[PATCH v2 0/5] arm64: dts: Add board support for Khadas Edge2

[PATCH v2 0/5] arm64: dts: Add board support for Khadas Edge2

[Jacobe Zang]

These patches add board support as following:

 - 2x TYPE-C port
 - BCM43438 bluetooth
 - 1x HDMI port
 - AP6275P Wi-Fi module
 - Adjust CPU frequency

Changes in v2:
 - Removed unrelated parts in Wi-Fi patch.
 - Fix indent error in USB-C patch.

 - Link to v1: https://patchwork.kernel.org/project/linux-rockchip/cover/20240617030537.3112356-1-jacobe.zang@wesion.com/

---
Jacobe Zang (5):
  arm64: dts: rockchip: Add USB-C to Khadas Edge 2
  arm64: dts: rockchip: Add bluetooth rfkill to Khadas Edge2
  arm64: dts: rockchip: Add HDMI & VOP2 to Khadas Edge 2
  arm64: dts: rockchip: Add AP6275P wireless support to Khadas Edge 2
  arm64: dts: rockchip: Add cpufreq support to Khadas Edge2

 .../dts/rockchip/rk3588s-khadas-edge2.dts     | 200 +++++++++++++++++-
 1 file changed, 197 insertions(+), 3 deletions(-)

-- 
2.34.1

[PATCH v2 1/5] arm64: dts: rockchip: Add USB-C to Khadas Edge 2

[Jacobe Zang]

Khadas Edge 2 has 2x Type-C port. One just supports PD and
controlled by MCU. The other one supports PD, DP Alt mode and DRD. This
commit adds support for DRD.

Signed-off-by: Jacobe Zang <jacobe.zang@wesion.com>
---
 .../dts/rockchip/rk3588s-khadas-edge2.dts     | 118 ++++++++++++++++++
 1 file changed, 118 insertions(+)

diff --git a/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts b/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
index dbddfc3bb4641..8c0bc675690dd 100644
--- a/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
+++ b/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
@@ -6,6 +6,7 @@
 #include <dt-bindings/input/input.h>
 #include <dt-bindings/pinctrl/rockchip.h>
 #include <dt-bindings/leds/common.h>
+#include <dt-bindings/usb/pd.h>
 #include "rk3588s.dtsi"
 
 / {
@@ -112,6 +113,18 @@ vcc5v0_sys: vcc5v0-sys-regulator {
 		regulator-max-microvolt = <5000000>;
 	};
 
+	vbus5v0_typec: vbus5v0-typec-regulator {
+		compatible = "regulator-fixed";
+		regulator-name = "vbus5v0_typec";
+		regulator-min-microvolt = <5000000>;
+		regulator-max-microvolt = <5000000>;
+		enable-active-high;
+		gpio = <&gpio3 RK_PA4 GPIO_ACTIVE_HIGH>;
+		vin-supply = <&vcc5v0_sys>;
+		pinctrl-names = "default";
+		pinctrl-0 = <&typec5v_pwren>;
+	};
+
 	vcc_1v1_nldo_s3: vcc-1v1-nldo-s3-regulator {
 		compatible = "regulator-fixed";
 		regulator-name = "vcc_1v1_nldo_s3";
@@ -224,6 +237,56 @@ regulator-state-mem {
 &i2c2 {
 	status = "okay";
 
+	usbc0: usb-typec@22 {
+		compatible = "fcs,fusb302";
+		reg = <0x22>;
+		interrupt-parent = <&gpio1>;
+		interrupts = <RK_PB5 IRQ_TYPE_LEVEL_LOW>;
+		pinctrl-names = "default";
+		pinctrl-0 = <&usbc0_int>;
+		vbus-supply = <&vbus5v0_typec>;
+		status = "okay";
+
+		usb_con: connector {
+			compatible = "usb-c-connector";
+			label = "USB-C";
+			data-role = "dual";
+			power-role = "dual";
+			try-power-role = "source";
+			op-sink-microwatt = <1000000>;
+			sink-pdos = <PDO_FIXED(5000, 3000, PDO_FIXED_USB_COMM)
+			PDO_FIXED(9000, 3000, PDO_FIXED_USB_COMM)
+			PDO_FIXED(12000, 3000, PDO_FIXED_USB_COMM)>;
+			source-pdos = <PDO_FIXED(5000, 3000, PDO_FIXED_USB_COMM)>;
+
+			ports {
+				#address-cells = <1>;
+				#size-cells = <0>;
+
+				port@0 {
+						reg = <0>;
+						usbc0_orien_sw: endpoint {
+						remote-endpoint = <&usbdp_phy0_orientation_switch>;
+					};
+				};
+
+				port@1 {
+						reg = <1>;
+						usbc0_role_sw: endpoint {
+						remote-endpoint = <&dwc3_0_role_switch>;
+					};
+				};
+
+				port@2 {
+						reg = <2>;
+						dp_altmode_mux: endpoint {
+						remote-endpoint = <&usbdp_phy0_dp_altmode_mux>;
+					};
+				};
+			};
+		};
+	};
+
 	hym8563: rtc@51 {
 		compatible = "haoyu,hym8563";
 		reg = <0x51>;
@@ -256,6 +319,16 @@ vcc5v0_host_en: vcc5v0-host-en {
 		};
 	};
 
