I've just added a 32MB RAM card to one of my diskless 3/260's running
NetBSD-1.2 GENERIC.  It now panics shortly after xdm gets going in
locore.s with "cpu_switch".  Xdm seems to dump core just before the
crash too.  The memory tests run A-OK.

The machine originaly had just the first 24 MB RAM as 3 8MB cards and
worked reasonably well (other than the X server dying every time
Netscape viewed a jpg, and overall being about half the speed of a 3/60
running SunOS-4.1.1U1).

/netbsd: Copyright (c) 1982, 1986, 1989, 1991, 1993
/netbsd: ^IThe Regents of the University of California.  All rights reserved.
/netbsd: 
/netbsd: NetBSD 1.2 (GENERIC) #89: Thu Oct  3 18:08:02 EDT 1996
/netbsd:     gwr@bach:/play/gwr/netbsd/build/sys/arch/sun3/compile/GENERIC
/netbsd: Model: Sun 3/260 (hostid 13005c06)
/netbsd: fpu: mc68881
/netbsd: real mem = 58703872
/netbsd: avail mem = 54026240
/netbsd: using 371 buffers containing 3039232 bytes of memory
/netbsd: cache enabled
/netbsd: mainbus0 (root)
/netbsd: obio0 at mainbus0
/netbsd: zsc0 at obio0 addr 0x0 level 6 (softpri 3)
/netbsd: kbd0 at zsc0 channel 0 (console)
/netbsd: ms0 at zsc0 channel 1
/netbsd: zsc1 at obio0 addr 0x20000 level 6 (softpri 3)
/netbsd: zstty0 at zsc1 channel 0
/netbsd: zstty1 at zsc1 channel 1
/netbsd: eeprom0 at obio0 addr 0x40000
/netbsd: clock0 at obio0 addr 0x60000 level 5
/netbsd: memerr0 at obio0 addr 0x80000 (ECC memory)
/netbsd: intreg0 at obio0 addr 0xa0000
/netbsd: ie0 at obio0 addr 0xc0000 level 3 hwaddr 08:00:20:07:d3:a4
/netbsd: obmem0 at mainbus0
/netbsd: bwtwo0 at obmem0 addr 0xff000000 (1152x900)
/netbsd: vmes0 at mainbus0
/netbsd: cgtwo0 at vmes0 addr 0xff400000 level 4 vector 0xa8 (1152x900)
/netbsd: vmel0 at mainbus0
/netbsd: root on ie0
/netbsd: swap on ie0
/netbsd: dump on ie0
/netbsd: nfs_boot: using network interface 'ie0'
/netbsd: ie0: TDR detected an open 8192 clocks away
/netbsd: nfs_boot: client_addr=0xcc5cfe08
/netbsd: ie0: TDR detected a short 4096 clocks away
/netbsd: nfs_boot: server_addr=0xcc5cfe02
/netbsd: nfs_boot: hostname=pretty.weird.com
/netbsd: root on most:/export/root/pretty
/netbsd: swap on most:/export/swap/pretty
/netbsd: ie0: TDR detected a short 4096 clocks away

Swap is only a 16392 block file on the server.  So, I thought this might
be part of the problem and I borrowed 64MB of space from another
partition on the server and tried again.

This time it ran about ten minutes longer.  However xdm dumped core,
then shortly after rlogging in to try and look at the xdm.core, the
panic cpu_switch happened again.  However it would appear the swap file
was still never touched.

Any ideas?  Should I pull the memory and try upgrading to -current?  Is
there any chance of a binary 1.2C-current sun3 snapshot showing up soon?

I have four identical boards in another 3/260 running SunOS-4.1.1U1
very happily (they are Sun RAM boards...).

-- 
							Greg A. Woods

+1 416 443-1734			VE3TCP			robohack!woods
Planix, Inc. <woods@planix.com>; Secrets of the Weird <woods@weird.com>