*BSD News Article 25127


Return to BSD News archive

Newsgroups: comp.os.386bsd.questions
Path: sserve!newshost.anu.edu.au!munnari.oz.au!news.Hawaii.Edu!ames!sgiblab!swrinde!cs.utexas.edu!uunet!brunix!cs.brown.edu!Mark_Weaver
From: Mark_Weaver@brown.edu
Subject: Re: Random seg. faults w/ NetBSD
In-Reply-To: simonm@dcs.glasgow.ac.uk's message of Fri, 17 Dec 1993 13:24:05 GMT
Message-ID: <MARK_WEAVER.93Dec18172426@tonto-slip9.cis.brown.edu>
Sender: news@cs.brown.edu
Reply-To: Mark_Weaver@brown.edu
Organization: Brown University Department of Computer Science
References: <CI6L86.3HJ@dcs.glasgow.ac.uk>
Date: Sat, 18 Dec 1993 22:24:26 GMT
Lines: 26

In article <CI6L86.3HJ@dcs.glasgow.ac.uk> simonm@dcs.glasgow.ac.uk (Simon Marlow) writes:
> I've just bought a new motherboard (486/40 8M 256k write-back cache).
> When running a compile job (compiling the kernel for instance) I get
> random segmentation faults from the compiler.  If I turn off the
> external cache using setup, everything works fine.  No amount of
> fiddling with wait states and cache cycles seems to help while the
> cache is on.  The exact same NetBSD installation used to work fine
> with my old 386/25 motherboard.
> 
> I thought I'd consult the group before I send the board back.  Does
> anyone have a clue as to what might cause this?

The problem is in the design of your new motherboard.  DMA transfers
do not automatically invalidate the cache on your board, which
effectively means that much of the data being read from your disk
seems to be getting corrupted.

I have the same problem with my motherboard, but I don't mind too much.
The speed difference isn't nearly as large as I'd expect, and this
may be fixed at some point in the future.  There is a fairly clean
way to work around it in software; it just hasn't been done yet.

	Mark
--------------------------------------------------------------------
Email: Mark_Weaver@brown.edu           | Brown University
PGP Key: finger mhw@cs.brown.edu       | Dept of Computer Science