+	usb-typec {
+		usbc0_int: usbc0-int {
+			rockchip,pins = <1 RK_PB5 RK_FUNC_GPIO &pcfg_pull_up>;
+		};
+
+		typec5v_pwren: typec5v-pwren {
+			rockchip,pins = <3 RK_PA4 RK_FUNC_GPIO &pcfg_pull_none>;
+		};
+	};
+
 	ir-receiver {
 		ir_receiver_pin: ir-receiver-pin {
 			rockchip,pins = <1  RK_PA7  RK_FUNC_GPIO  &pcfg_pull_none>;
@@ -681,6 +754,14 @@ &uart9 {
 	status = "okay";
 };
 
+&u2phy0 {
+	status = "okay";
+};
+
+&u2phy0_otg {
+	status = "okay";
+};
+
 &u2phy2 {
 	status = "okay";
 };
@@ -707,6 +788,43 @@ &usb_host0_ohci {
 	status = "okay";
 };
 
+&usbdp_phy0 {
+	orientation-switch;
+	mode-switch;
+	sbu1-dc-gpios = <&gpio4 RK_PA0 GPIO_ACTIVE_HIGH>;
+	sbu2-dc-gpios = <&gpio4 RK_PA1 GPIO_ACTIVE_HIGH>;
+	status = "okay";
+
+	port {
+		#address-cells = <1>;
+		#size-cells = <0>;
+
+		usbdp_phy0_orientation_switch: endpoint@0 {
+			reg = <0>;
+			remote-endpoint = <&usbc0_orien_sw>;
+		};
+
+		usbdp_phy0_dp_altmode_mux: endpoint@1 {
+			reg = <1>;
+			remote-endpoint = <&dp_altmode_mux>;
+		};
+	};
+};
+
+&usb_host0_xhci {
+	usb-role-switch;
+	status = "okay";
+
+	port {
+		#address-cells = <1>;
+		#size-cells = <0>;
+		dwc3_0_role_switch: endpoint@0 {
+			reg = <0>;
+			remote-endpoint = <&usbc0_role_sw>;
+		};
+	};
+};
+
 &usb_host1_ehci {
 	status = "okay";
 };
-- 
2.34.1

[PATCH v2 2/5] arm64: dts: rockchip: Add bluetooth rfkill to Khadas Edge2

[Jacobe Zang]

Add bluetooth node which managed by rfkill, bluetooth and
wlan controller on Khadas Edge2 was BCM43438. In uart9 add
RTS node in pinctrl.

Signed-off-by: Jacobe Zang <jacobe.zang@wesion.com>
---
 .../boot/dts/rockchip/rk3588s-khadas-edge2.dts  | 17 ++++++++++++++++-
 1 file changed, 16 insertions(+), 1 deletion(-)

diff --git a/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts b/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
index 8c0bc675690dd..a82f10312eacd 100644
--- a/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
+++ b/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
@@ -77,6 +77,14 @@ blue_led: led-2 {
 		};
 	};
 
+	bluetooth-rfkill {
+		compatible = "rfkill-gpio";
+		label = "rfkill-bluetooth";
+		radio-type = "bluetooth";
+		pinctrl-names = "default";
+		pinctrl-0 = <&bt_reset_pin>;
+	};
+
 	vcc3v3_pcie_wl: vcc3v3-pcie-wl-regulator {
 		compatible = "regulator-fixed";
 		enable-active-high;
@@ -750,8 +758,15 @@ &uart2 {
 
