Re: [PATCH v21 14/14] arm64: dts: qcom: kaanapali-mtp-camera-s5kjn5: Add DT overlay
From: Hangxiang Ma
Date: Fri Oct 23 2026 - 05:44:42 EST
On 10/21/2026 10:02 PM, Konrad Dybcio wrote:
On 10/12/26 8:45 AM, Hangxiang Ma wrote:
Enable the S5KJN5 image sensor via CCI on Kaanapali MTP board.
The Kaanapali MTP board connect S5KJN5 with QUP I2C bus by default. This
DTBO enables S5KJN5 on CCI-0 interface connected to the CSIPHY2 in four
lane mode on a rework camera module.
Signed-off-by: Hangxiang Ma <hangxiang.ma@xxxxxxxxxxxxxxxx>
---
[...]
+&camss_csiphy2_in_port {
+ camss_csiphy2_inep: endpoint {
+ data-lanes = <1 2 3 4>;
+ remote-endpoint = <&csiphy2_out_ep>;
+ };
+};
Please define the endpoint in the SoC DTSI like the glymur submission:
https://lore.kernel.org/linux-arm-msm/b5734dac-1285-452b-8107-9164e6a3e5fa@xxxxxxxxxxxxxxxx/T/#m55a2d04f3fe573dbbc86fe89a874514494eca345
[...]
+&csiphy2_in_port {
+ csiphy2_in_ep: endpoint {
+ bus-type = <MEDIA_BUS_TYPE_CSI2_DPHY>;
+ clock-lanes = <0>;
+ data-lanes = <1 2 3 4>;
+ remote-endpoint = <&s5kjn5_out_ep>;
+ };
+};
+
+&csiphy2_out_port {
+ csiphy2_out_ep: endpoint {
+ remote-endpoint = <&camss_csiphy2_inep>;
+ };
+};
likewise for these two
With that:
Reviewed-by: Konrad Dybcio <konrad.dybcio@xxxxxxxxxxxxxxxx>
Konrad
Hi Konrad, there is one subtle difference between the Glymur and Kaanapali bindings that I'd like to clarify.
Kaanapali follows the X1E80100 binding, where data-lanes is required. Therefore, the endpoints need to be defined in the board DTS, following the same pattern as Bryan's series.
<https://lore.kernel.org/linux-arm-msm/20260917-x1e-camss-csi2-phy-dtsi-v7-4-1a63eb35838b@xxxxxxxxxx/>
Moving these endpoint definitions to the SoC DTSI, as done for Glymur, would cause dtbs_check warnings. Would it be OK to keep the current endpoint definitions for Kaanapali?
---
Best Regards,
Hangxiang