Dear Forum Member
I’m noticing since (not 100% sure) kernel 5.18.8 issues with my RTL8168H PHY on MSI X570 Gaming Egde WIFI mainboard.
Up to Kernel 5.18.14 the PHY were form time to time unable to establish a link. Since Kernel 5.18.15 the link cannot established at all. Also now with Kernel 5.19.1 and the latest r8168 driver from August 11th.
Cross checking and changing beginning with Kernel 5.18.15 the r8168 driver to the built in r8169 driver does recognizes the PHY but did not start at the interface. Meaning that there is no attempt at all to start up the device and try to establish a link.
Both driver show the interface:
enp39s0: flags=4099<UP,BROADCAST,MULTICAST> mtu 1500
ether 42:ac:51:0b:b9:c6 txqueuelen 1000 (Ethernet)
RX packets 0 bytes 0 (0.0 B)
RX errors 0 dropped 0 overruns 0 frame 0
TX packets 133 bytes 34686 (33.8 KiB)
TX errors 0 dropped 0 overruns 0 carrier 0 collisions 0
device interrupt 101 base 0x1000
I’m noticing the the loaded mod’s are different for r8168 and r8169.
lsmod for r8168 shows the following:
r8168 225280 0
r8169 102400 0
mdio_devres 16384 1 r8169
libphy 172032 3 r8169,mdio_devres,realtek
r8168 225280 0
dmesg shows for r8168 this:
[ 4.071687] r8168 Gigabit Ethernet driver 8.050.03-NAPI loaded
[ 4.071705] r8168 0000:27:00.0: enabling device (0000 → 0003)
[ 4.084675] r8168: This product is covered by one or more of the following patents: US6,570,884, US6,115,776, and US6,327,625.
[ 4.086704] r8168 Copyright (C) 2022 Realtek NIC software team email@example.com
[ 4.231631] r8168 0000:27:00.0 enp39s0: renamed from eth0
[ 21.616882] r8168: enp39s0: link up
[ 22.640612] r8168: enp39s0: link down
[ 44.976806] r8168: enp39s0: link up
[ 46.000619] r8168: enp39s0: link down
[ 68.337117] r8168: enp39s0: link up
[ 69.361398] r8168: enp39s0: link down
but no message at all for r8169.
As the wiki say the driver has issues with IPv6 it was disabled specifically for the interface by creating the file 40-ipv6.conf in /usr/lib/modprode.d and added according to the wiki:
net.ipv6.conf.enp39s0.disable_ipv6 = 1
Does anyone have an idea how to further dissect this issue?