After the core 158 update, the green network card says goodbye after a few days

[ipfireuser87 (Stefan Schembera)]

I have noticed since today that my green nic has been failing since the core update 158

I suddenly had no more internet or no access to the firewall that the green network card said goodbye to

the following error message in the log

after the core 158 update i restarted it until today it ran without problems and now this serious problem has appeared

I think the firewall ran for 3-4 days without problems and now this error

16:27:40	kernel:	TDH <0>
16:27:40	kernel:	TDT <2>
16:27:40	kernel:	next_to_use <2>
16:27:40	kernel:	next_to_clean <0>
16:27:40	kernel:	buffer_info[next_to_clean]:
16:27:40	kernel:	time_stamp <1096c7bf8>
16:27:40	kernel:	next_to_watch <0>
16:27:40	kernel:	jiffies <1096c8480>
16:27:40	kernel:	next_to_watch.status <0>
16:27:40	kernel:	MAC Status <40080083>
16:27:40	kernel:	PHY Status <796d>
16:27:40	kernel:	PHY 1000BASE-T Status <3800>
16:27:40	kernel:	PHY Extended Status <3000>
16:27:40	kernel:	PCI Status <10>
16:27:42	kernel:	e1000e 0000:00:19.0 green0: Reset adapter unexpectedly
16:27:46	kernel:	e1000e: green0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: None
16:27:48	kernel:	e1000e 0000:00:19.0 green0: Detected Hardware Unit Hang:
16:27:48	kernel:	TDH <0>
16:27:48	kernel:	TDT <3>
16:27:48	kernel:	next_to_use <3>
16:27:48	kernel:	next_to_clean <0>
16:27:48	kernel:	buffer_info[next_to_clean]:
16:27:48	kernel:	time_stamp <1096c8b58>
16:27:48	kernel:	next_to_watch <0>
16:27:48	kernel:	jiffies <1096c8dc2>
16:27:48	kernel:	next_to_watch.status <0>
16:27:48	kernel:	MAC Status <40080083>
16:27:48	kernel:	PHY Status <796d>
16:27:48	kernel:	PHY 1000BASE-T Status <3800>
16:27:48	kernel:	PHY Extended Status <3000>
16:27:48	kernel:	PCI Status <10>
16:27:50	kernel:	e1000e 0000:00:19.0 green0: Detected Hardware Unit Hang:
16:27:50	kernel:	TDH <0>
16:27:50	kernel:	TDT <3>
16:27:50	kernel:	next_to_use <3>
16:27:50	kernel:	next_to_clean <0>
16:27:50	kernel:	buffer_info[next_to_clean]:
16:27:50	kernel:	time_stamp <1096c8b58>
16:27:50	kernel:	next_to_watch <0>
16:27:50	kernel:	jiffies <1096c9040>
16:27:50	kernel:	next_to_watch.status <0>
16:27:50	kernel:	MAC Status <40080083>
16:27:50	kernel:	PHY Status <796d>
16:27:50	kernel:	PHY 1000BASE-T Status <3800>
16:27:50	kernel:	PHY Extended Status <3000>
16:27:50	kernel:	PCI Status <10>
16:27:52	kernel:	e1000e 0000:00:19.0 green0: Detected Hardware Unit Hang:

so it runs in a continuous loop

after restarting the firewall, the log entries continue

e1000e 0000:00:19.0 green0: Reset adapter unexpectedly
16:53:38	kernel:	e1000e: green0 NIC Link is Up 1000 Mbps Full Duplex, Flow Control: None
16:59:02	kernel:	e1000e 0000:00:19.0 green0: Detected Hardware Unit Hang:
16:59:02	kernel:	TDH <6c>
16:59:02	kernel:	TDT <96>
16:59:02	kernel:	next_to_use <96>
16:59:02	kernel:	next_to_clean <6a>
16:59:02	kernel:	buffer_info[next_to_clean]:
16:59:02	kernel:	time_stamp <100064f61>
16:59:02	kernel:	next_to_watch <6c>
16:59:02	kernel:	jiffies <1000651c0>
16:59:02	kernel:	next_to_watch.status <0>
16:59:02	kernel:	MAC Status <40080083>
16:59:02	kernel:	PHY Status <796d>
16:59:02	kernel:	PHY 1000BASE-T Status <3800>
16:59:02	kernel:	PHY Extended Status <3000>
16:59:02	kernel:	PCI Status <10>
16:59:04	kernel:	e1000e 0000:00:19.0 green0: Detected Hardware Unit Hang:
16:59:04	kernel:	TDH <6c>
16:59:04	kernel:	TDT <96>
16:59:04	kernel:	next_to_use <96>
16:59:04	kernel:	next_to_clean <6a>
16:59:04	kernel:	buffer_info[next_to_clean]:
16:59:04	kernel:	time_stamp <100064f61>
16:59:04	kernel:	next_to_watch <6c>
16:59:04	kernel:	jiffies <100065400>
16:59:04	kernel:	next_to_watch.status <0>
16:59:04	kernel:	MAC Status <40080083>
16:59:04	kernel:	PHY Status <796d>
16:59:04	kernel:	PHY 1000BASE-T Status <3800>
16:59:04	kernel:	PHY Extended Status <3000>
16:59:04	kernel:	PCI Status <10>
16:59:06	kernel:	e1000e 0000:00:19.0 green0: Detected Hardware Unit Hang:

the internet and access to the firewall are currently running without problems the question is how long

The hanging entries of the green network card

is only since the core 158 update before these were not available

the type of network cards is
Intel® 82579LM and 82574L,
2x Gigabit Ethernet LAN ports

many people have the same problem should be a driver problem

e1000 driver hang

[pmueller (Peter Müller)]

Hi,

thanks for your detailed post.

Yes, this is a known driver problem indeed. We hope to solve it by updating to Linux kernel 5.10.x with the upcoming Core Update, but I am not an expert when it comes to NIC drivers at all, and vaguely recall Arne, our hardware/kernel expert, saying kernel 5.10.x won’t solve it. However, that was like half a year ago.

Although it comes with performance impacts, this command should solve this NIC issue:

ethtool -K eth0 gso off gro off tso off 

Please refer to this post on information for how to run it automatically every time IPFire boots up.

Thanks, and best regards,
Peter Müller