XPost: linux.debian.bugs.dist 
 From: pablo@seestieto.com 
  
 Package: src:linux 
 Version: 6.12.41-1 
 Severity: important 
  
 Dear Maintainer, 
  
 When utilizing darktable photo editing software (5.0.1-2) with GPU 
 acceleration on AMD Ryzen 5 3400G integrated graphics, the amdgpu kernel 
 driver crashes after a period of time. This results in the failure of 
 not only darktable itself but the entire GNOME/Wayland session. A core 
 dump from Xwayland is generated in $HOME. 
  
 I've installed at least the following trixie packages to enable opencl: 
 librocm-smi-dev librocm-smi64-1 rocm-cmake rocm-device-libs libc 
 ang-common-17-dev 
  
 I can consistently reproduce this issue by utilizing darktable and 
 performing certain operations that induce GPU load, such as generating 
 thumbnails while scrolling through a series of newly imported images. 
  
 Same thing happens with other ROCm usage scenarios as well, like 
 attempting to run local LLM models with ollama (integrated GPUs are 
 not really supported, but it could work without this crash). 
  
 -- Package-specific info: 
  
 ** Version: 
 Linux version 6.12.41+deb13-amd64 (debian-kernel@lists.debian.org) 
 (x86_64-linux-gnu-gcc-14 (Debian 14.2.0-19) 14.2.0, GNU ld (GNU Binutils for 
 Debian) 2.44) #1 SMP PREEMPT_DYNAMIC Debian 6.12.41-1 (2025-08-12) 
  
 ** Command line: 
 BOOT_IMAGE=/vmlinuz-6.12.41+deb13-amd64 root=/dev/mapper/XXXXXXXX--vg-root 
 ro 
 quiet 
  
 ** Not tainted 
  
 ** Kernel log: 
 Aug 28 15:36:45 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: Dumping IP 
 State 
 Aug 28 15:36:45 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: Dumping IP 
 State 
 Completed 
 Aug 28 15:36:45 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: ring sdma0 
 timeout, signaled seq=1312044, emitted seq=1312046 
 Aug 28 15:36:45 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: GPU reset 
 begin! 
 Aug 28 15:36:45 XXXXXXX kernel: amdgpu: Failed to suspend process 0x8010 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: MODE2 reset 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: GPU reset 
 succeeded, trying to resume 
 Aug 28 15:36:46 XXXXXXX kernel: [drm] PCIE GART of 1024M enabled. 
 Aug 28 15:36:46 XXXXXXX kernel: [drm] PTB located at 0x000000F47FC00000 
 Aug 28 15:36:46 XXXXXXX kernel: [drm] VRAM is lost due to GPU reset! 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: PSP is 
 resuming... 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: reserve 
 0x400000 
 from 0xf47f800000 for PSP TMR 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: RAS: optional 
 ras 
 ta ucode is not available 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: RAP: optional 
 rap 
 ta ucode is not available 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: SECUREDISPLAY: 
 securedisplay ta ucode is not available 
 Aug 28 15:36:46 XXXXXXX kernel: [drm] kiq ring mec 2 pipe 1 q 0 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: ring gfx uses 
 VM 
 inv eng 0 on hub 0 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: ring comp_1.0.0 
 uses VM inv eng 1 on hub 0 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: ring comp_1.1.0 
 uses VM inv eng 4 on hub 0 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: ring comp_1.2.0 
 uses VM inv eng 5 on hub 0 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: ring comp_1.3.0 
 uses VM inv eng 6 on hub 0 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: ring comp_1.0.1 
 uses VM inv eng 7 on hub 0 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: ring comp_1.1.1 
 uses VM inv eng 8 on hub 0 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: ring comp_1.2.1 
 uses VM inv eng 9 on hub 0 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: ring comp_1.3.1 
 uses VM inv eng 10 on hub 0 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: ring kiq_0.2.1. 
 0 
 uses VM inv eng 11 on hub 0 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: ring sdma0 uses 
 VM inv eng 0 on hub 8 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: ring vcn_dec 
 uses 
 VM inv eng 1 on hub 8 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: ring vcn_enc0 
 uses VM inv eng 4 on hub 8 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: ring vcn_enc1 
 uses VM inv eng 5 on hub 8 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: ring jpeg_dec 
 uses VM inv eng 6 on hub 8 
 Aug 28 15:36:46 XXXXXXX kernel: amdgpu 0000:0b:00.0: amdgpu: GPU reset(2) 
 succeeded! 
 Aug 28 15:36:46 XXXXXXX gnome-shell[2620]: amdgpu: The CS has cancelled 
 because the context is lost. This context is innocent. 
 Aug 28 15:36:47 XXXXXXX org.signal.Signal.desktop[5529]: [47:082 
 /153647.425551:ERROR:ui/gfx/x/connection.cc:65] X connection error received. 
  
 [continued in next message] 
  
 --- SoupGate-Win32 v1.05 
  * Origin: you cannot sedate... all the things you hate (1:229/2)