The exception you are getting is 0x200, machine check at address 0x2eb4a0.
According to gdb, this is in the uvm_km_kmemalloc routine.  It is likely
trying to access some invalid physical memory location while probing for
devices.  You might try poking around in ddb to find the trapframe for the
stack address, i'm not sure why trace isn't working at that point, but
it may be that memory translation is off or something.

If you can figure out which devices it is trying to probe, take them
out of the kernel and see if you can get enough devices working in
order to boot.  Then debug from there.

Darrin

Michael Wolfson <mw@costello.cnf.cornell.edu> writes:

> > NetBSD 1.5ZA (GENERIC) ...
> >    chs@pm2.chiq.com ...
> > [...]
> > obio0 at pci1 dev 23 function 0: addr 0x80000000
> > zsc0 at obio0 offset 0x13000: irq 22,23
> > zstty0 at zsc0 channel 0
> > zstty1 at zsc0 channel 1
> > awacs0 at obio0 offset 0x14000: irq 24,9,10
> > audio0 at awacs0: full duplex
> > trap type 200 at 2eb4a0
> > Press a key to panic.
> > panic: trap
> > Stopped in pid0 (swapper) at cpu_Debugger+0x10:   lwz r0, r1, 0x14,
> > db> trace
> > db>
> >
> >
> >