AMD Radeon iGPU randomly resets whilst playing games

Ask Fedora
, , , , , , ,
1

UPDATE: the issue was caused by an outdated BIOS and unrelated to Fedora Linux. If you’re experiencing issues almost identical to mine, see what I marked as solution. If you’re not in the same boat, you can still skim the thread for some useful insight if you’re having a similar AMDGPU problem.

The title is self-explanatory… my GPU randomly resets while gaming, crashing the game and restarting KWin effects.

Now follows some useful context. To begin, here is some relevant system information of mine, copy-pasted from fastfetch:
OS: Fedora Linux 43 (Forty Three) x86_64
Host: 83DD (IdeaPad Slim 5 16AHP9)
Kernel: Linux 7.1.5-101.fc43.x86_64
Display (AGO0001): 1920x1080 in 13", 60 Hz [External]
DE: KDE Plasma 6.7.3
WM: KWin (Wayland)
WM Theme: SMOD
Theme: Windows7Aero (Aero) [Qt], Windows-7-Better [GTK2/3]
Icons: Windows 7 Aero [Qt], Windows 7 Aero [GTK2/3/4]
Font: Noto Sans (10pt) [Qt], Noto Sans (10pt) [GTK2/3/4]
Cursor: aero-drop (24px)
CPU: AMD Ryzen 7 8845HS (16) @ 5.14 GHz
GPU: AMD Radeon 780M Graphics [Integrated]
Memory: 3.43 GiB / 27.20 GiB (13%)
Swap: 0 B / 8.00 GiB (0%)
Disk (/): 845.93 GiB / 952.28 GiB (89%) - btrfs
Local IP (wlp2s0): 192.168.0.203/24
Battery (L23M3PK1): 100% [AC Connected]
Locale: en_GB.UTF-8

Second of all, I must talk about the issue itself as well. I don’t remember exactly when it started or what could have caused it, as it began happening quite some time ago and I simply put up with it. I’m asking here to get some insight into what the most likely cause could be and how I could go about troubleshooting and fixing this annoying issue.

Here are some of the things that I’ve done to my system that I recall to be around when the issue started, sorted from what I think is the most likely cause to the least likely.

  • Installing AMD ROCm
  • Installing COSMIC over my existing install (which was Fedora Sway). Sometime after, I ditched COSMIC for KDE Plasma, but my GPU was already having issues before I installed KDE. Obviously, I’m also having the same problem on KDE, otherwise I wouldn’t be here. Right now, I have all three desktops installed on my computer.
  • Using an external monitor (that being an Acer V223HQ, a 1080p LED monitor like any other you’d find in early 2010s, it really doesn’t matter in contrast to the HDMI>VGA adapter which is what’s actually plugged into my laptop. Though I doubt this as I remember my games crashing on the internal display as well, even if not sure).
    What I do know for sure is that I experienced the same random resets when connected to my TV over HDMI.
  • Running local LLMs via LocalAI (related to point 1)

Those are the things that I’ve done personally that might’ve led me here. And, of course, I can’t rule out the dreaded possibility of bad hardware, which I’d hope is not the case…

I know this report is really vague but I’d like to be pointed into the right direction for troubleshooting this mess… I am also not reimaging my computer, I have so many configs and hundreds of gigabytes of data and programs collected on a single BTRFS partition, maybe I could figure out how to reinstall Fedora over my existing root partition without deleting any data?

Third of all, let me cover what finally tipped me over to open a forum thread about this… I had just installed the Windows version of Minecraft Bedrock via this launcher called BedrockOnLinux and it was my second time playing it. I was on a server when, as usual, my monitor went black for a few seconds, displaying the “no signal” message. Then it goes back to my desktop but the game is frozen, my mouse is completely gone when inside the window and I get an “application not responding” dialog on top of the game which I can only OK away using the keyboard, after which I am back to the launcher… sometimes, but not in this case (probably due to DND settings or something), this is accompanied by a “KWin effects were restarted due to a graphics reset” notification after the fact.

I was absolutely done for after this incident, so I went to my terminal and printed out the dmesg. Sure enough, a bunch of blah blah about the GPU and it having successfully reset, this log will be quoted at the bottom. Read on for now.

As I said before, this issue is not KDE-exclusive and also not game specific. It also has nothing to do with my Plasma theme as I was having identical problems on COSMIC as well without any custom theme.

Other games that have caused my GPU driver to crash, non-exhaustive list: Minecraft Bedrock (but through mcpelauncher), BeamNG.drive, Titanfall 2…

There are others but I can’t say for sure, I literally can’t remember on which games it happens as I’ve been trying so hard to ignore the blatant problem.

There is absolutely no pattern to this, it is completely random! Sometimes, Minecraft would lock up before I can even connect to a server. Or a few minutes after I join. I could have been playing BeamNG perfectly fine for 1-2 hours, then I get the dreaded black screen and a not responding dialog followed by the game’s error reporter. All of these are accompanied by the KWin effects in a synchronised harmony.

Another thing to mention is that this has never happened during non-gaming use of my computer, even during GPU-intensive operations like prompting a local LLM.

And finally, here is the dmesg collected right after I got kicked out of Minecraft for Windows… I know for sure it’s the same thing for all the other games even though I haven’t seen dmesg, as the same graphics reset is mentioned after KWin panics.

Though it is a great excuse for procrastination, playing every game on my computer until my GPU resets then collecting dmesg, as opposed to doing the important work I don’t feel like doing.

I tend to ramble a lot but I’m done now, here is the dmesg.

[   53.932175] amdgpu 0000:04:00.0: [drm] REG_WAIT timeout 1us * 100 tries - dcn31_program_compbuf_size line:141
[  331.410461] amdgpu 0000:04:00.0: Dumping IP State
[  331.412783] amdgpu 0000:04:00.0: Dumping IP State Completed
[  331.412794] amdgpu 0000:04:00.0: [drm] AMDGPU device coredump file has been created
[  331.412796] amdgpu 0000:04:00.0: [drm] Check your /sys/class/drm/card1/device/devcoredump/data
[  331.412798] amdgpu 0000:04:00.0: ring gfx_0.0.0 timeout, signaled seq=54225, emitted seq=54227
[  331.412800] amdgpu 0000:04:00.0:  Process MINECRAFT MAIN  pid 4990 thread vkd3d_queue pid 5081
[  331.412802] amdgpu 0000:04:00.0: Starting gfx_0.0.0 ring reset
[  333.416612] amdgpu 0000:04:00.0: MES failed to respond to msg=RESET
[  333.416619] amdgpu 0000:04:00.0: failed to reset legacy queue
[  333.416621] amdgpu 0000:04:00.0: reset via MES failed and try pipe reset -110
[  333.416623] amdgpu 0000:04:00.0: The CPFW hasn't support pipe reset yet.
[  333.416625] amdgpu 0000:04:00.0: Ring gfx_0.0.0 reset failed
[  333.416628] amdgpu 0000:04:00.0: GPU reset begin!. Source:  1
[  333.424239] amdgpu 0000:04:00.0: [drm] *ERROR* Failed to initialize parser -125!
[  335.463265] amdgpu 0000:04:00.0: MES failed to respond to msg=REMOVE_QUEUE
[  335.463279] amdgpu 0000:04:00.0: failed to unmap legacy queue
[  335.737206] [drm:gfx_v11_0_hw_fini [amdgpu]] *ERROR* failed to halt cp gfx
[  335.739173] amdgpu 0000:04:00.0: MODE2 reset
[  335.777176] amdgpu 0000:04:00.0: GPU reset succeeded, trying to resume
[  335.777808] amdgpu 0000:04:00.0: [drm] PCIE GART of 512M enabled (table at 0x00000080FFD00000).
[  335.777895] amdgpu 0000:04:00.0: SMU is resuming...
[  335.779928] amdgpu 0000:04:00.0: SMU is resumed successfully!
[  335.787014] amdgpu 0000:04:00.0: [drm] DMUB hardware initialized: version=0x08005D00
[  336.044962] amdgpu 0000:04:00.0: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
[  336.044968] amdgpu 0000:04:00.0: ring comp_1.0.0 uses VM inv eng 1 on hub 0
[  336.044971] amdgpu 0000:04:00.0: ring comp_1.1.0 uses VM inv eng 4 on hub 0
[  336.044972] amdgpu 0000:04:00.0: ring comp_1.2.0 uses VM inv eng 6 on hub 0
[  336.044974] amdgpu 0000:04:00.0: ring comp_1.3.0 uses VM inv eng 7 on hub 0
[  336.044975] amdgpu 0000:04:00.0: ring comp_1.0.1 uses VM inv eng 8 on hub 0
[  336.044977] amdgpu 0000:04:00.0: ring comp_1.1.1 uses VM inv eng 9 on hub 0
[  336.044978] amdgpu 0000:04:00.0: ring comp_1.2.1 uses VM inv eng 10 on hub 0
[  336.044979] amdgpu 0000:04:00.0: ring comp_1.3.1 uses VM inv eng 11 on hub 0
[  336.044981] amdgpu 0000:04:00.0: ring sdma0 uses VM inv eng 12 on hub 0
[  336.044982] amdgpu 0000:04:00.0: ring vcn_unified_0 uses VM inv eng 0 on hub 8
[  336.044984] amdgpu 0000:04:00.0: ring jpeg_dec uses VM inv eng 1 on hub 8
[  336.044986] amdgpu 0000:04:00.0: ring mes_kiq_3.1.0 uses VM inv eng 13 on hub 0
[  336.046784] amdgpu 0000:04:00.0: GPU reset(1) succeeded!
[  336.046798] amdgpu 0000:04:00.0: [drm] device wedged, but no recovery needed
[  343.697676] clocksource: Watchdog remote CPU 12 read timed out
[  382.501108] evm: overlay not supported
[  386.184202] usb 1-4: reset full-speed USB device number 3 using xhci_hcd

And to clarify, I have tried checking my /sys/class/drm/card1/device/devcoredump/data by opening the folder in Dolphin, and while there were some virtual files inside, the folder completely vanished seconds after opening it and I got kicked out… how am I supposed to check it then? There was also no chance to retry as the thing was already gone without a trace.

Please point me in the right direction… also excuse me for any potential mistakes or confusion in all the text above, English is not my first language.

Thanks in advance for your advice!

2

Second occurrence since starting this thread…

[ 2203.989317] process 'common/BeamNG.drive/BinLinux/BeamNG.drive.x64' started with executable stack
[ 3941.505931] amdgpu 0000:04:00.0: Dumping IP State
[ 3941.508670] amdgpu 0000:04:00.0: Dumping IP State Completed
[ 3941.508685] amdgpu 0000:04:00.0: [drm] AMDGPU device coredump file has been created
[ 3941.508687] amdgpu 0000:04:00.0: [drm] Check your /sys/class/drm/card1/device/devcoredump/data
[ 3941.508689] amdgpu 0000:04:00.0: ring gfx_0.0.0 timeout, signaled seq=1004833, emitted seq=1004835
[ 3941.508692] amdgpu 0000:04:00.0:  Process ===>>> BeamNGMa pid 16115 thread   --- Render -- pid 16134
[ 3941.508696] amdgpu 0000:04:00.0: Starting gfx_0.0.0 ring reset
[ 3943.512533] amdgpu 0000:04:00.0: MES failed to respond to msg=RESET
[ 3943.512541] amdgpu 0000:04:00.0: failed to reset legacy queue
[ 3943.512543] amdgpu 0000:04:00.0: reset via MES failed and try pipe reset -110
[ 3943.512546] amdgpu 0000:04:00.0: The CPFW hasn't support pipe reset yet.
[ 3943.512547] amdgpu 0000:04:00.0: Ring gfx_0.0.0 reset failed
[ 3943.512550] amdgpu 0000:04:00.0: GPU reset begin!. Source:  1
[ 3945.556763] amdgpu 0000:04:00.0: MES failed to respond to msg=REMOVE_QUEUE
[ 3945.556776] amdgpu 0000:04:00.0: failed to unmap legacy queue
[ 3945.829942] [drm:gfx_v11_0_hw_fini [amdgpu]] *ERROR* failed to halt cp gfx
[ 3945.831958] amdgpu 0000:04:00.0: MODE2 reset
[ 3945.867399] amdgpu 0000:04:00.0: GPU reset succeeded, trying to resume
[ 3945.868054] amdgpu 0000:04:00.0: [drm] PCIE GART of 512M enabled (table at 0x00000080FFD00000).
[ 3945.868132] amdgpu 0000:04:00.0: SMU is resuming...
[ 3945.870838] amdgpu 0000:04:00.0: SMU is resumed successfully!
[ 3945.877070] amdgpu 0000:04:00.0: [drm] DMUB hardware initialized: version=0x08005D00
[ 3946.139750] amdgpu 0000:04:00.0: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
[ 3946.139764] amdgpu 0000:04:00.0: ring comp_1.0.0 uses VM inv eng 1 on hub 0
[ 3946.139766] amdgpu 0000:04:00.0: ring comp_1.1.0 uses VM inv eng 4 on hub 0
[ 3946.139768] amdgpu 0000:04:00.0: ring comp_1.2.0 uses VM inv eng 6 on hub 0
[ 3946.139769] amdgpu 0000:04:00.0: ring comp_1.3.0 uses VM inv eng 7 on hub 0
[ 3946.139771] amdgpu 0000:04:00.0: ring comp_1.0.1 uses VM inv eng 8 on hub 0
[ 3946.139772] amdgpu 0000:04:00.0: ring comp_1.1.1 uses VM inv eng 9 on hub 0
[ 3946.139774] amdgpu 0000:04:00.0: ring comp_1.2.1 uses VM inv eng 10 on hub 0
[ 3946.139776] amdgpu 0000:04:00.0: ring comp_1.3.1 uses VM inv eng 11 on hub 0
[ 3946.139778] amdgpu 0000:04:00.0: ring sdma0 uses VM inv eng 12 on hub 0
[ 3946.139780] amdgpu 0000:04:00.0: ring vcn_unified_0 uses VM inv eng 0 on hub 8
[ 3946.139782] amdgpu 0000:04:00.0: ring jpeg_dec uses VM inv eng 1 on hub 8
[ 3946.139784] amdgpu 0000:04:00.0: ring mes_kiq_3.1.0 uses VM inv eng 13 on hub 0
[ 3946.141716] amdgpu 0000:04:00.0: GPU reset(1) succeeded!
[ 3946.141731] amdgpu 0000:04:00.0: [drm] device wedged, but no recovery needed
[ 3946.152478] amdgpu 0000:04:00.0: [drm] *ERROR* Failed to initialize parser -125!
[ 3965.637118] usb 1-4: reset full-speed USB device number 3 using xhci_hcd
3

