Qubes OS 4.3.1 - System Freezes When Launching Qubes

User Support
1

Hello,
Since upgrading from Qubes OS 4.2 to 4.3.1, I’ve been getting random system freezes. The freezes most often happen when launching a qube, though the behavior is inconsistent.

For example, starting sys-net will occasionally cause the entire system to become frozen (except for the cursor), while at other times the qube starts and functions normally, but the system freezes again after a while of casual use. I’ve also occasionally happen to encounter freezes outside the qubes, as, for example, running Xfce Terminal using interface in dom0. Interestingly, I can almost always boot the qubes normally if I first launch the Xfce Terminal in dom0, input any command in it and then start the fedora-43-xfce template.

Primarly, after decrypting the LUKS disk, the system would proceed to a black screen that only had a working cursor and no user login prompt. I was able to work around this by adding the qubes.skip_autostart parameter in the GRUB menu.

What I have tried (with no success):

  • At first, I suspected the Wi-Fi card (RTL8822BE) might be causing the issue. However, replacing it with a new one (AX210) did not resolve the problem, and the freezes continued.
  • I’ve reinstalled the system 3 times, including installation with the newer kernel version.
  • I’ve tried changing the qubes template from fedora-43-xfce to debian-13-xfce.

I haven’t been able to point a consistent pattern or a reliable way to reproduce the issue. Any kind of help is appreciated.

Specification:
Model: Thinkpad A485,
CPU: Ryzen 5 Pro 2500U,
RAM: 16GB DDR4,
Kernel Version: 7.0.9-1 (tried it with 6.18.31-1, same result)

2

Could you observe your system RAM/swap before/during the freeze with a tool such as htop, btop or free -h in dom0?

Also, logs from the journal in dom0 would be useful. Trigger a freeze and after a reboot, see whether there’s anything interesting at the end of the journal from this command: sudo journalctl -b -1. Feel free to post it here if there’s no information you’d rather not post.

3

Found repetitive errors at the end of the journalctl. Seems like the issue comes from the memory swap.

dom0 kernel: get_swap_device: Bad swap offset entry 3fffffffffffe
dom0 kernel: get_swap_device: Bad swap offset entry 3fffffffffffe
dom0 kernel: get_swap_device: Bad swap offset entry 3fffffffffffe
dom0 kernel: get_swap_device: Bad swap offset entry 3fffffffffffe
dom0 kernel: get_swap_device: Bad swap offset entry 3fffffffffffe
dom0 kernel: get_swap_device: Bad swap offset entry 3fffffffffffe
dom0 kernel: get_swap_device: Bad swap offset entry 3fffffffffffe
4

Errors with swap devices are, in my experience, one of the two:

  1. Your memory is so full that everything’s crashing down
  2. Actual errors with the swap device