[PATCH 0/5] arm: dts: mt7623: relocate gmacs, mt7530 switch, and mux phy

[PATCH 0/5] arm: dts: mt7623: relocate gmacs, mt7530 switch, and mux phy

[Arınç ÜNAL]

Yeia sou!

I initially just wanted to submit the phy muxing patch for BPI-R2 but after
looking around the bindings, I decided to do a bit of a cleanup.

The main takeaway of the cleanup is that the MT7530 switch is actually a
part of the multi-chip module on the MT7623AI SoC. I'm moving the bindings
for this switch to mt7623a.dtsi so they don't need to be defined on every
device with this SoC. This should ease it up for supporting more devices
with MT7623AI SoC on mainline.

I don't know if there's a pin wired for interrupt. The switch is going to
work with polling for the time being. Folks from MediaTek, feel free to
fill me in on this.

I've tested phy muxing on my BPI-R2. DTs compile fine.

Arınç ÜNAL (5):
  arm: dts: mt7623: add gmacs to mt7623.dtsi
  arm: dts: mt7623: add mt7530 switch to mt7623a.dtsi
  arm: dts: mt7623: change mt7530 switch address
  arm: dts: mt7623: mux phy0 on Bananapi BPI-R2
  arm: dts: mt7623: enable flow control on port@6

 arch/arm/boot/dts/mt7623.dtsi                 | 14 ++++
 arch/arm/boot/dts/mt7623a-rfb-emmc.dts        | 86 ++++++----------------
 arch/arm/boot/dts/mt7623a-rfb-nand.dts        | 86 ++++++----------------
 arch/arm/boot/dts/mt7623a.dtsi                | 74 +++++++++++++++++++
 arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts | 22 +++---
 arch/arm/boot/dts/mt7623n-rfb-emmc.dts        | 13 ++--
 6 files changed, 147 insertions(+), 148 deletions(-)

[PATCH 1/5] arm: dts: mt7623: add gmacs to mt7623.dtsi

[Arınç ÜNAL]

There are two gigabit MACs on the MT7623AI and MT7623NI SoCs. Add them to
mt7623.dtsi and adjust the DTs that call mt7623.dtsi.

Signed-off-by: Arınç ÜNAL <arinc.unal@arinc9.com>
---
 arch/arm/boot/dts/mt7623.dtsi                 | 14 ++++++++++++++
 arch/arm/boot/dts/mt7623a-rfb-emmc.dts        |  4 +---
 arch/arm/boot/dts/mt7623a-rfb-nand.dts        |  4 +---
 arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts |  4 +---
 arch/arm/boot/dts/mt7623n-rfb-emmc.dts        |  8 ++------
 5 files changed, 19 insertions(+), 15 deletions(-)

diff --git a/arch/arm/boot/dts/mt7623.dtsi b/arch/arm/boot/dts/mt7623.dtsi
index 25d31e40a553..5eb698a90d34 100644
--- a/arch/arm/boot/dts/mt7623.dtsi
+++ b/arch/arm/boot/dts/mt7623.dtsi
@@ -981,6 +981,20 @@ eth: ethernet@1b100000 {
 		#address-cells = <1>;
 		#size-cells = <0>;
 		status = "disabled";
+
+		gmac0: mac@0 {
+			compatible = "mediatek,eth-mac";
+			reg = <0>;
+			phy-mode = "trgmii";
+			status = "disabled";
+		};
+
+		gmac1: mac@1 {
+			compatible = "mediatek,eth-mac";
+			reg = <1>;
+			phy-mode = "rgmii";
+			status = "disabled";
+		};
 	};
 
 	crypto: crypto@1b240000 {
diff --git a/arch/arm/boot/dts/mt7623a-rfb-emmc.dts b/arch/arm/boot/dts/mt7623a-rfb-emmc.dts
index e8b4b6d30d19..8297123f9d09 100644
--- a/arch/arm/boot/dts/mt7623a-rfb-emmc.dts
+++ b/arch/arm/boot/dts/mt7623a-rfb-emmc.dts
@@ -116,9 +116,7 @@ &eth {
 	status = "okay";
 
 	gmac0: mac@0 {
-		compatible = "mediatek,eth-mac";
-		reg = <0>;
-		phy-mode = "trgmii";
+		status = "okay";
 
 		fixed-link {
 			speed = <1000>;
diff --git a/arch/arm/boot/dts/mt7623a-rfb-nand.dts b/arch/arm/boot/dts/mt7623a-rfb-nand.dts
index 61f5da68d4b0..9ee859b8eea7 100644
--- a/arch/arm/boot/dts/mt7623a-rfb-nand.dts
+++ b/arch/arm/boot/dts/mt7623a-rfb-nand.dts
@@ -120,9 +120,7 @@ &eth {
 	status = "okay";
 
 	gmac0: mac@0 {
-		compatible = "mediatek,eth-mac";
-		reg = <0>;
-		phy-mode = "trgmii";
+		status = "okay";
 
 		fixed-link {
 			speed = <1000>;
diff --git a/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts b/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
index 5008115d2494..a5800a524302 100644
--- a/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
+++ b/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
@@ -175,9 +175,7 @@ &eth {
 	status = "okay";
 
 	gmac0: mac@0 {
-		compatible = "mediatek,eth-mac";
-		reg = <0>;
-		phy-mode = "trgmii";
+		status = "okay";
 
 		fixed-link {
 			speed = <1000>;
diff --git a/arch/arm/boot/dts/mt7623n-rfb-emmc.dts b/arch/arm/boot/dts/mt7623n-rfb-emmc.dts
index bf67a8e9be59..2a869a6476eb 100644
--- a/arch/arm/boot/dts/mt7623n-rfb-emmc.dts
+++ b/arch/arm/boot/dts/mt7623n-rfb-emmc.dts
@@ -160,9 +160,7 @@ &eth {
 	status = "okay";
 
 	gmac0: mac@0 {
-		compatible = "mediatek,eth-mac";
-		reg = <0>;
-		phy-mode = "trgmii";
+		status = "okay";
 
 		fixed-link {
 			speed = <1000>;
@@ -172,9 +170,7 @@ fixed-link {
 	};
 
 	mac@1 {
-		compatible = "mediatek,eth-mac";
-		reg = <1>;
-		phy-mode = "rgmii";
+		status = "okay";
 		phy-handle = <&phy5>;
 	};
 
-- 
2.37.2

[PATCH 2/5] arm: dts: mt7623: add mt7530 switch to mt7623a.dtsi

[Arınç ÜNAL]

The MT7530 switch is included as a part of the multi-chip module on the
MT7623AI SoC. Add it to mt7623a.dtsi and adjust DTs that call mt7623a.dtsi.