 &uart9 {
 	pinctrl-names = "default";
-	pinctrl-0 = <&uart9m2_xfer &uart9m2_ctsn>;
+	pinctrl-0 = <&uart9m2_xfer &uart9m2_ctsn &uart9m2_rtsn>;
 	status = "okay";
+
+	bluetooth {
+		compatible = "brcm,bcm43438-bt";
+		pinctrl-names = "default";
+		pinctrl-0 = <&bt_reset_pin &bt_wake_host_irq &bt_wake_pin>;
+		shutdown-gpios = <&gpio0 RK_PD4 GPIO_ACTIVE_HIGH>;
+	};
 };
 
 &u2phy0 {
-- 
2.34.1

[PATCH v2 3/5] arm64: dts: rockchip: Add HDMI & VOP2 to Khadas Edge 2

[Jacobe Zang]

Add the HDMI, VP, VOP nodes which HDMI function needs.

Signed-off-by: Jacobe Zang <jacobe.zang@wesion.com>
---
 .../dts/rockchip/rk3588s-khadas-edge2.dts     | 36 +++++++++++++++++++
 1 file changed, 36 insertions(+)

diff --git a/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts b/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
index a82f10312eacd..233bab17bffd2 100644
--- a/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
+++ b/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
@@ -7,6 +7,7 @@
 #include <dt-bindings/pinctrl/rockchip.h>
 #include <dt-bindings/leds/common.h>
 #include <dt-bindings/usb/pd.h>
+#include <dt-bindings/soc/rockchip,vop2.h>
 #include "rk3588s.dtsi"
 
 / {
@@ -827,6 +828,7 @@ usbdp_phy0_dp_altmode_mux: endpoint@1 {
 };
 
 &usb_host0_xhci {
+	dr-mode = "otg";
 	usb-role-switch;
 	status = "okay";
 
@@ -851,3 +853,37 @@ &usb_host1_ohci {
 &usb_host2_xhci {
 	status = "okay";
 };
+
+&hdmi0 {
+	status = "okay";
+};
+
+&hdptxphy_hdmi0 {
+	status = "okay";
+};
+
+&display_subsystem {
+	clocks = <&hdptxphy_hdmi0>;
+	clock-names = "hdmi0_phy_pll";
+};
+
+&vop_mmu {
+	status = "okay";
+};
+
+&hdmi0_in {
+	hdmi0_in_vp0: endpoint {
+		remote-endpoint = <&vp0_out_hdmi0>;
+	};
+};
+
+&vop {
+	status = "okay";
+};
+
+&vp0 {
+	vp0_out_hdmi0: endpoint@ROCKCHIP_VOP2_EP_HDMI0 {
+		reg = <ROCKCHIP_VOP2_EP_HDMI0>;
+		remote-endpoint = <&hdmi0_in_vp0>;
+	};
+};
-- 
2.34.1

[PATCH v2 4/5] arm64: dts: rockchip: Add AP6275P wireless support to Khadas Edge 2

[Jacobe Zang]

Khadas Edge2 uses the PCI-e Ampak AP6275P 2T2R Wi-Fi 6 module.

Signed-off-by: Jacobe Zang <jacobe.zang@wesion.com>
---
 .../boot/dts/rockchip/rk3588s-khadas-edge2.dts  | 17 +++++++++++++++++
 1 file changed, 17 insertions(+)

diff --git a/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts b/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
index 233bab17bffd2..7d7cc3e76838c 100644
--- a/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
+++ b/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
@@ -365,6 +365,23 @@ &pcie2x1l2 {
 	reset-gpios = <&gpio3 RK_PD1 GPIO_ACTIVE_HIGH>;
 	vpcie3v3-supply = <&vcc3v3_pcie_wl>;
 	status = "okay";
+
+	pcie@0,0 {
+		reg = <0x400000 0 0 0 0>;
+		#address-cells = <3>;
+		#size-cells = <2>;
+		ranges;
+		device_type = "pci";
+		bus-range = <0x40 0x4f>;
+
+		wifi: wifi@0,0 {
+			compatible = "pci14e4,449d";
+			reg = <0x410000 0 0 0 0>;
+			clocks = <&hym8563>;
+			clock-names = "32k";
+		};
+	};
+
 };
 
 &pwm11 {
-- 
2.34.1

[PATCH v2 5/5] arm64: dts: rockchip: Add cpufreq support to Khadas Edge2

[Jacobe Zang]

This adjust CPU nodes on Khadas Edge2.

