Hello all!  I hope someone can help me with my latest alpha adventure:

I recently upgraded my Multia/UDB to the 961127 snapshot.  I've been
getting panics in the ncr driver since then (i.e. 5 times in 1 day of
total uptime).  Here's the latest message : 

	unexpected machine check :

	    mces    = 0x19
	    vector  = 0x670
	    param   = 0xfffffc0000006048
	    pc      = 0xfffffc0000284be4
	    ra      = 0xfffffc0000284bbc
	    curproc = 0xfffffe004a5ed400
	        pid = 101, comm = update

	panic: machine check
	ncr0:2: ERROR (a0:40) (e-aa-0) (8/13) @ (494:c0000000).
		script cmd = c000004
		reg:     da 10 80 13 47 08 02 1f 02 0e 82 aa 80 00 0a 00.
	ncr0: restart (fatal error).

At this point (apparently -- I didn't actually witness the crash), the 
screen is hatched with 3-1/4 diagonal lines of pixels (my guess is that 
they're '0xff's stored in <width of fb>+1 incremented locations.

I rebooted the machine, cleaned up the crashdump (/var/crash/*) so 
sendmail quit complaining and checked the addresses with 'gdb 
/netbsd.gdb' (netbsd.gdb from 961127 snapshot) :

	pc : sys_sync
	ra : sys_sync

This afternoon, the machine crashed also.  The PC/RA in that case was 
"ncr_timeout / softclock".  I have suspected that the Quantum XP32150W 
that came with the machine is flaky.  THere does not appear to be any 
noticeable correlation between crashes and specific system activity.

Hardware configuration is :

	Alpha PC AXPpci33 ("NoName"), 166MHz
	48 MB memory
	Ethernet on AUI connector (via 10-base2 to AUI xcvr)
	sd0: Toshiba MK1824FBV (300+ MB internal 2.5" drive)
	sd1: Quantum XP32150W (3.5" external)
	cd0: Compaq CR503BCQ (5.25" external)

All else is stock Multia/UDB from Starship / Computer Guys.

Any help/ideas would be most appreciated.

Thanks again!

	-rAT

O----O The Terminator rAT - rat@cynical.org - The Cynical Software Foundation
 \oO/
==\/==                            Oiwuetnbiset!