You are not logged in.

#1 Today 09:50:05

andrejr
Member
Registered: 2018-05-25
Posts: 13

Constant kernel panics and freezes on Lenovo Thinkpad T14 Gen5 AMD

Hi everyone, I've been having this issue for a while, and it's driving me nuts.

I constantly either get freezes (the machine completely freezes), or kernel panics (sometimes even w/ a QR code).

Here are the QR code variants: 1 and 2

It usually takes a while, but they can also repeat quite frequently.

I have a Thinkpad T14 Gen5 AMD, which is connected to a Lenovo  Thinkpad TB4 dock. The dock is connected to 2 monitors (daisy-chained) via usb-c; the monitors are slightly-above-4k.
I can't be sure, but I think the issues started appearing once I switched to USB connection for monitors, which requires MST.

These are my boot flags:

options     rw quiet loglevel=3
options     udev.log_priority=3 rd.udev.log_priority=3
options     vga=current
options     systemd.unified_cgroup_hierarchy=1 acpi.ec_no_wakeup=1
options     amdgpu.dcdebugmask=0x10
options     amdgpu.gpu_recovery=1
options     amdgpu.gfx_off=0 amdgpu.runpm=0 amdgpu.tmz=0

Offline

#2 Today 14:00:03

seth
Member
From: Won't reply 2 private help req
Registered: 2012-09-03
Posts: 78,812

Re: Constant kernel panics and freezes on Lenovo Thinkpad T14 Gen5 AMD

[10955.980424]  dsc_prepare_config+0x3bc/0x580 [amdgpu c05f8e65d103f2dc93eeac6c5b2458d8a423ec36]
[10955.980813]  dsc2_validate_stream+0x3a/0x60 [amdgpu c05f8e65d103f2dc93eeac6c5b2458d8a423ec36]
[10955.981140]  dcn20_validate_dsc+0x211/0x260 [amdgpu c05f8e65d103f2dc93eeac6c5b2458d8a423ec36]
[10955.981456]  dcn30_internal_validate_bw+0x9ec/0xae0 [amdgpu c05f8e65d103f2dc93eeac6c5b2458d8a423ec36]
[10955.981758]  dcn314_validate_bandwidth+0xc9/0x310 [amdgpu c05f8e65d103f2dc93eeac6c5b2458d8a423ec36]
[10955.982047]  create_validate_stream_for_sink+0x507/0x540 [amdgpu c05f8e65d103f2dc93eeac6c5b2458d8a423ec36]
[10955.982437]  pre_validate_dsc+0x38b/0x510 [amdgpu c05f8e65d103f2dc93eeac6c5b2458d8a423ec36]

dcn30_internal_validate_bw is commented "do not support self refresh only" so you're likely(?) hitting that path b/c of "amdgpu.dcdebugmask=0x10", https://elixir.bootlin.com/linux/v6.17. … ce.c#L1698

can also repeat quite frequently

Can you run w/o the dock for a while to see whether the frequency drops out in return?

started appearing once I switched to USB connection for monitors, which requires MST

That would fit the pattern of the backtrace - does either of the outputs support https://wiki.archlinux.org/title/Variable_refresh_rate ?
Have you tried to just swap master and slave?

Online

Board footer

Powered by FluxBB