Re: [PATCH 1/3] dt-bindings: mfd: twl: enable power button also for twl603x

[Krzysztof Kozlowski <krzk@kernel.org>]

On 21/10/2025 18:36, Andreas Kemnade wrote:
> On Tue, 21 Oct 2025 11:58:49 +0200
> Krzysztof Kozlowski <krzk@kernel.org> wrote:
> 
>> On 21/10/2025 10:45, Andreas Kemnade wrote:
>>> On Tue, 21 Oct 2025 09:10:28 +0200
>>> Krzysztof Kozlowski <krzk@kernel.org> wrote:
>>>   
>>>> On 20/10/2025 14:31, akemnade@kernel.org wrote:  
>>>>> From: Andreas Kemnade <andreas@kemnade.info>
>>>>>
>>>>> TWL603x has also a power button, so add the corresponding subnode.    
>>>>
>>>> No, we don't add subnodes just because there is a power button. This
>>>> needs broader explanation, see also my further comment.
>>>>  
>>> Hmm, what is the general pattern to follow if a mfd device has some
>>> functionality which depends on some optional external components?  
>>
>> Please describe it better - how these nodes depend on external
>> component? The power button logic/IC is in this device always. It is not
>> optional.
>>
> The power button logic is always there, yes, but it depends on an optional
> actual mechanical button connected to a pad of this device, which is
> not always there. The logic will not work if I just put my finger on the PMIC,
> but it will work if there is a mechanical button which I can press connected to
> the PMIC.


Hm... how do you represent this logic now? By adding status=disabled to
the pwrbutton node?


> 
>>> The might be a power button connected to it or not. I find it ugly
>>> to have non-existent stuff in the system.
>>> In general, yes I understand the argument against the subnode.
>>>   
>>>>>
>>>>> Signed-off-by: Andreas Kemnade <andreas@kemnade.info>
>>>>> ---
>>>>>  Documentation/devicetree/bindings/mfd/ti,twl.yaml | 40 ++++++++++++++++++-----
>>>>>  1 file changed, 32 insertions(+), 8 deletions(-)
>>>>>
>>>>> diff --git a/Documentation/devicetree/bindings/mfd/ti,twl.yaml b/Documentation/devicetree/bindings/mfd/ti,twl.yaml
>>>>> index 776b04e182cb2..3527fee32cb07 100644
>>>>> --- a/Documentation/devicetree/bindings/mfd/ti,twl.yaml
>>>>> +++ b/Documentation/devicetree/bindings/mfd/ti,twl.yaml
>>>>> @@ -55,6 +55,15 @@ allOf:
>>>>>  
>>>>>          gpadc: false
>>>>>  
>>>>> +        pwrbutton:
>>>>> +          properties:
>>>>> +            compatible:
>>>>> +              const: ti,twl4030-pwrbutton
>>>>> +            interrupts:
>>>>> +              items:
>>>>> +                - items:
>>>>> +                    const: 8    
>>>>
>>>> What is the point of defining const interrupts? If they are const, then
>>>> it is implied by compatible and defined in the driver.
>>>>
>>>> Anyway, double items does not look right here. This is an odd syntax.
>>>>  
>>> Quoting Rob:
>>> As 'interrupts' is a matrix, this needs to be:
>>>
>>> interrupts:
>>>   items:
>>>     - items:
>>>         - const: 8
>>>
>>> https://lore.kernel.org/linux-omap/20240318150750.GA4000895-robh@kernel.org/  
>>
>>
>> OK, this answers second part but I don't understand why even having this
>> in DT. If this is fixed, should be implied by the compatible?
>>
> correct, they do not need to come from DT. The same is true for all
> subnodes of the twl[46]03X. I just followed the usual
> pattern there, which is of course not recommended for new designs.


OK, please mention the reasons (so following established pattern for
this legacy device) in commit msg.

Best regards,
Krzysztof