[BUG] vrcompositor RenderThread triggers recurring amdgpu GPU page faults (RDNA2 / RADV), often requiring a hard reset

[dyle71]

vrcompositor RenderThread triggers recurring amdgpu GPU page faults (RDNA2 / RADV), often requiring a hard reset

This is a two day investigation of myself in co-op with Claude Opus 5.

Summary

On an RX 6800 XT with RADV, vrcompositor's RenderThread repeatedly triggers a GPU
page fault in a narrow, recurring VA range. The kernel usually fails to reset the graphics
ring afterwards and performs a full GPU reset, which kills the session and requires a
hard power cycle.

The behaviour is independent of Mesa version, kernel version, Proton version, the running
application, and the SteamVR version — all were varied, see the exclusion list below.
vrcompositor is the faulting process in every single occurrence.

The problem is not new: systemd holds 41 vrcompositor / vrserver coredumps going back
to 2026-07-30. Most compositor crashes are invisible to the user because SteamVR
restarts the compositor; it only becomes fatal when the crash coincides with a GPU hang.

System

GPU	AMD Radeon RX 6800 XT — Navi 21, 1002:73bf rev c1
Driver	RADV (Mesa), amdgpu
CPU	AMD Ryzen 9 5950X
RAM	125 GB
OS	Arch Linux, KDE Plasma on Wayland (kwin_wayland)
Kernel	tested 6.18.49-lts and 7.2.3-arch1
Mesa	tested 26.1.6 and 26.2.2
linux-firmware	20260810
HMD	Valve Index, DisplayPort, tested at 90 Hz and 120 Hz
SteamVR	tested 2.17.8 (beta 2026-09-01) and the 2026-06-18 beta branch

Symptom

The headset and both desktop monitors go black. The machine does not recover; a hard
reset is required. In a minority of cases the kernel manages Ring reset succeeded and
the session survives with visible rendering corruption.

Kernel log (representative)

amdgpu 0000:0c:00.0: [gfxhub] page fault (src_id:0 ring:24 vmid:6 pasid:182)
amdgpu 0000:0c:00.0:  Process vrcompositor pid 8990 thread RenderThread pid 9080
amdgpu 0000:0c:00.0:   in page starting at address 0x00000080013c5000 from client 0x1b (UTCL2)
amdgpu 0000:0c:00.0: GCVM_L2_PROTECTION_FAULT_STATUS:0x00601430
amdgpu 0000:0c:00.0:          Faulty UTCL2 client ID: SQC (data) (0xa)
amdgpu 0000:0c:00.0:          MORE_FAULTS: 0x0
amdgpu 0000:0c:00.0:          WALKER_ERROR: 0x0
amdgpu 0000:0c:00.0:          PERMISSION_FAULTS: 0x3
amdgpu 0000:0c:00.0:          MAPPING_ERROR: 0x0
amdgpu 0000:0c:00.0:          RW: 0x0
amdgpu 0000:0c:00.0: ring gfx_0.1.0 timeout, signaled seq=117748, emitted seq=117751
amdgpu 0000:0c:00.0:  Process kwin_wayland pid 2788 thread kwin_wayla:cs0 pid 2846
amdgpu 0000:0c:00.0: Starting gfx_0.1.0 ring reset
amdgpu 0000:0c:00.0: Ring gfx_0.1.0 reset failed
amdgpu 0000:0c:00.0: GPU reset begin!. Source:  1

The fault is always client 0x1b (UTCL2), SQC (data), PERMISSION_FAULTS: 0x3 — a shader
reading memory it is not permitted to touch. The ring timeout is subsequently attributed to
kwin_wayland because the compositor of the desktop is the next submitter to the hung ring.

Observed faults

13 occurrences over two days. All in vrcompositor / RenderThread.

