: document lan969x in sparx5 dt-bindings

[Krzysztof Kozlowski <krzk@kernel.org>]

On 09/09/2024 11:43, Daniel Machon wrote:
>>>>>    compatible:
>>>>> -    const: microchip,sparx5-serdes
>>>>> +    enum:
>>>>> +      - microchip,sparx5-serdes
>>>>> +      - microchip,lan969x-serdes
>>>>
>>>> It seems there is no lan969x SoC/chip. Are you sure you are using
>>>> correct naming, matching what kernel is using? Maybe you just sent
>>>> whatever you had in downstream (hint: that's never a good idea).
>>>
>>> You are right. There is no upstream support for lan969x SoC yet. The
>>> upstreaming of the lan969x SoC has just begun, and this series is part
>>> of that upstreaming effort. The lan969x switch driver (not submitted
>>> yet) will depend on this SERDES driver, however, their functionality is
>>> really independent of each other. That is why I am also upstreaming the
>>> SERDES- and switch driver series independent of each other.
>>
>> That's not exactly my point. Becayse lan969x appears. I claim you use
>> incorrect name, so are you sure you do not use wildcards?
>> Best regards,
>> Krzysztof
> 
> Ahh.
> 
> So the problem is the 'x' in lan969x, right? I think we have a habbit of
> documenting compatible strings like this in bindings. Anyway, what I can
> do is document the different part numbers in the bindings: lan9691,
> lan9692, lan9693, lan9694, lan9696 and lan9698.

Depends. I remember I was confused about such wildcards before and for
some cases wildcard was ok. For some not. I don't remember, I don't know
which case is here.

Best regards,
Krzysztof