Leap 16.0 freezes / AMDGPU pageflip timeouts

English Hardware
1

I’m running openSUSE Leap 16.0 on an HP ProDesk with an AMD Raven Ridge APU (Vega 11, amdgpu driver) and KDE Plasma 6.4.2. The system repeatedly hard-freezes (screen frozen, keyboard/mouse dead), requiring reboot. However, SSH access still works during the freeze, so the kernel is not fully dead.

Hardware / graphics

  • GPU: AMD Raven Ridge Vega 11 (amdgpu)
  • Kernel: 6.12.0-160000.5-default (also tested newer kernel)
  • Desktop: KDE Plasma 6.4.2
  • Display: tried X11 and Wayland
  • Two monitors (Dell E198WFP 1440x900)
  • Outputs available: 2× DisplayPort + 1× VGA
  • Currently using VGA + DP (no HDMI available)

Symptoms

Freezes occur randomly during use. After reboot, logs show AMDGPU display errors. Under Wayland KWin reports a pageflip timeout. Under X11, restarting the display manager fails/hangs.

Key log messages

From kernel journal:

[drm:amdgpu_dm_atomic_check [amdgpu]] *ERROR* [CRTC:xx:crtc-x] hw_done or flip_done timed out

From Wayland session:

kwin_wayland_drm: Pageflip timed out! This is a bug in the amdgpu kernel driver

During freezes, systemd watchdog reports soft lockups:

watchdog: BUG: soft lockup - CPU#0 stuck for 414s! [kworker/u32:8:175]

Also saw network watchdog timeouts (likely secondary):

r8169 ... NETDEV WATCHDOG: transmit queue 0 timed out

What I’ve tried

  • Switched between Plasma X11 and Plasma Wayland → freezes still occur.
  • Tested kernel boot parameters:
    • amdgpu.aspm=0 → no improvement
    • amdgpu.dc=0 → system fails to boot (never reaches Plymouth splash)
  • Attempted recovery during freeze:
    • sudo systemctl restart display-manager → service times out / fails; Xorg remains stuck
    • SSH remains available, so machine is still running but graphics stack wedged
  • Noted that this never happened on Leap 15.x on same hardware.

Current suspicion

This appears to be an amdgpu_dm display/pageflip hang (multi-monitor or VGA/DP output related), causing KWin/Xorg to block in kernel DRM calls, eventually leading to watchdog soft lockups.

Looking for advice on

  • known Leap 16 / kernel 6.12 amdgpu pageflip timeout issues on Raven Ridge
  • recommended kernel parameters/workarounds
  • whether VGA output path could be triggering this
  • best kernel/mesa repo to try for a fix

(Logs can be provided if needed: journalctl -b -1 -k | grep -i amdgpu, plus full dmesg around the freeze.)

2

Is it possible for you to avoid VGA entirely, even if it means testing with a single DP-connected monitor? The current monitor setup may be triggering amdgpu DC pageflip timeouts and compositor hangs you are observing.

It may also worth be testing behaviour with a newer stable kernel.

sudo zypper ar -f https://download.opensuse.org/repositories/Kernel:/stable:/Backport/16.0/Kernel:stable:Backport.repo

Examine the available kernel versions…

zypper refresh
zypper search -sx kernel-default
zypper install kernel-default-<listed-version-to-install>

Reboot into the new kernel after installation.

If it doesn’t help, you can simply select your previous kernel from the boot menu to revert.

3

Thank you. I am trying with just the one monitor on the DP. I think I’ve had one freeze, but I wasn’t in a position to grab the logs, so I can’t confirm that yet.

If it does happen again without running over VGA then I will try a newer kernel - thank you for the suggestion and the instructions.

4

This topic was automatically closed 30 days after the last reply. New replies are no longer allowed.