Update: A play of Garry’s Mod got me another reset but, this time around, I also got my hands on the core dump data:

/sys/class/drm/card1/device/devcoredump/data

dmesg:

[ 2898.181712] amdgpu 0000:04:00.0: Dumping IP State
[ 2898.184344] amdgpu 0000:04:00.0: Dumping IP State Completed
[ 2898.184356] amdgpu 0000:04:00.0: [drm] AMDGPU device coredump file has been created
[ 2898.184357] amdgpu 0000:04:00.0: [drm] Check your /sys/class/drm/card1/device/devcoredump/data
[ 2898.184359] amdgpu 0000:04:00.0: ring gfx_0.0.0 timeout, signaled seq=656637, emitted seq=656639
[ 2898.184361] amdgpu 0000:04:00.0:  Process gmod pid 7431 thread gmod:cs0 pid 7463
[ 2898.184363] amdgpu 0000:04:00.0: Starting gfx_0.0.0 ring reset
[ 2900.188172] amdgpu 0000:04:00.0: MES failed to respond to msg=RESET
[ 2900.188179] amdgpu 0000:04:00.0: failed to reset legacy queue
[ 2900.188181] amdgpu 0000:04:00.0: reset via MES failed and try pipe reset -110
[ 2900.188184] amdgpu 0000:04:00.0: The CPFW hasn't support pipe reset yet.
[ 2900.188186] amdgpu 0000:04:00.0: Ring gfx_0.0.0 reset failed
[ 2900.188189] amdgpu 0000:04:00.0: GPU reset begin!. Source:  1
[ 2902.232277] amdgpu 0000:04:00.0: MES failed to respond to msg=REMOVE_QUEUE
[ 2902.232290] amdgpu 0000:04:00.0: failed to unmap legacy queue
[ 2902.505846] [drm:gfx_v11_0_hw_fini [amdgpu]] *ERROR* failed to halt cp gfx
[ 2902.507837] amdgpu 0000:04:00.0: MODE2 reset
[ 2902.545028] amdgpu 0000:04:00.0: GPU reset succeeded, trying to resume
[ 2902.545676] amdgpu 0000:04:00.0: [drm] PCIE GART of 512M enabled (table at 0x00000080FFD00000).
[ 2902.545782] amdgpu 0000:04:00.0: SMU is resuming...
[ 2902.547734] amdgpu 0000:04:00.0: SMU is resumed successfully!
[ 2902.554807] amdgpu 0000:04:00.0: [drm] DMUB hardware initialized: version=0x08005D00
[ 2902.815188] amdgpu 0000:04:00.0: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
[ 2902.815199] amdgpu 0000:04:00.0: ring comp_1.0.0 uses VM inv eng 1 on hub 0
[ 2902.815201] amdgpu 0000:04:00.0: ring comp_1.1.0 uses VM inv eng 4 on hub 0
[ 2902.815203] amdgpu 0000:04:00.0: ring comp_1.2.0 uses VM inv eng 6 on hub 0
[ 2902.815204] amdgpu 0000:04:00.0: ring comp_1.3.0 uses VM inv eng 7 on hub 0
[ 2902.815206] amdgpu 0000:04:00.0: ring comp_1.0.1 uses VM inv eng 8 on hub 0
[ 2902.815208] amdgpu 0000:04:00.0: ring comp_1.1.1 uses VM inv eng 9 on hub 0
[ 2902.815210] amdgpu 0000:04:00.0: ring comp_1.2.1 uses VM inv eng 10 on hub 0
[ 2902.815212] amdgpu 0000:04:00.0: ring comp_1.3.1 uses VM inv eng 11 on hub 0
[ 2902.815214] amdgpu 0000:04:00.0: ring sdma0 uses VM inv eng 12 on hub 0
[ 2902.815216] amdgpu 0000:04:00.0: ring vcn_unified_0 uses VM inv eng 0 on hub 8
[ 2902.815218] amdgpu 0000:04:00.0: ring jpeg_dec uses VM inv eng 1 on hub 8
[ 2902.815220] amdgpu 0000:04:00.0: ring mes_kiq_3.1.0 uses VM inv eng 13 on hub 0
[ 2902.817237] amdgpu 0000:04:00.0: GPU reset(1) succeeded!
[ 2902.817253] amdgpu 0000:04:00.0: [drm] device wedged, but no recovery needed
[ 2902.822329] amdgpu 0000:04:00.0: [drm] *ERROR* Failed to initialize parser -125!
[ 2911.968008] usb 1-4: reset full-speed USB device number 3 using xhci_hcd
[ 3055.702619] usb 1-4: reset full-speed USB device number 3 using xhci_hcd
[ 3089.575740] usb 1-4: reset full-speed USB device number 3 using xhci_hcd
4

In search of a solution, I have come across this thread by a user having issues basically identical to mine, judging by the dmesg.
https://discussion.fedoraproject.org/t/fedora-43-gnome-desktop-crashing-periodically-since-kernel-6-17-8/178743/8

In the replies to this post, it’s suggested that the current version of amd-gpu-firmware is problematic and it must be downgraded.

So I went ahead and downgraded the package by running dnf downgrade amd-gpu-firmware, which downgraded the entirety of linux-firmware to 20251021 along with it.

Hopeful that I finally fixed my struggles, I fired up Minecraft Bedrock and joined a server, only to be kicked out in less than five minutes…

It might be the firmware to blame, as it’s the most reasonable explanation. But I cannot try downgrading any further, as there aren’t any older firmware versions available for Fedora 43…

The lowest available version of the "amd-gpu-firmware.noarch" package is already installed, cannot downgrade it.

Dmesg from this run:

[  270.633103] amdgpu 0000:04:00.0: Dumping IP State
[  270.635340] amdgpu 0000:04:00.0: Dumping IP State Completed
[  270.635353] amdgpu 0000:04:00.0: [drm] AMDGPU device coredump file has been created
[  270.635356] amdgpu 0000:04:00.0: [drm] Check your /sys/class/drm/card1/device/devcoredump/data
[  270.635358] amdgpu 0000:04:00.0: ring gfx_0.0.0 timeout, signaled seq=69518, emitted seq=69521
[  270.635361] amdgpu 0000:04:00.0:  Process MINECRAFT MAIN  pid 4668 thread vkd3d_queue pid 4741
[  270.635364] amdgpu 0000:04:00.0: Starting gfx_0.0.0 ring reset
[  272.639172] amdgpu 0000:04:00.0: MES failed to respond to msg=RESET
[  272.639180] amdgpu 0000:04:00.0: failed to reset legacy queue
[  272.639183] amdgpu 0000:04:00.0: reset via MES failed and try pipe reset -110
[  272.639187] amdgpu 0000:04:00.0: The CPFW hasn't support pipe reset yet.
[  272.639189] amdgpu 0000:04:00.0: Ring gfx_0.0.0 reset failed
[  272.639195] amdgpu 0000:04:00.0: GPU reset begin!. Source:  1
[  272.651827] amdgpu 0000:04:00.0: [drm] *ERROR* Failed to initialize parser -125!
[  274.693061] amdgpu 0000:04:00.0: MES failed to respond to msg=REMOVE_QUEUE
[  274.693074] amdgpu 0000:04:00.0: failed to unmap legacy queue
[  274.966327] [drm:gfx_v11_0_hw_fini [amdgpu]] *ERROR* failed to halt cp gfx
[  274.968334] amdgpu 0000:04:00.0: MODE2 reset
[  275.005360] amdgpu 0000:04:00.0: GPU reset succeeded, trying to resume
[  275.005937] amdgpu 0000:04:00.0: [drm] PCIE GART of 512M enabled (table at 0x00000080FFD00000).
[  275.006018] amdgpu 0000:04:00.0: SMU is resuming...
[  275.008116] amdgpu 0000:04:00.0: SMU is resumed successfully!
[  275.014263] amdgpu 0000:04:00.0: [drm] DMUB hardware initialized: version=0x08005D00
[  275.275736] amdgpu 0000:04:00.0: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
[  275.275744] amdgpu 0000:04:00.0: ring comp_1.0.0 uses VM inv eng 1 on hub 0
[  275.275746] amdgpu 0000:04:00.0: ring comp_1.1.0 uses VM inv eng 4 on hub 0
[  275.275748] amdgpu 0000:04:00.0: ring comp_1.2.0 uses VM inv eng 6 on hub 0
[  275.275750] amdgpu 0000:04:00.0: ring comp_1.3.0 uses VM inv eng 7 on hub 0
[  275.275752] amdgpu 0000:04:00.0: ring comp_1.0.1 uses VM inv eng 8 on hub 0
[  275.275754] amdgpu 0000:04:00.0: ring comp_1.1.1 uses VM inv eng 9 on hub 0
[  275.275756] amdgpu 0000:04:00.0: ring comp_1.2.1 uses VM inv eng 10 on hub 0
[  275.275758] amdgpu 0000:04:00.0: ring comp_1.3.1 uses VM inv eng 11 on hub 0
[  275.275760] amdgpu 0000:04:00.0: ring sdma0 uses VM inv eng 12 on hub 0
[  275.275763] amdgpu 0000:04:00.0: ring vcn_unified_0 uses VM inv eng 0 on hub 8
[  275.275765] amdgpu 0000:04:00.0: ring jpeg_dec uses VM inv eng 1 on hub 8
[  275.275767] amdgpu 0000:04:00.0: ring mes_kiq_3.1.0 uses VM inv eng 13 on hub 0
[  275.277963] amdgpu 0000:04:00.0: GPU reset(1) succeeded!
[  275.277976] amdgpu 0000:04:00.0: [drm] device wedged, but no recovery needed

Though I’ll go ahead and ask: does the system cache firmware in any way? Is there any chance that I’m not running the downgraded firmware just yet? To make sure, I’ll keep the package downgraded and see how the issue progresses over time, but I’ve already put all of my energy into this stupid bug already.

Maybe it is a hardware issue after all? It doesn’t seem like it, and I’d sure hope not… I’m not sure if I’m still in warranty…

I’ll keep updating this thread…

5

Downgraded to amd-gpu-firmware 20251021 and kernel 6.16.0-65.fc43.x86_64 (downloaded from Koji), but the issue won’t go away. I’ve also tried setting power_dpm_force_performance_level to low (locks clock speed to lowest value, 800MHz in my case) and high (which forces maximum clock speed), as I’ve seen both of these suggested around the internet for similar issues, but the crashes persist.

Other than Minecraft Bedrock, the issue is most easily reproducible in Counter-Strike 2, where I’d be consistently crashed out with VK_ERROR_DEVICE_LOST usually before the warm-up even ends. I’ve already gotten banned by anti-cheat on a Minecraft server due to debugging the non-stop crashes (though I have successfully appealed that one), I might get VAC banned next which would be extremely annoying…

I’d call hardware issue as AMD iGPUs have reported similar problems before, linked to consistent hardware failures in the case of the Steam Deck. While I’m not using one of those, it’s pretty close hardware-wise to my system. But I still have to test under Windows to conclude that possibility.

