Arma Reforger - Random amdgpu system crashes

vkd3dopen

HansKristian-Work/vkd3d-proton#2487 · opened 2025-06-01 by DanielGaaA · updated 2026-01-01 · 17 comments · github

DDanielGaaA 2025-06-01 github

Hello while playing Arma Reforger I am getting random amdgpu crashes that lock up system. There is small chance system will recover but most of the time I need to do hard reset. I was having this crashes for several months so drivers / proton / vkd3d changes over the time but crashes continues. The crash is with GE-Proton10-3. In kernel log it always mention vkd3d_queue caused the crash.

amdgpu 0000:0c:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:7 pasid:32806)
amdgpu 0000:0c:00.0: amdgpu:  in process enfMain pid 41900 thread vkd3d_queue pid 41970

I managed to get umr waves logs as mentioned in different issue: https://github.com/HansKristian-Work/vkd3d-proton/issues/2441#issuecomment-2823536562

Software information

Arma Reforger

System information

System:
  Host: username Kernel: 6.14.6-1-default arch: x86_64 bits: 64 compiler: gcc
    v: 14.2.1
  Desktop: KDE Plasma v: 6.3.5 tk: Qt v: N/A wm: kwin_wayland dm: SDDM
    Distro: openSUSE Tumbleweed 20250522
CPU:
  Info: 6-core model: AMD Ryzen 5 5600X bits: 64 type: MT MCP arch: Zen 3+
    rev: 0 cache: L1: 384 KiB L2: 3 MiB L3: 32 MiB
  Speed (MHz): avg: 4092 min/max: 550/4654 boost: enabled cores: 1: 4092
    2: 4092 3: 4092 4: 4092 5: 4092 6: 4092 7: 4092 8: 4092 9: 4092 10: 4092
    11: 4092 12: 4092 bogomips: 88632
  Flags: avx avx2 ht lm nx pae sse sse2 sse3 sse4_1 sse4_2 sse4a ssse3 svm
Graphics:
  Device-1: Advanced Micro Devices [AMD/ATI] Navi 21 [Radeon RX 6800/6800 XT
    / 6900 XT] driver: amdgpu v: kernel arch: RDNA-2 pcie: speed: 16 GT/s
    lanes: 16 ports: active: DP-2,HDMI-A-1 empty: DP-1,DP-3,Writeback-1
    bus-ID: 0c:00.0 chip-ID: 1002:73bf
  Display: wayland server: X.org v: 1.21.1.15 with: Xwayland v: 24.1.6
    compositor: kwin_wayland driver: X: loaded: modesetting unloaded: vesa
    alternate: fbdev dri: radeonsi gpu: amdgpu d-rect: 4480x2520 display-ID: 0
  Monitor-1: DP-2 pos: bottom-l model: VG27AQ1A res: 2560x1440 hz: 170
    dpi: 109 diag: 685mm (27")
  Monitor-2: HDMI-A-1 pos: top-right model: Samsung S24F350 res: 1920x1080
    hz: 72 dpi: 94 diag: 598mm (23.5")
  API: EGL v: 1.5 platforms: device: 0 drv: radeonsi device: 1 drv: swrast
    gbm: drv: kms_swrast surfaceless: drv: radeonsi wayland: drv: radeonsi x11:
    drv: radeonsi
  API: OpenGL v: 4.6 compat-v: 4.5 vendor: amd mesa v: 25.0.5 glx-v: 1.4
    direct-render: yes renderer: AMD Radeon RX 6800 (radeonsi navi21 LLVM
    20.1.4 DRM 3.61 6.14.6-1-default) device-ID: 1002:73bf display-ID: :0.0
  API: Vulkan v: 1.4.309 surfaces: xcb,xlib,wayland device: 0
    type: discrete-gpu driver: N/A device-ID: 1002:73bf device: 1 type: cpu
    driver: N/A device-ID: 10005:0000
  Info: Tools: api: clinfo, eglinfo, glxinfo, vulkaninfo
    de: kscreen-console,kscreen-doctor gpu: corectrl, radeontop, umr
    wl: wayland-info x11: xdpyinfo, xprop, xrandr

Wine version: GE-Proton10-3
VKD3D-Proton version: https://github.com/HansKristian-Work/vkd3d-proton/tree/7e829e88364f9429596b14d0c0ea0205e1337edc

Log files

Proton Logs: steam-1874880.log
dmesg: Arma-dmesg.txt
um waves: waves.log

Rreddragonmps 2025-06-03 github

Having this same problem on Elden Ring. Made it somewhat reproducible by running through gamescope, and resuming my playthrough. Then I alt+tab to another virtual desktop and trigger it consistently. My GPU is not overclocked and I have set power profile to high. I'm using GE-Proton10-3 also.

