I'm getting exactly the same thing, since about a week before the
signal changes. See pr port-mac68k/6411. pr6435 looks like it may be
related, as well.

The exact symptoms vary when you run disklabel after the machine's
been up a while. They include MMU fault traps and illegal instruction
errors. The only thing that's clear is that there's a serious problem.
I personally feel it's very significant that disklabel allocates pool
memory if and only if it's run over a disk that does not contain the
root fs, but I lack the skills to pursue that angle.

On Sat, 21 Nov 1998, Hauke Fath wrote:

> Trying to display a disklabel of an unmounted SCSI drive (a Fujitsu M2512
> MO here, but happens with an old Quantum 40M drive, too) panics (even) in
> single user:
> 
> # disklabel sd1
> panic: fstat
> Stopped in disklabel at _Debugger+0x6:  unlk    a6
> db> t
> _Debugger(128,77cf88,77cf44,1ef44,1eed2) + 6
> _panic(1eed2,7684d0,1a1ec,0,0) + 50
> _sys___fstat13(7684d0,77cf88,77cf80) + 66
> _syscall(117) + 116
> _trap0() + e