[ On Monday, April 30, 2001 at 13:57:29 (-0400), Matt Levin wrote: ]
> Subject: Re: help please? - dma/SCSI error
>
> so this turned out to be bad memory. removed the offending DIMM and everything is
> now okay. thanks to everyone who responded w/suggestions.

I was going to mention that but I forgot to go back and check myu logs
to confirm....  NetBSD-current (at least 1.5T) is much better at
decoding these and displaying the real cause.

I had a brand new new stick in my Axil 320 that was having problems and
which eventually caused the kernel to go into a loop spewing these
endlessly (instead of occasionally, that is):

Mar 27 20:30:00 sometimes /netbsd: NMI: system interrupts: 10000000<VME=0,SBUS=0,M>
Mar 27 20:30:00 sometimes /netbsd: memory error:
Mar 27 20:30:01 sometimes /netbsd:      EFSR: 231<CE,DW=3,SYNDROME=2>
Mar 27 20:30:01 sometimes /netbsd:      MBus transaction: 8fffcd30<VAH=0,TYPE=3,SIZE=5,C,VA=ff,S,MID=8>
Mar 27 20:30:01 sometimes /netbsd:      address: 0x0ef89600

Finding the offensive memory slot by its address is still a bit of a
mystery to me, at least in these modern machines....  The fact that in
my case it was not the old stick, which I had assumed was bad in the
beginning, but rather one of the new ones confused me for a long time.
I now need to understand the rules for mixing speeds in different banks
because I'm fairly sure one of the new sticks that I pulled is still OK,
but the old memory is of a different speed.  It works OK in a separate
bank from some new memory that's obviously OK too, but I don't know if I
can add the one remaining good new stick into the same bank as the old
stuff or not.  (On the Axil 320 you can add one stick at a time.)

-- 
							Greg A. Woods

+1 416 218-0098      VE3TCP      <gwoods@acm.org>     <woods@robohack.ca>
Planix, Inc. <woods@planix.com>;   Secrets of the Weird <woods@weird.com>