Signed-off-by: Arınç ÜNAL <arinc.unal@arinc9.com>
---

I don’t know if there’s a pin wired for interrupt. The switch is going to
work with polling for the time being. Folks from MediaTek, feel free to
fill me in on this.

Arınç

---
 arch/arm/boot/dts/mt7623a-rfb-emmc.dts | 84 +++++++-------------------
 arch/arm/boot/dts/mt7623a-rfb-nand.dts | 84 +++++++-------------------
 arch/arm/boot/dts/mt7623a.dtsi         | 73 ++++++++++++++++++++++
 3 files changed, 115 insertions(+), 126 deletions(-)

diff --git a/arch/arm/boot/dts/mt7623a-rfb-emmc.dts b/arch/arm/boot/dts/mt7623a-rfb-emmc.dts
index 8297123f9d09..5654284bab01 100644
--- a/arch/arm/boot/dts/mt7623a-rfb-emmc.dts
+++ b/arch/arm/boot/dts/mt7623a-rfb-emmc.dts
@@ -112,73 +112,31 @@ &crypto {
 	status = "okay";
 };
 
-&eth {
-	status = "okay";
+&switch0 {
+	ports {
+		port@0 {
+			status = "okay";
+			label = "lan0";
+		};
 
-	gmac0: mac@0 {
-		status = "okay";
+		port@1 {
+			status = "okay";
+			label = "lan1";
+		};
 
-		fixed-link {
-			speed = <1000>;
-			full-duplex;
-			pause;
+		port@2 {
+			status = "okay";
+			label = "lan2";
+		};
+
+		port@3 {
+			status = "okay";
+			label = "lan3";
 		};
-	};
 
-	mdio-bus {
-		#address-cells = <1>;
-		#size-cells = <0>;
-
-		switch@0 {
-			compatible = "mediatek,mt7530";
-			reg = <0>;
-			mediatek,mcm;
-			resets = <&ethsys MT2701_ETHSYS_MCM_RST>;
-			reset-names = "mcm";
-			core-supply = <&mt6323_vpa_reg>;
-			io-supply = <&mt6323_vemc3v3_reg>;
-
-			ports {
-				#address-cells = <1>;
-				#size-cells = <0>;
-
-				port@0 {
-					reg = <0>;
-					label = "lan0";
-				};
-
-				port@1 {
-					reg = <1>;
-					label = "lan1";
-				};
-
-				port@2 {
-					reg = <2>;
-					label = "lan2";
-				};
-
-				port@3 {
-					reg = <3>;
-					label = "lan3";
-				};
-
-				port@4 {
-					reg = <4>;
-					label = "wan";
-				};
-
-				port@6 {
-					reg = <6>;
-					label = "cpu";
-					ethernet = <&gmac0>;
-					phy-mode = "trgmii";
-
-					fixed-link {
-						speed = <1000>;
-						full-duplex;
-					};
-				};
-			};
+		port@4 {
+			status = "okay";
+			label = "wan";
 		};
 	};
 };
diff --git a/arch/arm/boot/dts/mt7623a-rfb-nand.dts b/arch/arm/boot/dts/mt7623a-rfb-nand.dts
index 9ee859b8eea7..afd177b3b516 100644
--- a/arch/arm/boot/dts/mt7623a-rfb-nand.dts
+++ b/arch/arm/boot/dts/mt7623a-rfb-nand.dts
@@ -116,73 +116,31 @@ &crypto {
 	status = "okay";
 };
 
-&eth {
-	status = "okay";
+&switch0 {
+	ports {
+		port@0 {
+			status = "okay";
+			label = "lan0";
+		};
 
-	gmac0: mac@0 {
-		status = "okay";
+		port@1 {
+			status = "okay";
+			label = "lan1";
+		};
 
-		fixed-link {
-			speed = <1000>;
-			full-duplex;
-			pause;
+		port@2 {
+			status = "okay";
+			label = "lan2";
 		};
-	};
 
-	mdio-bus {
-		#address-cells = <1>;
-		#size-cells = <0>;
-
-		switch@0 {
-			compatible = "mediatek,mt7530";
-			reg = <0>;
-			mediatek,mcm;
-			resets = <&ethsys MT2701_ETHSYS_MCM_RST>;
-			reset-names = "mcm";
-			core-supply = <&mt6323_vpa_reg>;
-			io-supply = <&mt6323_vemc3v3_reg>;
-
-			ports {
-				#address-cells = <1>;
-				#size-cells = <0>;
-
-				port@0 {
-					reg = <0>;
-					label = "lan0";
-				};
-
-				port@1 {
-					reg = <1>;
-					label = "lan1";
-				};
-
-				port@2 {
-					reg = <2>;
-					label = "lan2";
-				};
-
-				port@3 {
-					reg = <3>;
-					label = "lan3";
-				};
-
-				port@4 {
-					reg = <4>;
-					label = "wan";
-				};
-
-				port@6 {
-					reg = <6>;
-					label = "cpu";
-					ethernet = <&gmac0>;
-					phy-mode = "trgmii";
-
-					fixed-link {
-						speed = <1000>;
-						full-duplex;
-					};
-				};
-			};
+		port@3 {
+			status = "okay";
+			label = "lan3";
+		};
+
+		port@4 {
+			status = "okay";
+			label = "wan";
 		};
 	};
 };
diff --git a/arch/arm/boot/dts/mt7623a.dtsi b/arch/arm/boot/dts/mt7623a.dtsi
index d304b62d24b5..5207e0183f3a 100644
--- a/arch/arm/boot/dts/mt7623a.dtsi
+++ b/arch/arm/boot/dts/mt7623a.dtsi
@@ -18,7 +18,80 @@ &crypto {
 };
 
 &eth {
+	status = "okay";
 	power-domains = <&scpsys MT7623A_POWER_DOMAIN_ETH>;
+
+	gmac0: mac@0 {
+		status = "okay";
+
+		fixed-link {
+			speed = <1000>;
+			full-duplex;
+			pause;
+		};
+	};
+
+	mdio: mdio-bus {
+		#address-cells = <1>;
+		#size-cells = <0>;
+
+		switch0: switch@0 {
+			compatible = "mediatek,mt7530";
+			reg = <0>;
+			mediatek,mcm;
+			resets = <&ethsys MT2701_ETHSYS_MCM_RST>;
+			reset-names = "mcm";
+			core-supply = <&mt6323_vpa_reg>;
+			io-supply = <&mt6323_vemc3v3_reg>;
+
+			ports {
+				#address-cells = <1>;
+				#size-cells = <0>;
+
+				port@0 {
+					status = "disabled";
+					reg = <0>;
+					label = "swp0";
+				};
+
+				port@1 {
+					status = "disabled";
+					reg = <1>;
+					label = "swp1";
+				};
+
+				port@2 {
+					status = "disabled";
+					reg = <2>;
+					label = "swp2";
+				};
+
+				port@3 {
+					status = "disabled";
+					reg = <3>;
+					label = "swp3";
+				};
+
+				port@4 {
+					status = "disabled";
+					reg = <4>;
+					label = "swp4";
+				};
+
+				port@6 {
+					reg = <6>;
+					label = "cpu";
+					ethernet = <&gmac0>;
+					phy-mode = "trgmii";
+
+					fixed-link {
+						speed = <1000>;
+						full-duplex;
+					};
+				};
+			};
+		};
+	};
 };
 
 &nandc {
-- 
2.37.2

[PATCH 3/5] arm: dts: mt7623: change mt7530 switch address

[Arınç ÜNAL]

In the case of muxing phy0 of the MT7530 switch, the switch and the phy
will have the same address on the mdio bus, 0. This causes the ethernet
driver to fail since devices on the mdio bus cannot share an address.