Can you try this same steps and check if you can reproduce it consistently?

Start your game through gamescope (gamescope -f -- %command%)
Load a scene with gpu load, something that needs 3d rendering
Alt+tab and focus on any other window
"Profit"?

System
Operating System: Arch Linux
KDE Plasma Version: 6.3.5
KDE Frameworks Version: 6.14.0
Qt Version: 6.9.0
Kernel Version: 6.14.9-zen1-1-zen (64-bit)
Graphics Platform: Wayland
Processors: 4 × Intel® Core™ i5-6600K CPU @ 3.50GHz
Memory: 15.6 GiB of RAM
Graphics Processor: AMD Radeon RX 6700 XT
Manufacturer: Gigabyte Technology Co., Ltd.
Product Name: Z170M-D3H

Dmesg logs:

[  478.744053] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32807)
[  478.744060] amdgpu 0000:03:00.0: amdgpu:  in process eldenring.exe pid 5145 thread vkd3d_queue pid 5240
[  478.744062] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000453ff000 from client 0x1b (UTCL2)
[  478.744064] amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00601431
[  478.744066] amdgpu 0000:03:00.0: amdgpu:      Faulty UTCL2 client ID: SQC (data) (0xa)
[  478.744067] amdgpu 0000:03:00.0: amdgpu:      MORE_FAULTS: 0x1
[  478.744068] amdgpu 0000:03:00.0: amdgpu:      WALKER_ERROR: 0x0
[  478.744069] amdgpu 0000:03:00.0: amdgpu:      PERMISSION_FAULTS: 0x3
[  478.744070] amdgpu 0000:03:00.0: amdgpu:      MAPPING_ERROR: 0x0
[  478.744071] amdgpu 0000:03:00.0: amdgpu:      RW: 0x0
[  478.744074] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32807)
[  478.744076] amdgpu 0000:03:00.0: amdgpu:  in process eldenring.exe pid 5145 thread vkd3d_queue pid 5240
[  478.744077] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000453ff000 from client 0x1b (UTCL2)
[  478.744081] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32807)
[  478.744082] amdgpu 0000:03:00.0: amdgpu:  in process eldenring.exe pid 5145 thread vkd3d_queue pid 5240
[  478.744083] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000453ff000 from client 0x1b (UTCL2)
[  478.744087] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32807)
[  478.744088] amdgpu 0000:03:00.0: amdgpu:  in process eldenring.exe pid 5145 thread vkd3d_queue pid 5240
[  478.744089] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000453ff000 from client 0x1b (UTCL2)
[  488.749675] amdgpu 0000:03:00.0: amdgpu: Dumping IP State
[  488.751153] amdgpu 0000:03:00.0: amdgpu: Dumping IP State Completed
[  488.751219] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32807)
[  488.751223] amdgpu 0000:03:00.0: amdgpu:  in process eldenring.exe pid 5145 thread vkd3d_queue pid 5240
[  488.751225] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000453ff000 from client 0x1b (UTCL2)
[  488.751227] amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00601431
[  488.751228] amdgpu 0000:03:00.0: amdgpu:      Faulty UTCL2 client ID: SQC (data) (0xa)
[  488.751230] amdgpu 0000:03:00.0: amdgpu:      MORE_FAULTS: 0x1
[  488.751231] amdgpu 0000:03:00.0: amdgpu:      WALKER_ERROR: 0x0
[  488.751232] amdgpu 0000:03:00.0: amdgpu:      PERMISSION_FAULTS: 0x3
[  488.751233] amdgpu 0000:03:00.0: amdgpu:      MAPPING_ERROR: 0x0
[  488.751234] amdgpu 0000:03:00.0: amdgpu:      RW: 0x0
[  488.751239] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32807)
[  488.751240] amdgpu 0000:03:00.0: amdgpu:  in process eldenring.exe pid 5145 thread vkd3d_queue pid 5240
[  488.751242] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000453ff000 from client 0x1b (UTCL2)
[  488.751246] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32807)
[  488.751248] amdgpu 0000:03:00.0: amdgpu:  in process eldenring.exe pid 5145 thread vkd3d_queue pid 5240
[  488.751249] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000453ff000 from client 0x1b (UTCL2)
[  488.751254] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32807)
[  488.751256] amdgpu 0000:03:00.0: amdgpu:  in process eldenring.exe pid 5145 thread vkd3d_queue pid 5240
[  488.751257] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000453ff000 from client 0x1b (UTCL2)
[  488.751375] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32807)
[  488.751376] amdgpu 0000:03:00.0: amdgpu:  in process eldenring.exe pid 5145 thread vkd3d_queue pid 5240
[  488.751378] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000453ff000 from client 0x1b (UTCL2)
[  488.751379] amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00601431
[  488.751380] amdgpu 0000:03:00.0: amdgpu:      Faulty UTCL2 client ID: SQC (data) (0xa)
[  488.751381] amdgpu 0000:03:00.0: amdgpu:      MORE_FAULTS: 0x1
[  488.751382] amdgpu 0000:03:00.0: amdgpu:      WALKER_ERROR: 0x0
[  488.751383] amdgpu 0000:03:00.0: amdgpu:      PERMISSION_FAULTS: 0x3
[  488.751384] amdgpu 0000:03:00.0: amdgpu:      MAPPING_ERROR: 0x0
[  488.751385] amdgpu 0000:03:00.0: amdgpu:      RW: 0x0
[  488.751390] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32807)
[  488.751391] amdgpu 0000:03:00.0: amdgpu:  in process eldenring.exe pid 5145 thread vkd3d_queue pid 5240
[  488.751392] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000453ff000 from client 0x1b (UTCL2)
[  488.751398] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32807)
[  488.751399] amdgpu 0000:03:00.0: amdgpu:  in process eldenring.exe pid 5145 thread vkd3d_queue pid 5240
[  488.751400] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000453ff000 from client 0x1b (UTCL2)
[  488.751406] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32807)
[  488.751407] amdgpu 0000:03:00.0: amdgpu:  in process eldenring.exe pid 5145 thread vkd3d_queue pid 5240
[  488.751408] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000453ff000 from client 0x1b (UTCL2)
[  488.761192] amdgpu 0000:03:00.0: amdgpu: ring gfx_0.0.0 timeout, signaled seq=217999, emitted seq=218001
[  488.761198] amdgpu 0000:03:00.0: amdgpu: Process information: process eldenring.exe pid 5145 thread vkd3d_queue pid 5240
[  488.761200] amdgpu 0000:03:00.0: amdgpu: Starting gfx_0.0.0 ring reset
[  488.974242] amdgpu 0000:03:00.0: amdgpu: Ring gfx_0.0.0 reset failure

