[PATCH v8 04/18] dt-bindings: pinctrl: add Ambarella CV75 pinctrl

[zl020895]

Hi Krzysztof, Linus,

Sorry for the mess. 

Would this work?: I send a standalone
"dt-bindings: vendor-prefixes: add Ambarella" patch to Rob with only
the two lines in vendor-prefixes.yaml, and drop that hunk from the
SoC series. Once the prefix is in linux-next, the vendor-prefix
warnings should be fixed. The remaining example matches, such as
ambarella,drive-strength-syscon, still need the separate
GPIO/UART/pinctrl bindings.

Best regards,
Long Zhao

At 2026-10-13 17:22:56, "Krzysztof Kozlowski" <krzk at kernel.org> wrote:
>On 13/10/2026 10:44, Linus Walleij wrote:
>> On Mon, Oct 12, 2026 at 9:08 AM Krzysztof Kozlowski <krzk at kernel.org> wrote:
>>> On 12/10/2026 09:06, Linus Walleij wrote:
>>>> On Fri, Oct 9, 2026 at 8:17 AM Krzysztof Kozlowski <krzk at kernel.org> wrote:
>>>>> On 01/10/2026 09:03, Linus Walleij wrote:
>>>>
>>>>>> On Mon, Sep 21, 2026 at 6:23 AM Long Zhao via B4 Relay
>>>>>> <devnull+longzhao.ambarella.com at kernel.org> wrote:
>>>>>>
>>>>>>> From: Long Zhao <longzhao at ambarella.com>
>>>>>>>
>>>>>>> Document the CV75 pin controller binding using generic function and
>>>>>>> groups properties for pinmux configuration.
>>>>>>>
>>>>>>> Reviewed-by: Linus Walleij <linusw at kernel.org>
>>>>>>> Reviewed-by: Krzysztof Kozlowski <krzysztof.kozlowski at oss.qualcomm.com>
>>>>>>> Signed-off-by: Long Zhao <longzhao at ambarella.com>
>>>>>>
>>>>>> This patch 4/18 applied to the pinctrl tree.
>>>>>
>>>>> There is a dependency, you cannot take this patch alone.
>>>>
>>>> Oops, but is it a DTS dependency so if none of the other patches go in through
>>>
>>> No, it is a bindings dependency, introducing new DT bindings warning for
>>> both your tree and next.
>> 
>> Is there a problem with the bindings patch itself?
>> 
>> In that case Long can just send an incremental patch to
>> fix it right?
>> 
>>> Your tree is fine in this case as long as next
>>> is also fine... but next is not, because that patch did not make it there.
>> 
>> But the patch is in -next?
>> 
>> ~/linux-next$ ls
>> Documentation/devicetree/bindings/pinctrl/ambarella,cv75-pinctrl.yaml
>> Documentation/devicetree/bindings/pinctrl/ambarella,cv75-pinctrl.yaml
>> 
>>> We cannot have any new DT bindings warnings in next.
>> 
>> I guess I just don't get it :D
>> 
>> To me it seems like a fairly normal situation: ACKed bindings merged
>> together with the driver, no dependence on anything.
>
>But there is a dependency. Your commit uses undocumented vendor prefix
>which introduces warnings.
>
>New warnings:
>Documentation/devicetree/bindings/pinctrl/ambarella,cv75-pinctrl.example.dtb:
>pinctrl at e4010000 (ambarella,cv75-pinctrl):
>'ambarella,drive-strength-syscon', 'ambarella,pull-syscon' do not match
>any of the regexes: '^#.*',
>https://gitlab.com/robherring/linux-dt/-/jobs/17038310909
>
>
>
>BTW, this commit of yours:
>9c2932badcd3e3c0fc630eac5ff2bfd5fadf93f1
>
>does not even build, so a make error.
>
>> 
>> But if you tell me "revert the patch" I will simply revert the patch
>> because I trust you, it doesn't really matter why.
>
>I think Ambarella won't be picked up, because of total mess with
>patchsets, thus Rob should just take the vendor prefix... except there
>is no vendor prefix patch to take, since it is squashed with other work.
>
>I don't get why this became so messy and have no clue how this should be
>fixed, except revert which is sad churn-type of work.
>
>Best regards,
>Krzysztof