Re: [PATCH V1 1/3] dt-bindings: ufs: qcom: Add reg and reg-names

[Krzysztof Kozlowski <krzk@kernel.org>]

On 30/07/2025 12:27, Ram Kumar Dwivedi wrote:
> 
> 
> On 30-Jul-25 2:41 PM, Krzysztof Kozlowski wrote:
>> On 30/07/2025 10:22, Ram Kumar Dwivedi wrote:
>>> Update the Qualcomm UFS device tree bindings to support Multi-Circular
>>> Queue (MCQ) operation. This includes increasing the maximum number of
>>> register entries from 2 to 3 and extending the accepted values for
>>> reg-names to include "mcq_sqd" and "mcq_vs".
>>>
>>> These changes are required to enable MCQ support via Device Tree for
>>> platforms such as SM8650 and SM8750.
>>>
>>> Signed-off-by: Ram Kumar Dwivedi <quic_rdwivedi@quicinc.com>
>>> ---
>>>  .../devicetree/bindings/ufs/qcom,ufs.yaml     | 21 ++++++++++++-------
>>>  1 file changed, 13 insertions(+), 8 deletions(-)
>>>
>>> diff --git a/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml b/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml
>>> index 6c6043d9809e..de263118b552 100644
>>> --- a/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml
>>> +++ b/Documentation/devicetree/bindings/ufs/qcom,ufs.yaml
>>> @@ -86,12 +86,17 @@ properties:
>>>  
>>>    reg:
>>>      minItems: 1
>>> -    maxItems: 2
>>> +    maxItems: 3
>>>  
>>>    reg-names:
>>> -    items:
>>> -      - const: std
>>> -      - const: ice
>>> +    oneOf:
>>> +      - items:
>>> +          - const: std
>>> +          - const: ice
>>> +      - items:
>>> +          - const: ufs_mem
>>> +          - const: mcq_sqd
>>> +          - const: mcq_vs
>>
>> This is incompatible change and commit msg is inaccurate here. It says
>> "extending" but you are not extending at all.
>>
>> Recent qcom patches love to break ABI and impact users. No.
>>
>> Best regards,
>> Krzysztof
> 
> 
> Hi Krzysztof,
> 
> Thanks for your feedback.
> 
> Regarding your concern about this being an incompatible change — could you please clarify what specific aspect you believe breaks compatibility? 
> From my side, I’ve carefully tested the patch and verified that it does not break any existing DTs. I ran the following command to validate against the schema:


I missed that earlier list is not actually used for SM8550 and SM8650.
The syntax is a bit confusing after looking only at diff, which probably
means this binding is getting messy.

I think binding should be just split the constraints are easier to follow.

Best regards,
Krzysztof