Waves logs: waves.log

HHansKristian-Work maintainer 2025-06-03 github

The wave dump from Arma there is a clear descriptor heap out of bounds access, which is "fun".

HHansKristian-Work maintainer 2025-06-03 github

@reddragonmps that wave dump is bogus. The GPU was probably not in a proper hung state when that was taken.

DDanielGaaA 2025-06-03 github

I have new logs from today crash. I have to compress them because of waves file size.

logs.zip

Is this VKD3D issue or Mesa issue?
I also tried to do umr -RS log file but umr was segmentation faulting.

Rreddragonmps 2025-06-04 github

Sorry for posting some useless logs, I gathered them from vt so the gpu in fact did recover. Yesterday I installed Bazzite 41 and now I'm testing here. I just installed umr and nothing else in case there was any kind of package conflicting. Mesa version is 25.0.2. I tested on a native linux game (War Thunder) for over 10 minutes with no gpu halts. I have not experienced any gpu halts on Elden Ring Nightreign for the 10 hours I have played.

System:

Operating System: Bazzite 41
KDE Plasma Version: 6.3.4
KDE Frameworks Version: 6.12.0
Qt Version: 6.8.2
Kernel Version: 6.13.9-103.bazzite.fc41.x86_64 (64-bit)
Graphics Platform: Wayland
Processors: 4 × Intel® Core™ i5-6600K CPU @ 3.50GHz
Memory: 15.6 GiB of RAM
Graphics Processor: AMD Radeon RX 6700 XT
Manufacturer: Gigabyte Technology Co., Ltd.
Product Name: Z170M-D3H

I have dumped the logs correctly this time.
Waves logs: eldenring-waves.zip

Should I post dmesg logs also?

