Re: [PATCH v6 07/13] arm64: dts: qcom: x1e80100-crd: Add ov08x40 RGB sensor on CSIPHY4

From: Konrad Dybcio

Date: Wed Sep 16 2026 - 11:36:36 EST


On 9/16/26 4:58 PM, Bryan O'Donoghue wrote:
> On 09/09/2026 13:43, Konrad Dybcio wrote:
>>> +        port@3 {
>>> +            camss_csiphy4_inep: endpoint {
>>> +                data-lanes = <1 2 3 4>;
>>> +                remote-endpoint = <&csiphy4_out_ep>;
>>> +            };
>>> +        };
>> This port should be defined in the SoC DTSI
>
> 20260906-x1e-camss-csi2-phy-dtsi-v6-5-067f2ecc4630@xxxxxxxxxx
>
> +            ports {
> +                #address-cells = <1>;
> +                #size-cells = <0>;
> +
> +                port@0 {
> +                    reg = <0>;
> +                };
> +
> +                port@1 {
> +                    reg = <1>;
> +                };
> +
> +                port@2 {
> +                    reg = <2>;
> +                };
> +
> +                port@3 {
> +                    reg = <3>;
> +                };
> +            };
>
> I don't understand your ask.

I didn't notice it was. Even more so then, add a label to that
port and override the node's contents through a reference like
so:

&camss_csiphy4_port3 {
....
};

Perhaps moving the endpoint label and definition there too would
make sense, i.e.

-- hamoa.dtsi
ports {
#address-cells = <1>;
#size-cells = <0>;

port@0 {
reg = <0>;

camss_csiphy4_inep: endpoint {
};
};

...
};


-- x1-crd.dtsi
&camss_csiphy4_inep {
...
};

The current approach is super fragile and relies on the developer
never making any typos. DTC will enforce the existence of the
label at compile-time if you follow the approach I suggested.

Konrad