Signed-off-by: Jacobe Zang <jacobe.zang@wesion.com>
---
 .../arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts | 12 ++++++++++--
 1 file changed, 10 insertions(+), 2 deletions(-)

diff --git a/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts b/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
index 7d7cc3e76838c..5fb15d3dc23e9 100644
--- a/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
+++ b/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
@@ -160,34 +160,42 @@ vdd_3v3_sd: vdd-3v3-sd-regulator {
 
 &cpu_b0 {
 	cpu-supply = <&vdd_cpu_big0_s0>;
+	mem-supply = <&vdd_cpu_big0_mem_s0>;
 };
 
 &cpu_b1 {
 	cpu-supply = <&vdd_cpu_big0_s0>;
+	mem-supply = <&vdd_cpu_big0_mem_s0>;
 };
 
 &cpu_b2 {
 	cpu-supply = <&vdd_cpu_big1_s0>;
+	mem-supply = <&vdd_cpu_big1_mem_s0>;
 };
 
 &cpu_b3 {
 	cpu-supply = <&vdd_cpu_big1_s0>;
+	mem-supply = <&vdd_cpu_big1_mem_s0>;
 };
 
 &cpu_l0 {
 	cpu-supply = <&vdd_cpu_lit_s0>;
+	mem-supply = <&vdd_cpu_lit_mem_s0>;
 };
 
 &cpu_l1 {
 	cpu-supply = <&vdd_cpu_lit_s0>;
+	mem-supply = <&vdd_cpu_lit_mem_s0>;
 };
 
 &cpu_l2 {
 	cpu-supply = <&vdd_cpu_lit_s0>;
+	mem-supply = <&vdd_cpu_lit_mem_s0>;
 };
 
 &cpu_l3 {
 	cpu-supply = <&vdd_cpu_lit_s0>;
+	mem-supply = <&vdd_cpu_lit_mem_s0>;
 };
 
 &combphy0_ps {
@@ -208,7 +216,7 @@ &i2c0 {
 	pinctrl-0 = <&i2c0m2_xfer>;
 	status = "okay";
 
-	vdd_cpu_big0_s0: regulator@42 {
+	vdd_cpu_big0_s0: vdd_cpu_big0_mem_s0: regulator@42 {
 		compatible = "rockchip,rk8602";
 		reg = <0x42>;
 		fcs,suspend-voltage-selector = <1>;
@@ -225,7 +233,7 @@ regulator-state-mem {
 		};
 	};
 
-	vdd_cpu_big1_s0: regulator@43 {
+	vdd_cpu_big1_s0: vdd_cpu_big1_mem_s0: regulator@43 {
 		compatible = "rockchip,rk8603", "rockchip,rk8602";
 		reg = <0x43>;
 		fcs,suspend-voltage-selector = <1>;
-- 
2.34.1

Re: [PATCH v2 5/5] arm64: dts: rockchip: Add cpufreq support to Khadas Edge2

[Heiko Stübner]

Hi Jacobe Zang,

Am Montag, 17. Juni 2024, 09:11:12 CEST schrieb Jacobe Zang:
> This adjust CPU nodes on Khadas Edge2.
> 
> Signed-off-by: Jacobe Zang <jacobe.zang@wesion.com>
> ---
>  .../arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts | 12 ++++++++++--
>  1 file changed, 10 insertions(+), 2 deletions(-)
> 
> diff --git a/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts b/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
> index 7d7cc3e76838c..5fb15d3dc23e9 100644
> --- a/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
> +++ b/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
> @@ -160,34 +160,42 @@ vdd_3v3_sd: vdd-3v3-sd-regulator {
>  
>  &cpu_b0 {
>  	cpu-supply = <&vdd_cpu_big0_s0>;
> +	mem-supply = <&vdd_cpu_big0_mem_s0>;

as far as I remember there has not been any binding merged that declares
this supply. Thankfully following the double phandle below, the Edge2 is
designed to use the same regulator for the mem-supply, so special handling
isn't even needed.


