Affected by the same issue on a different machine and kernel.

## Environment

- **Ubuntu:** 24.04.4 LTS (Noble)
- **Kernel:** 7.0.0-28-generic
- **NIC:** Intel Corporation [8086:550a] rev 20 (I219-V), PCI 0000:00:1f.6
- **Driver:** e1000e, firmware 1.1-4

## Symptoms

Same Hardware Unit Hang pattern. The `e1000_watchdog_task` hogged CPU
for >10000us an increasing number of times (4 → 5 → 7 → 11 occurrences)
before the TX ring entered a permanent hang state. After the hang,
`tx_packets` stayed at 0 and `tx_dropped` climbed continuously. The NIC
did not recover — driver reload (`rmmod`/`modprobe`) did not clear the
hang. Only a full reboot restored functionality.

The hang occurred without any suspend/resume cycle, under normal network
load on a wired connection.

## Workaround (confirmed)

Disabling TSO/GSO/GRO offloading and adding `SmartPowerDownEnable=1` as
a module parameter prevented recurrence after reboot:

```bash
nmcli connection modify <uuid> \
  ethtool.feature-tso off \
  ethtool.feature-gso off \
  ethtool.feature-gro off

echo 'options e1000e SmartPowerDownEnable=1' | sudo tee 
/etc/modprobe.d/e1000e.conf
sudo update-initramfs -u
```

## Note

The device ID `[8086:550a]` (Meteor Lake / Arrow Lake era I219-V
stepping) may differ from the original report. Flagging in case this
affects triage or patch applicability.

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2121008

Title:
  e1000e 0000:00:1f.6 eno1: Detected Hardware Unit Hang:

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2121008/+subscriptions


-- 
ubuntu-bugs mailing list
[email protected]
https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs

Reply via email to