On 23/12/2022 11:31, Felipe Balbi wrote:
>
> Hi,
>
> Rob Herring <robh@xxxxxxxxxx> writes:
>>>> The Rockchip RK3399 DWC3 node has 'power-domain' property which isn't
>>>> allowed by the schema:
>>>>
>>>> usb@fe900000: Unevaluated properties are not allowed ('power-domains' was unexpected)
>>>>
>>>> Allow DWC3 nodes to have a single power-domains entry. We could instead
>>>> move the power-domains property to the parent wrapper node, but the
>>>> could be an ABI break (Linux shouldn't care). Also, we don't want to
>>>> encourage the pattern of wrapper nodes just to define resources such as
>>>> clocks, resets, power-domains, etc. when not necessary.
>>>>
>>>> Signed-off-by: Rob Herring <robh@xxxxxxxxxx>
>>>> ---
>>>> Documentation/devicetree/bindings/usb/snps,dwc3.yaml | 3 +++
>>>> 1 file changed, 3 insertions(+)
>>>>
>>>> diff --git a/Documentation/devicetree/bindings/usb/snps,dwc3.yaml b/Documentation/devicetree/bindings/usb/snps,dwc3.yaml
>>>> index 6d78048c4613..bcefd1c2410a 100644
>>>> --- a/Documentation/devicetree/bindings/usb/snps,dwc3.yaml
>>>> +++ b/Documentation/devicetree/bindings/usb/snps,dwc3.yaml
>>>> @@ -91,6 +91,9 @@ properties:
>>>> - usb2-phy
>>>> - usb3-phy
>>>>
>>>> + power-domains:
>>>> + maxItems: 1
>>>
>>> AFAICT this can be incorrect. Also, you could have Cc the dwc3
>>> maintainer to get comments.
>>
>> When we have a user with more and know what each one is, then we can
>> extend it. All the other users (upstream), put 'power-domains' in the
>
> Won't that be an ABI break at that point? You'll change the maximum
> number of power-domains.
Usually extending properties (in flexible way) is not an ABI break. What
would be broken here if it becomes three at some point? Does Linux or
other SW depends now on this being equal to 1?
Best regards,
Krzysztof
