*BSD News Article 9386


Return to BSD News archive

Received: by minnie.vk1xwt.ampr.org with NNTP
	id AA5684 ; Fri, 01 Jan 93 01:52:44 EST
Newsgroups: comp.unix.bsd
Path: sserve!manuel.anu.edu.au!munnari.oz.au!uunet!psinntp!dg-rtp!ponds.uucp!rivers
From: rivers@ponds.uucp (Thomas David Rivers)
Subject: Re: [386bsd]  NMI messages on the console and log.
References: <3806@tansei1.tansei.cc.u-tokyo.ac.jp>
Message-ID: <1992Dec29.013202.7291@ponds.uucp>
Date: Tue, 29 Dec 1992 01:32:02 GMT
Lines: 33


  These indicate memory problems (memory parity problems) with your
 motherboard.

  It is possible that such problems are caused by DMA devices you
 may have, and, if you have an older motherboard, its interaction with 
 the memory subsystem.

  Also, if you have a newer motherboard, it could be related to the
 BUS timings (which is configuarable in the setup of newer motherboards.)

  Also, some older motherboards have the habit of asserting this error
 if there is a parity problem on the AT-BUS instead of the memory.

  A card which is known to behave this way with older motherboards is
 the Adaptec 1542 SCSI controller.  Fortunately, that card contains a
 test for this situation in the card's ROM (see you manual for how to
 execute this test.)

  Finally, I owned such an old machine (a DTK KEEN 2000) which,
 although it passed the Adaptec test had the same problems.  I believe
 it just wasn't correct (it also had similar problems with ISC Unix) and
 fell into the situation where it asserted the AT-BUS parity problems.
 (That is, all memory tests showed everthing to be correct; but 386bsd
 and ISC Unix 2.0.2 would still demonstrate the problem.) Since they 
 are rather cheap now, I simply replaced the motherboard with a newer one.

  By the way, I did notice that linux didn't seem to have a problem like
 this, nor did 386bsd v0.0. However, it may be that at the time, these 
 OSs simply didn't log these problems.

	- Dave Rivers -
	(rivers@ponds.uucp)