Heiko


>  };
>  
>  &cpu_b1 {
>  	cpu-supply = <&vdd_cpu_big0_s0>;
> +	mem-supply = <&vdd_cpu_big0_mem_s0>;
>  };
>  
>  &cpu_b2 {
>  	cpu-supply = <&vdd_cpu_big1_s0>;
> +	mem-supply = <&vdd_cpu_big1_mem_s0>;
>  };
>  
>  &cpu_b3 {
>  	cpu-supply = <&vdd_cpu_big1_s0>;
> +	mem-supply = <&vdd_cpu_big1_mem_s0>;
>  };
>  
>  &cpu_l0 {
>  	cpu-supply = <&vdd_cpu_lit_s0>;
> +	mem-supply = <&vdd_cpu_lit_mem_s0>;
>  };
>  
>  &cpu_l1 {
>  	cpu-supply = <&vdd_cpu_lit_s0>;
> +	mem-supply = <&vdd_cpu_lit_mem_s0>;
>  };
>  
>  &cpu_l2 {
>  	cpu-supply = <&vdd_cpu_lit_s0>;
> +	mem-supply = <&vdd_cpu_lit_mem_s0>;
>  };
>  
>  &cpu_l3 {
>  	cpu-supply = <&vdd_cpu_lit_s0>;
> +	mem-supply = <&vdd_cpu_lit_mem_s0>;
>  };
>  
>  &combphy0_ps {
> @@ -208,7 +216,7 @@ &i2c0 {
>  	pinctrl-0 = <&i2c0m2_xfer>;
>  	status = "okay";
>  
> -	vdd_cpu_big0_s0: regulator@42 {
> +	vdd_cpu_big0_s0: vdd_cpu_big0_mem_s0: regulator@42 {
>  		compatible = "rockchip,rk8602";
>  		reg = <0x42>;
>  		fcs,suspend-voltage-selector = <1>;
> @@ -225,7 +233,7 @@ regulator-state-mem {
>  		};
>  	};
>  
> -	vdd_cpu_big1_s0: regulator@43 {
> +	vdd_cpu_big1_s0: vdd_cpu_big1_mem_s0: regulator@43 {
>  		compatible = "rockchip,rk8603", "rockchip,rk8602";
>  		reg = <0x43>;
>  		fcs,suspend-voltage-selector = <1>;
> 

Re: [PATCH v2 5/5] arm64: dts: rockchip: Add cpufreq support to Khadas Edge2

[Alexey Charkov]

On 17/06/2024 11:33, Heiko Stübner wrote:
> Hi Jacobe Zang,
>
> Am Montag, 17. Juni 2024, 09:11:12 CEST schrieb Jacobe Zang:
>> This adjust CPU nodes on Khadas Edge2.
>>
>> Signed-off-by: Jacobe Zang <jacobe.zang@wesion.com>
>> ---
>>   .../arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts | 12 ++++++++++--
>>   1 file changed, 10 insertions(+), 2 deletions(-)
>>
>> diff --git a/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts b/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
>> index 7d7cc3e76838c..5fb15d3dc23e9 100644
>> --- a/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
>> +++ b/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
>> @@ -160,34 +160,42 @@ vdd_3v3_sd: vdd-3v3-sd-regulator {
>>   
>>   &cpu_b0 {
>>   	cpu-supply = <&vdd_cpu_big0_s0>;
>> +	mem-supply = <&vdd_cpu_big0_mem_s0>;
> as far as I remember there has not been any binding merged that declares
> this supply. Thankfully following the double phandle below, the Edge2 is
> designed to use the same regulator for the mem-supply, so special handling
> isn't even needed.

Indeed, currently there isn't any user in the mainline tree (neither 
bindings nor drivers) for this separate regulator. Mainline cpufreq-dt 
only expects a single regulator, and as Heiko pointed out Edge2 uses the 
same physical regulator to power both the CPU core and its SRAM, so 
adding a separate mem-supply here isn't helpful.

Best regards, Alexey

Re: [PATCH v2 2/5] arm64: dts: rockchip: Add bluetooth rfkill to Khadas Edge2

[Alexey Charkov]

On 17/06/2024 10:11, Jacobe Zang wrote:
> Add bluetooth node which managed by rfkill, bluetooth and
> wlan controller on Khadas Edge2 was BCM43438. In uart9 add
> RTS node in pinctrl.

You refer to wlan in the commit message, but there is nothing wlan 
related in the patch itself. Update the commit message perhaps?

> Signed-off-by: Jacobe Zang <jacobe.zang@wesion.com>
> ---
>   .../boot/dts/rockchip/rk3588s-khadas-edge2.dts  | 17 ++++++++++++++++-
>   1 file changed, 16 insertions(+), 1 deletion(-)
>
> diff --git a/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts b/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
> index 8c0bc675690dd..a82f10312eacd 100644
> --- a/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
> +++ b/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
> @@ -77,6 +77,14 @@ blue_led: led-2 {
>   		};
>   	};
>   
> +	bluetooth-rfkill {
> +		compatible = "rfkill-gpio";
> +		label = "rfkill-bluetooth";
> +		radio-type = "bluetooth";
> +		pinctrl-names = "default";
> +		pinctrl-0 = <&bt_reset_pin>;

Does it actually work this way? I thought you'd need to also reference 
the respective GPIO to be able to trigger its state, not just switch the 
pinctrl configuration to GPIO.