Any address can be used for the switch, therefore, change the switch
address to 0x1f.

Signed-off-by: Arınç ÜNAL <arinc.unal@arinc9.com>
---
 arch/arm/boot/dts/mt7623a.dtsi                | 4 ++--
 arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts | 4 ++--
 arch/arm/boot/dts/mt7623n-rfb-emmc.dts        | 4 ++--
 3 files changed, 6 insertions(+), 6 deletions(-)

diff --git a/arch/arm/boot/dts/mt7623a.dtsi b/arch/arm/boot/dts/mt7623a.dtsi
index 5207e0183f3a..a1a393028629 100644
--- a/arch/arm/boot/dts/mt7623a.dtsi
+++ b/arch/arm/boot/dts/mt7623a.dtsi
@@ -35,9 +35,9 @@ mdio: mdio-bus {
 		#address-cells = <1>;
 		#size-cells = <0>;
 
-		switch0: switch@0 {
+		switch0: switch@1f {
 			compatible = "mediatek,mt7530";
-			reg = <0>;
+			reg = <0x1f>;
 			mediatek,mcm;
 			resets = <&ethsys MT2701_ETHSYS_MCM_RST>;
 			reset-names = "mcm";
diff --git a/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts b/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
index a5800a524302..94174cb2b7e1 100644
--- a/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
+++ b/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
@@ -188,9 +188,9 @@ mdio: mdio-bus {
 		#address-cells = <1>;
 		#size-cells = <0>;
 
-		switch@0 {
+		switch@1f {
 			compatible = "mediatek,mt7530";
-			reg = <0>;
+			reg = <0x1f>;
 			reset-gpios = <&pio 33 0>;
 			core-supply = <&mt6323_vpa_reg>;
 			io-supply = <&mt6323_vemc3v3_reg>;
diff --git a/arch/arm/boot/dts/mt7623n-rfb-emmc.dts b/arch/arm/boot/dts/mt7623n-rfb-emmc.dts
index 2a869a6476eb..7c3298a92887 100644
--- a/arch/arm/boot/dts/mt7623n-rfb-emmc.dts
+++ b/arch/arm/boot/dts/mt7623n-rfb-emmc.dts
@@ -183,9 +183,9 @@ phy5: ethernet-phy@5 {
 			phy-mode = "rgmii-rxid";
 		};
 
-		switch@0 {
+		switch@1f {
 			compatible = "mediatek,mt7530";
-			reg = <0>;
+			reg = <0x1f>;
 			reset-gpios = <&pio 33 0>;
 			core-supply = <&mt6323_vpa_reg>;
 			io-supply = <&mt6323_vemc3v3_reg>;
-- 
2.37.2

[PATCH 4/5] arm: dts: mt7623: mux phy0 on Bananapi BPI-R2

[Arınç ÜNAL]

Mux the MT7530 switch's phy0 to gmac5 which is wired to the SoC's gmac1.
This achieves 2 Gbps total bandwidth to the CPU using the second RGMII.

Tested-by: Arınç ÜNAL <arinc.unal@arinc9.com>
Signed-off-by: Arınç ÜNAL <arinc.unal@arinc9.com>
---
 arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts | 14 +++++++++-----
 1 file changed, 9 insertions(+), 5 deletions(-)

diff --git a/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts b/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
index 94174cb2b7e1..28051136c4e2 100644
--- a/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
+++ b/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
@@ -184,10 +184,19 @@ fixed-link {
 		};
 	};
 
+	gmac1: mac@1 {
+		status = "okay";
+		phy-handle = <&ethphy0>;
+	};
+
 	mdio: mdio-bus {
 		#address-cells = <1>;
 		#size-cells = <0>;
 
+		ethphy0: ethernet-phy@0 {
+			reg = <0>;
+		};
+
 		switch@1f {
 			compatible = "mediatek,mt7530";
 			reg = <0x1f>;
@@ -199,11 +208,6 @@ ports {
 				#address-cells = <1>;
 				#size-cells = <0>;
 
-				port@0 {
-					reg = <0>;
-					label = "wan";
-				};
-
 				port@1 {
 					reg = <1>;
 					label = "lan0";
-- 
2.37.2

[PATCH 5/5] arm: dts: mt7623: enable flow control on port@6

[Arınç ÜNAL]

Flow control needs to be enabled on both sides to work.
It is already enabled on gmac0, enable it on port@6 too.

Signed-off-by: Arınç ÜNAL <arinc.unal@arinc9.com>
---
 arch/arm/boot/dts/mt7623a.dtsi         | 1 +
 arch/arm/boot/dts/mt7623n-rfb-emmc.dts | 1 +
 2 files changed, 2 insertions(+)

diff --git a/arch/arm/boot/dts/mt7623a.dtsi b/arch/arm/boot/dts/mt7623a.dtsi
index a1a393028629..a522a00ab67e 100644
--- a/arch/arm/boot/dts/mt7623a.dtsi
+++ b/arch/arm/boot/dts/mt7623a.dtsi
@@ -87,6 +87,7 @@ port@6 {
 					fixed-link {
 						speed = <1000>;
 						full-duplex;
+						pause;
 					};
 				};
 			};
diff --git a/arch/arm/boot/dts/mt7623n-rfb-emmc.dts b/arch/arm/boot/dts/mt7623n-rfb-emmc.dts
index 7c3298a92887..96c46f766e6e 100644
--- a/arch/arm/boot/dts/mt7623n-rfb-emmc.dts
+++ b/arch/arm/boot/dts/mt7623n-rfb-emmc.dts
@@ -228,6 +228,7 @@ port@6 {
 					fixed-link {
 						speed = <1000>;
 						full-duplex;
+						pause;
 					};
 				};
 			};
-- 
2.37.2

Re: [PATCH 0/5] arm: dts: mt7623: relocate gmacs, mt7530 switch, and mux phy

[Matthias Brugger]

On 20/01/2023 21:53, Arınç ÜNAL wrote:
> Yeia sou!
> 

γεια σας!

