Re: [PATCH v2 3/3] dt-bindings: hwinfo: Add mtk-socinfo driver

[AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>]

Il 18/09/23 13:59, Krzysztof Kozlowski ha scritto:
> On 18/09/2023 10:47, AngeloGioacchino Del Regno wrote:
>> Il 17/09/23 10:26, Krzysztof Kozlowski ha scritto:
>>> On 15/09/2023 17:26, William-tw Lin wrote:
>>>> dt-binding documentation for mtk-socinfo driver.
>>>
>>> Here and in subject, drop driver and instead descrbe hardware.
>>>
>>>> mtk-socinfo driver provides SoC-related information.
>>>> Such information includes manufacturer information, SoC name,
>>>> SoC segment name, and SoC marketing name.
>>>>
>>>> Signed-off-by: William-tw Lin <william-tw.lin@mediatek.com>
>>>> ---
>>>>    .../bindings/hwinfo/mtk-socinfo.yaml          | 48 +++++++++++++++++++
>>>>    1 file changed, 48 insertions(+)
>>>>    create mode 100644 Documentation/devicetree/bindings/hwinfo/mtk-socinfo.yaml
>>>>
>>>> diff --git a/Documentation/devicetree/bindings/hwinfo/mtk-socinfo.yaml b/Documentation/devicetree/bindings/hwinfo/mtk-socinfo.yaml
>>>> new file mode 100644
>>>> index 000000000000..74f03f1dc404
>>>> --- /dev/null
>>>> +++ b/Documentation/devicetree/bindings/hwinfo/mtk-socinfo.yaml
>>>
>>> Nothing improved.
>>>
>>> This is a friendly reminder during the review process.
>>>
>>> It seems my previous comments were not fully addressed. Maybe my
>>> feedback got lost between the quotes, maybe you just forgot to apply it.
>>> Please go back to the previous discussion and either implement all
>>> requested changes or keep discussing them.
>>>
>>> Thank you.
>>>
>>>> @@ -0,0 +1,48 @@
>>>> +# SPDX-License-Identifier: (GPL-2.0-only OR BSD-2-Clause)
>>>> +%YAML 1.2
>>>> +---
>>>> +$id: http://devicetree.org/schemas/hwinfo/mtk-socinfo.yaml#
>>>> +$schema: http://devicetree.org/meta-schemas/core.yaml#
>>>> +
>>>> +title: MediaTek SoC ChipID
>>>> +
>>>> +maintainers:
>>>> +  - William Lin <william-tw.lin@mediatek.com>
>>>> +  - Matthias Brugger <matthias.bgg@gmail.com>
>>>> +  - Kevin Hilman <khilman@kernel.org>
>>>> +  - AngeloGioacchino Del Regno <angelogioacchino.delregno@collabora.com>
>>>> +
>>>> +description:
>>>> +  MediaTek SoCs store various product information in eFuses, including
>>>> +  Chip ID and Revision fields, usable to identify the manufacturer,
>>>> +  SoC version, plus segment and marketing names.
>>>> +
>>>> +properties:
>>>> +  compatible:
>>>> +    const: mediatek,socinfo
>>>
>>> What happened to compatibles? No, this is just wrong and no explained.
>>> You ignored other comments as well. Really, that's not the way to go.
>>>
>>
>> Practically, having different compatibles for each SoC is not needed, as
>> the only thing that changes between SoCs is the eFuse(s) that you read to
>> get the information - and that's all.
> 
> 
> And how do you guarantee that no future SoC will have any difference?
> How can you even predict it?
> 

You're right, it's too much long sighted and at some point in the future this
assumption will be *inevitably* wrong.

Thanks,
Angelo

>>
>> So ... we either use this driver with devicetree, giving it the right eFuses
> 
> I am talking about bindings, no driver.
> 
>> to read from, or we duplicate the mtk-efuse driver, or we statically assign
>> the eFuses in the driver itself and we set compatibles like
>> "mediatek,mt8195-socinfo" to select that... ideas?
> 
> Device specific compatibles followed by fallback, just like for every
> other review coming from me.
>  > Best regards,
> Krzysztof
>