PC freezes when I unplug and plug back in my network cable

Hello,

A very strange problem that I had never encountered before…
My configuration:

  • Dell Latitude 5400
  • Fedora 40 (but the problem existed under 39 as well)
  • Linux Fedora 6.8.9-300.fc40.x86_64 #1 SMP PREEMPT_DYNAMIC Thu May 2 18:59:06 UTC 2024 x86_64 GNU/Linux

$ lspci | grep -i net
00:14.3 Network controller: Intel Corporation Cannon Point-LP CNVi Wireless-AC
00:1f.6 Ethernet controller: Intel Corporation Ethernet Connection (6) I219-LM (rev 30)

When I unplug the network cable from my machine or turn off the device in front of it (a UDP acquisition card), my machine freezes… I then observe in ‘top’ a process of the type kworker/0:x that consumes 100% of the CPU. Very quickly, my machine becomes uncontrollable (no more keyboard, no more mouse, frozen display), and I have only one solution: a HARD reboot, restarting by holding down the power button for several seconds.
I tried with other devices and it seems to do the same… as if I couldn’t hotplug a network device.

It’s extremely disabling, and I don’t know what I can do…

Thanks in advance for your help.

Olivier

1 Like

I have the same problem with Fedora 39 on a Thinkpad T450s and T470s. I have reported this against kernel but not gotten any attention so far. May have to reassign the bug report to iwlwifi or so.
I experience the issue when I undock my Thinkpads from the docking station because that unplugs ethernet and I am expecting a switch over to wifi. I think I started observing this behavior around 3 to 4 months ago.

I’ll look for my bug report later and will link it here.

EDIT: Here is the bug I reported: https://bugzilla.redhat.com/show_bug.cgi?id=2269061

1 Like

Hello,

I have the same problem on my laptop after upgrading:

  • Dell Latitude 7480 - Fedora 40 (6.8.9-300.fc40.x86_64) , it was fine on Fedora 39.
  • Ethernet controller: Intel Corporation Ethernet Connection (4) I219-LM (rev 21

A snippet from the journal right before the hard reboot

I’m not sure what I’m seeing in you journal snippet is actually the same issue. Can we see more please?

Hi there,
I installed a 6.1 kernel and it looks like the problem is fixed… I’ll tell you more tomorrow.

While this is workaround, it can’t be the solution. If it works with an older kernel but not with recent kernels, then this is a regression that needs to be documented, reported and fixed. Help us collect logs and report it on the bugtracker.

Yes, I totally agree with you…
What should I do to provide more relevant and accurate logs ? journalctl, dmesg… ? Also I need to make sure I can find them after a hard reboot so I can post them here…
thank you for your help.

Hello I have attached more screenshots from the journal (Sorry I don’t know the right way to attach logs).
Thanks for your help.



Logs should be edited to remove extraneous detail, then posted as pre-formatted text using the </> button from the top line of the text entry box.

Journalctl captures enormous detail, so it is important to use the filter options to eliminate irrelevant entries. It can also generate very long lines. Once you have a good set of filter options, pipe the output to cat to get complete (wrapped) lines for forum posts. For your issue, I would start with journalctl -b -1 --no-hostname -g e1000e -g BUG | cat.

Searching for e1000e BUG searching while atomic, finds a trail of similar reports: https://bugzilla.redhat.com/show_bug.cgi?id=2276852 and https://bugzilla.redhat.com/show_bug.cgi?id=2276325 and https://patchwork.ozlabs.org/project/intel-wired-lan/patch/20240417190320.3159360-1-vitaly.lifshits@intel.com/.

Use journalctl output to check more details to confirm that you are seeing exactly the same bug.

1 Like