Time	Faulting VA	Outcome
2026-09-06 21:52	0x00000080013be000	GPU reset
2026-09-06 22:01	0x00000080013b3000	GPU reset
2026-09-06 22:27	0x00000080013be000	GPU reset
2026-09-07 13:42	0x00000080013b3000	GPU reset
2026-09-07 14:11	0x00000080013b3000	ring reset ×4, recovered
2026-09-07 14:57	0x00000080013c5000	GPU reset
2026-09-07 15:17	0x00000080013c5000	recovered
2026-09-07 15:19	0x00000080013c5000	GPU reset
2026-09-07 15:35	0x00000080013c5000	GPU reset
2026-09-07 16:39	0x00000080013c5000	GPU reset
2026-09-07 17:06	0x00000080013c5000	recovered
2026-09-07 17:12	0x00000080013c5000	GPU reset
2026-09-07 17:22	0x0000008001387000	GPU reset

Every faulting address falls inside a ~250 KB window (0x8001387000 – 0x80013c5000).
The address is stable across many runs and shifts only when the SteamVR build changes.

Second failure mode

Once, with no application running at all — SteamVR had just been started — the display
pipeline hung instead:

[drm] *ERROR* dc_stream_state is NULL for crtc '2'!     (×7)
[drm] *ERROR* [CRTC:100:crtc-2] flip_done timed out
[drm] *ERROR* [CRTC:100:crtc-2] hw_done or flip_done timed out

followed by vrcompositor aborting from its own ThreadWatchdog (SIGABRT). The system
survived. This confirms the fault is not tied to any particular application.

Ruled out

Each of the following was changed and the crash still reproduced:

• Application: Skyrim VR with and without mods, SKSE, Mod Organizer 2 / usvfs.
  One crash occurred with no application running at all.
• Proton: 11.0 and Experimental. Proton prefix rebuilt from scratch.
• Mesa: 1:26.1.6-1 and 1:26.2.2-1.
• Kernel: 6.18.49-lts and 7.2.3-arch1.
• SteamVR: 2.17.8 (2026-09-01 beta) and the 2026-06-18 beta branch.
• Render resolution: SteamVR supersample 1.5 and 1.0; in-app render scale 1.5 and 1.0.
• Refresh rate: 120 Hz and 90 Hz.
• Async reprojection: both enabled (CAP_SYS_NICE present) and disabled
  (Insufficient permission to create high priority queue). Crashes occur either way.
• DisplayPort link: no link-training errors in the kernel log at any point.

Other VR titles (Beat Saber, Hellblade: Senua's Sacrifice VR) are playable, but the
coredump history shows vrcompositor was crashing during those sessions too — silently,
because SteamVR restarts it.

Crash history

coredumpctl records for vrcompositor / vrserver, by day:

2026-07-30   6      2026-08-04   2      2026-08-09   2
2026-07-31   6      2026-08-05   6      2026-09-04   1
2026-08-01   1      2026-08-06   2      2026-09-06   1
2026-08-03   4                          2026-09-07   8

Attachments available on request

• Two amdgpu devcoredumps (/sys/class/drm/card1/device/devcoredump/data), captured
  under kernel 7.2.3 and under the June SteamVR build.
• dmesg for each.
• Two vrcompositor userspace coredumps with stack traces — one SIGSEGV, one SIGABRT
  from ThreadWatchdog. Offsets only; the binary is stripped.

Representative userspace trace (SIGSEGV, main thread), with a second thread inside RADV:

Stack trace of thread 9558:
#0  0x0000559872bce6d6 (vrcompositor + 0x22f6d6)
#1  0x0000559872c09c66 (vrcompositor + 0x26ac66)
#2  0x0000559872c279b8 (vrcompositor + 0x2889b8)
#3  0x0000559872c27be8 (vrcompositor + 0x288be8)
#4  0x00007f17d4427781 (libc.so.6 + 0x27781)
#5  0x00007f17d44278b9 (libc.so.6 + 0x278b9)
#6  0x0000559872aebe1a (vrcompositor + 0x14ce1a)

Stack trace of thread 9583:
#0  0x00007f17d44a0952 (libc.so.6 + 0xa0952)
#1  0x00007f17d449754c (libc.so.6 + 0x9754c)
#2  0x00007f17ce59351e (libvulkan_radeon.so + 0x39351e)
#3  0x00007f17ce55cfad (libvulkan_radeon.so + 0x35cfad)
#4  0x00007f17ce59348d (libvulkan_radeon.so + 0x39348d)
#5  0x00007f17d44980a2 (libc.so.6 + 0x980a2)
#6  0x00007f17d452080c (libc.so.6 + 0x12080c)

