My Framework is fully locking up whenever I try to play a video, such as YouTube or Twitter, after a long period of not playing videos. This usually happens when I open YouTube after playing a game, but I have managed to replicate it by idling then playing a YouTube video. This started to happen after I updated BIOS to 4.05.
My Framework 16 has 7840HS CPU and RS7700S GPU.
I am on CachyOS, with kernel ver 7.2.9-2-cachyos and linux-firmware-amdgpu ver 1:20260916-1.
journalctl -b -1 -p err outputs this after rebooting:
Oct 11 12:43:11 hydra kernel: ucsi_acpi USBC000:00: unknown error 256
Oct 11 13:34:35 hydra kernel: tpm tpm0: Operation Timed out
Oct 11 13:34:35 hydra kernel: tpm tpm0: tpm_try_transmit: send(): error -62
Oct 11 14:14:30 hydra kernel: amdgpu 0000:c4:00.0: ring vcn_unified_0 timeout, signaled seq=34, emitted seq=34
Oct 11 14:14:30 hydra kernel: amdgpu 0000:c4:00.0: Process helium pid 12685 thread helium:cs0 pid 12691
Oct 11 14:14:30 hydra kernel: amdgpu 0000:c4:00.0: Starting vcn_unified_0 ring reset
Oct 11 14:14:30 hydra kernel: amdgpu 0000:c4:00.0: vcn sram load failed -22
Oct 11 14:14:33 hydra kernel: amdgpu 0000:c4:00.0: vcn sram load failed -22
Oct 11 14:14:33 hydra kernel: amdgpu 0000:c4:00.0: Ring vcn_unified_0 reset failed
Oct 11 14:14:34 hydra kwin_wayland[1604]: Pageflip timed out! This is a bug in the amdgpu kernel driver
[... Browser systemd-coredump ...]
Oct 11 14:14:34 hydra kwin_wayland[1604]: Please report this at https://gitlab.freedesktop.org/drm/amd/-/issues
Oct 11 14:14:34 hydra kwin_wayland[1604]: With the output of 'sudo dmesg' and 'journalctl --user-unit plasma-kwin_wayland --boot 0'
I asked some clankers what this means and Gemini thinks this is because amdgpu is having trouble waking up the video decoding hardware from sleep. At its suggestion, I added amdgpu.ppfeaturemask=0xfff73fff amdgpu.gpu_recovery=1 amdgpu.dc=1 amdgpu.runpm=0 to my Limine boot parameters, but these parameters did not fix it.