[PATCH 2/2] ramips: mt7621-dts: we420223-99: mux phy0->gmac1

[Harm Berntsen]

This gives each port an individual link to the CPU. The advantage of
this is that it can now route packets faster between the ports (before
the CPU only had a single 1Gb link to the switch that has to be shared
between both ports. Another advantage is that in Linux 5.10 you can now
bridge a VLAN to a non-vlan port. Without this patch, you're not getting
any data across the bridge. That is fixed in Linux 5.15 but is still
handled by the CPU in any case. So therefore this patch is advantageous
in all cases except for when you need the device as a simple switch
without VLANs. For that case it's better to revert this and the switch
hardware will forward traffic without bothering the CPU.

Signed-off-by: Harm Berntsen <git at harmberntsen.nl>
---
 .../ramips/dts/mt7621_arcadyan_we420223-99.dts   | 16 ++++++++++++----
 1 file changed, 12 insertions(+), 4 deletions(-)

diff --git a/target/linux/ramips/dts/mt7621_arcadyan_we420223-99.dts b/target/linux/ramips/dts/mt7621_arcadyan_we420223-99.dts
index f68d79af15..88e823e382 100644
--- a/target/linux/ramips/dts/mt7621_arcadyan_we420223-99.dts
+++ b/target/linux/ramips/dts/mt7621_arcadyan_we420223-99.dts
@@ -188,16 +188,24 @@
 
 &switch0 {
 	ports {
-		port at 0 {
-			status = "okay";
-		};
-
 		port at 1 {
 			status = "okay";
 		};
 	};
 };
 
+&gmac1 {
+	status = "okay";
+	label = "lan0";
+	phy-handle = <&ethphy0>;
+};
+
+&mdio {
+	ethphy0: ethernet-phy at 0 {
+		reg = <0>;
+	};
+};
+
 &pcie {
 	status = "okay";
 };
-- 
2.37.2