Re: [PATCH 2/2] dt-bindings: arm: cpus: Add compatible qcom,oryon-1-5

[Krzysztof Kozlowski <krzk@kernel.org>]

On 13/04/2026 16:34, Shawn Guo wrote:
> On Mon, Apr 13, 2026 at 03:30:28PM +0200, Krzysztof Kozlowski wrote:
>> On 13/04/2026 15:10, Shawn Guo wrote:
>>> On Mon, Apr 13, 2026 at 02:40:18PM +0200, Krzysztof Kozlowski wrote:
>>>>>> Please organize your patchset correctly.
>>>>>
>>>>> Are you asking for a big series that consists of all the new bindings
>>>>> used by Nord DTS and DTS itself?  Unless this big series gets applied as
>>>>> one-go, there are still chances that bindings get into a kernel release
>>>>> without any users, e.g. subsystem maintainers pick up bindgins being
>>>>> reviewed, but DTS requires more iterations and thus misses the release.
>>>>
>>>> Please follow existing rules, communicated multiple times on the mailing
>>>> list. Qualcomm also has internal guideline clarifying this.
>>>>
>>>> Below are some upstream discussion clarifying this:
>>>> https://lore.kernel.org/linux-samsung-soc/CADrjBPq_0nUYRABKpskRF_dhHu+4K=duPVZX==0pr+cjSL_caQ@mail.gmail.com/T/#m2d9130a1342ab201ab49670fa6c858ee3724c83c
>>>> https://lore.kernel.org/all/49258645-d4d8-44a5-a4fc-b403c926a5d1@kernel.org/
>>>>
>>>> And how to do it:
>>>> https://lore.kernel.org/all/20231121-topic-sm8650-upstream-dt-v3-0-db9d0507ffd3@linaro.org/
>>>
>>> That's what I'm trying to do, posting bindings in prior to DTS, so that
>>
>> Hm? Nothing above claimed that. I am sorry, but which rule either
>> communicated publicly or privately encouraged or even allowed that?
>>
>>> when posting DTS, either bindings is already merged or we can refer to
>>> lore link of bindings.
>>>
>>> I still need to understand you comment "Bindings come with the user".
>>> Are you saying that bindings and DTS in different series should be posted
>>> at the same time to show bindings has an user?
>>
>> I am saying that you cannot post bindings alone where there is no user.
>> Why do want even such binding? I see no point in having it in the
>> kernel, unless you speak about very specific exception, but then please
>> clarify that exception and why it applies to this work.
> 
> I think I'm getting what the requirement is now, but still not sure what
> the point of the requirement is.
> 
> - Posting bindings and DTS at the same time doesn't guarantee they get
>   into the same release.  It can often happen that bindings are merged

No one claimed it will guarantee that and no one even suggested that.
Submitting patches even describes this...

>   alone into a release.  So it doesn't really help to avoid situation of
>   there is bindings in-tree with no users.

That was not the comment from me, either. When I say there is no user, I
mean no user at all. Nowhere in ecosystem considered usptream, including
mailing list.

> 
> - From what I can see, there are always DTS patches coming after bindings
>   changes, sooner or later.  No one would be bothered to submit a bindings
>   if the DTS using the bindings is only kept out of tree.

I would be bothered. I don't want to maintain unused (unused as in
upstream) ABI. I don't care about such ABI for downstream users and no
one in upstream should spend any precious cycles on reviewing something
which serves no upstream purposes.

> 
> In short, there will be Nord DTS using the binding coming, and I do not

Maybe there will, maybe there will not.

> think posting them at the same time should be a requirement.

Well, it is a requirement as I explained previously, said that
*multiple* times on the mailing list, documented expectations in
mentioned/linked email threads. It's also documented in submitting
patches in DT (although not with that strong wording).

Best regards,
Krzysztof