> Yep.  It happens on my IIci (68030) with NCR SCSI driver.  Kernel
> compiled with -O2...

So it appears to be SCSI driver and processor independant.  I
compiled a kernel several weeks ago with very little non-essential
stuff and it crashed in the same way.

Scott thinks it's unlikely that it's a pmap-related problem, which
is what I was kind of leaning toward.  I was making the assumption
that if this is a machine-independent bug, someone else would be seeing
similar behavior and that does not seem to be the case.

Hmm...  Maybe I can try this without COMPAT_13, booting diskless with a
-current snapshot.

My stripped kernel config:

include	"arch/mac68k/conf/std.mac68k"

maxusers	32		# estimated number of users

makeoptions	COPTS="-O2 -m68040 -pipe"
options 	DIAGNOSTIC
options 	DEBUG
options 	MALLOCLOG
options 	LOCKDEBUG

# CPU support.  At least one is REQUIRED.
options 	M68040

# CPU-related options.
options 	FPSP

# Standard system options
options 	RTC_OFFSET=0	# hardware clock is this many mins. west of GMT

# Diagnostic/debugging support options
options 	KMEMSTATS	# kernel memory statistics (vmstat -m)
options 	DDB		# in-kernel debugger

# Compatibility options
options 	COMPAT_13	# NetBSD 1.3,

# File systems
file-system 	FFS		# UFS

options 	FIFO		# FIFOs; RECOMMENDED

# Networking options
options 	INET		# IP + ICMP + TCP + UDP

# Kernel root file system and dump configuration.
config		netbsd	root on sd0a type ffs

#
# Device configuration
#

mainbus0 at root
fpu0 at mainbus?			# Floating-Point Coprocessor support

# Basic Bus Support

# On-board I/O bus support
obio0 at mainbus?
adb0	at obio?			# ADB bus
aed*	at adb?				# ADB event device
kbd*	at adb?				# ADB keyboard
intvid0	at obio?			# Internal video hardware
grf*	at intvid?
ite0	at grf?
esp0	at obio?			# SCSI NCR 53C9x
scsibus* at scsi?
sd*	at scsibus? target ? lun ?	# SCSI disk drives

pseudo-device	loop			# network loopback