Please describe your issue in as much detail as possible:
Launching a game after suspend/resume cycle freezes the desktop and crashes the steam client. Game runs very slow with corrupted graphics. Reproducible in about 20% of suspend/resume cycles.
Steps for reproducing this issue:
Launch Steam
Suspend computer
Wake computer
Launch a windows game (Crymachina in this case)
Expected result
Game launches normally.
Actual result
Desktop freezes
Steam client crashes
Game starts but runs very slow (0.1fps) with corrupted graphics
Native applications like glxgears / vkgears / supertuxkart run normally
or
Desktop freezes
Steam client suvives but shows errors in kernel log
Game starts but runs very slow (0.1fps) with corrupted graphics
Native applications like glxgears / vkgears / supertuxkart run normally
Workarounds
If the desktop freezes and the steam client crashes only a reboot helps. If only the game has corruption on startup restarting steam helps.
Messages observed
kernel: amdgpu 0000:03:00.0: amdgpu: ring gfx_0.0.0 timeout, but soft recovered
EDIT: After a steam update and one week without issues it happened again today, no steam crash but corrupted graphics in the game and low fps. The tools glxgears, vkgears and supertuxkart work flawlessly while Steam shows issues. This time logs were flooded with these messages:
[39121.545792] x86/split lock detection: #DB: CHTTPClientThre/3315 took a bus_lock trap at address: 0xf253c5a4
[39121.545809] x86/split lock detection: #DB: CHTTPClientThre/3315 took a bus_lock trap at address: 0xf253c5a4
[39122.195716] handle_bus_lock: 102 callbacks suppressed
[39122.195720] x86/split lock detection: #DB: CHTTPClientThre/3315 took a bus_lock trap at address: 0xf253c5a4
[39122.195883] x86/split lock detection: #DB: CHTTPClientThre/3315 took a bus_lock trap at address: 0xf253c5a4
[39122.195894] x86/split lock detection: #DB: CHTTPClientThre/3315 took a bus_lock trap at address: 0xf253c5a4
[39122.195905] x86/split lock detection: #DB: CHTTPClientThre/3315 took a bus_lock trap at address: 0xf253c5a4
[39122.343622] x86/split lock detection: #DB: CHTTPClientThre/3315 took a bus_lock trap at address: 0xf253c5a4
[39122.343693] x86/split lock detection: #DB: CHTTPClientThre/3315 took a bus_lock trap at address: 0xf253c5a4
[39122.593004] x86/split lock detection: #DB: CHTTPClientThre/3315 took a bus_lock trap at address: 0xf253c5a4
[39122.593080] x86/split lock detection: #DB: CHTTPClientThre/3315 took a bus_lock trap at address: 0xf253c5a4
[39122.593091] x86/split lock detection: #DB: CHTTPClientThre/3315 took a bus_lock trap at address: 0xf253c5a4
[39122.638130] x86/split lock detection: #DB: CHTTPClientThre/3315 took a bus_lock trap at address: 0xf253c5a4
[39127.412892] handle_bus_lock: 32 callbacks suppressed
[39127.412896] x86/split lock detection: #DB: CHTTPClientThre/3315 took a bus_lock trap at address: 0xf253c5a4
[39127.412978] x86/split lock detection: #DB: CHTTPClientThre/3315 took a bus_lock trap at address: 0xf253c5a4
[39131.929566] traps: steamwebhelper[32967] trap invalid opcode ip:79b6bc2cac6f sp:7ffdaea40f30 error:0 in libcef.so[66c9c6f,79b6b7fed000+a36d000]
[39132.572902] x86/split lock detection: #DB: CHTTPClientThre/3315 took a bus_lock trap at address: 0xf253c5a4
[39132.573010] x86/split lock detection: #DB: CHTTPClientThre/3315 took a bus_lock trap at address: 0xf253c5a4
[39137.257318] amdgpu 0000:03:00.0: amdgpu: Dumping IP State
[39137.258124] amdgpu 0000:03:00.0: amdgpu: Dumping IP State Completed
[39137.258646] amdgpu 0000:03:00.0: amdgpu: ring gfx_0.0.0 timeout, but soft recovered
EDIT2: Happened again today. The kernel messages were the same as last time. After killing the frozen game executable Steam is now showing a perpetual "Shutting down Steam..." window but doesn't close. Steam can be shutdown manually from the gnome tray icon however. After a Steam restart everything seems to work again but once inside the game the display corruption and flicker show up again. This issue persists through many game restarts until I run vkgears once. After that the game starts normally again.
EDIT3: Issue no longer happens with kernel 6.14.5.arch1-1 and Steam client 1745876290.
Your system information
Please describe your issue in as much detail as possible:
Launching a game after suspend/resume cycle freezes the desktop and crashes the steam client. Game runs very slow with corrupted graphics. Reproducible in about 20% of suspend/resume cycles.
Steps for reproducing this issue:
Expected result
Actual result
or
Workarounds
If the desktop freezes and the steam client crashes only a reboot helps. If only the game has corruption on startup restarting steam helps.
Messages observed
kernel: amdgpu 0000:03:00.0: amdgpu: ring gfx_0.0.0 timeout, but soft recoveredEDIT: After a steam update and one week without issues it happened again today, no steam crash but corrupted graphics in the game and low fps. The tools glxgears, vkgears and supertuxkart work flawlessly while Steam shows issues. This time logs were flooded with these messages:
EDIT2: Happened again today. The kernel messages were the same as last time. After killing the frozen game executable Steam is now showing a perpetual "Shutting down Steam..." window but doesn't close. Steam can be shutdown manually from the gnome tray icon however. After a Steam restart everything seems to work again but once inside the game the display corruption and flicker show up again. This issue persists through many game restarts until I run vkgears once. After that the game starts normally again.
EDIT3: Issue no longer happens with kernel 6.14.5.arch1-1 and Steam client 1745876290.