Indeed, I have had a dual-socket EPYC Zen4 machine (gigabyte mobo) that
ran Ubuntu 24.04 just fine with both the 6.8 kernel and 6.11 HWE kernel.
After installing Ubuntu 25.04 GA, it crashes during the bootup (at
different points, but never gets to finish the boot). One of the dmesg
warnings I managed to capture immediately before the crash was this:

"genirq: Flags mismatch ira 0. 00000000 (0000:06:00.1) vs. 00215a00
(timer)"

I do know that I have one faulty (Samsung 980 PRO) nvme SSD that
occasionally produces correctable PCIe AER evnets, but keeps working ok.
I did disable AER and ACS in the BIOS and additionally set both
pci=noaer and   pci_aspm=off at the boot time, but still ended up with
the same crash.

For now I just installed the Oracle Linux 9 (with latest UEK8 kernel
that's Linux 6.12) and it works ok, no hardware errors logged in dmesg
or ras-daemon-ctl MCE logs either.

Also, the beta version of Ubuntu 25.04 (don't remember from how many
weeks ago) worked ok too.

Edit: I may have used the boot-time GRUB v2 settings wrong. First I set
the parameters mentioned above (including mce=off) as "set param=value"
on the grub boot-time command line and maybe that didn't work. When I
booted up from the Ubuntu 25.04 GA installer USB disk and added these 3
parameters after the "linux /casper/vmlinuz" (but before the "---") then
the installer booted up OK.

-- 
You received this bug notification because you are a member of Ubuntu
Bugs, which is subscribed to Ubuntu.
https://bugs.launchpad.net/bugs/2106553

Title:
  Epyc Genoa system unable to boot starting with Kernel 6.14.0

To manage notifications about this bug go to:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2106553/+subscriptions


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

Reply via email to