Re: [PATCH v4 1/4] dt-bindings: arm-smmu: Document smmu-500 binding for SM6125

From: Marijn Suijten
Date: Thu Dec 22 2022 - 05:10:53 EST


On 2022-12-22 10:29:40, Krzysztof Kozlowski wrote:
> On 22/12/2022 09:23, Marijn Suijten wrote:
> > On 2022-12-20 10:52:49, Krzysztof Kozlowski wrote:
> >> On 19/12/2022 20:28, Marijn Suijten wrote:
> >>> On 2022-12-19 10:09:03, Krzysztof Kozlowski wrote:
> >>>> On 19/12/2022 10:07, Krzysztof Kozlowski wrote:
> >>>>> On 16/12/2022 22:58, Marijn Suijten wrote:
> >>>>>> From: Martin Botka <martin.botka@xxxxxxxxxxxxxx>
> >>>>>>
> >>>>>> Document smmu-500 compatibility with the SM6125 SoC.
> >>>>>>
> >>>>>
> >>>>> Acked-by: Krzysztof Kozlowski <krzysztof.kozlowski@xxxxxxxxxx>
> >>>>
> >>>> Wait, not entirely... no constraints for clocks and regs?
> >>>
> >>> Quite odd that there is no warning for my DT patch as it clearly
> >>> requires at least one clock...
> >
> > Again, any idea why there's no warning for this DT mismatching minItems:
> > 1 for clocks, clock-names and power-domains?
>
> I don't know what do you have in DT and what is mismatched. Why there
> should be a warning?

There is:

clock-names:
minItems: 1
maxItems: 7

clocks:
minItems: 1
maxItems: 7

But I did not provide _any_ (see patch 2 of this series). Shouldn't
that trigger a warning?

> >>> Irrespective of that downstream doesn't define any (nor power domains).
> >>> How should we proceed?
> >>
> >> Binding now has constraints for clocks so at least that should be added
> >> to your variant.
> >
> > And that should be:
> >
> > clock-names: false
> > clocks: false
> > power-domains: false
> >
> > Because this board does declare have any, at least not when going off of
> > downstream DT?
>
> I'll add it for existing platforms, so you can rebase on top.

Thanks, will do!

- Marijn