Whole series applied, thanks!

> I initially just wanted to submit the phy muxing patch for BPI-R2 but after
> looking around the bindings, I decided to do a bit of a cleanup.
> 
> The main takeaway of the cleanup is that the MT7530 switch is actually a
> part of the multi-chip module on the MT7623AI SoC. I'm moving the bindings
> for this switch to mt7623a.dtsi so they don't need to be defined on every
> device with this SoC. This should ease it up for supporting more devices
> with MT7623AI SoC on mainline.
> 
> I don't know if there's a pin wired for interrupt. The switch is going to
> work with polling for the time being. Folks from MediaTek, feel free to
> fill me in on this.
> 
> I've tested phy muxing on my BPI-R2. DTs compile fine.
> 
> Arınç ÜNAL (5):
>    arm: dts: mt7623: add gmacs to mt7623.dtsi
>    arm: dts: mt7623: add mt7530 switch to mt7623a.dtsi
>    arm: dts: mt7623: change mt7530 switch address
>    arm: dts: mt7623: mux phy0 on Bananapi BPI-R2
>    arm: dts: mt7623: enable flow control on port@6
> 
>   arch/arm/boot/dts/mt7623.dtsi                 | 14 ++++
>   arch/arm/boot/dts/mt7623a-rfb-emmc.dts        | 86 ++++++----------------
>   arch/arm/boot/dts/mt7623a-rfb-nand.dts        | 86 ++++++----------------
>   arch/arm/boot/dts/mt7623a.dtsi                | 74 +++++++++++++++++++
>   arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts | 22 +++---
>   arch/arm/boot/dts/mt7623n-rfb-emmc.dts        | 13 ++--
>   6 files changed, 147 insertions(+), 148 deletions(-)
> 
> 

Re: [PATCH 0/5] arm: dts: mt7623: relocate gmacs, mt7530 switch, and mux phy

[Matthias Brugger]

On 20/01/2023 21:53, Arınç ÜNAL wrote:
> Yeia sou!
> 

γεια σας!

Whole series applied, thanks!

> I initially just wanted to submit the phy muxing patch for BPI-R2 but after
> looking around the bindings, I decided to do a bit of a cleanup.
> 
> The main takeaway of the cleanup is that the MT7530 switch is actually a
> part of the multi-chip module on the MT7623AI SoC. I'm moving the bindings
> for this switch to mt7623a.dtsi so they don't need to be defined on every
> device with this SoC. This should ease it up for supporting more devices
> with MT7623AI SoC on mainline.
> 
> I don't know if there's a pin wired for interrupt. The switch is going to
> work with polling for the time being. Folks from MediaTek, feel free to
> fill me in on this.
> 
> I've tested phy muxing on my BPI-R2. DTs compile fine.
> 
> Arınç ÜNAL (5):
>    arm: dts: mt7623: add gmacs to mt7623.dtsi
>    arm: dts: mt7623: add mt7530 switch to mt7623a.dtsi
>    arm: dts: mt7623: change mt7530 switch address
>    arm: dts: mt7623: mux phy0 on Bananapi BPI-R2
>    arm: dts: mt7623: enable flow control on port@6
> 
>   arch/arm/boot/dts/mt7623.dtsi                 | 14 ++++
>   arch/arm/boot/dts/mt7623a-rfb-emmc.dts        | 86 ++++++----------------
>   arch/arm/boot/dts/mt7623a-rfb-nand.dts        | 86 ++++++----------------
>   arch/arm/boot/dts/mt7623a.dtsi                | 74 +++++++++++++++++++
>   arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts | 22 +++---
>   arch/arm/boot/dts/mt7623n-rfb-emmc.dts        | 13 ++--
>   6 files changed, 147 insertions(+), 148 deletions(-)
> 
> 

Re: [PATCH 0/5] arm: dts: mt7623: relocate gmacs, mt7530 switch, and mux phy

[Matthias Brugger]

On 20/01/2023 21:53, Arınç ÜNAL wrote:
> Yeia sou!
> 

γεια σας!

Whole series applied, thanks!

> I initially just wanted to submit the phy muxing patch for BPI-R2 but after
> looking around the bindings, I decided to do a bit of a cleanup.
> 
> The main takeaway of the cleanup is that the MT7530 switch is actually a
> part of the multi-chip module on the MT7623AI SoC. I'm moving the bindings
> for this switch to mt7623a.dtsi so they don't need to be defined on every
> device with this SoC. This should ease it up for supporting more devices
> with MT7623AI SoC on mainline.
> 
> I don't know if there's a pin wired for interrupt. The switch is going to
> work with polling for the time being. Folks from MediaTek, feel free to
> fill me in on this.
> 
> I've tested phy muxing on my BPI-R2. DTs compile fine.
> 
> Arınç ÜNAL (5):
>    arm: dts: mt7623: add gmacs to mt7623.dtsi
>    arm: dts: mt7623: add mt7530 switch to mt7623a.dtsi
>    arm: dts: mt7623: change mt7530 switch address
>    arm: dts: mt7623: mux phy0 on Bananapi BPI-R2
>    arm: dts: mt7623: enable flow control on port@6
> 
>   arch/arm/boot/dts/mt7623.dtsi                 | 14 ++++
>   arch/arm/boot/dts/mt7623a-rfb-emmc.dts        | 86 ++++++----------------
>   arch/arm/boot/dts/mt7623a-rfb-nand.dts        | 86 ++++++----------------
>   arch/arm/boot/dts/mt7623a.dtsi                | 74 +++++++++++++++++++
>   arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts | 22 +++---
>   arch/arm/boot/dts/mt7623n-rfb-emmc.dts        | 13 ++--
>   6 files changed, 147 insertions(+), 148 deletions(-)
> 
> 

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH 4/5] arm: dts: mt7623: mux phy0 on Bananapi BPI-R2

[Frank Wunderlich]

Am 20. Januar 2023 21:53:17 MEZ schrieb "Arınç ÜNAL" <arinc.unal@arinc9.com>:

> 		switch@1f {
> 			compatible = "mediatek,mt7530";
> 			reg = <0x1f>;
>@@ -199,11 +208,6 @@ ports {
> 				#address-cells = <1>;
> 				#size-cells = <0>;
> 
>-				port@0 {
>-					reg = <0>;
>-					label = "wan";
>-				};
>-

This will break existing userspace setups using wan as interface name.

