Network interface doesn't come up with kernel > kernel-5.4.12

My wired network interface doesn’t come up with any kernel > kernel-5.4.12
I submitted Bugzilla #1798245, but there has been no response. The details are all in the bug report. I cannot upgrade my kernel to include the latest security fixes due to this problem. Please help.


Hi @ldaptester1, welcome to the forum. Please take a minute to go through the introductory posts in the #start-here category if you’ve not had a chance.

Ah, well, this is an odd one, especially given your statement “Strangely, I have another machine configured virtually identically (hardware and software) does not exhibit this behavior and has no problem with 5.4.13”

So, that indicates that it isn’t necessarily a kernel issue.

Can you please grab the updated live image from here, update the kernel there to the latest available and see if your network comes up? This will help us confirm whether it is an issue limited to your installed system (and its configuration):

Thank you for your prompt response. It will take me quite a few days to try the test you suggested. The machine in question is a headless server in a remote location.

But the kernel is certainly involved here. I boot with 5.4.12 and the network comes up. I reboot with 5.4.13 and the network does not come up. I immediately reboot with 5.4.12 and the network comes up. Nothing on disk has changed. No files, no configurations, no software, nothing.

Now 5.4.13 obviously works on other machines. I have a number of other machines using that kernel and later without problem. So it must be some subtle interaction between later kernels and the configuration of the problem machine. But I have checked all the relevant configurations I can think of, and also compared them to working machines. I’m still clueless.

It might be faster to track down the problem using the evidence. The error message:
Connection activation failed: No suitable device found for this connection (device lo not available because device is strictly unmanaged).
should give a clue as to where to look. What could cause such a message?

1 Like

I’m not sure to be honest. Does the output of nmcli differ in the two kernels? Does the identifier for the interface change, for example? On my machine lo is also unmanaged. (The wired interface is referred to as enp4s0.)

Yes. The output of the nmcli commands looks quite different between kernels 5.4.12 and 5.4.13. Look carefully at my bugzilla 1798245 where I have documented this. There it clearly shows that under 5.4.13, the system thinks that the wired network interface is unmanaged. Why this would be is the mystery. The ifcfg- file clearly says NM_CONTROLLED=yes. Under 5.4.12, the device name does change from “enp1s0” to “p49p1” as it always has, and as it does on other working machines I have.

I believe the reference to device “lo” is a mistake in the error message. In the case of the command, “nmcli device connect enp1s0”, we get a more appropriate message:
Error: Failed to add/activate new connection: Connection ‘enp1s0’ is not available on device enp1s0 because device is strictly unmanaged

So what changed between 5.4.12 and 5.4.13 that might affect all this? Is there a way to find out?

What are the requirements for an interface to be considered managed?

I was able to boot in F31-XFCE-x86_64-LIVE-20200206.iso which comes with kernel-5.4.17-200.fc31.x86_64 and while the default DHCP configuration did not work, I was able to manually set a static IP and activate the wired network interface.

I don’t know what new information this gives us. We knew the hardware was OK. Obviously, newer kernels don’t like some configuration detail on my hard drive. 5.4.12 and older have no problem. But what configuration detail is it?