From owner-freebsd-questions Tue Jun 18 04:47:43 1996 Return-Path: owner-questions Received: (from root@localhost) by freefall.freebsd.org (8.7.5/8.7.3) id EAA20783 for questions-outgoing; Tue, 18 Jun 1996 04:47:43 -0700 (PDT) Received: from ra.dkuug.dk (ra.dkuug.dk [193.88.44.193]) by freefall.freebsd.org (8.7.5/8.7.3) with SMTP id EAA20764; Tue, 18 Jun 1996 04:47:37 -0700 (PDT) Received: (from sos@localhost) by ra.dkuug.dk (8.6.12/8.6.12) id NAA18166; Tue, 18 Jun 1996 13:45:58 +0200 Message-Id: <199606181145.NAA18166@ra.dkuug.dk> Subject: Re: FreeBSD works with Cy486DLC processors? To: alex@fa.tdktca.com (Alex Nash) Date: Tue, 18 Jun 1996 13:45:57 +0200 (MET DST) Cc: bmk@fta.com, Eloy.Paris@ven.ra.rockwell.com, questions@FreeBSD.org, hardware@FreeBSD.org, hal@wwa.com In-Reply-To: <31C69327.32639FD9@fa.tdktca.com> from "Alex Nash" at Jun 18, 96 06:29:43 am From: sos@FreeBSD.org Reply-to: sos@FreeBSD.org X-Mailer: ELM [version 2.4 PL24] Content-Type: text Sender: owner-questions@FreeBSD.org X-Loop: FreeBSD.org Precedence: bulk In reply to Alex Nash who wrote: > > Brant M. Katkansky wrote: > > > > > I installed FreeBSD 2.1.0-RELEASE a couple of weeks ago and since then I > > > have been having programs exiting with signals 10 and 11, making my system > > > too unstable to work as a dedicated e-mail server and as a PPP to Ethernet > > > gateway. > > > > [snip] > > > > I had one given to me not too long ago. Mine is plagued with various > > sig 10 and 11's, same as yours. > > > > Here's the interesting part - disabling the internal and external cache > > makes the problem worse. > > This should be fairly easy to explain: You have bad SIMMs. While your > program is running, erroneous results are returned from RAM and the > processor tries to execute them. Your program subsequently seg faults > due to an invalid instruction. If your cache works properly and your > SIMMs don't, the cache can mitigate these effects since RAM accesses are > less frequent (thus missing the odd inverted bit somewhere :) ). Disable > the caches and now you will be much more likely to see a bad SIMM in action. > Hmm, just to add another datapoint: I can run 2.1 on my 486DLC board, but -current will panic within 2 minutes with a page fault or semilar. I have a very strong feeling that our Cyrix/486DLC support leaves much to be desired, as I can make the system run better (not error free) if I disable the Cyrix code in locore.s. It seems that using the BIOS defaults (both caches on) runs alot better than what we are trying to do to it..... -=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Soren Schmidt (sos@FreeBSD.org) FreeBSD Core Team So much code to hack -- so little time.