> 				port@1 {
> 					reg = <1>;
> 					label = "lan0";


regards Frank

Re: [PATCH 4/5] arm: dts: mt7623: mux phy0 on Bananapi BPI-R2

[Frank Wunderlich]

Am 20. Januar 2023 21:53:17 MEZ schrieb "Arınç ÜNAL" <arinc.unal@arinc9.com>:

> 		switch@1f {
> 			compatible = "mediatek,mt7530";
> 			reg = <0x1f>;
>@@ -199,11 +208,6 @@ ports {
> 				#address-cells = <1>;
> 				#size-cells = <0>;
> 
>-				port@0 {
>-					reg = <0>;
>-					label = "wan";
>-				};
>-

This will break existing userspace setups using wan as interface name.

> 				port@1 {
> 					reg = <1>;
> 					label = "lan0";


regards Frank

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH 1/5] arm: dts: mt7623: add gmacs to mt7623.dtsi

[Frank Wunderlich]

Am 20. Januar 2023 21:53:14 MEZ schrieb "Arınç ÜNAL" <arinc.unal@arinc9.com>:

>index 25d31e40a553..5eb698a90d34 100644
>--- a/arch/arm/boot/dts/mt7623.dtsi
>+++ b/arch/arm/boot/dts/mt7623.dtsi
>@@ -981,6 +981,20 @@ eth: ethernet@1b100000 {
> 		#address-cells = <1>;
> 		#size-cells = <0>;
> 		status = "disabled";
>+
>+		gmac0: mac@0 {
>+			compatible = "mediatek,eth-mac";
>+			reg = <0>;
>+			phy-mode = "trgmii";
>+			status = "disabled";
>+		};
>+
>+		gmac1: mac@1 {
>+			compatible = "mediatek,eth-mac";
>+			reg = <1>;
>+			phy-mode = "rgmii";
>+			status = "disabled";
>+		};
> 	};
> 

>diff --git a/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts b/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
>index 5008115d2494..a5800a524302 100644
>--- a/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
>+++ b/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
>@@ -175,9 +175,7 @@ &eth {
> 	status = "okay";
> 
> 	gmac0: mac@0 {

Should node not be accessed with label (&gmac0) instead of defining it again and shadow the one from dtsi?

>-		compatible = "mediatek,eth-mac";
>-		reg = <0>;
>-		phy-mode = "trgmii";
>+		status = "okay";
> 
> 		fixed-link {
> 			speed = <1000>;

Same for other boards.

regards Frank

Re: [PATCH 1/5] arm: dts: mt7623: add gmacs to mt7623.dtsi

[Frank Wunderlich]

Am 20. Januar 2023 21:53:14 MEZ schrieb "Arınç ÜNAL" <arinc.unal@arinc9.com>:

>index 25d31e40a553..5eb698a90d34 100644
>--- a/arch/arm/boot/dts/mt7623.dtsi
>+++ b/arch/arm/boot/dts/mt7623.dtsi
>@@ -981,6 +981,20 @@ eth: ethernet@1b100000 {
> 		#address-cells = <1>;
> 		#size-cells = <0>;
> 		status = "disabled";
>+
>+		gmac0: mac@0 {
>+			compatible = "mediatek,eth-mac";
>+			reg = <0>;
>+			phy-mode = "trgmii";
>+			status = "disabled";
>+		};
>+
>+		gmac1: mac@1 {
>+			compatible = "mediatek,eth-mac";
>+			reg = <1>;
>+			phy-mode = "rgmii";
>+			status = "disabled";
>+		};
> 	};
> 

>diff --git a/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts b/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
>index 5008115d2494..a5800a524302 100644
>--- a/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
>+++ b/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
>@@ -175,9 +175,7 @@ &eth {
> 	status = "okay";
> 
> 	gmac0: mac@0 {

Should node not be accessed with label (&gmac0) instead of defining it again and shadow the one from dtsi?

>-		compatible = "mediatek,eth-mac";
>-		reg = <0>;
>-		phy-mode = "trgmii";
>+		status = "okay";
> 
> 		fixed-link {
> 			speed = <1000>;

Same for other boards.

regards Frank

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH 1/5] arm: dts: mt7623: add gmacs to mt7623.dtsi

[Arınç ÜNAL]

On 25.01.2023 19:45, Frank Wunderlich wrote:
> Am 20. Januar 2023 21:53:14 MEZ schrieb "Arınç ÜNAL" <arinc.unal@arinc9.com>:
> 
>> index 25d31e40a553..5eb698a90d34 100644
>> --- a/arch/arm/boot/dts/mt7623.dtsi
>> +++ b/arch/arm/boot/dts/mt7623.dtsi
>> @@ -981,6 +981,20 @@ eth: ethernet@1b100000 {
>> 		#address-cells = <1>;
>> 		#size-cells = <0>;
>> 		status = "disabled";
>> +
>> +		gmac0: mac@0 {
>> +			compatible = "mediatek,eth-mac";
>> +			reg = <0>;
>> +			phy-mode = "trgmii";
>> +			status = "disabled";
>> +		};
>> +
>> +		gmac1: mac@1 {
>> +			compatible = "mediatek,eth-mac";
>> +			reg = <1>;
>> +			phy-mode = "rgmii";
>> +			status = "disabled";
>> +		};
>> 	};
>>
> 
>> diff --git a/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts b/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
>> index 5008115d2494..a5800a524302 100644
>> --- a/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
>> +++ b/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
>> @@ -175,9 +175,7 @@ &eth {
>> 	status = "okay";
>>
>> 	gmac0: mac@0 {
> 
> Should node not be accessed with label (&gmac0) instead of defining it again and shadow the one from dtsi?

I think that's up to preference. I kept it the current way as it's 
cleaner than taking it out of &eth.

Arınç

Re: [PATCH 1/5] arm: dts: mt7623: add gmacs to mt7623.dtsi

[Arınç ÜNAL]

On 25.01.2023 19:45, Frank Wunderlich wrote:
> Am 20. Januar 2023 21:53:14 MEZ schrieb "Arınç ÜNAL" <arinc.unal@arinc9.com>:
> 
>> index 25d31e40a553..5eb698a90d34 100644
>> --- a/arch/arm/boot/dts/mt7623.dtsi
>> +++ b/arch/arm/boot/dts/mt7623.dtsi
>> @@ -981,6 +981,20 @@ eth: ethernet@1b100000 {
>> 		#address-cells = <1>;
>> 		#size-cells = <0>;
>> 		status = "disabled";
>> +
>> +		gmac0: mac@0 {
>> +			compatible = "mediatek,eth-mac";
>> +			reg = <0>;
>> +			phy-mode = "trgmii";
>> +			status = "disabled";
>> +		};
>> +
>> +		gmac1: mac@1 {
>> +			compatible = "mediatek,eth-mac";
>> +			reg = <1>;
>> +			phy-mode = "rgmii";
>> +			status = "disabled";
>> +		};
>> 	};
>>
> 
>> diff --git a/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts b/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
>> index 5008115d2494..a5800a524302 100644
>> --- a/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
>> +++ b/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
>> @@ -175,9 +175,7 @@ &eth {
>> 	status = "okay";
>>
>> 	gmac0: mac@0 {
> 
> Should node not be accessed with label (&gmac0) instead of defining it again and shadow the one from dtsi?