Rrunar-work 2025-07-07 github

@DanielGaaA How long does it usually take to reproduce a hang? Any other details you can give (like graphics settings, map type, locations, etc.) would help too.

DDanielGaaA 2025-07-14 github

Hello I recently switched to different distro (Bazzite) but I still have same crashes in Arma. The crash is very random I had not found the way to reproduce. It can take 20 mininutes but mostly it is after 1h play and sometimes it doesn't crash even after 3h.

I mostly play on W.C.S servers on map Serhiivka my last crashes were around airfield and in industrial zone
It also crashed on Everon map in the city Montignac next to the church.

During the weekend I tried to use VKD3D_CONFIG=breadcrumbs but game started memory leaking and after 45min it will crash due to running out of memory. This was with Proton 10 (debug) and Proton Experimental (debug)

Settings:

DDanielGaaA 2025-07-14 github

Should I create new issue for memory leak when I use VKD3D_CONFIG=breadcrumbs ?

I also managed to get new crash dump with RADV_DEBUG=hang

radv_dumps_73326_2025.07.14_21.51.52.zip

DDanielGaaA 2025-07-15 github

I made 100GB swapfile to avoid running out of memory.
New crash with RADV_DEBUG=hang,nocache VKD3D_CONFIG=breadcrumbs so there is spv file with RADV hang.
But I am unable to locate breadcrumbs log. Where is it located? I am also attaching Proton log.

radv_dumps_7161_2025.07.15_21.45.34.zip

Rrunar-work 2025-07-18 github

Thanks for the update. The breadcrumbs log will be included in the Proton log. If it works, you should see a line in the log with Device lost observed, analyzing breadcrumbs ... It didn't output anything in this case, so either Proton wasn't using a debug build of vkd3d-proton, or (in rare cases) it's not a hang that will produce any breadcrumbs.

If you haven't done this already, when you select Proton Experimental debug you can rename the .debug files in SteamLibrary/steamapps/common/Proton - Experimental/files/lib/wine/vkd3d-proton/x86_64-windows/ to replace the regular .dll files and see if that works better.

Some games use a lot of VRAM with breadcrumbs tracing, unfortunately. If your own workaround works well enough I'd just stick with that.

DDanielGaaA 2025-07-18 github

It didn't output anything in this case, so either Proton wasn't using a debug build of vkd3d-proton, or (in rare cases) it's not a hang that will produce any breadcrumbs.

Maybe RADV_DEBUG=hang prevented breadcrumbs from working? Right now I am just trying to use it without it. Right now I am using RADV_DEBUG=syncshaders VKD3D_CONFIG=breadcrumbs but didn't manage get gpu driver crash yet. but game did freeze for me twice. First time i forced closed it and don't have logs anymore but for second time I let game be and after few minutes game crashed. I am not sure if this log useful because it doesn't have breadcrumbs as there was no GPU crash.

steam-1874880.log

I will keep trying to get breadcrumb log but playing at 15-25 fps is annoying.

DDanielGaaA 2025-07-19 github

Finally I have breadcrumb log

steam-1874880.log

EEtmix 2025-08-23 github

@HansKristian-Work have you had a chance to look at the logs @DanielGaaA provided? I’m experiencing the same issue on RX 6800 XT / 5800X3D / 32 GB RAM with Proton GE 10-12 and mesa-git. Let me know if you need more logs or any other info.

EEtmix 2025-08-23 github

@DanielGaaA do you also see stuttering when joining a server or moving quickly? It feels like shader compilation, but it never seems to go away even after hours of gameplay. I was able to reduce it to more playable levels by lowering settings to low, but the stutter is still present, just less severe.

BBB-Harris 2025-09-22 github

Stopping by to say that I'm having the same issue with page faults while playing Reforger, although rare as it's happened twice over 54 hours of gameplay. I have also noticed that I occasionally (maybe once every couple hours? It varies) get artifacts while playing the game, namely yellow flashes.

In fact, I was playing Squad today and had a similar page fault, together with the same occasional yellow flashes, so I suspect the crashes are related.

