Re: [PATCH v4 1/2] dt-bindings: interconnect: qcom: Add SM7250 RPMh NoC

From: Ishu Brar

Date: Thu Nov 12 2026 - 00:15:47 EST


On Sat, 31 Oct 2026 05:19:35 +0000, Ishu Brar wrote:
> On Wed, 28 Oct 2026 09:10:21 +0100, Krzysztof Kozlowski wrote:
>> On 28/10/2026 08:59, Krzysztof Kozlowski wrote:
>>>> + clocks:
>>>> + items:
>>>> + - description: aggre UFS PHY AXI clock
>>>> + - description: aggre USB3 PRIM AXI clock
>>>> +
>>>> +patternProperties:
>>>> + '^interconnect-[0-9]+$':
>>>
>>> We actually do not need to enforce -0: '^interconnect(-[0-9]+)?$'
>>
>> Actually is this really a pattern? How many children are there?
>
> No, not a pattern. there is at most one child per provider:
> aggre2-noc -> compute-noc
> gem-noc -> mc-virt
> system-noc -> qup-virt
>
> The other providers have no child. v5 would replace the pattern with a
> single fixed "interconnect" child node.
>
> Should the binding also allow the child only under these three
> providers (and "interconnect: false" for the rest), or is the plain
> child node enough?

Gentle ping on the question above. If the plain child node is enough, I'll send v5 with it.