I think that's up to preference. I kept it the current way as it's 
cleaner than taking it out of &eth.

Arınç

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH 4/5] arm: dts: mt7623: mux phy0 on Bananapi BPI-R2

[Arınç ÜNAL]

On 25.01.2023 19:41, Frank Wunderlich wrote:
> Am 20. Januar 2023 21:53:17 MEZ schrieb "Arınç ÜNAL" <arinc.unal@arinc9.com>:
> 
>> 		switch@1f {
>> 			compatible = "mediatek,mt7530";
>> 			reg = <0x1f>;
>> @@ -199,11 +208,6 @@ ports {
>> 				#address-cells = <1>;
>> 				#size-cells = <0>;
>>
>> -				port@0 {
>> -					reg = <0>;
>> -					label = "wan";
>> -				};
>> -
> 
> This will break existing userspace setups using wan as interface name.

Yup. The OS being used will have to either rename the interface or adapt 
to the new name, eth1.

Arınç

Re: [PATCH 4/5] arm: dts: mt7623: mux phy0 on Bananapi BPI-R2

[Arınç ÜNAL]

On 25.01.2023 19:41, Frank Wunderlich wrote:
> Am 20. Januar 2023 21:53:17 MEZ schrieb "Arınç ÜNAL" <arinc.unal@arinc9.com>:
> 
>> 		switch@1f {
>> 			compatible = "mediatek,mt7530";
>> 			reg = <0x1f>;
>> @@ -199,11 +208,6 @@ ports {
>> 				#address-cells = <1>;
>> 				#size-cells = <0>;
>>
>> -				port@0 {
>> -					reg = <0>;
>> -					label = "wan";
>> -				};
>> -
> 
> This will break existing userspace setups using wan as interface name.

Yup. The OS being used will have to either rename the interface or adapt 
to the new name, eth1.

Arınç

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH 1/5] arm: dts: mt7623: add gmacs to mt7623.dtsi

[Matthias Brugger]

On 25/01/2023 17:52, Arınç ÜNAL wrote:
> On 25.01.2023 19:45, Frank Wunderlich wrote:
>> Am 20. Januar 2023 21:53:14 MEZ schrieb "Arınç ÜNAL" <arinc.unal@arinc9.com>:
>>
>>> index 25d31e40a553..5eb698a90d34 100644
>>> --- a/arch/arm/boot/dts/mt7623.dtsi
>>> +++ b/arch/arm/boot/dts/mt7623.dtsi
>>> @@ -981,6 +981,20 @@ eth: ethernet@1b100000 {
>>>         #address-cells = <1>;
>>>         #size-cells = <0>;
>>>         status = "disabled";
>>> +
>>> +        gmac0: mac@0 {
>>> +            compatible = "mediatek,eth-mac";
>>> +            reg = <0>;
>>> +            phy-mode = "trgmii";
>>> +            status = "disabled";
>>> +        };
>>> +
>>> +        gmac1: mac@1 {
>>> +            compatible = "mediatek,eth-mac";
>>> +            reg = <1>;
>>> +            phy-mode = "rgmii";
>>> +            status = "disabled";
>>> +        };
>>>     };
>>>
>>
>>> diff --git a/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts 
>>> b/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
>>> index 5008115d2494..a5800a524302 100644
>>> --- a/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
>>> +++ b/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
>>> @@ -175,9 +175,7 @@ &eth {
>>>     status = "okay";
>>>
>>>     gmac0: mac@0 {
>>
>> Should node not be accessed with label (&gmac0) instead of defining it again 
>> and shadow the one from dtsi?
> 
> I think that's up to preference. I kept it the current way as it's cleaner than 
> taking it out of &eth.
> 

I'd prefer to do use the label, so that we stay consistent in the source tree.
I'll remove the whole series for now. Please provide a v2.

Thanks,
Matthias

Re: [PATCH 1/5] arm: dts: mt7623: add gmacs to mt7623.dtsi

[Matthias Brugger]

On 25/01/2023 17:52, Arınç ÜNAL wrote:
> On 25.01.2023 19:45, Frank Wunderlich wrote:
>> Am 20. Januar 2023 21:53:14 MEZ schrieb "Arınç ÜNAL" <arinc.unal@arinc9.com>:
>>
>>> index 25d31e40a553..5eb698a90d34 100644
>>> --- a/arch/arm/boot/dts/mt7623.dtsi
>>> +++ b/arch/arm/boot/dts/mt7623.dtsi
>>> @@ -981,6 +981,20 @@ eth: ethernet@1b100000 {
>>>         #address-cells = <1>;
>>>         #size-cells = <0>;
>>>         status = "disabled";
>>> +
>>> +        gmac0: mac@0 {
>>> +            compatible = "mediatek,eth-mac";
>>> +            reg = <0>;
>>> +            phy-mode = "trgmii";
>>> +            status = "disabled";
>>> +        };
>>> +
>>> +        gmac1: mac@1 {
>>> +            compatible = "mediatek,eth-mac";
>>> +            reg = <1>;
>>> +            phy-mode = "rgmii";
>>> +            status = "disabled";
>>> +        };
>>>     };
>>>
>>
>>> diff --git a/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts 
>>> b/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
>>> index 5008115d2494..a5800a524302 100644
>>> --- a/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
>>> +++ b/arch/arm/boot/dts/mt7623n-bananapi-bpi-r2.dts
>>> @@ -175,9 +175,7 @@ &eth {
>>>     status = "okay";
>>>
>>>     gmac0: mac@0 {
>>
>> Should node not be accessed with label (&gmac0) instead of defining it again 
>> and shadow the one from dtsi?
> 
> I think that's up to preference. I kept it the current way as it's cleaner than 
> taking it out of &eth.
> 

I'd prefer to do use the label, so that we stay consistent in the source tree.
I'll remove the whole series for now. Please provide a v2.

