On 10/03/2024 14:46, Niklas Söderlund wrote:
> Hi Krzysztof,
>
> Thanks for your comments.
>
> On 2024-03-10 09:44:45 +0100, Krzysztof Kozlowski wrote:
>> On 09/03/2024 16:53, Niklas Söderlund wrote:
>>> The Renesas Ethernet AVB bindings do not allow the MDIO bus to be
>>> described. This has not been needed as only a single PHY is
>>> supported and no MDIO bus properties have been needed.
>>>
>>> Add an optional mdio node to the binding which allows the MDIO bus to be
>>> described and allow bus properties to be set.
>>>
>>> Signed-off-by: Niklas Söderlund <niklas.soderlund+renesas@xxxxxxxxxxxx>
>>> ---
>>
>> I believe this is v2. Mark your patchsets clearly (git format-patch -v2
>> or use b4) and provide changelog under --- or in the cover letter.
>>
>>
>>> Documentation/devicetree/bindings/net/renesas,etheravb.yaml | 4 ++++
>>> 1 file changed, 4 insertions(+)
>>>
>>> diff --git a/Documentation/devicetree/bindings/net/renesas,etheravb.yaml b/Documentation/devicetree/bindings/net/renesas,etheravb.yaml
>>> index de7ba7f345a9..5345ad8e1be4 100644
>>> --- a/Documentation/devicetree/bindings/net/renesas,etheravb.yaml
>>> +++ b/Documentation/devicetree/bindings/net/renesas,etheravb.yaml
>>> @@ -93,6 +93,10 @@ properties:
>>> description: Number of size cells on the MDIO bus.
>>> const: 0
>>>
>>> + mdio:
>>> + $ref: /schemas/net/mdio.yaml#
>>> + unevaluatedProperties: false
>>> +
>>
>> Please fixup the phy pattern, so it will be obvious it is for
>> ethernet-phy and probably deprecate it. The phy goes to mdio bus, right?
>
> Yes the PHY goes on the MDIO bus and the pattern is only needed for
> backward compatibility.
>
> The pattern was specific to ethernet-phy in the past, but Rob removed it
> in commit ac8fe40c3628 ("dt-bindings: net: renesas: Drop ethernet-phy
> node schema"). Have something changed and I should revert that as part
> of this patch?
Ah, indeed. Let it stay as is. I thought there would be conflict with
mdio, but pattern still looks for unit address, so it's fine to have
both: existing @[0-9a-f] and mdio.
>
> I agree it should be listed as deprecated, would this diff work for you?
>
> +# In older bindings there where no mdio child-node to describe the MDIO bus
> +# and the PHY. To not fail older bindings accept any node with an address. New
> +# users should describe the PHY inside the mdio child-node.
> patternProperties:
> "@[0-9a-f]$":
> type: object
> + deprecated: true
Looks fine.
>
> Depending on if Rob's patch should be reverted in whole or in part I
> could also try to revert the pattern to "^ethernet-phy@[0-9a-f]$" if you
> wish. Please let me know what looks best to you.
>
Best regards,
Krzysztof