[pink-fluffy-proto]

same issue (GCVM_L2_PROTECTION_FAULT_STATUS in dmesg)

was fine yesterday / day before, now after updates (steam + steamvr) getting the above issue (was on beta 2.17.x yesterday)

2.16.7 working OK

2.17 stable / 2.17.9 beta: also seeing weird lag in steamvr playspace setup (not present in previous 2.16.7) before steamvr dies, might be related

RX 7900 XTX fwiw

[ChristophHaag]

I suppose mine is close enough? R9700 on Archlinux. Though mine happened every time on startup. It rendered a gray frame but never managed to render an actual image frame. I have SteamVR Home disabled, so it would have rendered the dome void.

 amdgpu 0000:0c:00.0: [gfxhub] page fault (src_id:0 ring:24 vmid:7 pasid:1399)
 amdgpu 0000:0c:00.0:  Process vrcompositor pid 1519993 thread RenderThread pid 1520027
 amdgpu 0000:0c:00.0:   in page starting at address 0x000000800143c000 from client 10
 amdgpu 0000:0c:00.0: GCVM_L2_PROTECTION_FAULT_STATUS:0x00701431
 amdgpu 0000:0c:00.0:          Faulty UTCL2 client ID: SQC (data) (0xa)
 amdgpu 0000:0c:00.0:          MORE_FAULTS: 0x1
 amdgpu 0000:0c:00.0:          WALKER_ERROR: 0x0
 amdgpu 0000:0c:00.0:          PERMISSION_FAULTS: 0x3
 amdgpu 0000:0c:00.0:          MAPPING_ERROR: 0x0
 amdgpu 0000:0c:00.0:          RW: 0x0
 amdgpu 0000:0c:00.0: [gfxhub] page fault (src_id:0 ring:24 vmid:7 pasid:1399)
 amdgpu 0000:0c:00.0:  Process vrcompositor pid 1519993 thread RenderThread pid 1520027
 amdgpu 0000:0c:00.0:   in page starting at address 0x000000800143c000 from client 10
 amdgpu 0000:0c:00.0: [gfxhub] page fault (src_id:0 ring:24 vmid:7 pasid:1399)
 amdgpu 0000:0c:00.0:  Process vrcompositor pid 1519993 thread RenderThread pid 1520027
 amdgpu 0000:0c:00.0:   in page starting at address 0x000000800143c000 from client 10
 amdgpu 0000:0c:00.0: [gfxhub] page fault (src_id:0 ring:24 vmid:7 pasid:1399)
 amdgpu 0000:0c:00.0:  Process vrcompositor pid 1519993 thread RenderThread pid 1520027
 amdgpu 0000:0c:00.0:   in page starting at address 0x000000800143c000 from client 10

I didn't have patience to debug myself, so here is Astra's summary:

Astra

• The vertex shader VS_GEOM reads set 0, binding 1 as a uniform buffer, to populate gl_Layer.
• Its compiled descriptor offset is 0, overlapping image-descriptor data. Reading those bytes as a buffer descriptor produces 0x800144a000—exactly the address in the kernel GPU fault.
• That scalar shader-data read faults, the graphics queue hangs, and SteamVR subsequently aborts. Cache-disabled execution reproduces it too.

The validation-layer run also reports invalid SteamVR pipeline layouts: binding 9 declares three images in SPIR-V, but the layout allocates only one (VUID …07991). That supports an application-side layout problem, although it doesn’t independently prove who caused the binding-1 mismatch; a RADV layout/compiler bug remains possible.

And the script it gave me to work around the issue: aislop-steamvr-layer-fix.py. No promises that it won't break anything in SteamVR, but it makes it at least run for me.

[lkroll]

I have a similar kernel error on Void Linux.
First time it froze when opening Beat Saber.
Second time when starting SteamVR.
Had to force reset my system.

