Re: [PATCH v2 13/13] arm64: defconfig: Enable I3C and SPD5118 hwmon

[Akhil R <akhilrajeev@nvidia.com>]

On Mon, 13 Apr 2026 09:12:18 +0200, Krzysztof Kozlowski wrote:
> On 13/04/2026 08:57, Akhil R wrote:
>>>> Isn't I3C needed for SPD5118. Otherwise I understand even less from this
>>>> rationale - why I3C is being enabled here?
>>>>
>>>> And before author asks what do I want to here: no, it is author's job to
>>>> convince me to accept I3C in defconfig. Not mine.
>>>
>>> BTW, all this was asked at v1 and author did not improve the commit msg
>>> beside giving quite broad/unspecific "Vera".
>> 
>> If I am not wrong, the ask in v1 was to specify the product which this is
>> getting used - 'Vera' it is. I do not know why you would think it is
>> unspecific.
> 
> I already said why. Because I Googled it and Google told me it can be
> "architecture". And no, you do not use SPD5118 on architecture.

Vera is the CPU and it uses SOCAMM LPDDR5, which has an SPD5118 within.
I will write this down in the next version. 

> 
>> 
>> As Thierry and Guenter mentioned, the lack of policy and 'mix of both' in
>> the defconfig makes it quite difficult to understand what could genuinely
>> be convincing other than putting down every little detail or do a trial
>> and error.
> 
> I think the main problem is that people forgot that commits must answer
> WHY you are doing this. Now my assumption is that people sending
> defconfigs do not understand why they are doing it, therefore they
> cannot explain "why" in commit msg.
> 
> Look, find me in following nvidia patches any answers to why this change
> is needed:
> 
> https://lore.kernel.org/all/20260409131340.168556-7-pshete@nvidia.com/
> https://lore.kernel.org/all/20240829134252.49661-1-jonathanh@nvidia.com/
> https://lore.kernel.org/all/20240509215808.126217-1-bwicaksono@nvidia.com/
> 
> I understand that I can clarify what sort of answers to "why?" I expect.
> but lack of such clarification is not excuse to not provide ANY
> explanation in commit msg.
> 
> Basically it is logic like:
> "Uh, I don't how to explain this change, why do we need it, why is it
> good, why am I doing that...anyway, let's send it!"
> 
> Why doing something in the first place if one does not know the reason
> behind?

The answer I wrote to "why" is that "it is required to support the
SOCAMM in Vera". But I may have overlooked the ambiguity of the
details of Vera available online. I will make it more descriptive.

Regards,
Akhil