AMD GPU Timeout by WayVR Page fault

[Cubbix]

Hello fellow friends. This has been a known issue on my system for quite some time and was solved by switching to the LTS kernel, but now that the new LTS kernel has been released, I am now encountering the same issue.

Additional system information
OS: Arch Linux x86_64
Host: MS-7C91 (1.0)
Kernel: Linux 6.18.16-1-lts
Uptime: 15 mins
Packages: 1265 (pacman), 30 (flatpak)
Shell: bash 5.3.9
Display (100025283): 2560x1440 @ 1.35x in 27", 165 Hz [External]
DE: KDE Plasma 6.6.2
WM: KWin (Wayland)
WM Theme: Sweet-Dark-transparent
Theme: Sweet-transparent-toolbar (Sweet) [Qt], Breeze-Dark [GTK2], Breeze ]
Icons: candy-icons [Qt], candy-icons [GTK2/3/4]
Font: Noto Sans (10pt) [Qt], Noto Sans (10pt) [GTK2/3/4]
Cursor: Sweet (24px)
Terminal: konsole 25.12.3
CPU: AMD Ryzen 7 5800X (16) @ 5.09 GHz
GPU: AMD Radeon RX 6700 XT [Discrete]
Memory: 3.95 GiB / 31.26 GiB (13%)
Swap: Disabled
Disk (/): 354.79 GiB / 456.39 GiB (78%) - ext4
Local IP (enp42s0): 192.168. - - -
Locale: en_US.UTF-8

Symptoms
Everything that uses GPU stops, usually resulting into a system still ruining, but no displays. A restart us always required when this event occurs.

Logs
Mar 07 17:50:35 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: [gfxhub] page fault (src_id:0 ring:24 vmid:5 pasid:32795)
Mar 07 17:50:35 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: Process wayvr pid 22573 thread wayvr pid 22573
Mar 07 17:50:35 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: in page starting at address 0x0000800001702000 from client 0x1b (UTCL2)
Mar 07 17:50:35 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: GCVM_L2_PROTECTION_FAULT_STATUS:0x00500C30
Mar 07 17:50:35 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: Faulty UTCL2 client ID: CPG (0x6)
Mar 07 17:50:35 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: MORE_FAULTS: 0x0
Mar 07 17:50:35 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: WALKER_ERROR: 0x0
Mar 07 17:50:35 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: PERMISSION_FAULTS: 0x3
Mar 07 17:50:35 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: MAPPING_ERROR: 0x0
Mar 07 17:50:35 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: RW: 0x0

Mar 07 17:50:46 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: Dumping IP State
Mar 07 17:50:46 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: Dumping IP State Completed
Mar 07 17:50:46 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: [drm] AMDGPU device coredump file has been created
Mar 07 17:50:46 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: [drm] Check your /sys/class/drm/card1/device/devcoredump/data
Mar 07 17:50:46 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: ring gfx_0.0.0 timeout, signaled seq=1731625, emitted seq=1731627
Mar 07 17:50:46 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: Process wayvr pid 22573 thread wayvr pid 22573
Mar 07 17:50:46 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: Starting gfx_0.0.0 ring reset
Mar 07 17:50:46 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: Ring gfx_0.0.0 reset failed
Mar 07 17:50:46 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: GPU reset begin!. Source: 1
Mar 07 17:50:46 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: MODE1 reset
Mar 07 17:50:46 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: GPU mode1 reset
Mar 07 17:50:46 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: GPU smu mode1 reset

Mar 07 17:50:51 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: GPU mode1 reset failed
Mar 07 17:50:51 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: ASIC reset failed with error, -62 for drm dev, 0000:2d:00.0
Mar 07 17:50:51 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: GPU reset end with ret = -62
Mar 07 17:50:51 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: GPU Recovery Failed: -62

ar 07 17:51:02 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: ring gfx_0.0.0 timeout, signaled seq=1731627, emitted seq=1731627
Mar 07 17:51:02 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: Process wayvr pid 22573 thread wayvr pid 22573
Mar 07 17:51:02 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: Starting gfx_0.0.0 ring reset
Mar 07 17:51:02 walfB550 kernel: amdgpu 0000:2d:00.0: amdgpu: Ring gfx_0.0.0 reset failed

[galister]

hi, around when did you first encounter this issue? do you remember the kernel version that you first had to downgrade from?

also, what's your wayvr version?

[Cubbix]

Hello @galister
The issue had first started about 1 month ago. It would have been whatever the latest Kernel was at the time. It was suggested to me to roll back to an LTS kernel, at that time being 6.12.75. That fixed the issue, however now that we have moved on to the 6.18 LTS Kernel, it appears that the issue has also migrated to the new LTS kernel as well.

Currently running the latest Kernel at this time: 6.19.6

WayVR: 26.2.1-1
wivrn-server 26.2.3-1
wivrn-dashboard 26.2.3-1
opencomposite-git r1232.cff07db-1

[galister]

I'm not sure, I often had GPU hangs this last 4-5 months, but not necessarily just with WayVR.

Trying on Mesa 26 + kernel 6.19.x would cause gpu hangs for me as well.

After downgrading, however, current setup seems to be holding out well, with an uptime of 6+ days as of now:

• Kernel: 6.18.12-gentoo-dist
• Mesa: 25.3.6
• linux-firmware: 20260110

Maybe try LTS kernel with mesa 25.x, see if it works for you.

If that doesn't work, try run wayvr with VK_INSTANCE_LAYERS=VK_LAYER_KHRONOS_validation and see if there are any vulkan validation issues.

[galister]

If you're having a hard time capturing logs due to the screen freeze, try run wayvr in tmux/zellij.

Then even if the screen freezes, it's possible to log in to the computer via SSH and reattach to the tmux/zellij session and see the logs in real-time.

[galister]

hmm, most of the search results for GCVM_L2_PROTECTION_FAULT_STATUS come back with the solution to downgrade linux-firmware