GPU: RX 6900 XT
Kernel: Linux 6.18.49
SteamVR: 2.17.9
Mesa: mesa-26.1.8_1

2026-09-11T19:39:22.96925 kern.err: [13485.421102] amdgpu 0000:03:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:1 pasid:367)
2026-09-11T19:39:22.96928 kern.err: [13485.421107] amdgpu 0000:03:00.0: amdgpu:  Process vrcompositor pid 13649 thread RenderThread pid 13683
2026-09-11T19:39:22.96929 kern.err: [13485.421109] amdgpu 0000:03:00.0: amdgpu:   in page starting at address 0x0000008001440000 from client 0x1b (UTCL2)
2026-09-11T19:39:22.96930 kern.err: [13485.421111] amdgpu 0000:03:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00101430
2026-09-11T19:39:22.96930 kern.err: [13485.421112] amdgpu 0000:03:00.0: amdgpu:          Faulty UTCL2 client ID: SQC (data) (0xa)
2026-09-11T19:39:22.96931 kern.err: [13485.421112] amdgpu 0000:03:00.0: amdgpu:          MORE_FAULTS: 0x0
2026-09-11T19:39:22.96931 kern.err: [13485.421113] amdgpu 0000:03:00.0: amdgpu:          WALKER_ERROR: 0x0
2026-09-11T19:39:22.96932 kern.err: [13485.421114] amdgpu 0000:03:00.0: amdgpu:          PERMISSION_FAULTS: 0x3
2026-09-11T19:39:22.96932 kern.err: [13485.421114] amdgpu 0000:03:00.0: amdgpu:          MAPPING_ERROR: 0x0
2026-09-11T19:39:22.96932 kern.err: [13485.421115] amdgpu 0000:03:00.0: amdgpu:          RW: 0x0
2026-09-11T19:39:33.41534 kern.info: [13495.865944] amdgpu 0000:03:00.0: amdgpu: Dumping IP State
2026-09-11T19:39:33.41723 kern.info: [13495.868079] amdgpu 0000:03:00.0: amdgpu: Dumping IP State Completed
2026-09-11T19:39:33.41725 kern.info: [13495.868121] amdgpu 0000:03:00.0: amdgpu: [drm] AMDGPU device coredump file has been created
2026-09-11T19:39:33.41726 kern.info: [13495.868123] amdgpu 0000:03:00.0: amdgpu: [drm] Check your /sys/class/drm/card1/device/devcoredump/data
2026-09-11T19:39:33.41726 kern.err: [13495.868124] amdgpu 0000:03:00.0: amdgpu: ring gfx_0.0.0 timeout, signaled seq=5188273, emitted seq=5188275
2026-09-11T19:39:33.41727 kern.err: [13495.868127] amdgpu 0000:03:00.0: amdgpu:  Process vrcompositor pid 13649 thread RenderThread pid 13683
2026-09-11T19:39:33.41727 kern.err: [13495.868129] amdgpu 0000:03:00.0: amdgpu: Starting gfx_0.0.0 ring reset
2026-09-11T19:39:33.41728 kern.err: [13495.868221] amdgpu 0000:03:00.0: amdgpu: Ring gfx_0.0.0 reset succeeded
2026-09-11T19:39:33.41729 kern.info: [13495.868224] amdgpu 0000:03:00.0: [drm] device wedged, but recovered through reset

[Fell]

And the script it gave me to work around the issue: aislop-steamvr-layer-fix.py. No promises that it won't break anything in SteamVR, but it makes it at least run for me.

I tried the script. It prevented the crash for me but it introduced wrong/skewed rendering when interacting with the SteamVR menu while in a game. I also had the usual double-vision problem when turning my head. Not sure if that was caused by the patch.

