Re: [PATCH v5 1/9] dt-bindings: soc: qcom: eud: Add per-path child nodes for UTMI routing

On 8/7/2026 6:52 AM, Rob Herring wrote:
> On Fri, Aug 07, 2026 at 09:20:26AM +0200, Krzysztof Kozlowski wrote:
>> On Sun, Aug 02, 2026 at 08:15:26PM -0700, Elson Serrao wrote:
>>> The Qualcomm Embedded USB Debugger (EUD) can intercept one or two
>>> High-Speed UTMI paths depending on the SoC. For example, SC7280 supports
>>> EUD on the primary UTMI path only, while SM8350 supports EUD on both
>>> primary and secondary UTMI paths.
>>>
>>> Each UTMI path connects an independent USB controller/connector pair and
>>> forms its own role-switch domain. EUD routes traffic based on USB role.
>>> In device role the debug hub is inserted, while in host role it is
>>> bypassed.
>>>
>>> The existing binding models EUD using a flattened ports representation.
>>> While sufficient for fixed device-role configurations, it cannot naturally
>>> associate a controller, connector, and role switch with a specific
>>> EUD-routable path. This is problematic for role-switch capable ports and
>>> becomes more pronounced on SoCs with multiple paths. In such cases, each
>>> path forms an independent role-switch domain and requires per-path role
>>> awareness.
>>>
>>> Model each UTMI path as a separate child node with its own OF graph. Add
>>> an SM8350 compatible for dual-path EUD topologies while retaining the
>>> legacy ports representation for backward compatibility.
>>>
>>> Signed-off-by: Elson Serrao <[email protected]>
>>> ---
>>>  .../bindings/soc/qcom/qcom,eud.yaml           | 128 +++++++++++++++++-
>>>  1 file changed, 127 insertions(+), 1 deletion(-)
>>
>> Reviewed-by: Krzysztof Kozlowski <[email protected]>
> 
> sashiko correctly points out the oneOf logic is flawed.
> 
Thanks Rob and Krzysztof for the review.
 
I'll address this in v6 by making the two representations in oneOF section
mutually exclusive.

Thanks
Elson