Ok, I *think* it is the new rasops code anyway.. :)  If I telnet from the
console (a cg6), to another machine (NetBSD/i386 1.4.1) and run something like 'elm',
my sparc (SS10/Model 30) immediately stalls.  One time I was able to read the error,
something about 'MBUS memory error, ...' and the machine is hung.  Can't drop back
to the prom, etc.  If I don't use the console, the machine is very stable, so I
believe it is related to the console code.  The reason I am suspicious of the new
rasops code, is because this type of thing worked fine several weeks ago, and it
wasn't until I booted a new kernel (with the goofy colors on my cg6) that I had
the failure.  Otherwise my machine is very reliable, running the 'GENERIC.MP'
kernel.

How would I capture this crash info?  Setting the output-device, or input-device
to ttya prevents me from logging in on the framebuffer.  Does there exist a tty
type device on the cg6 when using the serial console?

-Andrew
-- 
-----------------------------------------------------------------
Andrew Gillham                            | This space left blank
gillham@whirlpool.com                     | inadvertently.
I speak for myself, not for my employer.  | Contact the publisher.