System Info

Game: VRChat (GE-Proton-RTSP 11.0_20260609)
Headset: Valve Index
SteamVR Version: 2.17.9
Operating System: Arch Linux
KDE Plasma Version: 6.7.5
KDE Frameworks Version: 6.30.0
Qt Version: 6.11.2
Kernel Version: 7.2.4-arch1-2 (64-bit)
Graphics Platform: Wayland
Processors: 16 × AMD Ryzen 7 5800X3D 8-Core Processor
Memory: 64 GB of RAM (62.7 GB usable)
Graphics Processor: AMD Radeon RX 7900 XT
Product Name: X570 Phantom Gaming 4

Vulkan Info

==========
VULKANINFO
==========

Vulkan Instance Version: 1.4.357


Instance Extensions: count = 26
-------------------------------
VK_EXT_acquire_drm_display             : extension revision 1
VK_EXT_acquire_xlib_display            : extension revision 1
VK_EXT_debug_report                    : extension revision 10
VK_EXT_debug_utils                     : extension revision 2
VK_EXT_direct_mode_display             : extension revision 1
VK_EXT_display_surface_counter         : extension revision 1
VK_EXT_headless_surface                : extension revision 1
VK_EXT_layer_settings                  : extension revision 2
VK_EXT_surface_maintenance1            : extension revision 1
VK_EXT_swapchain_colorspace            : extension revision 5
VK_KHR_device_group_creation           : extension revision 1
VK_KHR_display                         : extension revision 23
VK_KHR_external_fence_capabilities     : extension revision 1
VK_KHR_external_memory_capabilities    : extension revision 1
VK_KHR_external_semaphore_capabilities : extension revision 1
VK_KHR_get_display_properties2         : extension revision 1
VK_KHR_get_physical_device_properties2 : extension revision 2
VK_KHR_get_surface_capabilities2       : extension revision 1
VK_KHR_portability_enumeration         : extension revision 1
VK_KHR_surface                         : extension revision 25
VK_KHR_surface_maintenance1            : extension revision 1
VK_KHR_surface_protected_capabilities  : extension revision 1
VK_KHR_wayland_surface                 : extension revision 6
VK_KHR_xcb_surface                     : extension revision 6
VK_KHR_xlib_surface                    : extension revision 6
VK_LUNARG_direct_driver_loading        : extension revision 1

Instance Layers: count = 18
---------------------------
VK_LAYER_FROG_gamescope_wsi_x86_64    Gamescope WSI (XWayland Bypass) Layer (x86_64)               1.3.221  version 1
VK_LAYER_INTEL_nullhw                 INTEL NULL HW                                                1.1.73   version 1
VK_LAYER_KHRONOS_validation           Khronos Validation Layer                                     1.4.357  version 1
VK_LAYER_LUNARG_api_dump              LunarG API dump layer                                        1.4.357  version 2
VK_LAYER_LUNARG_monitor               Execution Monitoring Layer                                   1.4.357  version 1
VK_LAYER_LUNARG_screenshot            LunarG image capture layer                                   1.4.357  version 1
VK_LAYER_MANGOHUD_NEXT_overlay_x86_64 MangoHud                                                     1.3.221  version 1
VK_LAYER_MANGOHUD_overlay_x86_64      Vulkan Hud Overlay                                           1.3.0    version 1
VK_LAYER_MESA_anti_lag                Open-source implementation of the VK_AMD_anti_lag extension. 1.4.303  version 1
VK_LAYER_MESA_device_select           Linux device selection layer                                 1.4.303  version 1
VK_LAYER_MESA_overlay                 Mesa Overlay layer                                           1.4.303  version 1
VK_LAYER_MESA_screenshot              Mesa Screenshot layer                                        1.4.303  version 1
VK_LAYER_MESA_vram_report_limit       Limit reported VRAM                                          1.4.303  version 1
VK_LAYER_RENDERDOC_Capture            Debugging capture layer for RenderDoc                        1.4.324  version 45
VK_LAYER_VALVE_steam_fossilize_32     Steam Pipeline Caching Layer                                 1.3.207  version 1
VK_LAYER_VALVE_steam_fossilize_64     Steam Pipeline Caching Layer                                 1.3.207  version 1
VK_LAYER_VALVE_steam_overlay_32       Steam Overlay Layer                                          1.3.207  version 1
VK_LAYER_VALVE_steam_overlay_64       Steam Overlay Layer                                          1.3.207  version 1

Devices:
========
GPU0:
        apiVersion         = 1.4.354
        driverVersion      = 26.2.2
        vendorID           = 0x1002
        deviceID           = 0x744c
        deviceType         = PHYSICAL_DEVICE_TYPE_DISCRETE_GPU
        deviceName         = AMD Radeon RX 7900 XT (RADV NAVI31)
        driverID           = DRIVER_ID_MESA_RADV
        driverName         = radv
        driverInfo         = Mesa 26.2.2-arch1.1
        conformanceVersion = 1.4.5.3
        deviceUUID         = 00000000-0c00-0000-0000-000000000000
        driverUUID         = 414d442d-4d45-5341-2d44-525600000000

[Packetdancer]

We've already tracked this down and squashed it internally; there should be a new build within a day or two that'll have the fix.

