Return to BSD News archive
Xref: sserve comp.os.386bsd.bugs:2681 comp.os.386bsd.questions:14837
Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!yarrina.connect.com.au!werple.apana.org.au!otis.apana.org.au!serval.net.wsu.edu!netnews.nwnet.net!oracle.pnl.gov!osi-east2.es.net!cronkite.nersc.gov!dancer.ca.sandia.gov!overload.lbl.gov!dog.ee.lbl.gov!agate!howland.reston.ans.net!vixen.cso.uiuc.edu!newsrelay.iastate.edu!news.iastate.edu!news.iastate.edu!michaelv
From: michaelv@MindBender.HeadCandy.com (Michael L. VanLoon)
Newsgroups: comp.os.386bsd.bugs,comp.os.386bsd.questions
Subject: Re: NetBSD 1.0 w/ AMD (compile probs)
Date: 30 Nov 1994 02:01:40 GMT
Organization: HeadCandy Associates... Sweets for the lobes.
Lines: 44
Distribution: world
Message-ID: <MICHAELV.94Nov29200140@MindBender.HeadCandy.com>
References: <3bfada$a28@que.iphase.com>
NNTP-Posting-Host: mindbender.headcandy.iastate.edu
In-reply-to: dhollist@iphase.com's message of 29 Nov 1994 13:30:18 GMT
In article <3bfada$a28@que.iphase.com> dhollist@iphase.com (David Hollister) writes:
I was told by the guy who got me using NetBSD 1.0 that he seemed to recall
hearing about some problems with people who had AMD CPUs that during large
compiles, that the compiler would occasionally fail with "internal errors"
and signals 6 and 11 (at least that's what I get).
This has nothing to do with the AMD CPU. I have an AMD 486DX2/80, and
have gotten *none* of those. I have it plugged into a Nice (that's
the brand) SuperEISA motherboard, with SiS chipset. I got none of
them, either on an ALR Veisa system with a 386 and motherboard cache.
I did get them with a Cyrix 486DLC plugged into the ALR.
From most accounts, what causes this is marginal caching hardware on
the motherboard (some motherboards do it a lot, some a little, some
not at all). It doesn't matter what CPU you use -- it's cache
related. I don't know why gcc tickles this more than anything else,
but I have a feeling it's a gcc bug that triggers this.
After successive makes,
it does work, but it's quite annoying.
Yes, this is the only way to "fix" it, other than turning your cache
off.
He told me that he seemed to recall
seeing something about a BIOS option change or something to that effect,
perhaps something to do with floating point handling.
No, it has to do with turning off caching options in your BIOS.
If you have write-back, try running it in write-through mode. If that
doesn't work, try selectively turning off your motherboard cache,
while leaving on your CPU cache, then vice versa. If you still have
problems, try turning off all cache. Some PC motherboards are just
not well designed, and no amount of software can fix that.
--
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
Michael L. VanLoon michaelv@HeadCandy.com michaelv@iastate.edu
Free your mind and your machine -- NetBSD free un*x for PC/Mac/Amiga/etc.
Working NetBSD ports: 386+PC, Mac, Amiga, HP300, Sun3, Sun4c, PC532
In progress: DEC pmax (MIPS R2k/3k), VAX, Sun4m
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -