Re: [GIT PULL 1/7] dt-bindings: Changes for v7.1-rc1

[Krzysztof Kozlowski <krzk@kernel.org>]

On 31/03/2026 13:09, Thierry Reding wrote:
> On Tue, Mar 31, 2026 at 11:50:18AM +0200, Krzysztof Kozlowski wrote:
>> On 31/03/2026 10:25, Thierry Reding wrote:
>>>
>>>>>       dt-bindings: phy: tegra-xusb: Document Type C support
>>>>
>>>> No acks, but that is waiting for one month so it is fine.
>>>
>>> It's got a Reviewed-by from Rob and there were no corresponding driver
>>> changes associated with it. There's literally no reason for this to go
>>> in through a subsystem tree.
>>
>> There are reasons - process and technical:
>> 1. It's a subsystem maintainer's patch, not yours. You rather should
>> have reason to take someone else's patches.
>>
>> 2. Conflict in that file.
>>
>>
>>>
>>>>>       dt-bindings: clock: tegra124-dfll: Convert to json-schema
>>>>>       dt-bindings: interrupt-controller: tegra: Fix reg entries
>>>>>       dt-bindings: arm: tegra: Add missing compatible strings
>>>>>       dt-bindings: phy: tegra: Document Tegra210 USB PHY
>>>>>       dt-bindings: memory: Add Tegra210 memory controller bindings
>>>>>       dt-bindings: memory: tegra210: Mark EMC as cooling device
>>>>
>>>> That's even my subsystem and I did not ack it. You did not even sent it
>>>> to me as requested by MAINTAINERS file (+dt is ignore alias), so
>>>> obviously I did not even had a chance to ack it.
>>>
>>> Ugh... really? I was Cc'ed to you as a DT maintainer as well as the
>>> devicetree mailing list, so I'm sure you've seen it. This had also been
>>
>> Really, you are supposed to use get_maintainers.pl. Not invent the CC
>> list, not come with own filtered list.
>>
>> I understand that my non dt address disappears when running
>> get_maintainers.pl on entire set.
>>
>> But if you sent that patch targetting subsystem instead of combining 6
>> or 7 subsystems at once, it would pop up.
>>
>> And yes, you are not supposed to mix 7 different subsystems in single
>> patchset. That's basic!
> 
> Pft... you're being unreasonable. These are all relatively trivial
> changes that have at one point or another been already reviewed. Do you
> seriously expect me to turn these into 7 different patch series so that
> everyone gets to spend even more time dealing with these?

That's what everyone else is doing, including Qualcomm which is
upstreaming multiple new SoCs per year, multiple patchsets, multiple
patches.

But no, I do not expect you to split it but I expect to use
get_maintainers.pl, unless that would cause a patchbomb cc-ing 50
addresses (I don't suggest that you created such patchbomb -
hypothetical). Removing maintainers from Cc list is not the way to solve
patchbomb problem, if such arises. Splitting a hypothetical patchbomb
would be the solution.

> 
>> But sure, let's skip memory controllers patch cc list and:
>>
>>> reviewed by Rob a long time ago, and honestly, it's also quite trivial.
>>> It's been on the list for a month and there were no objections, so it
>>> does pass all of the criteria you mentioned before.
>>
>> No objections because you did not cc people. How can you claim "no
>> objections from person foo" if you do not cc person "foo", because you
>> do not use get_maintainers.pl?
>>
>> Look, find me here phy mantainers:
>>
>> https://lore.kernel.org/all/20260223143305.3771383-7-thierry.reding@kernel.org/
> 
> It's an extremely trivial, one-line change to an existing binding. There
> is no interaction with the driver whatsoever, why would anyone outside
> of the DT and Tegra maintainers even want to look at this?

I explained in the beginning why someone could want to look:

1. It's a subsystem maintainer's patch, not yours.

2. Conflict in that file.

But regardless why someone would or would not want, get_maintainers.pl
asked you to Cc them.

> 
>>>> And we even had few days ago talk were I explained you how these
>>>> bindings must go. Seeing pull request completely ignoring that
>>>> discussion is just huge surprise.
>>>>
>>>> No, it cannot go in. Send patches to proper maintainers first.
>>>
>>> Stop making these baseless accusations, Krzysztof. You were on Cc and
>>
>> Not using get_maintainers.pl so not ccing right address is not baseless.
>> This is the fact.
> 
> I use get_maintainers.pl but filter down based on common sense.

Filtering means you don't use. Following this logic (although not
exactly "common sense" but writing bindings point you that common sense
for driver bindings is driver subsystem):
1. I use get_maintainers.pl on my patch.
2. Then I remove all addresses except linux-kernel@vger.kernel.org.
3. All good and I claim I am using get_maintainers.pl.


> 
>> Neither phy, nor interrupts, nor clocks were sent to right people.
> 
> Again, these are patches that were all on the list previously, sent to
> the right people at the time. They hadn't been picked up by now, so I

No, they were not sent to the right people at the time. The right people
at the time are pointed by get_maintainers.pl which you filtered out.

Look again, this is the v1:
https://lore.kernel.org/all/20260223143305.3771383-7-thierry.reding@kernel.org/

Where is any phy maintainer or the phy mailing list Cc-ed?

Instead of again using "baseless" arguments, please provide the actual
fact for that claim that you cc-ed right list and right people on that
patch.

> resent them, to the DT maintainers only, to give you guys a chance to

This is not a resent, this is the v1. Do you see "RESEND" in the
subject? I don't.

I am pointing out the facts now. If you disagree, please point me to the
original posting of phy patch with the 'right' people in CC list, as
requested by get_maintainers.pl.


Best regards,
Krzysztof