Re: [PATCH net-next v8 4/9] dt-bindings: net: convert hisi-femac.txt to YAML

From: Krzysztof Kozlowski
Date: Thu Mar 07 2024 - 05:55:00 EST


On 07/03/2024 11:33, Yang Xiwen wrote:
> On 3/7/2024 3:48 PM, Krzysztof Kozlowski wrote:
>> On 06/03/2024 10:28, Yang Xiwen wrote:
>>> On 3/6/2024 3:39 PM, Krzysztof Kozlowski wrote:
>>>> On 05/03/2024 08:51, Yang Xiwen via B4 Relay wrote:
>>>>> From: Yang Xiwen <forbidden405@xxxxxxxxxxx>
>>>>>
>>>>> Convert the old text binding to new YAML.
>>>>>
>>>>> While at it, make some changes to the binding:
>>>>> - The version numbers are not documented publicly. The version also does
>>>>> not change programming interface. Remove it until it's really needed.
>>>>> - A few clocks are missing in old binding file. Add them to match the real
>>>>> hardware.
>>>>>
>>>>> Signed-off-by: Yang Xiwen <forbidden405@xxxxxxxxxxx>
>>>>> ---
>>>>> .../bindings/net/hisilicon,hisi-femac.yaml | 89 ++++++++++++++++++++++
>>>>> .../devicetree/bindings/net/hisilicon-femac.txt | 41 ----------
>>>>> 2 files changed, 89 insertions(+), 41 deletions(-)
>>>>>
>>>>> diff --git a/Documentation/devicetree/bindings/net/hisilicon,hisi-femac.yaml b/Documentation/devicetree/bindings/net/hisilicon,hisi-femac.yaml
>>>>> new file mode 100644
>>>>> index 000000000000..ba207f2c9ae4
>>>>> --- /dev/null
>>>>> +++ b/Documentation/devicetree/bindings/net/hisilicon,hisi-femac.yaml
>>>>> @@ -0,0 +1,89 @@
>>>>> +# SPDX-License-Identifier: GPL-2.0-only OR BSD-2-Clause
>>>>> +%YAML 1.2
>>>>> +---
>>>>> +$id: http://devicetree.org/schemas/net/hisilicon,hisi-femac.yaml#
>>>>> +$schema: http://devicetree.org/meta-schemas/core.yaml#
>>>>> +
>>>>> +title: Hisilicon Fast Ethernet MAC controller
>>>>> +
>>>>> +maintainers:
>>>>> + - Yang Xiwen <forbidden405@xxxxxxxxxxx>
>>>>> +
>>>>> +allOf:
>>>>> + - $ref: ethernet-controller.yaml
>>>>> +
>>>>> +properties:
>>>>> + compatible:
>>>>> + items:
>>>>> + - enum:
>>>>> + - hisilicon,hi3516cv300-femac
>>>>> + - const: hisilicon,hisi-femac
>>>> Drop this fallback, your later driver change does not use it, so neither
>>>> should have binding. Explain in commit msg, that old binding was
>>>> incorrect (we discussed it a lot) thus you are making such change during
>>>> conversion.
>>>
>>> What about deprecating "hisilicon,hisi-femac-vn" compatibles and
>>> introduce a new generic compatible "hisilicon,hisi-femac" instead? This
>>> way, We can keep backward compatibility.
>> What backward compatibility? Didn't you say bindings and driver are
>> broken and you are going to break ABI to fix them up?
>
>
> Yes, indeed the old binding is broken. I'm just wondering why you say
> "hisilicon,hisi-femac" should be removed. The FEMAC core on both SoCs
> are compatible afaik.

You don't use that fallback. What is the point of having it? You express
compatibility via SoC compatibles.


Best regards,
Krzysztof