I'm also wondering if bt_reset_pin is the right one to use here. On my 
Rock 5B I had to trigger bt_wake_pin to get Bluetooth up and running.


Best regards,

Alexey

Re: [PATCH v2 5/5] arm64: dts: rockchip: Add cpufreq support to Khadas Edge2

[Jacobe Zang]

Thanks Heiko and Alexey, I will delete this patch in next version.

---
Best Regards
Jacobe

Re: [PATCH v2 4/5] arm64: dts: rockchip: Add AP6275P wireless support to Khadas Edge 2

[Alexey Charkov]

On 17/06/2024 10:11, Jacobe Zang wrote:
> Khadas Edge2 uses the PCI-e Ampak AP6275P 2T2R Wi-Fi 6 module.
>
> Signed-off-by: Jacobe Zang <jacobe.zang@wesion.com>
> ---
>   .../boot/dts/rockchip/rk3588s-khadas-edge2.dts  | 17 +++++++++++++++++
>   1 file changed, 17 insertions(+)
>
> diff --git a/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts b/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
> index 233bab17bffd2..7d7cc3e76838c 100644
> --- a/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
> +++ b/arch/arm64/boot/dts/rockchip/rk3588s-khadas-edge2.dts
> @@ -365,6 +365,23 @@ &pcie2x1l2 {
>   	reset-gpios = <&gpio3 RK_PD1 GPIO_ACTIVE_HIGH>;
>   	vpcie3v3-supply = <&vcc3v3_pcie_wl>;
>   	status = "okay";
> +
> +	pcie@0,0 {
> +		reg = <0x400000 0 0 0 0>;
> +		#address-cells = <3>;
> +		#size-cells = <2>;
> +		ranges;
> +		device_type = "pci";
> +		bus-range = <0x40 0x4f>;
> +
> +		wifi: wifi@0,0 {
> +			compatible = "pci14e4,449d";

This doesn't seem to be listed in the bindings, nor is there a mainline 
driver that would match either this compatible or PCI ID 14e4:449d. 
Maybe submit either or both of those first, to make sure they are 
reviewed and acceptable for mainline inclusion, before this change lands 
in DTS and becomes part of the ABI?

I'm also wondering why would adding a DT node for a PCI device be needed 
in the first place, given that PCI supports device discovery? Does it 
require some sort of boot-time fixup by the bootloader? If so, it might 
be helpful to state that in comments.

Best regards,

Alexey

Re: [PATCH v2 2/5] arm64: dts: rockchip: Add bluetooth rfkill to Khadas Edge2

[Jacobe Zang]

> You refer to wlan in the commit message, but there is nothing wlan
> related in the patch itself. Update the commit message perhaps?

At that time, I want to describe that wlan and bluetooth were from the module named BCM43438. But, it's true that wlan do not be metioned in the code. I will update it in next version.

> I thought you'd need to also reference the respective GPIO to be able to trigger its 
> state, not just switch the pinctrl configuration to GPIO.

You means the shutdown-gpios? I configure it in the bluetooth node, it can't be configure in rfkill node at the same time.

> I'm also wondering if bt_reset_pin is the right one to use here. On my
> Rock 5B I had to trigger bt_wake_pin to get Bluetooth up and running.

Yes, I try to configure bt_reset_pin, bt_wake_host_irq and  bt_wake_pin in pinctrl separately. Each of them works well. After I connected a BT device to Edge2, I input rfkill block 1, bluetooth function failed. Then I input rfkill unblock 1, it can rework. So at last I only configure the bt_reset_pin.

---
Best Regards
Jacobe

Re: [PATCH v2 2/5] arm64: dts: rockchip: Add bluetooth rfkill to Khadas Edge2

[Alexey Charkov]

(sorry, the previous one got sent in HTML by accident)

On 17/06/2024 12:55, Jacobe Zang wrote:
>> You refer to wlan in the commit message, but there is nothing wlan
>> related in the patch itself. Update the commit message perhaps?
> At that time, I want to describe that wlan and bluetooth were from the module named BCM43438. But, it's true that wlan do not be metioned in the code. I will update it in next version.
>
>> I thought you'd need to also reference the respective GPIO to be able to trigger its
>> state, not just switch the pinctrl configuration to GPIO.
> You means the shutdown-gpios? I configure it in the bluetooth node, it can't be configure in rfkill node at the same time.

