Re: [RFC PATCH v2 1/6] dt-bindings: pinctrl: starfive: Add JH8100 pinctrl

[Krzysztof Kozlowski <krzysztof.kozlowski@xxxxxxxxxx>]

On 23/02/2024 01:24, Rob Herring wrote:
> On Wed, Feb 21, 2024 at 08:24:26AM +0100, Krzysztof Kozlowski wrote:
>> On 20/02/2024 20:10, Conor Dooley wrote:
>>> On Tue, Feb 20, 2024 at 09:11:43AM +0100, Krzysztof Kozlowski wrote:
>>>> On 20/02/2024 07:42, Alex Soo wrote:
>>>>> Add documentation and header file for JH8100 pinctrl driver.
>>>>>
>>>>> Signed-off-by: Alex Soo <yuklin.soo@xxxxxxxxxxxxxxxx>
>>>>> ---
>>>>
>>>>
>>>> RFC? Why isn't this patch ready for review?
>>>
>>> The TL;DR is that Emil and I didn't want to apply the dts patches to
>>> support a platform that hadn't actually been taped out yet. 
>>> For an SoC in that state, at least the bindings for, clock and pinctrl
>>> could be subject to changes before tapeou. I think putting RFC on those
>>> patches is a good idea, but of course the rationale should be mentioned.
>>
>> That would be useful information. We also could mark some bindings
>> unstable and accept breaking ABI under certain conditions, like that it
>> is early work without users for long time.
> 
> The challenge with that is when do things get marked stable? No one has 
> any motivation to do that (unless users complain). For example, We have 
> a couple of platforms that have an unstable bindings statement that has 
> been there "forever".

I see. Let's see what I can do for existing "unstable" platforms, but
your argument makes sense - rarely people remember to un-unstable
bindings and there aren't that many incentives for maintainer to do so.

> 
> I would like a solution though. The only idea I have is passing 
> SystemReady cert, but that's an Arm thing.


Best regards,
Krzysztof