Re: [PATCH 01/10] dt: bindings: clock: add mtmips SoCs clock device tree binding documentation

[Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>]

On 21/03/2023 09:33, Arınç ÜNAL wrote:
> On 21.03.2023 11:27, Krzysztof Kozlowski wrote:
>> On 21/03/2023 09:24, Arınç ÜNAL wrote:
>>>>>
>>>>> If we take the calling new things mediatek route, we will never get to
>>>>> the bottom of fixing the naming inconsistency.
>>>>
>>>> All new things, so new SoCs, should be called mediatek, because there is
>>>> no ralink and mediatek is already used for them. So why some new
>>>> Mediatek SoCs are "mediatek" but some other also new SoCs are "ralink"?
>>>>
>>>> You can do nothing (and no actual need) about existing inconsistency...
>>>
>>> I couldn't change ralink -> mediatek because company acquisitions don't
>>> grant the change. I don't see any reason to prevent changing mediatek ->
>>> ralink without breaking the ABI on the existing schemas.
>>
>> You cannot change mediatek->ralink without breaking the ABI for the same
>> reasons.
> 
> Then this is where I ask for an exception.
> 
> The current solution only complicates things more.
> 
> https://github.com/paraka/linux/pull/1/files#diff-0ae6c456898d08536ce987c32f23f2eb6f4a0f7c38bff9a61bdf3d0daa3f6549R21

Sorry, I don't understand what's under this link and how some Github
repo pull helps in this discussion. I don't see there any text, which
could help.

I also do not understand why this pull proves that you can change
existing mediatek compatibles (we talk also about ARM, which is shipped
to million of devices) to ralink without breaking the ABI.

I do not see how choosing one variant for compatibles having two
variants of prefixes, complicates things. Following this argument
choosing "ralink" also complicates!



Best regards,
Krzysztof