In Documentation/devicetree/bindings/net/rfkill-gpio.yaml shutdown-gpios 
is listed as a required property, did you try building the sources with 
CHECK_DTBS=1 in make arguments?

If you already control this GPIO from elsewhere (such as from the 
bluetooth driver), then perhaps you don't need to define a separate 
rfkill device at all.

>> I'm also wondering if bt_reset_pin is the right one to use here. On my
>> Rock 5B I had to trigger bt_wake_pin to get Bluetooth up and running.
> Yes, I try to configure bt_reset_pin, bt_wake_host_irq and  bt_wake_pin in pinctrl separately. Each of them works well. After I connected a BT device to Edge2, I input rfkill block 1, bluetooth function failed. Then I input rfkill unblock 1, it can rework. So at last I only configure the bt_reset_pin.

That doesn't necessarily mean that the hardware gets an rfkill signal, 
what you describe above could also come from the software alone. It 
would be great to somehow check if the physical pin state gets triggered 
once you switch rfkill state from userspace, and then that the bluetooth 
device handles it appropriately (e.g. gets into a hardware-disable state 
somehow).

Best regards,
Alexey

Re: [PATCH v2 4/5] arm64: dts: rockchip: Add AP6275P wireless support to Khadas Edge 2

[Jacobe Zang]

>I'm also wondering why would adding a DT node for a PCI device be needed
>in the first place, given that PCI supports device discovery?

In fact, I learn that PCIe bus devices do not need compatible to probe just now... Before sending this patch, I committed the code that added "pci14e4,449d" to vendor-prefix.yaml and net/wireless/brcm,brcm4329-fmac.yaml. Now I know the reason why my addition was rejected. By the way, except for the compatible binding, is there any other binding that I should remove??

---
Best Regards
Jacobe

Re: [PATCH v2 4/5] arm64: dts: rockchip: Add AP6275P wireless support to Khadas Edge 2

[Alexey Charkov]

On Mon, Jun 17, 2024 at 2:58 PM Jacobe Zang <jacobe.zang@wesion.com> wrote:
>
> >I'm also wondering why would adding a DT node for a PCI device be needed
> >in the first place, given that PCI supports device discovery?
>
> In fact, I learn that PCIe bus devices do not need compatible to probe just now... Before sending this patch, I committed the code that added "pci14e4,449d" to vendor-prefix.yaml and net/wireless/brcm,brcm4329-fmac.yaml. Now I know the reason why my addition was rejected. By the way, except for the compatible binding, is there any other binding that I should remove??

If your PCI bridge is functioning properly and if your WiFi adapter is
connected and physically enabled (in terms of power and RFKILL) I
believe it should be automatically discovered and you should see it in
lspci. No additional DT nodes needed - but check if you need any
additional DT property somewhere to keep the HYM8563 clock enabled.
I'm not sure your pcie@0,0 node is needed either.

Then it's up to the driver to recognize your adapter by its PCI ID and
attach. I guess you'd need to extend the hardware IDs table in the
brcmfmac driver for it to attach - similar to [1]

[1] https://github.com/armbian/build/blob/main/patch/kernel/archive/rockchip-rk3588-6.10/0801-wireless-add-bcm43752.patch

Best regards,
Alexey

Re: [PATCH v2 2/5] arm64: dts: rockchip: Add bluetooth rfkill to Khadas Edge2

[Jacobe Zang]

> If you already control this GPIO from elsewhere (such as from the
> bluetooth driver), then perhaps you don't need to define a separate
> rfkill device at all.

Yes, I missed the error log before. The rfkill driver didn't probe successfully. I will remove this rfkill node and reserve bluetooth node next time.

---
Best Regards
Jacobe

Re: [PATCH v2 3/5] arm64: dts: rockchip: Add HDMI & VOP2 to Khadas Edge 2

[Heiko Stübner]

Am Montag, 17. Juni 2024, 09:11:10 CEST schrieb Jacobe Zang:
> Add the HDMI, VP, VOP nodes which HDMI function needs.
> 
> Signed-off-by: Jacobe Zang <jacobe.zang@wesion.com>

The rk3588's hdmi controller was just posted in a v1 and from
comments received it looks like it'll still take a bit of time before
that is ready to land. So this patch would also be for later.


