kernel: [90070.981674] e1000e 0000:00:1f.6 eno2: Detected Hardware Unit Hang:

fds2610

New Member
Aug 25, 2025
4
0
1
i encounter a strange problem since a couple of weeks. After two years of running fine, my Lenovo ThinkCentre suddenly starts displaying an error after a while (one or more days of operation) and stops being accessible via network:

2025-10-01T08:55:47.939730+02:00 pvehost2 kernel: [90070.981674] e1000e 0000:00:1f.6 eno2: Detected Hardware Unit Hang:
2025-10-01T08:55:47.941867+02:00 pvehost2 kernel: [90070.981674] TDH <b6>
2025-10-01T08:55:47.941876+02:00 pvehost2 kernel: [90070.981674] TDT <c8>
2025-10-01T08:55:47.941877+02:00 pvehost2 kernel: [90070.981674] next_to_use <c8>
2025-10-01T08:55:47.941878+02:00 pvehost2 kernel: [90070.981674] next_to_clean <b5>
2025-10-01T08:55:47.941878+02:00 pvehost2 kernel: [90070.981674] buffer_info[next_to_clean]:
2025-10-01T08:55:47.942614+02:00 pvehost2 kernel: [90070.981674] time_stamp <10559c7da>
2025-10-01T08:55:47.942627+02:00 pvehost2 kernel: [90070.981674] next_to_watch <b6>
2025-10-01T08:55:47.942628+02:00 pvehost2 kernel: [90070.981674] jiffies <10559d2c0>
2025-10-01T08:55:47.942632+02:00 pvehost2 kernel: [90070.981674] next_to_watch.status <0>
2025-10-01T08:55:47.942632+02:00 pvehost2 kernel: [90070.981674] MAC Status <40080083>
2025-10-01T08:55:47.942633+02:00 pvehost2 kernel: [90070.981674] PHY Status <796d>
2025-10-01T08:55:47.942634+02:00 pvehost2 kernel: [90070.981674] PHY 1000BASE-T Status <3800>
2025-10-01T08:55:47.942634+02:00 pvehost2 kernel: [90070.981674] PHY Extended Status <3000>
2025-10-01T08:55:47.942637+02:00 pvehost2 kernel: [90070.981674] PCI Status <10>
2025-10-01T08:55:49.920173+02:00 pvehost2 kernel: [90072.965635] e1000e 0000:00:1f.6 eno2: Detected Hardware Unit Hang:

... and so on...

Does it hint, the hardware network interface is broken and needs to be replaced? after hours of working properly? any ideas welcome.
 
Hi!

This is probably a known problem (also see kernel bug) with the e1000e driver. Have you tried turning off TCP segmentation offloading and generic segmentation offloading?

Code:
ethtook -K <interface> tso off gso off