On 24/12/2025 04:16, Jingyi Wang wrote:
> 
> 
> On 12/23/2025 9:29 PM, Krzysztof Kozlowski wrote:
>> On Tue, Dec 23, 2025 at 01:13:49AM -0800, Jingyi Wang wrote:
>>> diff --git 
>>> a/Documentation/devicetree/bindings/remoteproc/qcom,pas-common.yaml 
>>> b/Documentation/devicetree/bindings/remoteproc/qcom,pas-common.yaml
>>> index 63a82e7a8bf8..149e993282bb 100644
>>> --- a/Documentation/devicetree/bindings/remoteproc/qcom,pas-common.yaml
>>> +++ b/Documentation/devicetree/bindings/remoteproc/qcom,pas-common.yaml
>>> @@ -33,16 +33,22 @@ properties:
>>>        - description: Handover interrupt
>>>        - description: Stop acknowledge interrupt
>>>        - description: Shutdown acknowledge interrupt
>>> +      - description: Pong interrupt
>>> +      - description: Wake acknowledge interrupt
>>>  
>>>    interrupt-names:
>>>      minItems: 5
>>> +    maxItems: 7
>>>      items:
>>> -      - const: wdog
>>> -      - const: fatal
>>> -      - const: ready
>>> -      - const: handover
>>> -      - const: stop-ack
>>> -      - const: shutdown-ack
>>> +      enum:
>>
>> No, no. Stop doing random changes. NAK
>>
>> Now you remove strict order (see writing bindings) and claim every
>> device like SM8550 ADSP PAS has any order.
>>
>> And it is now de-synced with interrupts. Read writing bindings - this is
>> clearly described there at line 90!
>>
>> You can only grow existing list.
>>
>>
>>
> 
> the interrupt for soccp is defined as 
> "wdog","fatal","ready","handover","stop-ack","pong","wake-ack"
> while other pas could be:
> "wdog","fatal","ready","handover","stop-ack","shutdown-ack"
> 
> so grow existing list is not work for this,
> 
> In the v1, got your comments to adjust pas-common.yaml for the interrupt:
> https://lore.kernel.org/all/[email protected]/
> 
> and in v2, interrupt are moved as part to the "allOf" with if-else and
> also got NAK:
> https://lore.kernel.org/all/20251030-venomous-apricot-falcon-b3fd64@kuoka/
> 
> Could you please share a example for us to understand how to maintain it in
> pas-common.yaml, not define if-else and has strict order at the same time?
> That will be very helpful.
> 

Can you read the docs first? This is not a ping pong, where you keep
sending till it passes the review. If I provide you idea, then you will
send something whatever based on that because you still did not read the
rules governing bindings.

Best regards,
Krzysztof

Reply via email to