Another dmesg which is slightly different for some reason, but same shtick:

[  324.449994] amdgpu 0000:04:00.0: amdgpu: Dumping IP State
[  324.452494] amdgpu 0000:04:00.0: amdgpu: Dumping IP State Completed
[  324.452550] amdgpu 0000:04:00.0: amdgpu: [drm] AMDGPU device coredump file has been created
[  324.452552] amdgpu 0000:04:00.0: amdgpu: [drm] Check your /sys/class/drm/card1/device/devcoredump/data
[  324.462556] amdgpu 0000:04:00.0: amdgpu: ring gfx_0.0.0 timeout, signaled seq=51325, emitted seq=51327
[  324.462560] amdgpu 0000:04:00.0: amdgpu: Process information: process cs2 pid 5321 thread VKRenderThread pid 5355
[  324.462562] amdgpu 0000:04:00.0: amdgpu: Starting gfx_0.0.0 ring reset
[  326.466374] amdgpu 0000:04:00.0: amdgpu: MES failed to respond to msg=RESET
[  326.466380] [drm:amdgpu_mes_reset_legacy_queue [amdgpu]] *ERROR* failed to reset legacy queue
[  326.466584] amdgpu 0000:04:00.0: amdgpu: reset via MES failed and try pipe reset -110
[  326.466587] amdgpu 0000:04:00.0: amdgpu: The CPFW hasn't support pipe reset yet.
[  326.466588] amdgpu 0000:04:00.0: amdgpu: Ring gfx_0.0.0 reset failure
[  326.466591] amdgpu 0000:04:00.0: amdgpu: GPU reset begin!
[  328.515109] amdgpu 0000:04:00.0: amdgpu: MES failed to respond to msg=REMOVE_QUEUE
[  328.515116] [drm:amdgpu_mes_unmap_legacy_queue [amdgpu]] *ERROR* failed to unmap legacy queue
[  328.726828] [drm:gfx_v11_0_hw_fini [amdgpu]] *ERROR* failed to halt cp gfx
[  328.728429] amdgpu 0000:04:00.0: amdgpu: MODE2 reset
[  328.764168] amdgpu 0000:04:00.0: amdgpu: GPU reset succeeded, trying to resume
[  328.764807] [drm] PCIE GART of 512M enabled (table at 0x00000080FFD00000).
[  328.764907] amdgpu 0000:04:00.0: amdgpu: SMU is resuming...
[  328.766908] amdgpu 0000:04:00.0: amdgpu: SMU is resumed successfully!
[  328.773000] amdgpu 0000:04:00.0: amdgpu: [drm] DMUB hardware initialized: version=0x08005D00
[  329.149880] amdgpu 0000:04:00.0: amdgpu: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
[  329.149889] amdgpu 0000:04:00.0: amdgpu: ring comp_1.0.0 uses VM inv eng 1 on hub 0
[  329.149892] amdgpu 0000:04:00.0: amdgpu: ring comp_1.1.0 uses VM inv eng 4 on hub 0
[  329.149895] amdgpu 0000:04:00.0: amdgpu: ring comp_1.2.0 uses VM inv eng 6 on hub 0
[  329.149897] amdgpu 0000:04:00.0: amdgpu: ring comp_1.3.0 uses VM inv eng 7 on hub 0
[  329.149900] amdgpu 0000:04:00.0: amdgpu: ring comp_1.0.1 uses VM inv eng 8 on hub 0
[  329.149902] amdgpu 0000:04:00.0: amdgpu: ring comp_1.1.1 uses VM inv eng 9 on hub 0
[  329.149905] amdgpu 0000:04:00.0: amdgpu: ring comp_1.2.1 uses VM inv eng 10 on hub 0
[  329.149908] amdgpu 0000:04:00.0: amdgpu: ring comp_1.3.1 uses VM inv eng 11 on hub 0
[  329.149910] amdgpu 0000:04:00.0: amdgpu: ring sdma0 uses VM inv eng 12 on hub 0
[  329.149913] amdgpu 0000:04:00.0: amdgpu: ring vcn_unified_0 uses VM inv eng 0 on hub 8
[  329.149916] amdgpu 0000:04:00.0: amdgpu: ring jpeg_dec uses VM inv eng 1 on hub 8
[  329.149919] amdgpu 0000:04:00.0: amdgpu: ring mes_kiq_3.1.0 uses VM inv eng 13 on hub 0
[  329.152169] amdgpu 0000:04:00.0: amdgpu: GPU reset(1) succeeded!
[  329.152178] amdgpu 0000:04:00.0: [drm] device wedged, but recovered through reset
[  333.465699] show_signal_msg: 25 callbacks suppressed
[  333.465704] VKRenderThread[5355]: segfault at 0 ip 00007fd79940fab0 sp 00007fd7673fd490 error 6 in libtier0.so[20eab0,7fd79932e000+3cc000] likely on CPU 2 (core 1, socket 0)
[  333.465723] Code: 00 00 4a 8b 74 e3 08 48 85 f6 74 60 31 c0 4c 89 ef 49 83 c4 01 e8 a0 fc f1 ff 8b 03 44 39 e0 7f e1 66 0f 1f 84 00 00 00 00 00 <c7> 04 25 00 00 00 00 00 00 00 00 0f 0b ba 03 00 00 00 f0 0f b1 15
[  357.821830] usb 1-4: reset full-speed USB device number 3 using xhci_hcd
6

I’ve seen the Zen or CachyOS kernels suggested as being more stable surrounding AMD, so I’ve installed the latter from this Copr repository. While it does seem to improve general system smoothness (or maybe just placebo effect, dunno), it does not improve the crashing situation.
I could try wiping everything but /home and reinstalling Fedora on top, to see if my current system is to blame. I could maybe run Steam from a Fedora live USB (some small game obviously that would fit in a 13 GB ramdisk) or I’ll set up a Windows to Go installation and booting that from USB.
Installing Windows on this computer is not an option unless I get another 1 TB SSD and do some disk cloning shenanigans, as there’s no other way to hold my 900 GB of /home in limbo. I’ll hopefully get around to trying those first two things this week and I’d be absolutely screwed if it happened under Windows as it would mean my GPU is dying :frowning:

Even if my debugging concludes this way, I don’t think I would just install Windows and call it a day… apparently Windows is much better at gracefully handling these types of failures and rapidly recovering the GPU without crashing the system or active programs, while great in of itself, it could also be great at masking worsening hardware problems until it’s too late… somehow would have to confirm that this is not happening by checking the system event log or somewhere else.

Either way, given everything that I’ve tried, I’ll probably end up, in a meh case scenario, back on Windows or have to replace my laptop entirely in the worst case… darn it. If I unluckily end up down the second path, I’m avoiding AMD at all costs.

Issue is consistently reproducible within first 10 minutes of launching CS2. It used to be much less frequent in the past, again pointing towards hardware rather than software. Other games (even non-Steam, but I don’t have many of those) exhibit the issue as well, but are ever-so-slightly more stable.

I’m still keeping this open as a last-ditch effort to catch this as a Linux problem and fix it through software, which I’m still hopeful about as there are soo many similar reports involving AMD GPUs.

Please stop by and drop your two cents if you have any change on you.

Some dmesg, but from cachyos kernel as opposed to default:

[  274.232904] amdgpu 0000:04:00.0: Dumping IP State
[  274.235571] amdgpu 0000:04:00.0: Dumping IP State Completed
[  274.235582] amdgpu 0000:04:00.0: [drm] AMDGPU device coredump file has been created
[  274.235584] amdgpu 0000:04:00.0: [drm] Check your /sys/class/drm/card1/device/devcoredump/data
[  274.235586] amdgpu 0000:04:00.0: ring gfx_0.0.0 timeout, signaled seq=70686, emitted seq=70688
[  274.235588] amdgpu 0000:04:00.0:  Process cs2 pid 5622 thread VKRenderThread pid 5656
[  274.235590] amdgpu 0000:04:00.0: Starting gfx_0.0.0 ring reset
[  276.239385] amdgpu 0000:04:00.0: MES failed to respond to msg=RESET
[  276.239391] amdgpu 0000:04:00.0: failed to reset legacy queue
[  276.239393] amdgpu 0000:04:00.0: reset via MES failed and try pipe reset -110
[  276.239395] amdgpu 0000:04:00.0: The CPFW hasn't support pipe reset yet.
[  276.239396] amdgpu 0000:04:00.0: Ring gfx_0.0.0 reset failed
[  276.239399] amdgpu 0000:04:00.0: GPU reset begin!. Source:  1
[  278.290828] amdgpu 0000:04:00.0: MES failed to respond to msg=REMOVE_QUEUE
[  278.290835] amdgpu 0000:04:00.0: failed to unmap legacy queue
[  278.498184] [drm:gfx_v11_0_hw_fini [amdgpu]] *ERROR* failed to halt cp gfx
[  278.499776] amdgpu 0000:04:00.0: MODE2 reset
[  278.538225] amdgpu 0000:04:00.0: GPU reset succeeded, trying to resume
[  278.538872] amdgpu 0000:04:00.0: [drm] PCIE GART of 512M enabled (table at 0x00000080FFD00000).
[  278.538957] amdgpu 0000:04:00.0: SMU is resuming...
[  278.541641] amdgpu 0000:04:00.0: SMU is resumed successfully!
[  278.547576] amdgpu 0000:04:00.0: [drm] DMUB hardware initialized: version=0x08005400
[  278.802956] amdgpu 0000:04:00.0: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
[  278.802965] amdgpu 0000:04:00.0: ring comp_1.0.0 uses VM inv eng 1 on hub 0
[  278.802967] amdgpu 0000:04:00.0: ring comp_1.1.0 uses VM inv eng 4 on hub 0
[  278.802968] amdgpu 0000:04:00.0: ring comp_1.2.0 uses VM inv eng 6 on hub 0
[  278.802970] amdgpu 0000:04:00.0: ring comp_1.3.0 uses VM inv eng 7 on hub 0
[  278.802972] amdgpu 0000:04:00.0: ring comp_1.0.1 uses VM inv eng 8 on hub 0
[  278.802974] amdgpu 0000:04:00.0: ring comp_1.1.1 uses VM inv eng 9 on hub 0
[  278.802975] amdgpu 0000:04:00.0: ring comp_1.2.1 uses VM inv eng 10 on hub 0
[  278.802977] amdgpu 0000:04:00.0: ring comp_1.3.1 uses VM inv eng 11 on hub 0
[  278.802979] amdgpu 0000:04:00.0: ring sdma0 uses VM inv eng 12 on hub 0
[  278.802981] amdgpu 0000:04:00.0: ring vcn_unified_0 uses VM inv eng 0 on hub 8
[  278.802983] amdgpu 0000:04:00.0: ring jpeg_dec uses VM inv eng 1 on hub 8
[  278.802985] amdgpu 0000:04:00.0: ring mes_kiq_3.1.0 uses VM inv eng 13 on hub 0
[  278.805137] amdgpu 0000:04:00.0: GPU reset(1) succeeded!
[  278.805151] amdgpu 0000:04:00.0: [drm] device wedged, but no recovery needed
[  278.810891] amdgpu 0000:04:00.0: [drm] *ERROR* Failed to initialize parser -125!
[  282.512906] show_signal_msg: 24 callbacks suppressed
[  282.512910] VKRenderThread[5656]: segfault at 0 ip 00007f8adbc0fab0 sp 00007f8ab51f9490 error 6 in libtier0.so[20eab0,7f8adbb2e000+3cc000] likely on CPU 3 (core 1, socket 0)
[  282.512922] Code: 00 00 4a 8b 74 e3 08 48 85 f6 74 60 31 c0 4c 89 ef 49 83 c4 01 e8 a0 fc f1 ff 8b 03 44 39 e0 7f e1 66 0f 1f 84 00 00 00 00 00 <c7> 04 25 00 00 00 00 00 00 00 00 0f 0b ba 03 00 00 00 f0 0f b1 15
[  348.908287] usb 1-4: reset full-speed USB device number 3 using xhci_hcd
7

I’ve added kernel parameters amdgpu.aspm=0, amdgpu.runpm=0, amdgpu.bapm=0 and pcie_aspm=off (not sure if applicable for an iGPU, but still tried it) as suggested by some forum threads but they did not have any effect. They disable some power management features or something like that…

I’ve also added the explosive parameter that is amdgpu.dpm=0 which, in some cases, stops the system from outputting anything. For me, it worked fine but, once I experienced an in-game GPU reset, HDMI-out died and never recovered (had to forcefully reboot by holding the power button…). Haven’t tried pulling out the cable though, maybe it would’ve gone back to the eDP.

Soon I’ll try reproducing the issues on the laptop’s eDP now that they’ve gotten so foolishly easy to reproduce. If I fail to reproduce, my hardware is fine and it’s actually the unjustifiably expensive HDMI>VGA converter that I wasted my money on. Maybe it’s behaving erratically and causing issues downstream with the GPU?

