Re: [PATCH 12/12] arm64: defconfig: Enable I3C and SPD5118 hwmon

[Krzysztof Kozlowski]

On 25/03/2026 13:58, Thierry Reding wrote:
> On Wed, Mar 25, 2026 at 12:03:37PM +0100, Krzysztof Kozlowski wrote:
>> On 25/03/2026 11:59, Krzysztof Kozlowski wrote:
>>> On 25/03/2026 11:31, Thierry Reding wrote:
>>>> On Thu, Mar 19, 2026 at 06:15:14PM +0100, Krzysztof Kozlowski wrote:
>>>>> On 19/03/2026 18:09, Akhil R wrote:
>>>>>> On Thu, 19 Mar 2026 10:40:34 +0100, Krzysztof Kozlowski wrote:
>>>>>>> On Wed, Mar 18, 2026 at 10:57:25PM +0530, Akhil R wrote:
>>>>>>>> Add I3C subsystem support, DesignWare I3C master controller, and
>>>>>>>> SPD5118 hwmon sensor as modules to the defconfig.
>>>>>>>
>>>>>>> Why? If there is no user of that, why would we want it? Your commit msg
>>>>>>> should explain that.
>>>>>>
>>>>>> Ack. This is for Tegra410 which has a DesignWare I3C host controller.
>>>>>> I will add this in the commit message.
>>>>>
>>>>> Board or products. Not SoCs.
>>>>
>>>> Is this a new requirement? I see a bit of both in defconfig changes.
>>>
>>> Almost every review from me has it for 2-3 years... And it is a known
>>
>> And I already explained this to *you* 3 years ago:
>>
>> https://lore.kernel.org/all/ac8f30a7-fc72-9a44-74b3-a69001bfdaaf@xxxxxxxxxx/
>>
>> So how this could be a new requirement *now* if three years ago we had
>> exactly same discussion.
>>
>> I understand question for the first time, but why this being brought up
>> as "why is this a new thing" again?
> 
> I have to admit I did not remember what we discussed, so I had to go
> read that exchange again. It sounds to me like we were not discussing
> the specific issue of a missing description as to which particular
> product needed this, but you were instead rejecting the idea of
> enabling drivers that were not strictly necessary like those for PCI
> devices because they were making your life more difficult by building
> drivers by default that you were not interested in.
> 
> Here you're arguing that you want proof that this is going to be used
> by some upstream-supported device, which are two different things,
> because they might very well be drivers that you're not interested in
> but end up building if documented properly.
> 
> So I find it a little hard to keep track of what is acceptable to you
> and what isn't. Are you objecting to this on the grounds of it bloating
> the kernel build or because you want documentation for what platforms a
> driver is being used on?
> 
> Our action items will be different depending on what your answer is: if
> you want documentation about what device this will be used for, we'll
> get you that information. If your concern is that it bloats the build we
> drop the patch and will have to ask users to build their own
> configurations.
> 
> Maybe to avoid these kinds of discussions in the past you can write down
> your rules about what should go into defconfig and what should not. And
> maybe we can eventually find consensus and find something that people
> can use as a reference.

I think answer is pretty simple and comes from the reason WHAT is the
purpose of defconfig. It's purpose is only for us.

Therefore defconfig can have anything anyone will find useful, when
building and running vanilla upstream kernel on their devices, with
exception of explicit needs for pluggable devices because then it bloats
the kernel to impossible stage (otherwise look for me sending all USB,
PCI, MEDIA, whatever devices for defconfig...).

For example:
1. Something used only by a DT board not enabled upstream: no, because
you cannot run upstream kernel on it,
2. Something used only by a ACPI platform, which require some out of
tree patches to build: no, because you cannot run upstream kernel on it
3. Something not being part of the device but pluggable: depends,
explain why upstream contributors would want it.

Look at the commit here:

"Add I3C subsystem support, DesignWare I3C master controller, and
SPD5118 hwmon sensor as modules to the defconfig."

Does it say why doing it? No.

Nowhere in this posting I was objecting to actual change. At least not
yet. I only ask WHY you are doing it.

Why is it so hard for a contributor to know and express why they are
doing something? If one does not know WHY they are doing it, then why
the heck they are doing it?

Best regards,
Krzysztof