Re: [PATCH v24 4/4] dt-binding: clock: remove nuvoton npcm845-clk bindings

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

On 22/05/2024 18:34, Tomer Maimon wrote:
> Hi Rob,
> 
> Kind reminder about the question in the mail thread below.
> Your response would be greatly appreciated.
> 
> Thanks,
> 
> Tomer
> 
> On Thu, 16 May 2024 at 12:44, Tomer Maimon <tmaimon77@gmail.com> wrote:
>>
>> Hi Rob,
>>
>> Thanks for your comment.
>>
>> On Mon, 13 May 2024 at 18:53, Rob Herring <robh@kernel.org> wrote:
>>>
>>> On Thu, May 09, 2024 at 10:24:11PM +0300, Tomer Maimon wrote:
>>>> Remove nuvoton,npcm845-clk binding since the NPCM8xx clock driver
>>>> using the auxiliary device framework and not the device tree framework.
>>>
>>> Again, this is an ABI break. Changing driver architecture for 1 OS is
>>> not a reason to change DT.
>> Is it an ABI break even if the NPCM8xx clock driver hasn't upstream
>> the kernel vanilla yet?
>>
>> I thought that since the NPCM8xx clock driver hasn't upstream the
>> kernel vanilla yet and and in the latest NPCM8xx clock driver patch
>> the NPCM8xx clock driver.
>> using auxiliary device framework instead of DT we should remove the
>> nuvoton,npcm845-clk.yaml file.
>> https://patchwork.kernel.org/project/linux-clk/patch/20240509192411.2432066-4-tmaimon77@gmail.com/
>>

Binding goes with the driver, so I wonder how did it happen that driver
is not in mainline but binding is?

Anyway, once binding is released other users might start to use it, so
yeah, could still be ABI break.


Best regards,
Krzysztof