[Packetdancer]

This should now be rolled out in 2.17.10, which is on the beta branch for SteamVR. Give that a try if you can; we want to make sure there's no other lingering issues before flipping .10 into the main release branch.

[Fell]

The crash is definitely fixed for me. I had it 100% on startup, now it's gone.
Double-vision remains, but that's a different issue.

[amalon]

For me (on Arch with 2.17.10) it no longer crashes, but instead I'm getting "Failed to connect to window manager" "SteamVR failed to connect to the window manager. You may need to restart your session or SteamVR (497)". It definitely worked for me on 2.17.8.

[Spacefish]

I still get page faults in this region when trying to enter Theater mode (still the same crash as #906 i guess..
I traced it back to a shader trying to access a page that´s not mapped. At least the jump faulting code was binary identical to a compiled shader... It´s hard to debug it though without any source or such.

Thu Sep 17 2026 02:31:13.938571 [Info] - Not starting NewFrameHandler thread on Linux
Thu Sep 17 2026 02:31:13.961551 [Info] - External connection from /home/spacy/.local/share/Steam/steamapps/common/SteamVR/bin/linux64/vrdashboard 54050
Thu Sep 17 2026 02:31:13.973074 [Error] - WaitForPendingPresent: failed to wait for present
Thu Sep 17 2026 02:31:13.991528 [Error] - WaitForPendingPresent: failed to wait for present
Thu Sep 17 2026 02:31:14.000736 [Info] - External connection from /home/spacy/.local/share/Steam/steamapps/common/SteamVR/bin/vrwebhelper/linux64/vrwebhelper 54052
Thu Sep 17 2026 02:31:14.013421 [Error] - WaitForPendingPresent: failed to wait for present
Thu Sep 17 2026 02:31:14.060139 [Info] - External connection from /home/spacy/.local/share/Steam/ubuntu12_64/steamwebhelper 14682
Thu Sep 17 2026 02:31:14.718508 [Info] - External connection from /home/spacy/.local/share/Steam/steamapps/common/SteamVR/bin/vrwebhelper/linux64/vrwebhelper 54226
Thu Sep 17 2026 02:31:14.723981 [Info] - External connection from /home/spacy/.local/share/Steam/steamapps/common/SteamVR/bin/vrwebhelper/linux64/vrwebhelper 54227
Thu Sep 17 2026 02:31:14.724037 [Info] - External connection from /home/spacy/.local/share/Steam/steamapps/common/SteamVR/bin/vrwebhelper/linux64/vrwebhelper 54233
Thu Sep 17 2026 02:31:15.194771 [Info] - External connection from /home/spacy/.local/share/Steam/steamapps/common/SteamVR/bin/linux64/vrmonitor 54057
Thu Sep 17 2026 02:31:15.448966 [Info] - Socket closed
Thu Sep 17 2026 02:31:15.449002 [Info] - Unable to read message from socket: 0
Thu Sep 17 2026 02:31:15.449020 [Info] - Process vrstartup (52679) disconnected (Thread(0x0x7e0908000c00/0x0x7e)
Thu Sep 17 2026 02:31:15.449028 [Info] - Lost pipe connection from vrstartup (52679)
Thu Sep 17 2026 02:31:16.444742 [Info] - External connection from /home/spacy/.local/share/Steam/steamapps/common/SteamVR/bin/vrwebhelper/linux64/vrwebhelper 54406
Thu Sep 17 2026 02:31:16.551238 [Info] - UpdateInputFocus: Failed to give focus to panel with SGID 20028 after 1 seconds.
Thu Sep 17 2026 02:31:16.561019 [Info] - VREvent_ChaperoneUniverseHasChanged -> StandingCoordinateSystemUpdate
Thu Sep 17 2026 02:31:41.097080 [Error] - WaitForPendingPresent: failed to wait for present
Thu Sep 17 2026 02:31:41.113082 [Error] - WaitForPendingPresent: failed to wait for present
Thu Sep 17 2026 02:31:43.101243 [Error] - vkGetFenceStatus failed with -4
Thu Sep 17 2026 02:31:43.101334 [Error] - Failed WaitEvent(CompositorPresent)! event=1 vkerror=-4
Thu Sep 17 2026 02:31:43.101344 [Error] - Failed to wait for GPU work
Thu Sep 17 2026 02:31:43.101371 [Error] - vkGetFenceStatus failed with -4
Thu Sep 17 2026 02:31:43.101392 [Error] - Failed GetDeltas(Compositor): vkerror=-4
Thu Sep 17 2026 02:31:43.101401 [Error] - Failed GetDeltas(CompositorPresent): vkerror=-4
Thu Sep 17 2026 02:31:43.101639 [Error] - WaitForPendingPresent: failed to wait for present
Thu Sep 17 2026 02:31:43.123651 [Error] - vkGetFenceStatus failed with -4
Thu Sep 17 2026 02:31:43.123702 [Error] - Failed WaitEvent(CompositorPresent)! event=1 vkerror=-4
Thu Sep 17 2026 02:31:43.123712 [Error] - Failed to wait for GPU work
Thu Sep 17 2026 02:31:43.123728 [Error] - vkGetFenceStatus failed with -4
Thu Sep 17 2026 02:31:43.123739 [Error] - Failed GetDeltas(Compositor): vkerror=-4
Thu Sep 17 2026 02:31:43.123745 [Error] - Failed GetDeltas(CompositorPresent): vkerror=-4
Thu Sep 17 2026 02:31:43.123882 [Error] - WaitForPendingPresent: failed to wait for present
Thu Sep 17 2026 02:31:49.981961 [Info] - Failed Watchdog timeout in thread Render in RenderDistort after 6.815166 seconds. Aborting.

