*BSD News Article 40408


Return to BSD News archive

Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!yarrina.connect.com.au!news.mira.net.au!otis.apana.org.au!serval.net.wsu.edu!netnews.nwnet.net!oracle.pnl.gov!osi-east2.es.net!lll-winken.llnl.gov!uwm.edu!vixen.cso.uiuc.edu!sdd.hp.com!news.cs.indiana.edu!usenet.ucs.indiana.edu!bigbang.astro.indiana.edu!ahabig
From: ahabig@bigbang.astro.indiana.edu (Alec Habig)
Newsgroups: comp.os.386bsd.questions
Subject: [FreeBSD] - new motherboard woes
Date: 5 Jan 1995 17:21:14 GMT
Organization: Indiana University Astrophysics, Bloomington, IN
Lines: 32
Message-ID: <3eh9qa$1ie@usenet.ucs.indiana.edu>
NNTP-Posting-Host: bigbang.astro.indiana.edu

Hi all,
	I joyfully installed a nice new 386DX-40 motherboard last night (I got
it on the cheap to retire my old 386DX-25), and encountered two problems with
FreeBSD (version 1.1.5 - lack of an upgrade option for 2.0 and a lack of a 
place to backup user stuff for the obligatory reinstall has trapped me like 
a fly in amber).

1) During periods of intense CPU and disk I/O (like a compile), the machine
randomly hangs.  Back before 1.1, this was due to an interrupt timing problem
in wd.c for the IDE controller.  The kernal would somehow miss the drive's "I'm
done" flag, and sit there waiting for it forever.  However, this was fixed for
1.1+ - at least the upgrade cured the problem on my system.  However, the new
clock speed on my new board seems to have brought out the problem again.  Does
anybody have a patch that really fixes this?  Can I grab the latest wd.c from
./current and stand any chance of it working properly with v1.1.5?

2) My new board's main attraction was the external cache for the CPU.  If I
boot DOS, it improves my machine's speed by almost a factor of 2.  However, if
I try and boot FreeBSD with the cache enabled, it crashes just at the beginning
of the boot sequence (right after you hit return, before the memory test) and
complains "page fault in kernal mode".  Side note - the cache makes OS/2 spew 
as well, although DOS thinks it's peachy.  What gives?  Where should I look for
a potential hardware problem?  Or does the kernal need to know about the cache
somehow?

	Thanks,
	Alec
-- 
	    Alec Habig, Indiana University High Energy Astrophysics
		       ahabig@bigbang.astro.indiana.edu
	      http://astrowww.astro.indiana.edu/personnel/ahabig/
	    Ted Kennedy's car has killed more people than my guns.