Re: [PATCH v8 10/10] arm64: dts: rockchip: rk3399-evb-ind: Add support for DisplayPort
From: Chaoyi Chen
Date: Thu Oct 30 2025 - 02:50:45 EST
On 10/30/2025 2:13 PM, Peter Chen wrote:
On Thu, Oct 30, 2025 at 11:14 AM Chaoyi Chen <chaoyi.chen@xxxxxxxxxxxxxx> wrote:
On 10/30/2025 10:50 AM, Peter Chen wrote:Hi Chaoyi,
Yes, I have tried both the v2 and v8 approaches, and both can work.Have you tried debugging based on upstream code?Okay. My question is basic: USB2 PHY supplies DP/DM, and the DP/DM isFrom the perspective of Type-C, this should not be added. Is the approach in v2 correct [0] ?
short for Type-C connector,
and no control is needed for Type-C application.
Why is there a remote-endpoint connection between USB2 PHY and Type-C connector?
v2 is correct, but the dts needs to improve.In v8 patch5, we implemented typec_mux and typec_switch in the USB/DP PHY.
- There is a remote-endpoint connection for USB role switch between
Type-C connector
device and USB controller device
- There is a remote-endpoint connection for orientation and lane configuration
between Type-C connector device and USB/DP PHY device.
I think the current remote-endpoint connections are all child node of the USB/DP PHY. That is:
&tcphy0_dp {
mode-switch;
...
};
&tcphy0_usb3 {
orientation-switch;
...
};
Does this still need to be improved? Thank you.
There are two questions I have still not seen the answer to:
- Why USB2 PHY is related to your Type-C patch?
I was just following other people's approach. Sorry, this should be removed from the dts.
- How does the USB role switch event notify the USB controller driver, eg dwc3?
Sorry, I misunderstood what you said before. There is indeed a missing usb-role-switch now. I referred to the approach in rk3588-evb1-v10.dts. Is the following way of writing correct?
&usbc_connector {
ports {
#address-cells = <1>;
#size-cells = <0>;
port@0 {
reg = <0>;
usbc_orien_sw: endpoint {
remote-endpoint = <&tcphy0_typec_orien_sw>;
};
};
port@1 {
reg = <1>;
usbc_role_sw: endpoint {
remote-endpoint = <&dwc3_0_role_switch>;
};
};
port@2 {
reg = <2>;
usbc_dp: endpoint {
remote-endpoint = <&tcphy0_typec_dp>;
};
};
};
};
&usbdrd_dwc3_0 {
status = "okay";
usb-role-switch;
port {
#address-cells = <1>;
#size-cells = <0>;
dwc3_0_role_switch: endpoint@0 {
reg = <0>;
remote-endpoint = <&usbc_role_sw>;
};
};
};
&tcphy0_usb3 {
orientation-switch;
port {
tcphy0_typec_orien_sw: endpoint {
remote-endpoint = <&usbc_orien_sw>;
};
};
};
&tcphy0_dp {
mode-switch;
port {
#address-cells = <1>;
#size-cells = <0>;
tcphy0_typec_dp: endpoint@0 {
reg = <0>;
remote-endpoint = <&usbc_dp>;
};
};
};
Peter--
Peter--
[0]: https://lore.kernel.org/all/20250715112456.101-6-kernel@xxxxxxxxxx/
Or is the following approach correct?
port@0 {
reg = <0>;
usbc_hs: endpoint {
remote-endpoint = <&tcphy0>;
};
};
port@1 {
reg = <1>;
usbc_ss: endpoint {
remote-endpoint = <&tcphy0>;
};
};
port@2 {
reg = <2>;
usbc_dp: endpoint {
remote-endpoint = <&tcphy0_typec_dp>;
};
};
I assume the Micro-USB connector does not use Type-C/PD IC, is itWait a minute, actually we have multiple hardware interfaces, one of which is Type-C, eventually connected to USBDPPHY, and the other is micro-usb connected to U2PHY.I checked the phy-rockchip-inno-usb2.c but did not find any switch or mux. Does this mean that we need to implement them? Thank you.+ port@1 {.....
+ reg = <1>;
+
+ usbc_ss: endpoint {
+ remote-endpoint = <&tcphy0_typec_ss>;
+ };
+ };
+
+ port@2 {
+ reg = <2>;
+
+ usbc_dp: endpoint {
+ remote-endpoint = <&tcphy0_typec_dp>;
+ };
+ };
+ };
+ };
+ };
+};
+
&u2phy0 {There is no switch and mux, how to co-work with Type-C?
status = "okay";
+
+ port {
+ u2phy0_typec_hs: endpoint {
+ remote-endpoint = <&usbc_hs>;
+ };
+ };
};
right? Does it relate to this patch?
Best regards,
Peter
Best,
Chaoyi
Best,
Chaoyi