Re: 10.99.7 panic: defibrillate

[Taylor R Campbell <riastradh%NetBSD.org@localhost>]

> Date: Sun, 13 Aug 2023 08:03:00 +0200
> From: Thomas Klausner <wiz%NetBSD.org@localhost>
> 
> So it happened again, no bulk build this time, just qt5-qtwebengine in
> a sandbox.
> 
> panic: cpu0: softints stuck for 16 seconds

You could try `no options HEARTBEAT' in your kernel and see if
anything is stuck after a day of builds -- the same crash commands
might show something is up.  Maybe also enter ddb with C-A-ESC or
hw.cnmagic and, for each CPU N, do `mach cpu N' and then `bt' in case
one core is stuck at raised IPL or something.

> Using an older gdb I get:
> (gdb) bt
> [...]
> #7  0xffffffff8023c2a4 in trap (frame=0xffffd3a110933ad0) at /usr/src/sys/arch/amd64/amd64/trap.c:315
> #8  0xffffffff80234ad4 in alltraps ()
> #9  0x0000000000000003 in ?? ()

We need to fix gdb stack traces through kernel trap frames...

Can you try with crash(8)?

# crash -N netbsd.40 -M netbsd.40.core
crash> bt
crash> ps
crash> ps/w
crash> show all tstiles