Thanks,
Matthias

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: [PATCH 4/5] arm: dts: mt7623: mux phy0 on Bananapi BPI-R2

[Matthias Brugger]

On 25/01/2023 17:54, Arınç ÜNAL wrote:
> On 25.01.2023 19:41, Frank Wunderlich wrote:
>> Am 20. Januar 2023 21:53:17 MEZ schrieb "Arınç ÜNAL" <arinc.unal@arinc9.com>:
>>
>>>         switch@1f {
>>>             compatible = "mediatek,mt7530";
>>>             reg = <0x1f>;
>>> @@ -199,11 +208,6 @@ ports {
>>>                 #address-cells = <1>;
>>>                 #size-cells = <0>;
>>>
>>> -                port@0 {
>>> -                    reg = <0>;
>>> -                    label = "wan";
>>> -                };
>>> -
>>
>> This will break existing userspace setups using wan as interface name.
> 
> Yup. The OS being used will have to either rename the interface or adapt to the 
> new name, eth1.
> 

In that case please explain in the commit message why this is needed. We are 
trying really hard to not break userspace by a device-tree update.

Regards,
Matthias

Re: [PATCH 4/5] arm: dts: mt7623: mux phy0 on Bananapi BPI-R2

[Matthias Brugger]

On 25/01/2023 17:54, Arınç ÜNAL wrote:
> On 25.01.2023 19:41, Frank Wunderlich wrote:
>> Am 20. Januar 2023 21:53:17 MEZ schrieb "Arınç ÜNAL" <arinc.unal@arinc9.com>:
>>
>>>         switch@1f {
>>>             compatible = "mediatek,mt7530";
>>>             reg = <0x1f>;
>>> @@ -199,11 +208,6 @@ ports {
>>>                 #address-cells = <1>;
>>>                 #size-cells = <0>;
>>>
>>> -                port@0 {
>>> -                    reg = <0>;
>>> -                    label = "wan";
>>> -                };
>>> -
>>
>> This will break existing userspace setups using wan as interface name.
> 
> Yup. The OS being used will have to either rename the interface or adapt to the 
> new name, eth1.
> 

In that case please explain in the commit message why this is needed. We are 
trying really hard to not break userspace by a device-tree update.

Regards,
Matthias

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Aw: Re: [PATCH 4/5] arm: dts: mt7623: mux phy0 on Bananapi BPI-R2

[Frank Wunderlich]

Hi
> Gesendet: Dienstag, 31. Januar 2023 um 12:55 Uhr
> Von: "Matthias Brugger" <matthias.bgg@gmail.com>
> An: "Arınç ÜNAL" <arinc.unal@arinc9.com>, frank-w@public-files.de, "Rob Herring" <robh+dt@kernel.org>, "Krzysztof Kozlowski" <krzysztof.kozlowski+dt@linaro.org>
> Cc: devicetree@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-mediatek@lists.infradead.org, linux-kernel@vger.kernel.org, erkin.bozoglu@xeront.com, "Sean Wang" <sean.wang@mediatek.com>, "DENG Qingfang" <dqfext@gmail.com>
> Betreff: Re: [PATCH 4/5] arm: dts: mt7623: mux phy0 on Bananapi BPI-R2
>
> 
> 
> On 25/01/2023 17:54, Arınç ÜNAL wrote:
> > On 25.01.2023 19:41, Frank Wunderlich wrote:
> >> Am 20. Januar 2023 21:53:17 MEZ schrieb "Arınç ÜNAL" <arinc.unal@arinc9.com>:
> >>
> >>>         switch@1f {
> >>>             compatible = "mediatek,mt7530";
> >>>             reg = <0x1f>;
> >>> @@ -199,11 +208,6 @@ ports {
> >>>                 #address-cells = <1>;
> >>>                 #size-cells = <0>;
> >>>
> >>> -                port@0 {
> >>> -                    reg = <0>;
> >>> -                    label = "wan";
> >>> -                };
> >>> -
> >>
> >> This will break existing userspace setups using wan as interface name.
> > 
> > Yup. The OS being used will have to either rename the interface or adapt to the 
> > new name, eth1.
> > 
> 
> In that case please explain in the commit message why this is needed. We are 
> trying really hard to not break userspace by a device-tree update.

I understand the reason why Arınç wants to change it:

currently all ports including wan-port are routed via dsa over gmac0 (trgmii), Arınç enables the second gmac and port5 of switch to route wan-traffic over the other gmac. But unfortunately there is no mainline-way to rename the gmac1 interface (eth1) to wan...this would be a way to not break the userspace.

i had an older patch for mtk_eth_soc [1] which do it at driver level for mediatek ethernet driver, but afair a generic way to (re)name interfaces in dts (vendor independ) was requested as the right way.

regards Frank

[1] https://github.com/frank-w/BPI-Router-Linux/commit/5088c9c81e90014622a950fca38cfb29e0421155

Aw: Re: [PATCH 4/5] arm: dts: mt7623: mux phy0 on Bananapi BPI-R2

[Frank Wunderlich]

Hi
> Gesendet: Dienstag, 31. Januar 2023 um 12:55 Uhr
> Von: "Matthias Brugger" <matthias.bgg@gmail.com>
> An: "Arınç ÜNAL" <arinc.unal@arinc9.com>, frank-w@public-files.de, "Rob Herring" <robh+dt@kernel.org>, "Krzysztof Kozlowski" <krzysztof.kozlowski+dt@linaro.org>
> Cc: devicetree@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-mediatek@lists.infradead.org, linux-kernel@vger.kernel.org, erkin.bozoglu@xeront.com, "Sean Wang" <sean.wang@mediatek.com>, "DENG Qingfang" <dqfext@gmail.com>
> Betreff: Re: [PATCH 4/5] arm: dts: mt7623: mux phy0 on Bananapi BPI-R2
>
> 
> 
> On 25/01/2023 17:54, Arınç ÜNAL wrote:
> > On 25.01.2023 19:41, Frank Wunderlich wrote:
> >> Am 20. Januar 2023 21:53:17 MEZ schrieb "Arınç ÜNAL" <arinc.unal@arinc9.com>:
> >>
> >>>         switch@1f {
> >>>             compatible = "mediatek,mt7530";
> >>>             reg = <0x1f>;
> >>> @@ -199,11 +208,6 @@ ports {
> >>>                 #address-cells = <1>;
> >>>                 #size-cells = <0>;
> >>>
> >>> -                port@0 {
> >>> -                    reg = <0>;
> >>> -                    label = "wan";
> >>> -                };
> >>> -
> >>
> >> This will break existing userspace setups using wan as interface name.
> > 
> > Yup. The OS being used will have to either rename the interface or adapt to the 
> > new name, eth1.
> > 
> 
> In that case please explain in the commit message why this is needed. We are 
> trying really hard to not break userspace by a device-tree update.

