Re: [PATCH 07/18] MIPS: DTS: jz4780: fix otg node as reported by dtbscheck

From: "H. Nikolaus Schaller" <hns@goldelico.com>
To: Zhou Yanjie <zhouyanjie@wanyeetech.com>
Cc: Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>,
	Rob Herring <robh+dt@kernel.org>,
	Paul Cercueil <paul@crapouillou.net>,
	Thomas Bogendoerfer <tsbogend@alpha.franken.de>,
	linux-kernel@vger.kernel.org, devicetree@vger.kernel.org,
	linux-mips@vger.kernel.org, letux-kernel@openphoenux.org
Subject: Re: [PATCH 07/18] MIPS: DTS: jz4780: fix otg node as reported by dtbscheck
Date: Sun, 10 Apr 2022 21:13:52 +0200	[thread overview]
Message-ID: <8FA36CE4-CD39-4767-B49E-D112E4313036@goldelico.com> (raw)
In-Reply-To: <b6090a27-12b7-0c01-6d33-ae32b31b4b74@wanyeetech.com>

Hi,

> Am 10.04.2022 um 18:32 schrieb Zhou Yanjie <zhouyanjie@wanyeetech.com>:
> 
> Hi folks,
> 
> On 2022/4/9 下午9:53, H. Nikolaus Schaller wrote:
>> 
>>> Am 09.04.2022 um 15:44 schrieb Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>:
>>> 
>>> On 09/04/2022 15:32, H. Nikolaus Schaller wrote:
>>>> 
>>>>> Am 09.04.2022 um 15:15 schrieb Krzysztof Kozlowski <krzysztof.kozlowski@linaro.org>:
>>>>> 
>>>>> On 09/04/2022 15:05, H. Nikolaus Schaller wrote:
>>>>>>> This looks wrong, the block usually should have a specific compatible.
>>>>>>> Please mention why it does not.
>>>>>> Well, I did not even have that idea that it could need an explanation.
>>>>>> 
>>>>>> There is no "ingenic,jz4780-otg" and none is needed here to make it work.
>>>>> Make it work in what terms? We talk about hardware description, right?
>>>> Yes.
>>>> 
>>>>>> Therefore the generic "snps,dwc2" is sufficient.
>>>>> No, you are mixing now driver behavior (is sufficient) with hardware
>>>>> description.
>>>> No. "snps,dwc2" is a hardware description for a licensed block.
>>>> Not a driver behavior.
>>> snps,dwc2 matches the original block, not necessarily this
>>> implementation. Unless you are sure?
>> I assume. Nobody has reported an issue without having any specific jz4780 driver in place.
>> Well, that is only evidence, not bullet proof.
>> 
>>>>> Most of licensed blocks require the specific compatible to
>>>>> differentiate it.
>>>> If there is a need to differentiate.
>>> No, regardless whether there is a need currently, most of them have
>>> specific compatibles, because there are some minor differences. Even if
>>> difference is not visible from programming model or wiring, it might
>>> justify it's own specific compatible. For example because maybe once
>>> that tiny difference will require some changes.
>>> 
>>> Someone added the ingenic compatible, so why do you assume that one tool
>>> (bindings) is correct but other piece of code (using specific
>>> compatible) is not? You use the argument "bindings warning" which is not
>>> enough. Argument that blocks are 100% same, is good enough, if you are
>>> sure. Just use it in commit msg. But are you sure that these are the
>>> same? Same pins, same programming model (entire model, not used by Linux)?
>> The compatible ingenic,jz4780-otg was introduced in 158c774d3c64859e84dd20e04d5fb18c8d3d318e.
>> Hence I have added Yanjie for clarification why he added it in the .dts and not in the bindings.
> 
> 
> It's my fault, last year I made an OTG driver for Ingenic SoCs and sent it
> to the mailing list, and then I received some revision comments, but for
> some personal reasons I didn't continue to improve it.
> 
> I'll finish these modifications as soon as possible and send them out.
> Then after they merge into the mainline, this problem will be solved.

No need to apologize.

If you agree I can add "ingenic,jz4780-otg" to the schema file and keep
the .dts in the v2 of my series.

And I'll add you to the list of reviewers, so you can please comment v2
if it is correct or if we are still missing something.

Best regards and thanks,
Nikolaus