Framework Laptop 13 AMD Ryzen AI 300 is repeatedly hard-freezing on Ubuntu 25.10, and the pattern looks like an AMDGPU kernel-driver hang rather than a normal app crash.
System
- Framework Laptop 13 AMD Ryzen AI 300
- Radeon 840M / 860M class iGPU
- Ubuntu 25.10
- Kernel reproducing now: 6.17.0-35-generic
User-visible behavior
- screen stops updating
- mouse still moves
- clicks stop working
- caps lock still responds
Consistent kernel signature
- amdgpu: MES failed to respond to msg=MISC (WAIT_REG_MEM)
- amdgpu: failed to reg_write_reg_wait
- workqueue: amdgpu_tlb_fence_work [amdgpu] hogged CPU
- sometimes also: MES ring buffer is full
What I already tested
- amdgpu.mes=0
- amdgpu.cwsr_enable=0
- PSR/display-state test via:
options amdgpu dcdebugmask=0x10
However, the machine still froze on 2026-06-10 after resume with all of those already active.
Latest reproduced crash
- boot started: 2026-06-09 22:45:39 PDT
- suspend exit: 2026-06-10 08:40:45 PDT
- first fatal GPU errors: 2026-06-10 08:41:07 PDT
- boot ended: 2026-06-10 08:41:13 PDT
So suspend/resume looks like a strong trigger or amplifier.
Same failure class has reproduced for me on:
- 6.17.0-23-generic
- 6.17.0-29-generic
- 6.17.0-35-generic
Related threads
- AMD GPU MES Timeouts Causing System Hangs on Framework Laptop 13 (AMD AI 300 Series)
- Framework 13 Ryzen AI 7 340 - MES ring buffer full / system freeze on Linux Mint 22.3
Questions
- Is Framework tracking this as a known AI 300 / Radeon 840M-860M suspend/resume amdgpu issue?
- Is there a recommended kernel, firmware, or BIOS combination for this exact pattern?
- Has anyone actually stabilized this on similar hardware without giving up suspend/resume?