Though I have had this same breed of crashes when connected to a modern TV directly over HDMI, so either HDMI altogether is to blame or that speculation is a dead end.

I’ll report back after I’ve tried some of the above, most importantly I’ll try to reproduce the issue with nothing in the HDMI port.

8

I’ve run dracut --force to ensure that the downgraded firmware is being loaded into the initramfs. It seemed to have fixed the issue, but that’s not the case. The Minecraft server I joined forcefully disabled Bedrock’s Vibrant Visuals graphics mode, which is nicer graphics similar to installing shaders on Java edition of the game.

Falling back to the regular “Fancy” graphics reduced GPU load and allowed the game to remain stable. When I joined a different server, within the same Minecraft client run, which did not restrict the graphics, it only took 2-3 minutes for a GPU reset to occur.

While this could be blamed on the unsupported edition of Minecraft running through an unsupported launcher, the issue is also reproducible by playing (Linux-native build) CS2 and, less often, on other games.

Given that lower graphics settings in certain games fix the issue altogether, it has something to do with high GPU load… high GPU load that my system was handling with no problems whatsoever last year. Interestingly, underclocking does not (!) lead to the same effect as lowering graphics settings, as it does not fix the crashing.

Also, while Minecraft Bedrock as well as other games have been having the issue, the Java edition of Minecraft has been unaffected until now, apparently? I haven’t been playing Java until recently on this computer, but it ran without any GPU resets even when the GPU was under maximum load - game running at 15-20 FPS with a 128x texture pack and Complementary Unbound shader pack set to ULTRA preset.

Today, I fired up the same configuration expecting to write about its reliability, but it started pounding sand as well. I only got to fly around in Creative mode for a few minutes, taking some screenshots and switching to windowed mode and back to fullscreen, shortly after that, GPU reset.

Minecraft running windowed right after I entered the world, showing what graphics settings I’m onto here.

Screenshot_20260811_144529
Screenshot_20260811_1445291920×1080 322 KB

LACT running on top of Minecraft, showing clock speed between 2500-2700MHz and high temps (but again, my system could handle these just fine less than a year back).

Screenshot_20260811_144628
Screenshot_20260811_1446281920×1080 871 KB

Fullscreen this time, screenshot coincidentally taken seconds before the GPU reset.

Screenshot_20260811_144810
Screenshot_20260811_1448101920×1080 455 KB

All of it usually ends in a core dump depending on the game, but only now did I get an actual crash report notification, report of which attached at the end.

Note that Prism Launcher is installed as a Flatpak, as you might be able to tell from the quoted block below, but BedrockOnLinux launcher is an AppImage and causing the same issues, so are games from Steam which is installed from RPM Fusion (so native).

I’ll probably try to reproduce in a live environment, though I’ll have to pick a viable game for the test.

The thing:

           PID: 6401 (java)
           UID: 1000 (asicosilomu)
           GID: 1000 (asicosilomu)
        Signal: 6 (ABRT)
     Timestamp: Tue 2026-08-11 14:48:17 EEST (21s ago)
  Command Line: /home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/java/java-runtime-epsilon/bin/java -Duser.language=en -Xms512m -Xmx16384m -Djava.library.path=/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/instances/26.2/natives -cp /app/share/PrismLauncher/NewLaunch.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-freetype-natives-linux/3.4.1/lwjgl-freetype-natives-linux-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-freetype/3.4.1/lwjgl-freetype-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-jemalloc-natives-linux/3.4.1/lwjgl-jemalloc-natives-linux-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-jemalloc/3.4.1/lwjgl-jemalloc-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-natives-linux/3.4.1/lwjgl-natives-linux-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-openal-natives-linux/3.4.1/lwjgl-openal-natives-linux-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-openal/3.4.1/lwjgl-openal-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-opengl-natives-linux/3.4.1/lwjgl-opengl-natives-linux-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-opengl/3.4.1/lwjgl-opengl-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-sdl-natives-linux/3.4.1/lwjgl-sdl-natives-linux-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-sdl/3.4.1/lwjgl-sdl-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-shaderc-natives-linux/3.4.1/lwjgl-shaderc-natives-linux-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-shaderc/3.4.1/lwjgl-shaderc-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-spvc-natives-linux/3.4.1/lwjgl-spvc-natives-linux-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-spvc/3.4.1/lwjgl-spvc-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-stb-natives-linux/3.4.1/lwjgl-stb-natives-linux-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-stb/3.4.1/lwjgl-stb-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-tinyfd-natives-linux/3.4.1/lwjgl-tinyfd-natives-linux-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-tinyfd/3.4.1/lwjgl-tinyfd-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-vma-natives-linux/3.4.1/lwjgl-vma-natives-linux-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-vma/3.4.1/lwjgl-vma-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-vulkan/3.4.1/lwjgl-vulkan-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl/3.4.1/lwjgl-3.4.1-unsafe.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-glfw-natives-linux/3.4.1/lwjgl-glfw-natives-linux-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/lwjgl/lwjgl-glfw/3.4.1/lwjgl-glfw-3.4.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/at/yawk/lz4/lz4-java/1.10.1/lz4-java-1.10.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/com/azure/azure-json/1.4.0/azure-json-1.4.0.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/com/github/oshi/oshi-core/6.9.0/oshi-core-6.9.0.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/com/google/code/gson/gson/2.14.0/gson-2.14.0.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/com/google/guava/failureaccess/1.0.3/failureaccess-1.0.3.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/com/google/guava/guava/33.6.0-jre/guava-33.6.0-jre.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/com/ibm/icu/icu4j/78.3/icu4j-78.3.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/com/microsoft/azure/msal4j/1.24.1/msal4j-1.24.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/com/mojang/authlib/9.0.75/authlib-9.0.75.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/com/mojang/blocklist/1.0.10/blocklist-1.0.10.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/com/mojang/brigadier/1.3.10/brigadier-1.3.10.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/com/mojang/datafixerupper/10.0.21/datafixerupper-10.0.21.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/com/mojang/jtracy/1.0.37/jtracy-1.0.37.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/com/mojang/jtracy-natives-linux/1.0.37/jtracy-natives-linux-1.0.37.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/com/mojang/logging/1.7.12/logging-1.7.12.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/com/mojang/patchy/2.2.10/patchy-2.2.10.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/com/mojang/text2speech/1.19.12/text2speech-1.19.12.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/commons-codec/commons-codec/1.22.0/commons-codec-1.22.0.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/commons-io/commons-io/2.20.0/commons-io-2.20.0.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/io/netty/netty-buffer/4.2.15.Final/netty-buffer-4.2.15.Final.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/io/netty/netty-codec-base/4.2.15.Final/netty-codec-base-4.2.15.Final.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/io/netty/netty-codec-compression/4.2.15.Final/netty-codec-compression-4.2.15.Final.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/io/netty/netty-codec-http/4.2.15.Final/netty-codec-http-4.2.15.Final.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/io/netty/netty-common/4.2.15.Final/netty-common-4.2.15.Final.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/io/netty/netty-handler/4.2.15.Final/netty-handler-4.2.15.Final.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/io/netty/netty-resolver/4.2.15.Final/netty-resolver-4.2.15.Final.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/io/netty/netty-transport-classes-epoll/4.2.15.Final/netty-transport-classes-epoll-4.2.15.Final.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/io/netty/netty-transport-classes-kqueue/4.2.15.Final/netty-transport-classes-kqueue-4.2.15.Final.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/io/netty/netty-transport-native-epoll/4.2.15.Final/netty-transport-native-epoll-4.2.15.Final-linux-aarch_64.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/io/netty/netty-transport-native-epoll/4.2.15.Final/netty-transport-native-epoll-4.2.15.Final-linux-x86_64.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/io/netty/netty-transport-native-unix-common/4.2.15.Final/netty-transport-native-unix-common-4.2.15.Final.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/io/netty/netty-transport/4.2.15.Final/netty-transport-4.2.15.Final.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/it/unimi/dsi/fastutil/8.5.18/fastutil-8.5.18.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/net/java/dev/jna/jna-platform/5.17.0/jna-platform-5.17.0.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/net/java/dev/jna/jna/5.17.0/jna-5.17.0.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/net/sf/jopt-simple/jopt-simple/5.0.4/jopt-simple-5.0.4.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/apache/commons/commons-compress/1.28.0/commons-compress-1.28.0.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/apache/commons/commons-lang3/3.20.0/commons-lang3-3.20.0.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/apache/logging/log4j/log4j-api/2.26.0/log4j-api-2.26.0.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/apache/logging/log4j/log4j-core/2.26.0/log4j-core-2.26.0.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/apache/logging/log4j/log4j-slf4j2-impl/2.26.0/log4j-slf4j2-impl-2.26.0.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/jcraft/jorbis/0.0.17/jorbis-0.0.17.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/joml/joml/1.10.8/joml-1.10.8.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/jspecify/jspecify/1.0.0/jspecify-1.0.0.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/slf4j/slf4j-api/2.0.17/slf4j-api-2.0.17.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/net/fabricmc/intermediary/26.2/intermediary-26.2.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/ow2/asm/asm/9.10.1/asm-9.10.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/ow2/asm/asm-analysis/9.10.1/asm-analysis-9.10.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/ow2/asm/asm-commons/9.10.1/asm-commons-9.10.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/ow2/asm/asm-tree/9.10.1/asm-tree-9.10.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/org/ow2/asm/asm-util/9.10.1/asm-util-9.10.1.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/net/fabricmc/sponge-mixin/0.17.3+mixin.0.8.7/sponge-mixin-0.17.3+mixin.0.8.7.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/net/fabricmc/fabric-loader/0.19.3/fabric-loader-0.19.3.jar:/home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/libraries/com/mojang/minecraft/26.2/minecraft-26.2-client.jar org.prismlauncher.EntryPoint
    Executable: /home/asicosilomu/.var/app/org.prismlauncher.PrismLauncher/data/PrismLauncher/java/java-runtime-epsilon/bin/java
 Control Group: /user.slice/user-1000.slice/user@1000.service/app.slice/app-flatpak-org.prismlauncher.PrismLauncher-2186416583.scope
          Unit: user@1000.service
     User Unit: app-flatpak-org.prismlauncher.PrismLauncher-2186416583.scope
         Slice: user-1000.slice
     Owner UID: 1000 (asicosilomu)
       Boot ID: d6723b49e82246b1aeb1356a1a016cdc
    Machine ID: fa8d13a3f5be474fa861ecaa502a38b4
      Hostname: fedora
       Storage: /var/lib/systemd/coredump/core.java.1000.d6723b49e82246b1aeb1356a1a016cdc.6401.1786448897000000.zst (present)
  Size on Disk: 604.1M
       Message: Process 6401 (java) of user 1000 dumped core.
                
                Module libextnet.so without build-id.
                Module libsyslookup.so without build-id.
                Module libverify.so without build-id.
                Module libmanagement_ext.so without build-id.
                Module libmanagement.so without build-id.
                Module libnet.so without build-id.
                Module libnio.so without build-id.
                Module libsimdsort.so without build-id.
                Module libzip.so without build-id.
                Module libjava.so without build-id.
                Module libjimage.so without build-id.
                Module libjvm.so without build-id.
                Module libjli.so without build-id.
                Module java without build-id.
                Stack trace of thread 129:
                #0  0x00007faa759e8952 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x29952)
                #1  0x00007fa960823820 n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0xc23820)
                #2  0x00007fa960827a9f n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0xc27a9f)
                #3  0x00007fa960252df2 n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0x652df2)
                #4  0x00007fa96029d69c n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0x69d69c)
                #5  0x00007faa75a5966a n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x9a66a)
                #6  0x00007faa75adcf94 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x11df94)
                
                Stack trace of thread 142:
                #0  0x00007faa75a61ce2 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0xa2ce2)
                #1  0x00007faa75a55e3c n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x96e3c)
                #2  0x00007faa75a564ac n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x974ac)
                #3  0x00007faa75a58b7e n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x99b7e)
                #4  0x00007fa96029d76d n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0x69d76d)
                #5  0x00007fa960252d1c n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0x652d1c)
                #6  0x00007fa96029d69c n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0x69d69c)
                #7  0x00007faa75a5966a n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x9a66a)
                #8  0x00007faa75adcf94 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x11df94)
                
                Stack trace of thread 141:
                #0  0x00007faa75a61ce2 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0xa2ce2)
                #1  0x00007faa75a55e3c n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x96e3c)
                #2  0x00007faa75a564ac n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x974ac)
                #3  0x00007faa75a58b7e n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x99b7e)
                #4  0x00007fa96029d76d n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0x69d76d)
                #5  0x00007fa960252d1c n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0x652d1c)
                #6  0x00007fa96029d69c n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0x69d69c)
                #7  0x00007faa75a5966a n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x9a66a)
                #8  0x00007faa75adcf94 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x11df94)
                