[ 9731.697288] amdgpu 0000:0d:00.0: [gfxhub] page fault (src_id:0 ring:24 vmid:5 pasid:365)
[ 9731.697298] amdgpu 0000:0d:00.0:  Process vrcompositor pid 53981 thread RenderThread pid 54045
[ 9731.697300] amdgpu 0000:0d:00.0:   in page starting at address 0x00000080013c3000 from client 10
[ 9731.697302] amdgpu 0000:0d:00.0: GCVM_L2_PROTECTION_FAULT_STATUS:0x00501430
[ 9731.697304] amdgpu 0000:0d:00.0:      Faulty UTCL2 client ID: SQC (data) (0xa)
[ 9731.697305] amdgpu 0000:0d:00.0:      MORE_FAULTS: 0x0
[ 9731.697306] amdgpu 0000:0d:00.0:      WALKER_ERROR: 0x0
[ 9731.697308] amdgpu 0000:0d:00.0:      PERMISSION_FAULTS: 0x3
[ 9731.697309] amdgpu 0000:0d:00.0:      MAPPING_ERROR: 0x0
[ 9731.697310] amdgpu 0000:0d:00.0:      RW: 0x0
[ 9733.710547] amdgpu 0000:0d:00.0: Dumping IP State
[ 9733.711754] amdgpu 0000:0d:00.0: Dumping IP State Completed
[ 9733.711776] amdgpu 0000:0d:00.0: [drm] AMDGPU device coredump file has been created
[ 9733.711778] amdgpu 0000:0d:00.0: [drm] Check your /sys/class/drm/card1/device/devcoredump/data
[ 9733.711780] amdgpu 0000:0d:00.0: ring gfx_0.0.0 timeout, signaled seq=4470481, emitted seq=4470484
[ 9733.711784] amdgpu 0000:0d:00.0:  Process vrcompositor pid 53981 thread RenderThread pid 54045
[ 9733.711786] amdgpu 0000:0d:00.0: Starting gfx_0.0.0 ring reset
[ 9733.712197] amdgpu 0000:0d:00.0: Ring gfx_0.0.0 reset succeeded
[ 9733.712199] amdgpu 0000:0d:00.0: [drm] device wedged, but no recovery needed

Full logs:
SteamVR-2026-09-17-AM_02_40_44.txt

[dyle71]

Mhm, my bug seems to be fixed. Well, at least this severe crash forcing me to reboot the whole system does not appear any longer.