I understand the reason why Arınç wants to change it:

currently all ports including wan-port are routed via dsa over gmac0 (trgmii), Arınç enables the second gmac and port5 of switch to route wan-traffic over the other gmac. But unfortunately there is no mainline-way to rename the gmac1 interface (eth1) to wan...this would be a way to not break the userspace.

i had an older patch for mtk_eth_soc [1] which do it at driver level for mediatek ethernet driver, but afair a generic way to (re)name interfaces in dts (vendor independ) was requested as the right way.

regards Frank

[1] https://github.com/frank-w/BPI-Router-Linux/commit/5088c9c81e90014622a950fca38cfb29e0421155

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel

Re: Aw: Re: [PATCH 4/5] arm: dts: mt7623: mux phy0 on Bananapi BPI-R2

[Arınç ÜNAL]

On 31.01.2023 19:39, Frank Wunderlich wrote:
> Hi
>> Gesendet: Dienstag, 31. Januar 2023 um 12:55 Uhr
>> Von: "Matthias Brugger" <matthias.bgg@gmail.com>
>> An: "Arınç ÜNAL" <arinc.unal@arinc9.com>, frank-w@public-files.de, "Rob Herring" <robh+dt@kernel.org>, "Krzysztof Kozlowski" <krzysztof.kozlowski+dt@linaro.org>
>> Cc: devicetree@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-mediatek@lists.infradead.org, linux-kernel@vger.kernel.org, erkin.bozoglu@xeront.com, "Sean Wang" <sean.wang@mediatek.com>, "DENG Qingfang" <dqfext@gmail.com>
>> Betreff: Re: [PATCH 4/5] arm: dts: mt7623: mux phy0 on Bananapi BPI-R2
>>
>>
>>
>> On 25/01/2023 17:54, Arınç ÜNAL wrote:
>>> On 25.01.2023 19:41, Frank Wunderlich wrote:
>>>> Am 20. Januar 2023 21:53:17 MEZ schrieb "Arınç ÜNAL" <arinc.unal@arinc9.com>:
>>>>
>>>>>          switch@1f {
>>>>>              compatible = "mediatek,mt7530";
>>>>>              reg = <0x1f>;
>>>>> @@ -199,11 +208,6 @@ ports {
>>>>>                  #address-cells = <1>;
>>>>>                  #size-cells = <0>;
>>>>>
>>>>> -                port@0 {
>>>>> -                    reg = <0>;
>>>>> -                    label = "wan";
>>>>> -                };
>>>>> -
>>>>
>>>> This will break existing userspace setups using wan as interface name.
>>>
>>> Yup. The OS being used will have to either rename the interface or adapt to the
>>> new name, eth1.
>>>
>>
>> In that case please explain in the commit message why this is needed. We are
>> trying really hard to not break userspace by a device-tree update.
> 
> I understand the reason why Arınç wants to change it:
> 
> currently all ports including wan-port are routed via dsa over gmac0 (trgmii), Arınç enables the second gmac and port5 of switch to route wan-traffic over the other gmac. But unfortunately there is no mainline-way to rename the gmac1 interface (eth1) to wan...this would be a way to not break the userspace.
> 
> i had an older patch for mtk_eth_soc [1] which do it at driver level for mediatek ethernet driver, but afair a generic way to (re)name interfaces in dts (vendor independ) was requested as the right way.
> 
> regards Frank
> 
> [1] https://github.com/frank-w/BPI-Router-Linux/commit/5088c9c81e90014622a950fca38cfb29e0421155

I tried this [0] before. It's not going to happen on mainline, and 
that's fine. I'll just clearly mention on the patch log that the 
interface name will change with this patch.

[0] 
https://lore.kernel.org/netdev/20220404114000.3549-1-arinc.unal@arinc9.com/t/#u

Arınç

Re: Aw: Re: [PATCH 4/5] arm: dts: mt7623: mux phy0 on Bananapi BPI-R2

[Arınç ÜNAL]

On 31.01.2023 19:39, Frank Wunderlich wrote:
> Hi
>> Gesendet: Dienstag, 31. Januar 2023 um 12:55 Uhr
>> Von: "Matthias Brugger" <matthias.bgg@gmail.com>
>> An: "Arınç ÜNAL" <arinc.unal@arinc9.com>, frank-w@public-files.de, "Rob Herring" <robh+dt@kernel.org>, "Krzysztof Kozlowski" <krzysztof.kozlowski+dt@linaro.org>
>> Cc: devicetree@vger.kernel.org, linux-arm-kernel@lists.infradead.org, linux-mediatek@lists.infradead.org, linux-kernel@vger.kernel.org, erkin.bozoglu@xeront.com, "Sean Wang" <sean.wang@mediatek.com>, "DENG Qingfang" <dqfext@gmail.com>
>> Betreff: Re: [PATCH 4/5] arm: dts: mt7623: mux phy0 on Bananapi BPI-R2
>>
>>
>>
>> On 25/01/2023 17:54, Arınç ÜNAL wrote:
>>> On 25.01.2023 19:41, Frank Wunderlich wrote:
>>>> Am 20. Januar 2023 21:53:17 MEZ schrieb "Arınç ÜNAL" <arinc.unal@arinc9.com>:
>>>>
>>>>>          switch@1f {
>>>>>              compatible = "mediatek,mt7530";
>>>>>              reg = <0x1f>;
>>>>> @@ -199,11 +208,6 @@ ports {
>>>>>                  #address-cells = <1>;
>>>>>                  #size-cells = <0>;
>>>>>
>>>>> -                port@0 {
>>>>> -                    reg = <0>;
>>>>> -                    label = "wan";
>>>>> -                };
>>>>> -
>>>>
>>>> This will break existing userspace setups using wan as interface name.
>>>
>>> Yup. The OS being used will have to either rename the interface or adapt to the
>>> new name, eth1.
>>>
>>
>> In that case please explain in the commit message why this is needed. We are
>> trying really hard to not break userspace by a device-tree update.
> 
> I understand the reason why Arınç wants to change it:
> 
> currently all ports including wan-port are routed via dsa over gmac0 (trgmii), Arınç enables the second gmac and port5 of switch to route wan-traffic over the other gmac. But unfortunately there is no mainline-way to rename the gmac1 interface (eth1) to wan...this would be a way to not break the userspace.
> 
> i had an older patch for mtk_eth_soc [1] which do it at driver level for mediatek ethernet driver, but afair a generic way to (re)name interfaces in dts (vendor independ) was requested as the right way.
> 
> regards Frank
> 
> [1] https://github.com/frank-w/BPI-Router-Linux/commit/5088c9c81e90014622a950fca38cfb29e0421155

I tried this [0] before. It's not going to happen on mainline, and 
that's fine. I'll just clearly mention on the patch log that the 
interface name will change with this patch.

[0] 
https://lore.kernel.org/netdev/20220404114000.3549-1-arinc.unal@arinc9.com/t/#u

Arınç

_______________________________________________
linux-arm-kernel mailing list
linux-arm-kernel@lists.infradead.org
http://lists.infradead.org/mailman/listinfo/linux-arm-kernel