Heiko

Re: [PATCH v2 4/5] arm64: dts: rockchip: Add AP6275P wireless support to Khadas Edge 2

[Sebastian Reichel]

[-- Attachment #1: Type: text/plain, Size: 2114 bytes --]

Hi,

On Mon, Jun 17, 2024 at 03:17:50PM GMT, Alexey Charkov wrote:
> On Mon, Jun 17, 2024 at 2:58 PM Jacobe Zang <jacobe.zang@wesion.com> wrote:
> >
> > >I'm also wondering why would adding a DT node for a PCI device be needed
> > >in the first place, given that PCI supports device discovery?
> >
> > In fact, I learn that PCIe bus devices do not need compatible to probe just now... Before sending this patch, I committed the code that added "pci14e4,449d" to vendor-prefix.yaml and net/wireless/brcm,brcm4329-fmac.yaml. Now I know the reason why my addition was rejected. By the way, except for the compatible binding, is there any other binding that I should remove??
> 
> If your PCI bridge is functioning properly and if your WiFi adapter is
> connected and physically enabled (in terms of power and RFKILL) I
> believe it should be automatically discovered and you should see it in
> lspci. No additional DT nodes needed - but check if you need any
> additional DT property somewhere to keep the HYM8563 clock enabled.
> I'm not sure your pcie@0,0 node is needed either.
> 
> Then it's up to the driver to recognize your adapter by its PCI ID and
> attach. I guess you'd need to extend the hardware IDs table in the
> brcmfmac driver for it to attach - similar to [1]
> 
> [1] https://github.com/armbian/build/blob/main/patch/kernel/archive/rockchip-rk3588-6.10/0801-wireless-add-bcm43752.patch

RK3588 EVB1 has the same WLAN/BT module and I started looking
into adding support at a low priority. I think this is very
similar to the Qualcomm chip Bartosz Golaszewski worked on and
thus should get a pwrseq driver, which handles the enable GPIOs,
regulators and clocks. If any of them are missing the PCIe device
is not discovered. Note, that this will actually require describing
the PCIe device in DT to reference the pwrseq device.

Additionally the brcmfmac need to be extended to actually have
working wlan.

I put some links and notes about this here:

https://gitlab.collabora.com/hardware-enablement/rockchip-3588/linux/-/issues/1

Greetings,

-- Sebastian

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH v2 2/5] arm64: dts: rockchip: Add bluetooth rfkill to Khadas Edge2

[Sebastian Reichel]

[-- Attachment #1: Type: text/plain, Size: 550 bytes --]

Hi,

On Mon, Jun 17, 2024 at 11:22:57AM GMT, Jacobe Zang wrote:
> > If you already control this GPIO from elsewhere (such as from the
> > bluetooth driver), then perhaps you don't need to define a separate
> > rfkill device at all.
> 
> Yes, I missed the error log before. The rfkill driver didn't probe
> successfully. I will remove this rfkill node and reserve bluetooth
> node next time.

Does it work with the RTC clock disabled? Otherwise this also needs
the pwrseq work I mentioned in the WLAN patch.

Greetings,

-- Sebastian

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]

Re: [PATCH v2 2/5] arm64: dts: rockchip: Add bluetooth rfkill to Khadas Edge2

[Jacobe Zang]

Hi Sebastian,

> Does it work with the RTC clock disabled? Otherwise this also needs
> the pwrseq work I mentioned in the WLAN patch.

After I disabled the RTC clock, bluetooth stopped working. I also check the Amlogic chip, some SOCs have the pwrseq defined in DT, maybe rk3588 need to add it as well.

---
Best Regards
Jacobe

Re: [PATCH v2 3/5] arm64: dts: rockchip: Add HDMI & VOP2 to Khadas Edge 2

[Jacobe Zang]

Hi Heiko,

> The rk3588's hdmi controller was just posted in a v1 and from
> comments received it looks like it'll still take a bit of time before
> that is ready to land.

All right. I will recommit after the posted patch was merged by mainline. By the way, could you provide information on the specific patch or share a link to it?

---
Best Regards
Jacobe

Re: [PATCH v2 4/5] arm64: dts: rockchip: Add AP6275P wireless support to Khadas Edge 2

[Jacobe Zang]

> If your PCI bridge is functioning properly and if your WiFi adapter is
> connected and physically enabled (in terms of power and RFKILL) I
> believe it should be automatically discovered and you should see it in
> lspci.

I'm sure that wifi@0,0 node is not needed. 

> I'm not sure your pcie@0,0 node is needed either.

I think pcie@0,0 node is useful. I have tried to comment it, after that Wi-Fi works properly but bluetooth can't work.

---
Best Regards
Jacobe