Now just Steam Home keeps crashing every time...

         PID: 576039 (steamtours)
         TID: 576039 (steamtours)
         UID: 1000 (dyle)
         GID: 1000 (dyle)
      Signal: 11 (SEGV) si_code: SEGV_MAPERR
   Timestamp: Thu 2026-09-17 20:22:06 CEST (7s ago)
    Command Line: /home/dyle/.local/share/Steam/steamapps/common/SteamVR/tools/steamvr_environments/game/bin/linuxsteamrt64/steamtours -vr -retail -vulkan -useappid SteamVRAppID -nowindow -vconport 29009
      Executable: /home/dyle/.local/share/Steam/steamapps/common/SteamVR/tools/steamvr_environments/game/bin/linuxsteamrt64/steamtours
   Control Group: /user.slice/user-1000.slice/user@1000.service/app.slice/app-steam@autostart.service
            Unit: user@1000.service
       User Unit: app-steam@autostart.service
           Slice: user-1000.slice
       Owner UID: 1000 (dyle)
         Boot ID: b071aec93a0340eb8fee11f08388938c
      Machine ID: a1882297a3f040009ada7b2f3d771564
        Hostname: novigrad
         Storage: /var/lib/systemd/coredump/core.steamtours.1000.b071aec93a0340eb8fee11f08388938c.576039.1789669326000000.zst (present)
    Size on Disk: 62.5M
         Message: Process 576039 (steamtours) of user 1000 dumped core.

              Stack trace of thread 576039:
              #0  0x00007f7237b785c3 n/a (libc.so.6 + 0x1785c3)
              #1  0x00007f71fdd6f895 n/a (libpanorama.so + 0x16f895)
              #2  0x00007f71fdd7030d n/a (libpanorama.so + 0x17030d)
              #3  0x00007f71fdc933af n/a (libpanorama.so + 0x933af)
              #4  0x00007f71fde0c5ed n/a (libpanorama.so + 0x20c5ed)
              #5  0x00007f7201f5ad90 n/a (libclient.so + 0x55ad90)
              #6  0x00007f72371fee73 n/a (libengine2.so + 0x3fee73)
              #7  0x00007f72370c0b8d n/a (libengine2.so + 0x2c0b8d)
              #8  0x00007f723738a6e9 n/a (libengine2.so + 0x58a6e9)
              #9  0x00007f723738addc n/a (libengine2.so + 0x58addc)
              #10 0x00007f72370673b4 n/a (libengine2.so + 0x2673b4)
              #11 0x00007f72373c0a02 n/a (libengine2.so + 0x5c0a02)
              #12 0x00007f7237065389 n/a (libengine2.so + 0x265389)
              #13 0x00007f723738a3f3 n/a (libengine2.so + 0x58a3f3)
              #14 0x00007f723738b123 n/a (libengine2.so + 0x58b123)
              #15 0x00007f72370602e3 n/a (libengine2.so + 0x2602e3)
              #16 0x00007f7237060f45 n/a (libengine2.so + 0x260f45)
              #17 0x00007f72370612f3 Source2Main (libengine2.so + 0x2612f3)
              #18 0x00005590aec03080 n/a (steamtours + 0x3080)
              #19 0x00005590aec02b0a n/a (steamtours + 0x2b0a)
              #20 0x00007f7237a27781 n/a (libc.so.6 + 0x27781)
              #21 0x00007f7237a278b9 __libc_start_main (libc.so.6 + 0x278b9)
              #22 0x00005590aec02d39 n/a (steamtours + 0x2d39)
              ELF object binary architecture: AMD x86-64

I guess, that's another issue then, right?