[BUG] Xid 109 (CTX_SWITCH_TIMEOUT) Regression in 575 Drivers - World of Warcraft

Graphics / Linux Linux
1

Hello,

I am experiencing a consistent and reproducible driver crash with the 575 series drivers when playing World of Warcraft. This issue is a regression, as the game runs perfectly with the 570 series drivers.

Summary of Bug

  • Symptom: Xid 109: CTX SWITCH TIMEOUT crash.
  • Trigger: Loading a new zone in World of Warcraft.
  • Regression: Occurs on all 575.x drivers, does not occur on 570.x drivers.
  • Unusual Behavior: On crash, the display reverts to a video frame from ~1 minute prior to the event.

System Information

  • OS: EndeavourOS x86_64 (Arch Linux)
  • Kernel: 6.15.7-arch1-1
  • CPU: AMD Ryzen 7 9800X3D
  • GPU: NVIDIA GeForce RTX 5080
  • DE: KDE Plasma 6.4.3 (X11 Session)
  • Display: 3440x1440 @ 175 Hz

Driver Versions

  • Failing Driver Series: 575.xx.xx (The crash has occurred with all 575 packages released in Arch so far).
  • Last Known Good Driver: 570.153.02 (Arch packages below)
    • lib32-nvidia-utils-570.153.02-1
    • nvidia-open-dkms-570.153.02-2
    • nvidia-utils-570.153.02-2

Steps to Reproduce

  1. Launch World of Warcraft through Steam using GE-Proton10-8.
  2. Load into the game world, specifically the city of Dornogal.
  3. Travel to and enter the sub-zone “The Ringing Deeps”.
  4. The game will reliably crash during the loading screen for the new zone.

Relevant Log Entries

dmesg:

NVRM: Xid (PCI:0000:01:00): 109, pid=5607, name=WoW.exe, Ch 00000020, errorString CTX SWITCH TIMEOUT, Info 0xc02a

Xorg.0.log:

(WW) NVIDIA: Wait for channel idle timed out.

(Note: There is an ~600ms delay between the Xorg warning and the NVRM Xid error).

Additional Context

I have typically used Proton Experimental when testing this. I have Steam Overlay, Steam Input, and Game Recording disabled. Launch options are set to: gamemoderun prime-run %command%

Interestingly, I encountered the exact same symptoms (Xid 109 and display reversion to an old frame) about a year ago under a different Proton/Lutris configuration, although the crash could happen at any point during gameplay. At the time, I was incorrectly using GE-Proton as a runner, not using UMU, which was later integrated into Lutris, according the GE-proton’s github page.

The standard nvidia-bug-report.log.gz is attached. Please let me know if you need a verbose log or any further testing. Thank you!

nvidia-bug-report.log.gz (880.1 KB)

1 Like
2

My reproducible crashes in World of Warcraft appear to have stopped after following a suggestion from user @pphop (in another thread).

The workaround is to disable “Compute Effects” within WoW’s graphics settings.

For reference, here are the in-game descriptions for the setting:

Compute Effects: Controls the quality of Compute-based effects such as Volumetric Fog and some particle effects.

When Disabled: Volume Fog disabled. Compute-based particle collision disabled.

This seems to pinpoint the crashes to a specific function handled by the driver. I am no longer experiencing the issue after making this change.

Testing Environment:

  • GPU: NVIDIA GeForce RTX 5080
  • NVIDIA Driver: 575.64.05
  • OS: EndeavourOS x86_64
  • Kernel: 6.15.9
  • Platform: Steam with Proton Experimental (build from August 4th, 2025)