Adding an Arch Linux per-boot bisect that confirms the same root cause on the same device `1002:15bf` (Phoenix1 Radeon 780M, gfx1103).
## Bisect (kernel version vs `ring gfx_0.0.0 timeout` per boot) ``` 7.0.12-arch1-1 10 boots, 0 hangs 7.0.13-arch1-1 0, 0, 3 hangs <- regression introduced 7.0.13-arch1-2 0 hangs (1 boot) 7.0.14-arch1-1 2,10,0,0,0,1,1,1 hangs 7.1.2-arch3-1 3, 13 hangs ``` Regression appears with the 7.0.12 -> 7.0.13 bump. Latent and uptime/state-dependent — a clean boot can run hours before the first hang, consistent with TLB-fence accumulation in the MES ring. ## Firmware ruled out Phoenix MES blobs (`gc_11_0_0_mes.bin`, `gc_11_0_0_mes1.bin`, `gc_11_0_0_mes_2.bin`) are byte-identical between `linux-firmware` 20260519-1 and 20260622-1. A boot on 2026-06-26 ran 7.0.13-1 + the newer firmware for ~13h with 0 hangs; the next boot (same kernel + same firmware) hung. The kernel bump is the trigger, not the firmware. ## Kernel log signature ``` ring gfx_0.0.0 timeout, signaled seq=..., emitted seq=... Process kwin_wayland pid ... thread kwin_wayla:cs0 Starting gfx_0.0.0 ring reset MES failed to respond to msg=RESET failed to reset legacy queue reset via MES failed and try pipe reset -110 Ring gfx_0.0.0 reset failed GPU reset begin!. Source: 1 MES failed to respond to msg=REMOVE_QUEUE failed to unmap legacy queue [drm:gfx_v11_0_hw_fini] *ERROR* failed to halt cp gfx MODE2 reset GPU reset succeeded, trying to resume [drm] device wedged, but no recovery needed ``` KWin (Wayland) then loses its GL context (`GL_CONTEXT_LOST`, `A graphics reset not attributable to the current GL context occurred`). 30 ring timeouts / 28 MES failures / 28 "GPU reset succeeded" in 10 days. ## Workaround in use `amdgpu.cwsr_enable=0` on the kernel command line — effective on this system, keeps the desktop fully functional. Confirmed across ROCm#5590, Framework community#71364/52. This bisect supports the TLB-fence/MES-ring-flooding root cause and the `e9f58ff991dd` fix direction. Thanks for the work on this. --- *Reported from an Arch Linux workstation — AMD Phoenix1 (Radeon 780M, gfx1103), kernel 7.1.2-arch3-1, KDE Plasma 6.7 / KWin Wayland. Adding this bisect as cross-distro confirmation of the same root cause on device 1002:15bf.* -- You received this bug notification because you are a member of Ubuntu Bugs, which is subscribed to Ubuntu. https://bugs.launchpad.net/bugs/2148538 Title: Kernel lockup on 6.17.0-1017-oem Lenovo P14s gen 6 AMD Ryzen AI 7 pro 350 To manage notifications about this bug go to: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/2148538/+subscriptions -- ubuntu-bugs mailing list [email protected] https://lists.ubuntu.com/mailman/listinfo/ubuntu-bugs
