Fatal Error in GC (SuspendThread Loop failed)
available in future release
Log In
aGothLizard
Using Linux Mint 22.3 Cinnamon. Was fine October 30th, am now getting this issue today. Tried a multitude of fixes including rebooting, reseating USB devices and my headset, verifying files, etc. I'm using GE-Proton-12 which is what I was using prior.
lexibigcheese
i'm having this issue on the stable AND open beta, but AFTER i've joined a world!
it just keeps happening, including after rebooting!
木头木风
Native Windows 11 (no Proton). ~1 in 3 launches succeeds. Measurements:
ENV: Xeon E5-2698 v3 (16C/32T), 32GB, RTX 3080 (617.14), Win11 22631.4112, Unity 6000.0.67f1-DWR (VRChat.exe 6000.0.67.59584), launched via Steam/EAC. No antivirus (Defender realtime OFF), no mods or injected DLLs (WER module list only contains Steam overlay, NVIDIA drivers, Defender AMSI).
SYMPTOM: freezes ~13.5s after process start, then either recovers after 17-19s, or aborts with "Fatal error in GC - SuspendThread loop failed". VRChat.gc.log contains only that line. WER: GameAssembly.dll, exception 0x80000003 (Mono's deliberate breakpoint), offset 0xc3eb25.
This machine's WER archive holds 36 AppCrash entries for VRChat.exe since 2026-09-16.
NOTE: the process keeps showing the dialog until OK is clicked, so WER/AppCrash timestamps = click time, not crash time.
TIMING (sampled every 100-250ms from launch):
- t~13.4s: 15-19 threads created in one ~10ms batch (priority Idle, CPU 0), from ProcessThread.StartTime
- t~13.5-14.5s: window Not Responding, process CPU time frozen at 13.69s and flat for 30s+ => retry loop is sleeping, not spinning
- thread count at freeze: 165 -> 175-178
- success: logging resumes 17-19s later
- 2nd stall of 10-18s at "Found 141 time zones" -> "Loaded HRTF"
- failure: dialog appears ~30s after the freeze
- both stalls also occur in sessions that succeed, so they are a constant, not the trigger
FREQUENCY: ~25 launches, 4 succeeded / ~21 aborted, interleaved within minutes. Once in, stable (6h session). Always before login: the log never gets past "OvrLipSync Awake", so not account/network/server.
RULED OUT (each tested): reinstall + EAC repair; default audio device (VRChat only selects its mic at t~40s, after both stalls); Steam overlay (verified NOT injected via gameoverlay_renderer.txt); CPU affinity 0xFF and 0xFFFF confirmed applied, thread count unchanged; SteamVR running vs not; the older Unity 2022 client also crashes (since 2026-09-16).
Looks like a race in Mono/Boehm GC stop-the-world at startup. Nothing local changes the outcome, only timing. WER reports, output_log and gc.log available.
ΔSTRΔLIX
VRChat suddenly stopped launching on my PC. It opens to a white screen, becomes Not Responding, and then crashes with a SuspendThread error.
Specs:
CPU: Ryzen 7 5800XT
GPU: RTX 5070
OS: Windows 11
Error details:
Faulting module: GameAssembly.dll
Exception code: 0x80000003
I’ve tried reinstalling VRChat, updating drivers and Windows, clearing game data, testing beta and older versions, clean boots, and even resetting Windows. Nothing has fixed it.
VRChat used to work perfectly on this same PC. Any ideas what could be causing this?
Rexen²
I cleared my cached and while it hasn't fix the issue fully that seems to make the crash happen less.
StormRel
Merged in a post:
Fatal error in GC SuspendThread loop failed
King․ Ddraig
I open VRChat, and within 1-2 minutes I get this error pop up, and crashes to desktop, playing through Steam Proton on linux mint
Khloe_vr
The latest version of VRChat crashes with the error “SuspendThread loop failed,” I'm using Unity 6000.0.67f1-DWR_6158d48dc694.
On Linux Ubuntu 26.04.1
It crashes in GE-proton, Proton RTSP, and Proton Standard 11.0
Before the update, there were no problems with the game.
Photo Viewer
View photos in a modal
DerpyEagle
Was suffering this issue bad today on a 9950X3D2 with Arch, native Steam, latest kernel & drivers, with any versions of Proton. Couldn't consistently change worlds, open menus, or stay in a world for too long before getting the SuspendThread crash.
Disabling Esync and Fsync seemed to fix it for me. Set your VRC launch options to:
PROTON_NO_ESYNC=1 PROTON_NO_FSYNC=1 %command%
This will likely cause a performance hit but I had no crashes in 3 hours after setting this when previously I would crash within minutes. Hoping it helps someone else. Also hoping we get whatever release this fix is coming to soon...
arxmodeus
DerpyEagle This allowed me to use VRChat long enough to go to other worlds and muck around a bit at least. Will give longer test later. Many thanks if this is the silver bullet!
Ｆｅｌｌ
DerpyEagle This inspired me to check if NTSync is enabled. NTSync is the newer in-kernel replacement for ESync/FSync. It wasn't enabled.
To check:
ls -l /dev/ntsync
(if this file exists it's enabled)To enable it:
sudo modprobe ntsync
To enable it on future reboots:
echo ntsync | sudo tee /etc/modules-load.d/ntsync.conf
With this, I could play for over an hour without issues.
Ｆｅｌｌ
I don't actually think it's related to X3D chips anymore. We're getting it consistently on a 3900X (Mint Cinnamon) but not at all on a 5800X3D (Arch KDE Plasma)
AlexTaco-Anton
This issue appeared firstly in the
unity-6
beta branch and then continued to progress in the open-beta
branch after merging the unity-6
beta branch after announcing it was ready for merging.Distribution doesn't matter in this case because fault is the same on different setups and use cases.
I can only suggest (and guess) using the
git bisect
5 patches before merging unity-6
into open-beta
and 2 patches after merging.Distribution:
Debian Forky/sid
.Graphical environment: KDE on Wayland.
Steam installation: Non-flatpak.
Proton: 11 RTSP Fork.
SLR: steamrt4.
Architecture: AMD64
All information about the fault has already been sent by the Unity crash handler.
The issue is already in the
release
branch.Load More
→