Kernel panic.

GlitchyDot · Aug 9, 2025

Having strange kernel panics during video playback. ( firefox youtube )
Running 14.2-RELEASE but with beckhof patched kernel for nVidia passtrough.

Code:

uname -UKr
14.2-RELEASE 1402000 1402000

My Xorg runs on AMD Radeon PRO W5500 which is somewhat unsupported but sometimes it does work ( like now ).
Can it be its due to my GPU ? I have jellyfin running but using CPU for video encoding/transcoding.
Now testing Firefox without hardware acceleration to see if its fine as crashes are random - sometimes takes 30min sometimes takes 4h. Also have bhyve running and sometimes jail or two.
/var/crash/core
[CORE]
cat /var/crash/core.txt.2
'version' has unknown type; cast it to its declared type
'version' has unknown type; cast it to its declared type
Unable to find matching kernel for /var/crash/vmcore.2
[/CODE]
How to check/read vmcore ?
P.s. i installed 14.2-RELEASE from dvd iso, but never did freebsd-update and then installed beckhofs kernel and never did freebsd-update. Can i do freebsd-update fetch | freebsd-update install and still have my patched kernel ? All i did was pkg update / pkg upgrade last night.

nxjoseph · Aug 9, 2025

GlitchyDot said:
Can i do freebsd-update fetch | freebsd-update install and still have my patched kernel ?

I don't think so, you need to rebuild the kernel from source with applied patch (if it doesn't fail, though).

GlitchyDot said:
All i did was pkg update / pkg upgrade last night.

So this problem occured after an pkg(8) update? Then, i'd recommend checking version of your DRM kernel module packages' version.

% pkg info -x drm

For example, my system:

Rich (BB code):

% pkg info -x drm
drm-61-kmod-6.1.128.1403000_5 ---> OSVERSION is matching: 140300 -> 14.3

The binary package repositories you use may be having inappropriate (for your use case) OSVERSION'ed DRM kernel module packages so that it would cause problems. If so, build DRM packages from ports.

GlitchyDot · Aug 9, 2025

nxjoseph said:
I don't think so, you need to rebuild the kernel from source with applied patch (if it doesn't fail, though).

So this problem occured after an pkg(8) update? Then, i'd recommend checking version of your DRM kernel module packages' version.

% pkg info -x drm

For example, my system:

Rich (BB code):

% pkg info -x drm drm-61-kmod-6.1.128.1403000_5 ---> OSVERSION is matching: 140300 -> 14.3

The binary package repositories you use may be having inappropriate (for your use case) OSVERSION'ed DRM kernel module packages so that it would cause problems. If so, build DRM packages from ports.

no, pre pkg update was same, drm kmod matches kernel version.
gpu issues is my initial thought as in the past only Xorg or Wayland would crash but now whole kernel panic and no other crashes

nxjoseph · Aug 10, 2025

GlitchyDot said:
gpu issues is my initial thought as in the past only Xorg or Wayland would crash but now whole kernel panic and no other crashes

Maybe try the GPU in another os like Linux?

Maturin · Aug 10, 2025

GlitchyDot said:
All i did was pkg update / pkg upgrade last night.

Maybe it's a good idea to update freebsd, too.
Yesterday some security mails were posted here in the forums which normally are provided by the according mailing list. When those mails come it's a good sign for new patches for FreeBSD are out.
Besides it's always a good idea to keep FreeBSD and the packages congruent I recommend you also update the system.
If you are not running a default kernel, you need to compile and install a new one, which may take some minutes (according to your machine and config file) but be a simple task.

GlitchyDot · Aug 10, 2025

Ke

nxjoseph said:
Maybe try the GPU in another os like Linux?

GPU with linux works fine.
My gpu is recognised as Navi 14 but is actually Navi 14 PRO. but during install Firmware gets recognized as Navi 14 and installed.

Code:

 6    1 0xffffffff83600000   502958 amdgpu.ko
 7    2 0xffffffff83b03000    7b050 drm.ko
13    1 0xffffffff83b96000    2f050 amdgpu_navi14_sos_bin.ko
14    1 0xffffffff83bc6000    2c2e0 amdgpu_navi14_asd_bin.ko
15    1 0xffffffff83bf3000     a3e0 amdgpu_navi14_ta_bin.ko
16    1 0xffffffff83bfe000    42a70 amdgpu_navi14_smc_bin.ko
17    1 0xffffffff83c41000    425e0 amdgpu_navi14_pfp_wks_bin.ko
18    1 0xffffffff83c84000    425e0 amdgpu_navi14_me_wks_bin.ko
19    1 0xffffffff83cc7000    42560 amdgpu_navi14_ce_wks_bin.ko
20    1 0xffffffff83d0a000     c848 amdgpu_navi14_rlc_bin.ko
21    1 0xffffffff83d17000    43a10 amdgpu_navi14_mec_wks_bin.ko
22    1 0xffffffff83d5b000    43a10 amdgpu_navi14_mec2_wks_bin.ko
23    1 0xffffffff83d9f000     a4e0 amdgpu_navi14_sdma_bin.ko
24    1 0xffffffff83daa000     a4e0 amdgpu_navi14_sdma1_bin.ko
25    1 0xffffffff83db5000    643a0 amdgpu_navi14_vcn_bin.ko

But maybe this is not GPU fault .. i dont know. Had one more kernel panic last night without playing any videos, anything and no panic since.
So i dont know exact issues with panics.
In the past - i was not able to use Xorg or use with minimal time and the n Xorg crashes and no displays found.

Maturin said:
Maybe it's a good idea to update freebsd, too.
Yesterday some security mails were posted here in the forums which normally are provided by the according mailing list. When those mails come it's a good sign for new patches for FreeBSD are out.
Besides it's always a good idea to keep FreeBSD and the packages congruent I recommend you also update the system.
If you are not running a default kernel, you need to compile and install a new one, which may take some minutes (according to your machine and config file) but be a simple task.

I really dont wanna do it now because i have GPU pass trough which is non official FreeBSD Kernel.

But if i try to use debug tool for kernel - i get error: Cant open vmcore without a kernel or "/var/crash/vmcore.3": not in executable format: file format not recognized.
Does it mean my kernel not found or its not configured properly for debugging ?