Hello,

I installed Armbian Hirsute xfce bleeding edge kernel 5.13.y on my RockPro64 4GB and updated it right after that. After restarting after the update, ethernet doesn't work anymore. (Wifi connecting icon is shown and no connection to ethernet) dsmesg says that the link is up. 
armbianmonitor: https://paste.ubuntu.com/p/9Yz8SJVrzS/
apt/history.log: https://paste.ubuntu.com/p/GQbfZp3Xch/
For apt history only the last two operations seem relevant. I assume the previous ones are made by whoever/whatever created the image.

So eth got lost somewhere on move from 5.13.y to 5.14.y

Could be here: https://github.com/armbian/build/pull/3125

or upstream. Dont know

I haven't noticed any eth related problems on 5.13.y and RK3399, but when moving to 5.14.y I also found that eth no longer worked. I corrected the issue by reverting the following commit: https://github.com/torvalds/linux/commit/2d26f6e39afb88d32b8f39e76a51b542c3c51674

 

I understand its not a true fix. This was tested on the NanoPC-T4.

 

--- a/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c	2021-09-12 03:01:00.000000000 -0400
+++ b/drivers/net/ethernet/stmicro/stmmac/dwmac-rk.c	2021-09-14 07:19:24.402736613 -0400
@@ -21,6 +21,7 @@
 #include <linux/delay.h>
 #include <linux/mfd/syscon.h>
 #include <linux/regmap.h>
+#include <linux/pm_runtime.h>

 #include "stmmac_platform.h"

@@ -1528,6 +1529,9 @@
 		return ret;
 	}

+	pm_runtime_enable(dev);
+	pm_runtime_get_sync(dev);
+
 	if (bsp_priv->integrated_phy)
 		rk_gmac_integrated_phy_powerup(bsp_priv);

@@ -1536,9 +1540,14 @@

 static void rk_gmac_powerdown(struct rk_priv_data *gmac)
 {
+	struct device *dev = &gmac->pdev->dev;
+
 	if (gmac->integrated_phy)
 		rk_gmac_integrated_phy_powerdown(gmac);

+	pm_runtime_put_sync(dev);
+	pm_runtime_disable(dev);
+
 	phy_power_on(gmac, false);
 	gmac_clk_enable(gmac, false);
 }

 

https://github.com/armbian/build/pull/3149

Checkout PR, build image and test. Let me know if this does the trick.

I found this on lore.kernel.org: https://lore.kernel.org/linux-rockchip/YS0v9UbzoHkiU9om@sashalap/T/#t

Apparently they're looking into a proper fix, but nothing has come of it yet.

Confirm this patch works for me on nanopc t4 on kernel 5.14.5, applied  and built locally.