Re: [EXTERNAL] Re: [PATCH v3 3/3] arm64: dts: marvell: cp11x: Add reset controller node

[Krzysztof Kozlowski <krzk@kernel.org>]

On 06/03/2025 18:42, Wilson Ding wrote:
> 
> 
>> -----Original Message-----
>> From: Krzysztof Kozlowski <krzk@kernel.org>
>> Sent: Wednesday, March 5, 2025 11:29 PM
>> To: Wilson Ding <dingwei@marvell.com>; linux-kernel@vger.kernel.org;
>> devicetree@vger.kernel.org; linux-arm-kernel@lists.infradead.org;
>> robh@kernel.org
>> Cc: andrew@lunn.ch; gregory.clement@bootlin.com;
>> sebastian.hesselbarth@gmail.com; krzk+dt@kernel.org; conor+dt@kernel.org;
>> p.zabel@pengutronix.de; Sanghoon Lee <salee@marvell.com>;
>> Geethasowjanya Akula <gakula@marvell.com>
>> Subject: [EXTERNAL] Re: [PATCH v3 3/3] arm64: dts: marvell: cp11x: Add reset
>> controller node
>>
>> On 04/03/2025 20:08, Wilson Ding wrote:
>>>
>>> I did consider shrinking the syscon's register address range to make
>>> the reset-controller node to be independent from the syscon node.
>>> However, I found the syscon node is also referred by some devices for
>>> miscellaneous configurations . The reset configuration register
>>> happens to be located in between these registers and clock/GPIO
>>> registers.
>>>
>>>> drop offset in your patch or unify everything into 'reg'.
>>>>
>>>
>>> This is exactly what I proposed in v3 patch. Do I misunderstand you?
>>>
>>> CP11X_LABEL(swrst): reset-controller@268 {
>>> 	compatible = "marvell,armada8k-reset";
>>> 	reg = <0x268 0x4>;
>>> 	#reset-cells = <1>;
>>> };
>>
>> I don't see the other device being fixed here. How did you unify them?
> 
> This patch series is about the proposal of Armada8K's reset controller
> dt-binding. The dt-bindings issues of clock/GPIO controllers have been
> there for years. Having to say, it is not just a simple patch to fix it. It


I understand, you just want to throw your patch here over the wall. It's
reasonable, I feel it. Just like previous cases for this binding -
everyone wanted one subnode at a time, ignoring bigger picture, each
time making it franken-node or franken-binding.

Please listen to Greg's talk from years ago about upstreaming. "I don't
want your code":
https://www.youtube.com/watch?v=fMeH7wqOwXA


Best regards,
Krzysztof