                Stack trace of thread 70:
                #0  0x00007faa75a61ce2 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0xa2ce2)
                #1  0x00007faa75a55e3c n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x96e3c)
                #2  0x00007faa75a564ac n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x974ac)
                #3  0x00007faa75a61be8 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0xa2be8)
                #4  0x00007faa750de362 _ZN14PosixSemaphore4waitEv (libjvm.so + 0xede362)
                #5  0x00007faa753d14db _ZN12WorkerThread3runEv (libjvm.so + 0x11d14db)
                #6  0x00007faa75303578 _ZN6Thread8call_runEv (libjvm.so + 0x1103578)
                #7  0x00007faa75020a0b _ZL19thread_native_entryP6Thread (libjvm.so + 0xe20a0b)
                #8  0x00007faa75a5966a n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x9a66a)
                #9  0x00007faa75adcf94 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x11df94)
                
                Stack trace of thread 65:
                #0  0x00007faa75a61ce2 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0xa2ce2)
                #1  0x00007faa75a55e3c n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x96e3c)
                #2  0x00007faa75a564ac n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x974ac)
                #3  0x00007faa75a61be8 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0xa2be8)
                #4  0x00007faa750de362 _ZN14PosixSemaphore4waitEv (libjvm.so + 0xede362)
                #5  0x00007faa753d14db _ZN12WorkerThread3runEv (libjvm.so + 0x11d14db)
                #6  0x00007faa75303578 _ZN6Thread8call_runEv (libjvm.so + 0x1103578)
                #7  0x00007faa75020a0b _ZL19thread_native_entryP6Thread (libjvm.so + 0xe20a0b)
                #8  0x00007faa75a5966a n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x9a66a)
                #9  0x00007faa75adcf94 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x11df94)
                
                Stack trace of thread 75:
                #0  0x00007faa75a61ce2 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0xa2ce2)
                #1  0x00007faa75a55e3c n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x96e3c)
                #2  0x00007faa75a564ac n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x974ac)
                #3  0x00007faa75a61be8 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0xa2be8)
                #4  0x00007faa750de362 _ZN14PosixSemaphore4waitEv (libjvm.so + 0xede362)
                #5  0x00007faa753d14db _ZN12WorkerThread3runEv (libjvm.so + 0x11d14db)
                #6  0x00007faa75303578 _ZN6Thread8call_runEv (libjvm.so + 0x1103578)
                #7  0x00007faa75020a0b _ZL19thread_native_entryP6Thread (libjvm.so + 0xe20a0b)
                #8  0x00007faa75a5966a n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x9a66a)
                #9  0x00007faa75adcf94 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x11df94)
                
                Stack trace of thread 130:
                #0  0x00007faa75a61ce2 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0xa2ce2)
                #1  0x00007faa75a55e3c n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x96e3c)
                #2  0x00007faa75a564ac n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x974ac)
                #3  0x00007faa75a58b7e n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x99b7e)
                #4  0x00007fa96029d76d n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0x69d76d)
                #5  0x00007fa960252d1c n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0x652d1c)
                #6  0x00007fa96029d69c n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0x69d69c)
                #7  0x00007faa75a5966a n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x9a66a)
                #8  0x00007faa75adcf94 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x11df94)
                
                Stack trace of thread 135:
                #0  0x00007faa75a61ce2 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0xa2ce2)
                #1  0x00007faa75a55e3c n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x96e3c)
                #2  0x00007faa75a564ac n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x974ac)
                #3  0x00007faa75a58b7e n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x99b7e)
                #4  0x00007fa96029d76d n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0x69d76d)
                #5  0x00007fa960252d1c n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0x652d1c)
                #6  0x00007fa96029d69c n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0x69d69c)
                #7  0x00007faa75a5966a n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x9a66a)
                #8  0x00007faa75adcf94 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x11df94)
                
                Stack trace of thread 161:
                #0  0x00007faa75a61ce2 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0xa2ce2)
                #1  0x00007faa75a55e3c n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x96e3c)
                #2  0x00007faa75a55e84 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x96e84)
                #3  0x00007faa75acfdae n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x110dae)
                #4  0x00007fa962936ec8 n/a (/usr/lib/x86_64-linux-gnu/libpulse.so.0.24.3 + 0x34ec8)
                #5  0x00007fa962920b41 n/a (/usr/lib/x86_64-linux-gnu/libpulse.so.0.24.3 + 0x1eb41)
                #6  0x00007fa96292af91 n/a (/usr/lib/x86_64-linux-gnu/libpulse.so.0.24.3 + 0x28f91)
                #7  0x00007fa96292b050 n/a (/usr/lib/x86_64-linux-gnu/libpulse.so.0.24.3 + 0x29050)
                #8  0x00007fa96293bc2d n/a (/usr/lib/x86_64-linux-gnu/libpulse.so.0.24.3 + 0x39c2d)
                #9  0x00007fa9628df30b n/a (/usr/lib/x86_64-linux-gnu/pulseaudio/libpulsecommon-17.0.so + 0x5930b)
                #10 0x00007faa75a5966a n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x9a66a)
                #11 0x00007faa75adcf94 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x11df94)
                
                Stack trace of thread 132:
                #0  0x00007faa75a61ce2 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0xa2ce2)
                #1  0x00007faa75a55e3c n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x96e3c)
                #2  0x00007faa75a564ac n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x974ac)
                #3  0x00007faa75a58b7e n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x99b7e)
                #4  0x00007fa96029d76d n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0x69d76d)
                #5  0x00007fa960252d1c n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0x652d1c)
                #6  0x00007fa96029d69c n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0x69d69c)
                #7  0x00007faa75a5966a n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x9a66a)
                #8  0x00007faa75adcf94 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x11df94)
                
                Stack trace of thread 144:
                #0  0x00007faa75a61ce2 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0xa2ce2)
                #1  0x00007faa75a55e3c n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x96e3c)
                #2  0x00007faa75a564ac n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x974ac)
                #3  0x00007faa75a58b7e n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x99b7e)
                #4  0x00007fa96029d76d n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0x69d76d)
                #5  0x00007fa960252d1c n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0x652d1c)
                #6  0x00007fa96029d69c n/a (/usr/lib/x86_64-linux-gnu/GL/default/lib/libgallium-26.1.5.so + 0x69d69c)
                #7  0x00007faa75a5966a n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x9a66a)
                #8  0x00007faa75adcf94 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x11df94)
                
                Stack trace of thread 49:
                #0  0x00007faa75a61ce2 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0xa2ce2)
                #1  0x00007faa75a55e3c n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x96e3c)
                #2  0x00007faa75a564ac n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x974ac)
                #3  0x00007faa75a58d88 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x99d88)
                #4  0x00007faa7502d647 _ZN15PlatformMonitor4waitEm (libjvm.so + 0xe2d647)
                #5  0x00007faa74fcc539 _ZN7Monitor28wait_without_safepoint_checkEm (libjvm.so + 0xdcc539)
                #6  0x00007faa74fe7fed _ZNK13WatcherThread5sleepEv (libjvm.so + 0xde7fed)
                #7  0x00007faa74fe80f1 _ZN13WatcherThread3runEv (libjvm.so + 0xde80f1)
                #8  0x00007faa75303578 _ZN6Thread8call_runEv (libjvm.so + 0x1103578)
                #9  0x00007faa75020a0b _ZL19thread_native_entryP6Thread (libjvm.so + 0xe20a0b)
                #10 0x00007faa75a5966a n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x9a66a)
                #11 0x00007faa75adcf94 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x11df94)
                
                Stack trace of thread 56:
                #0  0x00007faa75a61ce2 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0xa2ce2)
                #1  0x00007faa75a55e3c n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x96e3c)
                #2  0x00007faa75a564ac n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x974ac)
                #3  0x00007faa75a58d88 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x99d88)
                #4  0x00007faa7502d647 _ZN15PlatformMonitor4waitEm (libjvm.so + 0xe2d647)
                #5  0x00007faa74fcc5be _ZN7Monitor4waitEm (libjvm.so + 0xdcc5be)
                #6  0x00007faa7492ca46 _ZN12CompileQueue3getEP14CompilerThread (libjvm.so + 0x72ca46)
                #7  0x00007faa7492fe6b _ZN13CompileBroker20compiler_thread_loopEv (libjvm.so + 0x72fe6b)
                #8  0x00007faa74c0be98 _ZN10JavaThread17thread_main_innerEv (libjvm.so + 0xa0be98)
                #9  0x00007faa75303578 _ZN6Thread8call_runEv (libjvm.so + 0x1103578)
                #10 0x00007faa75020a0b _ZL19thread_native_entryP6Thread (libjvm.so + 0xe20a0b)
                #11 0x00007faa75a5966a n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x9a66a)
                #12 0x00007faa75adcf94 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x11df94)
                
                Stack trace of thread 46:
                #0  0x00007faa75a61ce2 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0xa2ce2)
                #1  0x00007faa75a55e3c n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x96e3c)
                #2  0x00007faa75a564ac n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x974ac)
                #3  0x00007faa75a61be8 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0xa2be8)
                #4  0x00007faa750de362 _ZN14PosixSemaphore4waitEv (libjvm.so + 0xede362)
                #5  0x00007faa753d14db _ZN12WorkerThread3runEv (libjvm.so + 0x11d14db)
                #6  0x00007faa75303578 _ZN6Thread8call_runEv (libjvm.so + 0x1103578)
                #7  0x00007faa75020a0b _ZL19thread_native_entryP6Thread (libjvm.so + 0xe20a0b)
                #8  0x00007faa75a5966a n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x9a66a)
                #9  0x00007faa75adcf94 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x11df94)
                
                Stack trace of thread 191:
                #0  0x00007faa75adaf0d n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x11bf0d)
                #1  0x00007faa0824479c n/a (/tmp/lwjgl_asicosilomu/3.4.1-snapshot/x64/libopenal.so + 0x4479c)
                #2  0x00007faa08afef44 n/a (/usr/lib/x86_64-linux-gnu/libstdc++.so.6.0.34 + 0xfef44)
                #3  0x00007faa75a5966a n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x9a66a)
                #4  0x00007faa75adcf94 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x11df94)
                
                Stack trace of thread 184:
                #0  0x00007faa75a61ce2 n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0xa2ce2)
                #1  0x00007faa75a55e3c n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x96e3c)
                #2  0x00007faa75a564ac n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x974ac)
                #3  0x00007faa75a58b7e n/a (/usr/lib/x86_64-linux-gnu/libc.so.6 + 0x99b7e)
                #4  0x00007faa7502d313 _ZN6Parker4parkEbl (libjvm.so + 0xe2d313)
                #5  0x00007faa753412e3 Unsafe_Park (libjvm.so + 0x11412e3)
                #6  0x00007faa61386f72 n/a (n/a + 0x0)
                #7  0x00007faa60467426 n/a (n/a + 0x0)
                #8  0x00007faa60467426 n/a (n/a + 0x0)
                #9  0x00007faa5b095bec n/a (n/a + 0x0)
                #10 0x00007faa59a88c84 n/a (n/a + 0x0)
                ELF object binary architecture: AMD x86-64
9

Did some more testing and figured out a way to reliably reproduce the GPU reset. It is, in fact, triggered by high GPU load, so all that’s needed is a reliable way to put the GPU under load without the entropy introduced by games.
I’ve tried other GPU stress tests, such as Unigine Heaven and FurMark 2 and, at least for relatively short runs, they have caused no problems.
On the other hand, GravityMark is not like the others. By installing the Linux-native version and firing up the browser-based UI, I can leave everything default but make sure resolution matches the monitor (1080p) and, most importantly, set the number of asteroids to 4 MILLION. Once I click start demo, it loads after a few seconds and renders in full screen at about 5 fps (expected for such a high number of objects).
After 1-2 minutes of running GravityMark with these outrageous settings, the scene will freeze, display output cuts afterwards for a few seconds, then it returns and the scene is still frozen on screen. The program itself handles the fault gracefully, the benchmark is already stopped by that point and alt-tab to another window clears the full-screen render away.
The first time I tried the above, I got everything to crash including KWin, but subsequent runs have not produced such results. However, every run ends by crashing all hardware accelerated applications running in the user session (Firefox being a prime example).

What’s interesting is that GravityMark can also reliably reproduce a payload that only happened occasionally with in-game resets - the battery icon on the taskbar vanishes. Going to System Settings > Power Management to investigate, it does not work because “Power Management settings could not be loaded; The Power Management Service appears not to be running.”.

Screenshot_20260811_165035
Screenshot_20260811_165035959×699 139 KB

This also happened on latest stock kernel a few times, as I said it’s less frequent with games. My guess is that the “power management service” (whatever that is, I’m not sure but it must be whatever comes with Fedora/systemd by default) completely freaks out sometimes due to whatever the GPU is doing… which further points towards it being a power management issue, reported many times with AMD iGPUs as well as dGPUs on Linux.

Note that I’ve tried a bunch of different kernel parameters that claim to fix these supposed power management issues on AMD, and to no avail, the crashes persist.

