Thread (17 messages) 17 messages, 3 authors, 9d ago

Re: [PATCH v8 8/8] arm64: dts: qcom: kodiak: Desribe EUD UTMI routing and connector mapping

From: Elson Serrao <hidden>
Date: 2026-09-23 04:40:48
Also in: linux-arm-msm, linux-usb, lkml


On 9/21/2026 4:37 AM, Konrad Dybcio wrote:
On 9/13/26 12:53 AM, Elson Serrao wrote:
quoted
The existing EUD description uses a legacy single-path representation
that assumes uniform UTMI routing. This was sufficient for minimal
configurations where the USB port operated only in device mode and role
switching was not considered.

Update the description to explicitly model the topology using eud path
child node and add a role-switch declaration to reflect role-dependent
UTMI routing. Also on this SoC, the EUD hardware intercepts only the
primary UTMI path. So remove the USB endpoint mapping to the secondary
controller and associate it with the primary controller.

With the intermediate EUD node now in place, update the Kodiak board
descriptions to map the USB connector endpoint to the EUD instead of
directly to the USB controller, reflecting the intended hardware
topology. Enable the EUD by default now that the complete topology is
described.

Signed-off-by: Elson Serrao <redacted>
---
[...]
quoted
diff --git a/arch/arm64/boot/dts/qcom/qcm6490-fairphone-fp5.dts b/arch/arm64/boot/dts/qcom/qcm6490-fairphone-fp5.dts
index f41ad446a39e..b1f1ff6b99e6 100644
--- a/arch/arm64/boot/dts/qcom/qcm6490-fairphone-fp5.dts
+++ b/arch/arm64/boot/dts/qcom/qcm6490-fairphone-fp5.dts
@@ -98,7 +98,7 @@ port@0 {
 					reg = <0>;
 
 					pmic_glink_hs_in: endpoint {
-						remote-endpoint = <&usb_1_dwc3_hs>;
+						remote-endpoint = <&eud_con>;
 					};
 				};
 
@@ -1530,10 +1530,6 @@ &usb_1 {
 	status = "okay";
 };
 
-&usb_1_dwc3_hs {
-	remote-endpoint = <&pmic_glink_hs_in>;
-};
-
 &usb_1_hsphy {
 	vdda-pll-supply = <&vreg_l10c>;
 	vdda18-supply = <&vreg_l1c>;
@@ -1573,3 +1569,7 @@ &wifi {
 	qcom,calibration-variant = "Fairphone_5";
 	status = "okay";
 };
+
+&eud_con {
+	remote-endpoint = <&pmic_glink_hs_in>;
+};
Label references should be sorted alphabetically within a file,
there may be slight deviations because the tooling to enforce
that is still under development, but let's try to not add more
Ack. I will update this in v9

Thanks
Elson
Keyboard shortcuts
hback out one level
jnext message in thread
kprevious message in thread
ldrill in
Escclose help / fold thread tree
?toggle this help