Re: [PATCH v3] dt-bindings: gpio: add common consumer GPIO lines

[Krzysztof Kozlowski]

On 01/04/2022 17:01, Rob Herring wrote:
> On Fri, Apr 1, 2022 at 8:27 AM Krzysztof Kozlowski
> <krzysztof.kozlowski@xxxxxxxxxx> wrote:
>>
>> On 01/04/2022 15:13, Rob Herring wrote:
>>> On Fri, 01 Apr 2022 09:27:14 +0200, Krzysztof Kozlowski wrote:
>>>> Typical GPIO lines like enable, powerdown, reset or wakeup are not
>>>> documented as common, which leads to new variations of these (e.g.
>>>> pwdn-gpios).  Add a common schema which serves also as a documentation
>>>> for preferred naming.
>>>>
>>>> Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@xxxxxxxxxx>
>>>>
>>>> ---
>>>>
>>>> Changes since v2:
>>>> 1. Correct email.
>>>>
>>>> Changes since v1:
>>>> 1. Select-true, add maxItems and description for each entry (Rob).
>>>> 2. Mention ACTIVE_LOW in bindings description (Linus).
>>>> 3. Add allOf for pwrseq reset-gpios case.
>>>>
>>>> Signed-off-by: Krzysztof Kozlowski <krzysztof.kozlowski@xxxxxxxxxx>
>>>> ---
>>>>  .../bindings/gpio/gpio-consumer-common.yaml   | 64 +++++++++++++++++++
>>>>  1 file changed, 64 insertions(+)
>>>>  create mode 100644 Documentation/devicetree/bindings/gpio/gpio-consumer-common.yaml
>>>>
>>>
>>> My bot found errors running 'make DT_CHECKER_FLAGS=-m dt_binding_check'
>>> on your patch (DT_CHECKER_FLAGS is new in v5.13):
>>>
>>> yamllint warnings/errors:
>>>
>>> dtschema/dtc warnings/errors:
>>> /builds/robherring/linux-dt-review/Documentation/devicetree/bindings/regulator/richtek,rt4801-regulator.example.dt.yaml: rt4801@73: enable-gpios: [[4294967295, 2, 0], [4294967295, 3, 0]] is too long
>>>       From schema: /builds/robherring/linux-dt-review/Documentation/devicetree/bindings/gpio/gpio-consumer-common.yaml
>>>
>>
>> Hi Rob,
>>
>> With v1, you proposed to use maxItems for all these standard gpios, but
>> as we see here there are two exceptions:
>> 1. pwrseq might have up to 32 reset-gpios,
>> 2. richtek,rt4801 uses up to 2 enable-gpios.
> 
> There's always an outlier...
> 
>> One way is to add exceptions in gpio-consumer-common.yaml, like I did
>> for reset-gpios and pwrseq. However this scales poor if more of such
>> usages appear.
> 
> I'd reject any new cases, but even just 2 I don't really like.

The richtek,rt4801 enable-gpios are for controlling two separate
regulators, so it should have been under regulator subnodes/children.
Some other regulators follow this pattern, so only this one is done that
way.

That driver could be converted to enable-gpios per regulator, so if you
are sure about rejection of new cases, how about keeping current
exceptions in allOf:if?

> 
>> Maybe let's drop the maxItems for all of them?
> 
> Let's just drop it at least for now (though it seems we can keep it
> for powerdown-gpios).
> 
> A possible solution here may be adding 'maxItems: 1' automatically to
> schemas if not specified. I've been thinking of doing this on standard
> unit properties. That's another case of
> 99% of cases are a single entry with a few outliers.
> 


Best regards,
Krzysztof