As for the dmesg, it is actually quite interesting as, while until now it said GPU reset(1) succeeded! on the line before [drm] device wedged, when GravityMark resets the GPU it says GPU reset(4) succeeded!, so not quite the same thing. Also, there is more useful insight leading up to the actual reset, so I’ll attach this new dmesg below.

I’ll include multiple occurrences as it seems that every run of GravityMark was slightly different (I did multiple in one session).

[  176.442044] amdgpu 0000:04:00.0: Dumping IP State
[  176.444198] amdgpu 0000:04:00.0: Dumping IP State Completed
[  176.444209] amdgpu 0000:04:00.0: [drm] AMDGPU device coredump file has been created
[  176.444211] amdgpu 0000:04:00.0: [drm] Check your /sys/class/drm/card1/device/devcoredump/data
[  176.444213] amdgpu 0000:04:00.0: ring gfx_0.0.0 timeout, signaled seq=34891, emitted seq=34893
[  176.444216] amdgpu 0000:04:00.0:  Process MINECRAFT MAIN  pid 3785 thread vkd3d_queue pid 3839
[  176.444218] amdgpu 0000:04:00.0: Starting gfx_0.0.0 ring reset
[  178.448040] amdgpu 0000:04:00.0: MES failed to respond to msg=RESET
[  178.448044] amdgpu 0000:04:00.0: failed to reset legacy queue
[  178.448050] amdgpu 0000:04:00.0: reset via MES failed and try pipe reset -110
[  178.448052] amdgpu 0000:04:00.0: The CPFW hasn't support pipe reset yet.
[  178.448053] amdgpu 0000:04:00.0: Ring gfx_0.0.0 reset failed
[  178.448056] amdgpu 0000:04:00.0: GPU reset begin!. Source:  1
[  178.454614] amdgpu 0000:04:00.0: [drm] *ERROR* Failed to initialize parser -125!
[  180.491082] amdgpu 0000:04:00.0: MES failed to respond to msg=REMOVE_QUEUE
[  180.491092] amdgpu 0000:04:00.0: failed to unmap legacy queue
[  180.764352] [drm:gfx_v11_0_hw_fini [amdgpu]] *ERROR* failed to halt cp gfx
[  180.766336] amdgpu 0000:04:00.0: MODE2 reset
[  180.803362] amdgpu 0000:04:00.0: GPU reset succeeded, trying to resume
[  180.803976] amdgpu 0000:04:00.0: [drm] PCIE GART of 512M enabled (table at 0x00000080FFD00000).
[  180.804091] amdgpu 0000:04:00.0: SMU is resuming...
[  180.806069] amdgpu 0000:04:00.0: SMU is resumed successfully!
[  180.813375] amdgpu 0000:04:00.0: [drm] DMUB hardware initialized: version=0x08005400
[  181.068975] amdgpu 0000:04:00.0: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
[  181.068980] amdgpu 0000:04:00.0: ring comp_1.0.0 uses VM inv eng 1 on hub 0
[  181.068983] amdgpu 0000:04:00.0: ring comp_1.1.0 uses VM inv eng 4 on hub 0
[  181.068984] amdgpu 0000:04:00.0: ring comp_1.2.0 uses VM inv eng 6 on hub 0
[  181.068985] amdgpu 0000:04:00.0: ring comp_1.3.0 uses VM inv eng 7 on hub 0
[  181.068987] amdgpu 0000:04:00.0: ring comp_1.0.1 uses VM inv eng 8 on hub 0
[  181.068988] amdgpu 0000:04:00.0: ring comp_1.1.1 uses VM inv eng 9 on hub 0
[  181.068989] amdgpu 0000:04:00.0: ring comp_1.2.1 uses VM inv eng 10 on hub 0
[  181.068991] amdgpu 0000:04:00.0: ring comp_1.3.1 uses VM inv eng 11 on hub 0
[  181.068993] amdgpu 0000:04:00.0: ring sdma0 uses VM inv eng 12 on hub 0
[  181.068994] amdgpu 0000:04:00.0: ring vcn_unified_0 uses VM inv eng 0 on hub 8
[  181.068996] amdgpu 0000:04:00.0: ring jpeg_dec uses VM inv eng 1 on hub 8
[  181.068997] amdgpu 0000:04:00.0: ring mes_kiq_3.1.0 uses VM inv eng 13 on hub 0
[  181.071073] amdgpu 0000:04:00.0: GPU reset(1) succeeded!
[  181.071087] amdgpu 0000:04:00.0: [drm] device wedged, but no recovery needed
[  200.178389] clocksource: Watchdog remote CPU 13 read timed out
[  222.057433] overlayfs: upper fs does not support file handles, falling back to index=off.
[  222.057438] overlayfs: fs on '/home/asicosilomu/.local/share/containers/storage/overlay/compat2143321808/lower1' does not support file handles, falling back to xino=off.
[  222.086317] overlayfs: upper fs does not support file handles, falling back to index=off.
[  222.086322] overlayfs: fs on '/home/asicosilomu/.local/share/containers/storage/overlay/metacopy-check2906637626/l1' does not support file handles, falling back to xino=off.
[  222.096402] overlayfs: upper fs does not support file handles, falling back to index=off.
[  222.096406] overlayfs: fs on '/home/asicosilomu/.local/share/containers/storage/overlay/opaque-bug-check2437390010/l2' does not support file handles, falling back to xino=off.
[  224.172571] usb 1-4: reset full-speed USB device number 3 using xhci_hcd
[ ... ] a bunch of lines pertaining to "rtw89_8852ce", which is the NIC acting up so unrelated to the issue at hand [ ... ]
[  794.913816] usb 1-4: reset full-speed USB device number 3 using xhci_hcd
[ 1268.893851] amdgpu 0000:04:00.0: [drm] *ERROR* amdgpu_vm_validate() failed.
[ 1268.893856] amdgpu 0000:04:00.0: [drm] *ERROR* Not enough memory for command submission!
[ 1268.907424] amdgpu 0000:04:00.0: [drm] *ERROR* amdgpu_vm_validate() failed.
[ 1268.907428] amdgpu 0000:04:00.0: [drm] *ERROR* Not enough memory for command submission!
[ 1268.908614] amdgpu 0000:04:00.0: [drm] *ERROR* amdgpu_vm_validate() failed.
[ 1268.908616] amdgpu 0000:04:00.0: [drm] *ERROR* Not enough memory for command submission!
[ 1268.909789] amdgpu 0000:04:00.0: [drm] *ERROR* amdgpu_vm_validate() failed.
[ 1268.909791] amdgpu 0000:04:00.0: [drm] *ERROR* Not enough memory for command submission!
[ 1268.911025] amdgpu 0000:04:00.0: [drm] *ERROR* amdgpu_vm_validate() failed.
[ 1268.911028] amdgpu 0000:04:00.0: [drm] *ERROR* Not enough memory for command submission!
[ 1268.912183] amdgpu 0000:04:00.0: [drm] *ERROR* amdgpu_vm_validate() failed.
[ 1268.912185] amdgpu 0000:04:00.0: [drm] *ERROR* Not enough memory for command submission!
[ 1268.913387] amdgpu 0000:04:00.0: [drm] *ERROR* amdgpu_vm_validate() failed.
[ 1268.913389] amdgpu 0000:04:00.0: [drm] *ERROR* Not enough memory for command submission!
[ 1268.916321] amdgpu 0000:04:00.0: [drm] *ERROR* amdgpu_vm_validate() failed.
[ 1268.916323] amdgpu 0000:04:00.0: [drm] *ERROR* Not enough memory for command submission!
[ 1268.993731] amdgpu 0000:04:00.0: [drm] *ERROR* amdgpu_vm_validate() failed.
[ 1268.993736] amdgpu 0000:04:00.0: [drm] *ERROR* Not enough memory for command submission!
[ 1270.478361] amdgpu 0000:04:00.0: VM memory stats for proc RDD Process(6117) task browser 4 :cs0(5417) is non-zero when fini
[ 1270.580844] amdgpu 0000:04:00.0: VM memory stats for proc firefox(5337) task firefox:cs0(4766) is non-zero when fini
[ 1386.114182] amdgpu 0000:04:00.0: Dumping IP State
[ 1386.117614] amdgpu 0000:04:00.0: Dumping IP State Completed
[ 1386.117626] amdgpu 0000:04:00.0: [drm] AMDGPU device coredump file has been created
[ 1386.117628] amdgpu 0000:04:00.0: [drm] Check your /sys/class/drm/card1/device/devcoredump/data
[ 1386.117629] amdgpu 0000:04:00.0: ring gfx_0.0.0 timeout, signaled seq=217236, emitted seq=217238
[ 1386.117631] amdgpu 0000:04:00.0:  Process GravityMark.x64 pid 10749 thread GravityMark.x64 pid 10749
[ 1386.117633] amdgpu 0000:04:00.0: Starting gfx_0.0.0 ring reset
[ 1388.121436] amdgpu 0000:04:00.0: MES failed to respond to msg=RESET
[ 1388.121442] amdgpu 0000:04:00.0: failed to reset legacy queue
[ 1388.121444] amdgpu 0000:04:00.0: reset via MES failed and try pipe reset -110
[ 1388.121446] amdgpu 0000:04:00.0: Ring gfx_0.0.0 reset failed
[ 1388.121449] amdgpu 0000:04:00.0: GPU reset begin!. Source:  1
[ 1388.136288] amdgpu 0000:04:00.0: [drm] *ERROR* Failed to initialize parser -125!
[ 1390.180744] amdgpu 0000:04:00.0: MES failed to respond to msg=REMOVE_QUEUE
[ 1390.180756] amdgpu 0000:04:00.0: failed to unmap legacy queue
[ 1390.454781] [drm:gfx_v11_0_hw_fini [amdgpu]] *ERROR* failed to halt cp gfx
[ 1390.456734] amdgpu 0000:04:00.0: MODE2 reset
[ 1390.493753] amdgpu 0000:04:00.0: GPU reset succeeded, trying to resume
[ 1390.494498] amdgpu 0000:04:00.0: [drm] PCIE GART of 512M enabled (table at 0x00000080FFD00000).
[ 1390.494555] amdgpu 0000:04:00.0: SMU is resuming...
[ 1390.496766] amdgpu 0000:04:00.0: SMU is resumed successfully!
[ 1390.503710] amdgpu 0000:04:00.0: [drm] DMUB hardware initialized: version=0x08005400
[ 1390.764412] amdgpu 0000:04:00.0: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
[ 1390.764417] amdgpu 0000:04:00.0: ring comp_1.0.0 uses VM inv eng 1 on hub 0
[ 1390.764419] amdgpu 0000:04:00.0: ring comp_1.1.0 uses VM inv eng 4 on hub 0
[ 1390.764420] amdgpu 0000:04:00.0: ring comp_1.2.0 uses VM inv eng 6 on hub 0
[ 1390.764421] amdgpu 0000:04:00.0: ring comp_1.3.0 uses VM inv eng 7 on hub 0
[ 1390.764423] amdgpu 0000:04:00.0: ring comp_1.0.1 uses VM inv eng 8 on hub 0
[ 1390.764424] amdgpu 0000:04:00.0: ring comp_1.1.1 uses VM inv eng 9 on hub 0
[ 1390.764425] amdgpu 0000:04:00.0: ring comp_1.2.1 uses VM inv eng 10 on hub 0
[ 1390.764427] amdgpu 0000:04:00.0: ring comp_1.3.1 uses VM inv eng 11 on hub 0
[ 1390.764428] amdgpu 0000:04:00.0: ring sdma0 uses VM inv eng 12 on hub 0
[ 1390.764429] amdgpu 0000:04:00.0: ring vcn_unified_0 uses VM inv eng 0 on hub 8
[ 1390.764431] amdgpu 0000:04:00.0: ring jpeg_dec uses VM inv eng 1 on hub 8
[ 1390.764432] amdgpu 0000:04:00.0: ring mes_kiq_3.1.0 uses VM inv eng 13 on hub 0
[ 1390.765798] amdgpu 0000:04:00.0: GPU reset(2) succeeded!
[ 1390.765807] amdgpu 0000:04:00.0: [drm] device wedged, but no recovery needed
[ 1392.086687] amdgpu 0000:04:00.0: VM memory stats for proc kwin_wayland(1966) task kwin_wayla:cs0(1904) is non-zero when fini
[ 1392.090214] amdgpu 0000:04:00.0: VM memory stats for proc Xwayland(2032) task Xwayland:cs0(2022) is non-zero when fini
[ 1392.108064] amdgpu 0000:04:00.0: VM memory stats for proc Browser.x64(8394) task Browser.x64(8394) is non-zero when fini
[ 1392.134157] amdgpu 0000:04:00.0: VM memory stats for proc xwaylandvideobr(2811) task xwaylandvi:cs0(2547) is non-zero when fini
[ 1392.134475] amdgpu 0000:04:00.0: VM memory stats for proc plasmashell(2311) task plasmashel:cs0(2110) is non-zero when fini
[ 1394.367700] usb 1-4: reset full-speed USB device number 3 using xhci_hcd
[ 1871.682483] amdgpu 0000:04:00.0: Dumping IP State
[ 1871.685290] amdgpu 0000:04:00.0: Dumping IP State Completed
[ 1871.685319] amdgpu 0000:04:00.0: [drm] AMDGPU device coredump file has been created
[ 1871.685322] amdgpu 0000:04:00.0: [drm] Check your /sys/class/drm/card1/device/devcoredump/data
[ 1871.685325] amdgpu 0000:04:00.0: ring gfx_0.0.0 timeout, signaled seq=273842, emitted seq=273844
[ 1871.685328] amdgpu 0000:04:00.0:  Process GravityMark.x64 pid 12603 thread GravityMark.x64 pid 12603
[ 1871.685330] amdgpu 0000:04:00.0: Starting gfx_0.0.0 ring reset
[ 1873.689188] amdgpu 0000:04:00.0: MES failed to respond to msg=RESET
[ 1873.689192] amdgpu 0000:04:00.0: failed to reset legacy queue
[ 1873.689194] amdgpu 0000:04:00.0: reset via MES failed and try pipe reset -110
[ 1873.689195] amdgpu 0000:04:00.0: Ring gfx_0.0.0 reset failed
[ 1873.689198] amdgpu 0000:04:00.0: GPU reset begin!. Source:  1
[ 1873.704364] amdgpu 0000:04:00.0: [drm] *ERROR* Failed to initialize parser -125!
[ 1875.747973] amdgpu 0000:04:00.0: MES failed to respond to msg=REMOVE_QUEUE
[ 1875.747983] amdgpu 0000:04:00.0: failed to unmap legacy queue
[ 1876.022304] [drm:gfx_v11_0_hw_fini [amdgpu]] *ERROR* failed to halt cp gfx
[ 1876.024284] amdgpu 0000:04:00.0: MODE2 reset
[ 1876.061139] amdgpu 0000:04:00.0: GPU reset succeeded, trying to resume
[ 1876.061737] amdgpu 0000:04:00.0: [drm] PCIE GART of 512M enabled (table at 0x00000080FFD00000).
[ 1876.061821] amdgpu 0000:04:00.0: SMU is resuming...
[ 1876.063891] amdgpu 0000:04:00.0: SMU is resumed successfully!
[ 1876.071138] amdgpu 0000:04:00.0: [drm] DMUB hardware initialized: version=0x08005400
[ 1876.333941] amdgpu 0000:04:00.0: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
[ 1876.333951] amdgpu 0000:04:00.0: ring comp_1.0.0 uses VM inv eng 1 on hub 0
[ 1876.333954] amdgpu 0000:04:00.0: ring comp_1.1.0 uses VM inv eng 4 on hub 0
[ 1876.333955] amdgpu 0000:04:00.0: ring comp_1.2.0 uses VM inv eng 6 on hub 0
[ 1876.333956] amdgpu 0000:04:00.0: ring comp_1.3.0 uses VM inv eng 7 on hub 0
[ 1876.333958] amdgpu 0000:04:00.0: ring comp_1.0.1 uses VM inv eng 8 on hub 0
[ 1876.333960] amdgpu 0000:04:00.0: ring comp_1.1.1 uses VM inv eng 9 on hub 0
[ 1876.333962] amdgpu 0000:04:00.0: ring comp_1.2.1 uses VM inv eng 10 on hub 0
[ 1876.333964] amdgpu 0000:04:00.0: ring comp_1.3.1 uses VM inv eng 11 on hub 0
[ 1876.333965] amdgpu 0000:04:00.0: ring sdma0 uses VM inv eng 12 on hub 0
[ 1876.333967] amdgpu 0000:04:00.0: ring vcn_unified_0 uses VM inv eng 0 on hub 8
[ 1876.333968] amdgpu 0000:04:00.0: ring jpeg_dec uses VM inv eng 1 on hub 8
[ 1876.333970] amdgpu 0000:04:00.0: ring mes_kiq_3.1.0 uses VM inv eng 13 on hub 0
[ 1876.336253] amdgpu 0000:04:00.0: GPU reset(3) succeeded!
[ 1876.336267] amdgpu 0000:04:00.0: [drm] device wedged, but no recovery needed
[ 2087.747629] amdgpu 0000:04:00.0: Dumping IP State
[ 2087.751608] amdgpu 0000:04:00.0: Dumping IP State Completed
[ 2087.751622] amdgpu 0000:04:00.0: [drm] AMDGPU device coredump file has been created
[ 2087.751624] amdgpu 0000:04:00.0: [drm] Check your /sys/class/drm/card1/device/devcoredump/data
[ 2087.751626] amdgpu 0000:04:00.0: ring gfx_0.0.0 timeout, signaled seq=289999, emitted seq=290001
[ 2087.751628] amdgpu 0000:04:00.0:  Process GravityMark.x64 pid 13669 thread GravityMark.x64 pid 13669
[ 2087.751630] amdgpu 0000:04:00.0: Starting gfx_0.0.0 ring reset
[ 2089.755464] amdgpu 0000:04:00.0: MES failed to respond to msg=RESET
[ 2089.755470] amdgpu 0000:04:00.0: failed to reset legacy queue
[ 2089.755472] amdgpu 0000:04:00.0: reset via MES failed and try pipe reset -110
[ 2089.755475] amdgpu 0000:04:00.0: Ring gfx_0.0.0 reset failed
[ 2089.755477] amdgpu 0000:04:00.0: GPU reset begin!. Source:  1
[ 2089.770234] amdgpu 0000:04:00.0: [drm] *ERROR* Failed to initialize parser -125!
[ 2091.813974] amdgpu 0000:04:00.0: MES failed to respond to msg=REMOVE_QUEUE
[ 2091.813984] amdgpu 0000:04:00.0: failed to unmap legacy queue
[ 2092.088819] [drm:gfx_v11_0_hw_fini [amdgpu]] *ERROR* failed to halt cp gfx
[ 2092.090803] amdgpu 0000:04:00.0: MODE2 reset
[ 2092.128699] amdgpu 0000:04:00.0: GPU reset succeeded, trying to resume
[ 2092.129328] amdgpu 0000:04:00.0: [drm] PCIE GART of 512M enabled (table at 0x00000080FFD00000).
[ 2092.129413] amdgpu 0000:04:00.0: SMU is resuming...
[ 2092.132180] amdgpu 0000:04:00.0: SMU is resumed successfully!
[ 2092.138554] amdgpu 0000:04:00.0: [drm] DMUB hardware initialized: version=0x08005400
[ 2092.398704] amdgpu 0000:04:00.0: ring gfx_0.0.0 uses VM inv eng 0 on hub 0
[ 2092.398715] amdgpu 0000:04:00.0: ring comp_1.0.0 uses VM inv eng 1 on hub 0
[ 2092.398717] amdgpu 0000:04:00.0: ring comp_1.1.0 uses VM inv eng 4 on hub 0
[ 2092.398718] amdgpu 0000:04:00.0: ring comp_1.2.0 uses VM inv eng 6 on hub 0
[ 2092.398720] amdgpu 0000:04:00.0: ring comp_1.3.0 uses VM inv eng 7 on hub 0
[ 2092.398722] amdgpu 0000:04:00.0: ring comp_1.0.1 uses VM inv eng 8 on hub 0
[ 2092.398724] amdgpu 0000:04:00.0: ring comp_1.1.1 uses VM inv eng 9 on hub 0
[ 2092.398725] amdgpu 0000:04:00.0: ring comp_1.2.1 uses VM inv eng 10 on hub 0
[ 2092.398727] amdgpu 0000:04:00.0: ring comp_1.3.1 uses VM inv eng 11 on hub 0
[ 2092.398728] amdgpu 0000:04:00.0: ring sdma0 uses VM inv eng 12 on hub 0
[ 2092.398730] amdgpu 0000:04:00.0: ring vcn_unified_0 uses VM inv eng 0 on hub 8
[ 2092.398731] amdgpu 0000:04:00.0: ring jpeg_dec uses VM inv eng 1 on hub 8
[ 2092.398733] amdgpu 0000:04:00.0: ring mes_kiq_3.1.0 uses VM inv eng 13 on hub 0
[ 2092.400950] amdgpu 0000:04:00.0: GPU reset(4) succeeded!
[ 2092.400966] amdgpu 0000:04:00.0: [drm] device wedged, but no recovery needed
[ 2107.782868] usb 1-4: reset full-speed USB device number 3 using xhci_hcd
[ 2402.279277] ideapad_acpi VPC2004:00: conservation_mode attribute has been deprecated, see charge_types.
[ 3466.112675] usb 1-4: reset full-speed USB device number 3 using xhci_hcd
10

