KDE popped “Desktop effects were restarted due to a graphics reset” on Manjaro Plasma Wayland (hybrid Intel Arc + NVIDIA). That notification is a symptom, not a diagnosis. The useful work is in the journal and the i915 error state — before reboot, because /sys/class/drm/card*/error is gone after.

1
2
3
4
5
6
7
8

# Kernel side: DRM / GPU / hang / reset
journalctl -k -b | grep -Ei 'drm|gpu|i915|xe|nvidia|reset|hang|fault'

# Compositor / shell reaction
journalctl -b | grep -Ei 'kwin|plasmashell|GL_CONTEXT_LOST|graphics reset|device lost'

# After a hang, capture the crash dump Intel wants (cardN from the dmesg line)
sudo cat /sys/class/drm/card1/error | bzip2 > i915-error.bz2

What showed up: fence timeouts on kwin_wayland, then an Intel hang and chip reset — not an NVIDIA reset:

1
2
3
4

Fence expiration time out i915-0000:00:02.0:kwin_wayland[…]:…
i915 0000:00:02.0: [drm] GPU HANG: ecode 12:0:00000000
i915 0000:00:02.0: [drm] GPU error state saved to /sys/class/drm/card1/error
i915 0000:00:02.0: [drm] GT0: Resetting chip for stopped heartbeat on rcs0

Same second, fences also expired for plasmashell’s QSGRenderThread and wined3d_cs. KWin/Plasma then lost GL and recovered:

1
2
3

kwin_wayland: A graphics reset not attributable to the current GL context occurred.
kwin_wayland: GL_CONTEXT_LOST in context lost
plasmashell: Graphics device lost, cleaning up scenegraph and releasing RHI

~8s later a second hang named Wine:

1
2

i915 … GPU HANG: ecode 12:1:85dffdfb, in wined3d_cs […]
i915 … wined3d_cs[…] context reset due to GPU hang

Wine/Proton may have stressed the GPU, but the failure is in the Intel Linux graphics stack under Wayland: kernel i915, GuC/DMC/HuC firmware, Mesa, power management, KWin’s DRM path. Distro choice only changes package versions of that same upstream stack — Kubuntu does not ship a different Intel driver.

Report path and reading: how to file i915 bugs, GPU error state, drm/i915 issues, i915 docs. Hybrid session context on this machine: Ubuntu, Manjaro, and the Linux Desktop.