Re: [PATCH 2/3] dt-bindings: i2c: renesas,rzv2m: Fix SoC specific string

[Krzysztof Kozlowski]

On 04/11/2022 04:55, Geert Uytterhoeven wrote:
> Hi Krzysztof,
> 
> On Fri, Nov 4, 2022 at 2:48 AM Krzysztof Kozlowski
> <krzysztof.kozlowski@xxxxxxxxxx> wrote:
>> On 03/11/2022 19:06, Fabrizio Castro wrote:
>>> s/renesas,i2c-r9a09g011/renesas,r9a09g011-i2c/g for consistency.
>>>
>>> renesas,i2c-r9a09g011 is not actually used by the driver, therefore
>>> changing this doesn't cause any harm.
>>
>> And what about other users of DTS? One chosen driver implementation
>> might not be enough...
>>
>>>
>>> Fixes: ba7a4d15e2c4 ("dt-bindings: i2c: Document RZ/V2M I2C controller")
>>
>> You need to explain the bug - where is the issue, how it affects users.
>> Otherwise it is not a bug and there is nothing to fix.
> 
> Yep.
> 
>     The preferred form is "<vendor>,<family>-<module>", blah blah ...
> 
>>> Signed-off-by: Fabrizio Castro <fabrizio.castro.jz@xxxxxxxxxxx>
>>> ---
>>>  Documentation/devicetree/bindings/i2c/renesas,rzv2m.yaml | 4 ++--
>>>  1 file changed, 2 insertions(+), 2 deletions(-)
>>>
>>> diff --git a/Documentation/devicetree/bindings/i2c/renesas,rzv2m.yaml b/Documentation/devicetree/bindings/i2c/renesas,rzv2m.yaml
>>> index c46378efc123..92e899905ef8 100644
>>> --- a/Documentation/devicetree/bindings/i2c/renesas,rzv2m.yaml
>>> +++ b/Documentation/devicetree/bindings/i2c/renesas,rzv2m.yaml
>>> @@ -16,7 +16,7 @@ properties:
>>>    compatible:
>>>      items:
>>>        - enum:
>>> -          - renesas,i2c-r9a09g011  # RZ/V2M
>>> +          - renesas,r9a09g011-i2c  # RZ/V2M
>>
>> No.
>>
>> Deprecate instead old compatible. There are already users of it, at
>> least in kernel. Not sure about other OS/bootloaders/firmwares.
> 
> As stated in the cover letter:
> 
>     Since it's early days for r9a09g011.dtsi, and compatible
>     renesas,i2c-r9a09g011 isn't being actively used at the moment,
>     I think it's safe to change to make compatible strings less
>     confusing.
> 
> The bindings entered in v6.0.
> The first user in the kernel is not yet in a released kernel, it will
> be in v6.1. So it can still be fixed in v6.1...
> Even if we don't fix it before v6.2, I don't think there is much harm in
> making this change.
> 

ABI break reasons should be explained in the commit because cover letter
does not get merged.

Best regards,
Krzysztof