Seems like “power management service” is referring to powerdevil, which has crashed in my case:

○ plasma-powerdevil.service - Powerdevil
     Loaded: loaded (/usr/lib/systemd/user/plasma-powerdevil.service; static)
    Drop-In: /usr/lib/systemd/user/service.d
             └─10-timeout-abort.conf
     Active: inactive (dead) since Tue 2026-08-11 16:32:27 EEST; 54min ago
   Duration: 22min 49.342s
 Invocation: 75839e04aa7d403d9f462ab486c39238
    Process: 2238 ExecStart=/usr/libexec/org_kde_powerdevil (code=killed, signal=TERM)
   Main PID: 2238 (code=killed, signal=TERM)
   Mem peak: 13.5M
        CPU: 925ms

aug 11 16:09:38 fedora org_kde_powerdevil[2238]: Failed to register with host portal QDBusError("org.freedesktop.portal.Error.Failed", "Could not register app ID: App info not found for 'org.kde.org_kde_powerdevil'")
aug 11 16:09:38 fedora org_kde_powerdevil[2238]: Failed to register with host portal QDBusError("org.freedesktop.portal.Error.Failed", "Could not register app ID: App info not found for 'org.kde.org_kde_powerdevil'")
aug 11 16:32:27 fedora org_kde_powerdevil[2238]: The Wayland connection broke. Did the Wayland compositor die?
aug 11 16:32:27 fedora org_kde_powerdevil[2238]: Attempting wayland reconnect
aug 11 16:32:27 fedora org_kde_powerdevil[2238]: Failed to write to the pipe: Bad file descriptor.
aug 11 16:32:27 fedora org_kde_powerdevil[2238]: There are no outputs - creating placeholder screen
aug 11 16:32:27 fedora org_kde_powerdevil[2238]: X connection to :0 broken (explicit kill or server shutdown).
aug 11 16:32:27 fedora org_kde_powerdevil[2238]: QSocketNotifier: Socket notifiers cannot be enabled or disabled from another thread
aug 11 16:32:27 fedora org_kde_powerdevil[2238]: QtDBus: cannot relay signals from parent QObject(0x56156158c4a0 "") unless they are emitted in the object's thread QThread(0x56156151c8e0 "Qt mainThread"). Current thread is QThread(0x5615615973d0 "").
aug 11 16:32:27 fedora org_kde_powerdevil[2238]: [  2419] Watch thread terminated.

Restarting it with systemctl --user restart plasma-powerdevil.service succeeds and causes the battery icon and power settings and everything to reappear. Looking at the dates, now it seems that the crash here is a side effect of the GPU reset, on the rare occasion that it successfully knocks out Wayland and stuns plasmashell and kwin.

Everything is just so random…

11

I’ve ruled out Wayland being at fault by running GravityMark under openbox, which failed in the same way, no image for a few seconds then kicked me back to the TTY I had started the desktop from (startx /usr/bin/openbox).

12

I’VE SOLVED IT! Fedora and its derivatives are normally super stable for gaming so an issue like this would be odd. Even more odd that no amount of downgrades made it go away. In fact, Fedora Linux is not to blame at all, nothing is wrong at the software level!

However… that means there’s only two other places where things could go wrong:

  • the firmware
  • the hardware (or, without the sugar-coating, your GPU is dying)

I’m not messing around with GPU microcode… so only firmware I could think of was the BIOS (UEFI really, but whatever).

Thing is, fwupd never helped me with any of that. My BIOS seemed a little buggy since day one, since I tried to install Fedora, but I didn’t mind. Especially odd as the BIOS was the same but the issues worsened.

However, I’m 99% that the BIOS was, in fact, to blame all along. While looking for something else for this issue, I stumbled upon some Lenovo page where I could put my serial number. So, I did, and looked around.

Found the section for drivers, under which there was BIOS/UEFI. Clicked and saw, the latest firmware for my system was dated Feb. 2026. I got this laptop in Sept. 2024 and it’s been running factory firmware since then, no updates.

