Re: [PATCH v2 2/2] arm64: dts: meson-t7-a311d2-khadas-vim4: add initial device-tree

[Neil Armstrong <neil.armstrong@linaro.org>]

On 21/06/2023 10:01, Conor Dooley wrote:
> On Wed, Jun 21, 2023 at 08:37:02AM +0100, Lucas Tanure wrote:
>> On Wed, Jun 21, 2023 at 7:02 AM Krzysztof Kozlowski
>> <krzysztof.kozlowski@linaro.org> wrote:
>>>
>>> On 21/06/2023 00:09, Yixun Lan wrote:
>>>>> +            apb4: bus@fe000000 {
>>>>> +                    compatible = "simple-bus";
>>>>> +                    reg = <0x0 0xfe000000 0x0 0x480000>;
>>>>> +                    #address-cells = <2>;
>>>>> +                    #size-cells = <2>;
>>>>> +                    ranges = <0x0 0x0 0x0 0xfe000000 0x0 0x480000>;
>>>>> +
>>>>> +                    uart_A: serial@78000 {
>>>>> +                            compatible = "amlogic,meson-t7-uart",
>>>>                                                ~~~~~~~~~~~~~~~~~
>>>> if you introduce new compatible string, then at least you need to document it
>>>> so Documentation/devicetree/bindings/serial/amlogic,meson-uart.yaml need to be updated
>>>>
>>>> but my qeustion here, why bother introducing new compatible string if nothing
>>>> changed with the compatible data? given the uart is same IP with g12a, can't we just
>>>> use "amlogic,meson-g12-uart" for this? no only it will reduce the structure length of
>>>> meson_uart_dt_match[], but also relieve maintainer's review burden?
>>>
>>> https://elixir.bootlin.com/linux/v6.1-rc1/source/Documentation/devicetree/bindings/writing-bindings.rst#L42
>>>
>>> Best regards,
>>> Krzysztof
>>>
>> Hi, I did not understand the recommendation here.
>> Can I add "amlogic,meson-t7-uart" without Documentation changes?
> 
> No, you can't.
> 
>> I think Yes, as I can see a few compatible strings in dts that don't
>> exist anywhere else.
> 
> Aye, but we do not want to propagate that. New stuff should not be
> adding undocumented compatibles, and those that are undocumented should
> be documented.

Documentation is ongoing, it takes time !

Neil

> 
>> My idea here is to add "amlogic,meson-t7-uart" for future use if ever
>> created, like if we find a bug in the future that is only relevant to
>> T7 soc.
>> But for now, fallback to s4 uart, as it seems to be the same controller.
>>
>> >From Krzysztof said in the writing-bindings.rst, I am following the rules.
>>
>> So, what's the path forward here?
> 
> You are following the rules from the dts point of view, you just need a
> 3rd patch in which you document the pattern you have added here in
> amlogic,meson-uart.yaml. It is probably something like:
> +      - items:
> +          - const: amlogic,meson-t7-uart
> +          - const: amlogic,meson-s4-uart
> But I have not tested that, I just wrote that in my mail client.
> 
> Cheers,
> Conor.