Re: [PATCH v3 1/2] dt-bindings: soc: qcom: Add qcom,kaanapali-imem compatible

[Konrad Dybcio]

On 11/4/25 4:02 PM, Krzysztof Kozlowski wrote:
> On 04/11/2025 15:58, Konrad Dybcio wrote:
>> On 11/4/25 3:52 PM, Krzysztof Kozlowski wrote:
>>> On 04/11/2025 15:38, Konrad Dybcio wrote:
>>>> On 11/4/25 3:37 PM, Krzysztof Kozlowski wrote:
>>>>> On 04/11/2025 15:35, Konrad Dybcio wrote:
>>>>>> On 11/4/25 3:26 PM, Krzysztof Kozlowski wrote:
>>>>>>> This I got, but nothing here explains why you need generic compatible.
>>>>>>> To re-iterate: there was no generic compatible before, now there is.
>>>>>>> Writing bindings and numerous reviews from DT maintainers ask not to use
>>>>>>> generic compatibles.
>>>>>>
>>>>>> OK so let's not worry about a generic compatible. IMEM exists since
>>>>>> MSM8974 and it only had major hw updates with SM8550. They don't
>>>>>> impact the software interface though, so qcom,msm8974-imem is OK.
>>>>>>
>>>>>> There's a separate control/status register address space for each
>>>>>> instance of this IP (usually far apart from the actual SRAM pool),
>>>>>> which Linux doesn't have to care about.
>>>>>
>>>>> Just use qcom,kaanapali-imem - that's the first device here without syscons.
>>>>
>>>> So we don't want to move the existing ones over?
>>>
>>> This was never discussed and this patch did not do it. You cannot move
>>> them, that's ABI.
>>
>> I see, I implicitly assumed this would be a sweeping change.
>>
>> So should the Kaanapali submitters simply send a version of this
>> patch with:
>>
>> - oneOf:
>>   - const: qcom,kaanapali-imem
>>   - items:
>>     # existing big list
>>
>> ?
>>
>> I'm not a huge fan of using kaanapali as the fallback-going-forward
>> since it's literally the newest platform on the shelves (or perhaps
>> not even on the shelves yet..) so it's going to look funny when
>> someone comes up with support for another 2013 soc.. but perhaps
>> that's just how things are supposed to be
> 
> 
> Yes. Feel free to choose other fully compatible device as the fallback,
> like you mentioned in previous email, I proposed Kaanapali as the easiest.

Ehhh it's not super convenient given the available list

I see that msm8994 isn't described yet. If we don't need to care about
the pre-/post 8550 split (which we would only for the aforementioned control
register space which is NOT what this binding describes), let's go with that
as the fallback.

Konrad