vkd3d-proton-2.12 causes ring gfx_0.0.0 timeout on EVE Online

[ExecutorElassus]

Please describe your issue as accurately as possible. Include screenshots or videos if relevant

Software information

EVE Online (local install, not Steam)

System information

• GPU: Sapphire Radeon RX5700 XT
• Driver: amdgpu
• Wine version: wine-proton-8.0.5c
• VKD3D-Proton version: 2.12

Log files

dmesg output during crash:

[ 9508.100576] traps: glxgears[869] trap invalid opcode ip:5573bdda7855 sp:7ffe267f42d0 error:0 in glxgears[5573bdda7000+3000]
[10132.434965] [drm:amdgpu_job_timedout] *ERROR* ring gfx_0.0.0 timeout, signaled seq=2729093, emitted seq=2729095
[10132.434976] [drm:amdgpu_job_timedout] *ERROR* Process information: process exefile.exe pid 8730 thread dxvk-submit pid 8801
[10132.434981] amdgpu 0000:09:00.0: amdgpu: GPU reset begin!
[10136.435163] amdgpu 0000:09:00.0: amdgpu: failed to suspend display audio
[10136.753866] amdgpu 0000:09:00.0: amdgpu: BACO reset
[10138.821097] [drm:amdgpu_dm_atomic_check] *ERROR* [CRTC:85:crtc-0] hw_done or flip_done timed out
[10139.861254] amdgpu 0000:09:00.0: amdgpu: GPU reset succeeded, trying to resume
[10139.861552] [drm] PCIE GART of 512M enabled (table at 0x0000008000900000).
[10139.861589] [drm] VRAM is lost due to GPU reset!
[10139.861590] [drm] PSP is resuming...
[10139.912886] [drm] reserve 0x900000 from 0x81fd000000 for PSP TMR
[10139.980885] amdgpu 0000:09:00.0: amdgpu: RAS: optional ras ta ucode is not available
[10139.988884] amdgpu 0000:09:00.0: amdgpu: RAP: optional rap ta ucode is not available
[10139.988887] amdgpu 0000:09:00.0: amdgpu: SECUREDISPLAY: securedisplay ta ucode is not available
[10139.988891] amdgpu 0000:09:00.0: amdgpu: SMU is resuming...
[10139.988932] amdgpu 0000:09:00.0: amdgpu: use vbios provided pptable
[10139.988935] amdgpu 0000:09:00.0: amdgpu: smc_dpm_info table revision(format.content): 4.5
[10139.991776] amdgpu 0000:09:00.0: amdgpu: SMU is resumed successfully!
[10140.557794] [drm] kiq ring mec 2 pipe 1 q 0
[10140.561590] [drm] VCN decode and encode initialized successfully(under DPG Mode).
[10140.561735] [drm] JPEG decode initialized successfully.
[10140.561752] amdgpu 0000:09:00.0: amdgpu: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
[10140.561755] amdgpu 0000:09:00.0: amdgpu: ring comp_1.0.0 uses VM inv eng 1 on hub 0
[10140.561757] amdgpu 0000:09:00.0: amdgpu: ring comp_1.1.0 uses VM inv eng 4 on hub 0
[10140.561760] amdgpu 0000:09:00.0: amdgpu: ring comp_1.2.0 uses VM inv eng 5 on hub 0
[10140.561762] amdgpu 0000:09:00.0: amdgpu: ring comp_1.3.0 uses VM inv eng 6 on hub 0
[10140.561764] amdgpu 0000:09:00.0: amdgpu: ring comp_1.0.1 uses VM inv eng 7 on hub 0
[10140.561766] amdgpu 0000:09:00.0: amdgpu: ring comp_1.1.1 uses VM inv eng 8 on hub 0
[10140.561769] amdgpu 0000:09:00.0: amdgpu: ring comp_1.2.1 uses VM inv eng 9 on hub 0
[10140.561771] amdgpu 0000:09:00.0: amdgpu: ring comp_1.3.1 uses VM inv eng 10 on hub 0
[10140.561773] amdgpu 0000:09:00.0: amdgpu: ring kiq_0.2.1.0 uses VM inv eng 11 on hub 0
[10140.561776] amdgpu 0000:09:00.0: amdgpu: ring sdma0 uses VM inv eng 12 on hub 0
[10140.561778] amdgpu 0000:09:00.0: amdgpu: ring sdma1 uses VM inv eng 13 on hub 0
[10140.561780] amdgpu 0000:09:00.0: amdgpu: ring vcn_dec uses VM inv eng 0 on hub 8
[10140.561783] amdgpu 0000:09:00.0: amdgpu: ring vcn_enc0 uses VM inv eng 1 on hub 8
[10140.561785] amdgpu 0000:09:00.0: amdgpu: ring vcn_enc1 uses VM inv eng 4 on hub 8
[10140.561787] amdgpu 0000:09:00.0: amdgpu: ring jpeg_dec uses VM inv eng 5 on hub 8
[10140.564414] amdgpu 0000:09:00.0: amdgpu: recover vram bo from shadow start
[10140.569843] amdgpu 0000:09:00.0: amdgpu: recover vram bo from shadow done
[10140.569866] [drm] Skip scheduling IBs!
[10140.569868] amdgpu 0000:09:00.0: amdgpu: GPU reset(2) succeeded!

Starting EVE Online from the launcher, the screen freezes midway through loading the character-selection screen, and then the whole desktop shuts down, restarting the X server and kicking me to the system login screen. Downgrading to vkd3d-proton-2.11.1 removed the issue, and EVE starts normally.

[Blisto91]

We likely need to know the driver version. Post a Wine or Proton log if you can

[ExecutorElassus]

If you mean the GPU driver, the system reports it as "4.6 Mesa 24.0.2". I'll see if I can get EVE to write to a log that persists after the WM restarts.

[ExecutorElassus]

I'm not sure how much help this will be, but this is my Proton log from trying to start Horizon Forbidden West through the Steam launcher using GE-Proton-9-2 (which also uses vkd3d-2.12). On that platform, it gives me an error about my card not having support for Shader Model 6.6, and quitting. Maybe it's related? steam-2420110.log

[Blisto91]

You seem to have the amdvlk driver installed as both it and radv (mesa) gets listed with your GPU. Don't know if that is intended from you. amdvlk is known to not be fully compatible with vkd3d-proton

[ExecutorElassus]

hmm … so, can I safely remove that driver entirely? Nothing seems to depend on it.

[Blisto91]

If you also have both 32bit and 64bit radv (radeon-vulkan) installed then yes that should be fine. They will get used instead. The name of the driver depends on the distros package manager

[ExecutorElassus]

once I remove amdvlk, do I just need to restart the WM to use it, or reboot? Are there any settings I need to change?

[Blisto91]

Should be fine to just try again without a reboot when it is just the Vulkan driver. There shouldn't be any settings you need to mess with.

[ExecutorElassus]

OMG That worked! Removing amdvlk got Horizon Forbidden West to work under Steam, holy crap! (I know that's not the OP, but I'll check that in the morning. There's a good chance that proton-2.12 conflicts with amdvlk in a way that the previous version didn't) Amazing, thank you so much!