On Wed, 14 May 2025, Andrew Randrianasulu wrote:
> well, not sure how this might help?
>
> I trued to disable all acceleration via radeon.conf file copied and edited
> from my Linux disk, but still I see in NetBSD's /var/log/messages:
>
>
> May 14 04:13:23 netbsd10 savecore: reboot after panic: [ 43.6486328]
> panic: unlocking unlocked wait/wound mutex: 0xffff8fd3383ff520
>
> May 14 04:33:18 netbsd10 savecore: reboot after panic: [ 824.3507701]
> panic: unlocking unlocked wait/wound mutex: 0xffff8e7d24fa0d20
> May 14 04:33:18 netbsd10 savecore: system went down at Wed May 14 04:16:17
> 2025 May 14 04:33:18 netbsd10 savecore: writing
> compressed core to /var/crash/netbsd.1.core.gz May 14 04:34:00
> netbsd10 savecore: writing compressed kernel to /var/crash/netbsd.1.gz
>
That `panic: unlocking unlocked wait/wound mutex' looks very reproducible:
https://gnats.netbsd.org/cgi-bin/query-pr-single.pl?number=59275
and, also the possibly related,
https://gnats.netbsd.org/cgi-bin/query-pr-single.pl?number=58032
> netbsd.1.core.gz is 128 MB so it will not fit into email ;) :/
>
What you should do is compile a new -HEAD GENERIC kernel with the
amdgpu drivers and these lines uncommented:
options DIAGNOSTIC
options DEBUG
options LOCKDEBUG
That should help track down this issue considerably. Then, if there's kernel
crash dump, upload it somewhere public and post the URL and crash messages
to the 1st PR.
Ah, thanks a lot! Right now my machine busy with Linux/qemu but I'll look into -current netbsd kernel compilation with noted options set sometime next day.