Has anyone else noticed page faults in other games, or artifacts while playing games? I'm running my 7900XTX Hellhound completely stock and always have, so I very much hope it's not hardware instability.

Using RADV_DEBUG=hang with UMR might help with debug, but it causes games to run at 30FPS and crashes happen rarely enough that it could be 10s of hours before one occurs. Still, I might just have to put up with that.

My system:
OS: Arch Linux
Kernel: various default Arch distributed kernels
Mesa: various up-to-date versions because Arch
DE: KDE Plasma (various versions)
CPU: AMD Ryzen 7 7800X3D
GPU: AMD Radeon RX 7900 XTX Hellhound (running completely stock)

?ghost 2025-09-25 github

Passing by just to say I'm having the same issues listed here with the occasional random crash.

OS: Fedora Linux 42 (Workstation Edition) x86_64 PLASMA KDE running on wayland
Linux 6.16.7-200.fc42.x86_64
CPU : AMD Ryzen 7 5700X3D (16) @ 4.15 GHz
GPU: AMD Radeon RX 7800 XT [Discrete]
Ram : 102.08 GiB

Xx1yizhuo 2026-01-01 github

I experienced this today. The first and only crash (so far) in about 20 hours of gameplay.

Fedora Linux 43 Workstation (GNOME, Wayland)
Linux 6.17.12-300.fc43.x86_64
12th Gen Intel® Core™ i9-12900K × 24
AMD Radeon™ RX 9070 XT

 3:56:47 PM kernel: amdgpu 0000:03:00.0: amdgpu: Ring gfx_0.0.0 reset succeeded
 3:56:47 PM kernel: amdgpu 0000:03:00.0: amdgpu: Ring gfx_0.0.0 reset succeeded
 3:56:47 PM kernel: amdgpu 0000:03:00.0: amdgpu: Starting gfx_0.0.0 ring reset
 3:56:47 PM kernel: amdgpu 0000:03:00.0: amdgpu:  Process enfMain pid 35805 thread vkd3d_queue pid 35835
 3:56:47 PM kernel: amdgpu 0000:03:00.0: amdgpu: ring gfx_0.0.0 timeout, signaled seq=3790043, emitted seq=3790045
 3:56:37 PM kernel: amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000bbae8000 from client 10
 3:56:37 PM kernel: amdgpu 0000:03:00.0: amdgpu:  Process enfMain pid 35805 thread vkd3d_queue pid 35835
 3:56:37 PM kernel: amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32772)
 3:56:37 PM kernel: amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000bbae8000 from client 10
 3:56:37 PM kernel: amdgpu 0000:03:00.0: amdgpu:  Process enfMain pid 35805 thread vkd3d_queue pid 35835
 3:56:37 PM kernel: amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32772)
 3:56:37 PM kernel: amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000bbae8000 from client 10
 3:56:37 PM kernel: amdgpu 0000:03:00.0: amdgpu:  Process enfMain pid 35805 thread vkd3d_queue pid 35835
 3:56:37 PM kernel: amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32772)
 3:56:37 PM kernel: amdgpu 0000:03:00.0: amdgpu: 	 RW: 0x0
 3:56:37 PM kernel: amdgpu 0000:03:00.0: amdgpu: 	 MAPPING_ERROR: 0x0
 3:56:37 PM kernel: amdgpu 0000:03:00.0: amdgpu: 	 PERMISSION_FAULTS: 0x3
 3:56:37 PM kernel: amdgpu 0000:03:00.0: amdgpu: 	 WALKER_ERROR: 0x0
 3:56:37 PM kernel: amdgpu 0000:03:00.0: amdgpu: 	 MORE_FAULTS: 0x1
 3:56:37 PM kernel: amdgpu 0000:03:00.0: amdgpu: 	 Faulty UTCL2 client ID: SQC (data) (0xa)
 3:56:37 PM kernel: amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00601431
 3:56:37 PM kernel: amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x00008000bbae8000 from client 10
 3:56:37 PM kernel: amdgpu 0000:03:00.0: amdgpu:  Process enfMain pid 35805 thread vkd3d_queue pid 35835
 3:56:37 PM kernel: amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:32772)

Proton versions

proton experimentalx2 2025-07
ge-proton10-3x2 2025-06

Launch options

Launch lines

gamescope -f -- %command%x1 2025-06

Upstream links

HansKristian-Work/vkd3d-proton/issues/2441x1 2025-06