Anders Magnusson <ragge@ludd.luth.se> writes:

> I have just uploaded a new kernel to the last snapshot. The
> "vget/vref" bug is found and fixed now and it seems as it swaps OK
> for me anyway.

Unfortunately, it doesn't help me.  :-(  With your new kernel, I get
different crashes, but it still does crash.  My KA650 now experiences
"stray interrupt scb0, vec 0x60" events, which means a memory write
timed out -- but as I've said, NetBSD/vax 1.3F is rock solid, and the
MicroVAX Diagnostic Monitor is unable to find a memory problem.  An
example of a crash is seen in this dmesg output (as before, it boots
fine, but when i go and do "make obj" in an NFS-mounted /usr/src with
a local /usr/obj, it goes down within minutes):

NetBSD 1.3I (GENERIC) #2: Sat Dec 19 21:23:38 CET 1998
    ragge@subzero:/multi/src/sys/arch/vax/compile/GENERIC

MicroVAX 3500/3600
realmem = 16736256
avail mem = 13090816
Using 204 buffers containing 835584 bytes of memory.
[...]
scheduler: no room for pid 142(cron), free 1
stray interrupt scb 0, vec 0x60
stray interrupt scb 0, vec 0x60
stray interrupt scb 0, vec 0x60
stray interrupt scb 0, vec 0x60
machine check 10
	vap 80 istate1 86f49fd8 istate2 e8e0fc00 pc fcc07802 psl 8000031c
dmaser=0x88<QBNXM,LOSTERR> qbear=0xc dmaear=0x0
 parity error:	flushing cache
memory err!
panic: mchk
syncing disks... Trap: type 2, code 0, pc 8000066a, psl 4080000
panic: trap
NetBSD 1.3I (GENERIC) #2: Sat Dec 19 21:23:38 CET 1998
    ragge@subzero:/multi/src/sys/arch/vax/compile/GENERIC

MicroVAX 3500/3600
realmem = 16736256
panic: Segv in kernel mode: pc 86f47010 addr 86f47010
syncing disks... panic: ptelen fault in system space: addr ffffffb0 pc 80027e5e
NetBSD 1.3I (GENERIC) #2: Sat Dec 19 21:23:38 CET 1998
    ragge@subzero:/multi/src/sys/arch/vax/compile/GENERIC

MicroVAX 3500/3600
realmem = 16736256
avail mem = 13090816
[...]

Note that it still crashes again on the first reboot after a "real"
crash, with a "Segv in kernel mode" panic -- it always seems to do
this with a post-1.3F kernel.  On the off chance that this is a
hardware problem in my KA650/MS650 combination, I swapped the boards
for a KA630/MS630 kit, and, as before, it didn't even finish booting
before bombing out with an "access fault with no process".

Maybe I have some other hardware that it doesn't like?  Again, this
would be something that doesn't affect 1.3F, but screws up 1.3I...

Anything I can do to gather more information?  Please let me know!

-tih
-- 
Popularity is the hallmark of mediocrity.  --Niles Crane, "Frasier"