From owner-freebsd-stable Mon Oct 26 01:06:51 1998 Return-Path: Received: (from majordom@localhost) by hub.freebsd.org (8.8.8/8.8.8) id BAA19956 for freebsd-stable-outgoing; Mon, 26 Oct 1998 01:06:51 -0800 (PST) (envelope-from owner-freebsd-stable@FreeBSD.ORG) Received: from owl.org (owl.org [198.206.215.91]) by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id BAA19949 for ; Mon, 26 Oct 1998 01:06:47 -0800 (PST) (envelope-from cgull@owl.org) Received: (from cgull@localhost) by owl.org (8.9.1/8.9.1/cgull) id EAA03726; Mon, 26 Oct 1998 04:04:38 -0500 (EST) Message-ID: <19981026040438.38079@owl.org> Date: Mon, 26 Oct 1998 04:04:38 -0500 From: john hood To: Burkard Meyendriesch Cc: freebsd-stable@FreeBSD.ORG Subject: Re: ECC memory support References: <199810222157.XAA04898@Reineke.malepartus.de> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mailer: Mutt 0.89.1 In-Reply-To: <199810222157.XAA04898@Reineke.malepartus.de>; from Burkard Meyendriesch on Thu, Oct 22, 1998 at 11:57:01PM +0200 Sender: owner-freebsd-stable@FreeBSD.ORG Precedence: bulk X-Loop: FreeBSD.ORG On Thu, Oct 22, 1998 at 11:57:01PM +0200, Burkard Meyendriesch wrote: > thanks for the hint. Yes indeed, I run an AMD K6/233 with ECC enabled; > last night I upgraded my BIOS. Maybe it will help my spurious errors to > disappear... > > But my original questions are still pending: Isn't there any possibility > to detect 1-bit errors even if they are detected and corrected by my > ECC electronics? Isn't there any register of the 82430HX chip set which > can be examined after an memory error NMI? The only thing I could find is I coded this hack up a few weeks ago. It runs from cron. I can't be sure it works because it hasn't reported a single-bit error yet :) It's not exactly user-friendly (you'll need Intel doco to make sense of the ECC register), but you have to know how your motherboard is wired up to make sense of the error report anyway. I'm running it on a '440FX-based board, but my docs on the '430HX indicate it handles ECC reporting basically the same way, through this same register. --jh #!/usr/local/bin/bash PATH=/sbin:/usr/sbin:/bin:/usr/bin i=`pciconf -r pci0:0:0 0x90` SELF=$(( ( $i / 0x100 ) % 2 )) if [ $SELF -ne 0 ] then pciconf -w pci0:0:0 0x90 $i msg="ECC register is $i, register cleared" logger $i echo $i | mail -s "ECC error" root fi -- Mr. Belliveau said, "the difference was the wise, John Hood, cgull intelligent look on the face of the cow." He was @ *so* right. --Ofer Inbar owl.org To Unsubscribe: send mail to majordomo@FreeBSD.org with "unsubscribe freebsd-stable" in the body of the message