Let’s look at the before again, everything was cooked:

  • GravityMark - guaranteed GPU reset at the third cutscene.
  • Minecraft Bedrock (with Vibrant Visuals enabled) - crash in less than 5 min.
  • Counter Strike 2 - VK_ERROR_DEVICE_LOST right after the warmup ended.
  • Minecraft Java with insane shaders and settings - flying around in creative for a few minutes = GPU reset.

Today I successfully updated the BIOS and, once I booted back into Fedora, I fired up Gravity Mark. It ran for a non-stop 15-20 minutes, after which I stopped it manually by clicking Quit. This was due to time constraints, it could’ve easily run for much longer.

Then another run of GravityMark of about the same duration. No issue.

Still no GPU reset and haven’t rebooted either, all within one session.

Then I hopped on gmod with a friend and was able to play for, like, half an hour after which I exited the game normally. Again, that’s just how long we played for, no GPU reset yet.

For gmod I first used the Linux-native version, but then switched to Proton due to bugs specific to the Linux build, unrelated to the GPU issues. Just bugs everyone is having on the Linux version.

Still no reboot. No GPU reset.

Now, my setup consists of Minecraft Java on one half of the screen, running on max shaders as before. On the other half, Firefox, where I’m typing this. In the background, Prism Launcher main menu and an open Steam window downloading a game (the latter being an extra Chromium burden).

The system sometimes gets really sluggish with the Minecraft running, and this is simply due to bad scheduling. Switching to a better kernel, such as the CachyOS kernel I was using before, would certainly improve the situation. Obviously not 100%, you can’t expect the rest of the system to be buttery smooth when Minecraft is on screen and keeping the GPU at 99% non stop, but definitely another kernel could improve it.

I’ve removed that kernel for now though, it wouldn’t boot due to updated Secure Boot configuration, but a kbx update came in Discover and I installed that after. Maybe CachyOS kernel can be reinstalled now that I updated the signatures for secure boot.

Conclusion: as far as the current situation goes, this issue is unrelated to Fedora Linux and was, rather, a conflict between Mesa, probably, and the outdated, buggy BIOS.

I’m so glad it’s fixed and it wasn’t hardware after all!

Important takeaway: Keep your firmware up to date, kids!

Now, here’s a bonus section for further visitors coming from search results: how I updated the Lenovo BIOS without a Windows installation on my laptop.

PROCEED WITH EXTREME CAUTION! These instructions are provided as-is and solely based on my own experience. If anything goes wrong during this process, YOU’LL BE LEFT WITH AN EXPENSIVE PAPERWEIGHT! Even on identical configurations to mine, I don’t guarantee success! You might BRICK YOUR MOTHERBOARD and I’m not gonna be held liable with it.

I KNOW THESE ARE USUAL BIOS UPDATE WARNINGS BUT, AS YOU’LL BE UPDATING THROUGH AN UNSUPPORTED SOFTWARE STACK, FAILURE RATES ARE WAY HIGHER THAN OFFICIAL UPDATE METHODS!

To clarify, I am NOT RESPONSIBLE for lost data, fried GPUs, bricked motherboards, explosions or thermonuclear war! (or any other things of this nature)

To begin, I’ll assume that you’re booted into your Linux system and, obviously, you don’t have any Windows installation (if you do, then just go update the BIOS normally).

You’ll need the following:

  • stable power source;
  • internet access;
  • Windows 11 x86_64 ISO, language doesn’t matter, but must be 25H2 or later;
  • two USB drives, only one of which must be blank and big enough to have Windows installation media written to it;
  • knowledge of the keys for entering UEFI setup and boot order, in short, you must know how to successfully boot from USB.

The steps might suggest otherwise, but you aren’t gonna be installing Windows. Nothing will be written to your drive as a result of this procedure.

Step 1: obtaining the BIOS update

(this step is helpful even if you’re on Windows, but aren’t getting the update through Windows Update and don’t know where to download it. Once you’ve downloaded, follow the instructions on the Lenovo page or in the downloadable README)
If you’re following along on Linux only, then you must go through with my steps.

Head over to https://smartsupport.lenovo.com/us/en/documents/tvsu-update. Where it says to enter your serial number, do so, then click Submit.

Once you’ve been taken to your model-specific page, click Drivers & Software in the left sidebar, then click Select Drivers under Manual Update.

Click BIOS/UEFI. Then pick the latest version (if there are multiple listed, for me there was only one). Then click Download next to BIOS Update. This will download an .exe file, for me it was ngcn34ww.exe but it will differ based on BIOS version.

VERY IMPORTANT! Click View Checksum under BIOS Update and check them. You can never be too sure!

Update acquired!

Step 2: create Windows 11 installation media

Using the ISO that I assumed you already have (if not, get it from Microsoft, it’s pretty easy), write Windows installation media to your blank USB stick. I have Ventoy on my drive so I just copied over the ISO, but you can create the media “traditionally” as well - balenaEtcher is first to come to mind.

Once your Windows installation media is written, safely (!) remove the drive from your computer. On to step 3.

Step 3: prepare the update

Put the .exe file you downloaded from Lenovo onto any USB drive you’ve got. Doesn’t matter as it won’t delete the files, but I wouldn’t use a very important drive either.

Step 4: boot into Windows PE

Boot from the USB you created. Do not plug in the second file USB yet.
Do note that some Lenovos will show a faux BIOS updating screen when turned on with a drive inserted, it’s just looking for a BIOS update on the USB and bailing out to 100% when it can’t find one. No update is taking place.
While you could try updating over USB, Lenovo doesn’t make it clear how to do so… better follow along with me.

Once you’re booted, you should be asked for your language and region by Windows Setup. This screen should have a white background and no big Windows logo banner. If it has a coloured background and a big “Windows” banner, you’re on Windows <10 or Windows 11 pre-25H2. Use the correct version and repeat.

Step 5: we’re not installing Windows

That’s right! We aren’t here to install Windows! Windows installation media is just the easiest way to get yourself into the Windows PE (Preinstalled Environment), a stripped-down version of Windows (in our case, 11 25H2) that’s meant for servicing and maintenance and stuff like that. It’s also the environment running the Windows Setup you’re looking at and, normally, would use to install Windows.

We aren’t doing that today! Press Shift + F10 (or Shift + Fn + F10) and a command prompt will magically appear. If you’ve never done anything remotely close to this before, I might sound crazy, but please follow along.

Type “notepad” into the command prompt and get into, well, Notepad. Go to File > Open. In the dialog box, set the file type to All files. Now, from the left, click This PC. You should see a bunch of drives.

Now, plug in the second USB, where you put the update program. Then, right click in an empty space in the Drives pane and click Refresh. A new drive will appear, that is your file drive. Double-click it, find the BIOS update file, then right-click and Copy.

Now go back to This PC. Enter the Boot (X:) drive, this is the ramdisk that Windows PE is running off of. We’re putting the update file here for more stability. Right click in an empty space and click Paste. Then Refresh or it mightn’t show up.

Once the update exe shows up in drive (X:), close the Open dialog and close Notepad. You’re now back at the command prompt, and the current directory should be X:/sources. Type cd .. to return to the drive’s root, then dir to confirm that the BIOS update file is there. Now, type the name of the exe without the extension, then press Enter to run it (no preceding ./ needed because this is Windows).

You should see an Extracting dialog, then Lenovo application opens and explains what a BIOS update entails. Click Next until it tells you to save your work and close programs, or something along those lines.

To reiterate, I am NOT RESPONSIBLE for lost data, fried GPUs, bricked motherboards, explosions or thermonuclear war! (or any other things of this nature)

Clicking Next once more will immediately start the BIOS update. It’ll say that your PC will restart in a few moments to begin the update and it’ll warn you to not turn it off or disconnect power even at this early point.

By this point, you’ve already gone all-in. There’s no way back.

Your computer will restart, fans will start at full whack (this is normal!) and a BIOS update screen will appear with the dreaded warning about turning it off. A progress bar will slowly move towards 100%.

Then it may restart 1-2 times, after which you’ll be left at a blank screen with only the power button lit and maybe Esc/FnLock light, with fans starting to spin again. Don’t be worried, it’s normal for it to stay at a blank screen for upwards of a few minutes, don’t panic. And, of course, don’t disconnect power! Or forcefully power it off!

Once it’s done, you’ll be kicked either into UEFI setup or you’ll boot into Linux and arrive at your login screen. Or be met by a GRUB error if you were using a custom kernel with Secure Boot support, you’ll need to update the kdbx via fwupd I think to fix that. Just use stock kernel.

Obviously, if you’re in UEFI setup just press Esc or go to “exit discarding changes”.

Once you’re back in Linux, fire up a terminal and run sudo dmidecode to check the BIOS revision date. Go to the very beginning of the output, that’s where you’ll see it.

It’ll say “Release Date”, it’s usually about a month earlier than publication date on Lenovo website. But definitely a newer date than before the update.

Example output from after my update:

Getting SMBIOS data from sysfs.
SMBIOS 3.6.0 present.
Table at 0xB8E6E000.

Handle 0x0000, DMI type 0, 26 bytes
BIOS Information
        Vendor: LENOVO
        Version: NGCN34WW
        Release Date: 01/16/2026
        Address: 0xE0000
        Runtime Size: 128 kB
        ROM Size: 32 MB
        Characteristics:
                PCI is supported
                BIOS is upgradeable
                BIOS shadowing is allowed
                Boot from CD is supported
                Selectable boot is supported
                EDD is supported
                Japanese floppy for NEC 9800 1.2 MB is supported (int 13h)
                Japanese floppy for Toshiba 1.2 MB is supported (int 13h)
                5.25"/360 kB floppy services are supported (int 13h)
                5.25"/1.2 MB floppy services are supported (int 13h)
                3.5"/720 kB floppy services are supported (int 13h)
                3.5"/2.88 MB floppy services are supported (int 13h)
                8042 keyboard services are supported (int 9h)
                CGA/mono video services are supported (int 10h)
                ACPI is supported
                USB legacy is supported
                BIOS boot specification is supported
                Targeted content distribution is supported
                UEFI is supported
        BIOS Revision: 1.34
        Firmware Revision: 1.34

So there you have it, issue fixed. And a thing or two about updating the BIOS.

I’ll keep this thread here for future reference but I’ll go ahead and mark it as solved.

13

I’ve got a Lenovo laptop as well, but I can update the firmware without needing Windows PE at all. After fetching the exe from Lenovo, unzip the exe (it’s really just a zip archive!), then look for a file with a name matching LenovoBiosUpdateTool/Rfs/Fw/*/BIOS.cap. Run sudo fwupdmgr get-devices and look for UEFI Device Firmware. Note the GUID. Run sudo fwupdtool install-blob [CAP_file] [GUID] to update the firmware. Use with extreme caution, as noted above. If anything doesn’t seem quite right, don’t run that last command!

14

Yeah… I was thinking there must be some way to extract the BIOS and flash it from within Linux using fwupd. However, the prospect of attempting something like that is too risky for my tastes: it steers even farther away from the official update method and you’re writing firmware to the motherboard directly from the running Linux system. Any conflict of installed software or random crash, and the flash could get interrupted midway and leave behind a bricked system.

Though I was not sure if this is possible, it’s great that you pointed it out for those that might arrive here and prefer this method. However, using the official update tool under Windows PE is safer (though prob not as safe as running under a supported Windows 11 environment) as it’s a two-step process. First the update tool does an undocumented something, then restarts the system and on next boot, the BIOS sees the prepared file and updates itself offline. This prevents any conflicts that might come from the user’s operating system, Lenovo wouldn’t trust random users’ Windows installs to remain stable when writing firmware live. And I wouldn’t trust my own Linux system with that, I do all my regular system updates offline, so I’m cautious in these situations

Unless I’m wrong and the command you suggested prepares an offline BIOS update in the same way as the official tool… it looks to me like a direct flash command but I’m not familiar with fwupd whatsoever.

As a footnote, I’ve also noticed that other random bugs I was having on my computer, and was attributing to Fedora itself or to Wine, disappeared after updating the BIOS. Not to talk about the GPU, which was perfectly stable under every benchmark and game I tested.

Really caught me off-guard, these issues that seem like something is going wrong in the software stack (kernel, drivers, userspace), but actually aren’t. After all, the kernel accesses hardware via BIOS/UEFI, so if the factory BIOS is a buggy mess, so will be the user-facing system.

Second footnote: in the case of random driver misbehaviours or other weird bugs like these, a BIOS update is to be attempted before anything from the kernel and above is suspected. Really, updating your BIOS every now and then is just good practice, but many people (including me) don’t do so…

All the rambling above is just leaving everything I can for future visitors… I, myself, don’t like stumbling upon vague forum threads so I wrote as verbosely as possible.

Thanks for your input and providing the exact fwupd commands, saved me a bunch of web searches. I might try updating from Linux next time an updated BIOS is released, but I’m still kinda wary of doing so… given how high the stakes are. The BIOS update was already pretty nerve-racking doing it through my method, especially that blank screen at the end which Lenovo likes to stay on for some reason… live update is riskier but doesn’t take all of those useless round-trips.