From owner-freebsd-smp  Wed Nov  4 11:21:13 1998
Return-Path: <owner-freebsd-smp@FreeBSD.ORG>
Received: (from majordom@localhost)
          by hub.freebsd.org (8.8.8/8.8.8) id LAA16243
          for freebsd-smp-outgoing; Wed, 4 Nov 1998 11:21:13 -0800 (PST)
          (envelope-from owner-freebsd-smp@FreeBSD.ORG)
Received: from irbs.irbs.com (irbs.irbs.com [209.36.62.129])
          by hub.freebsd.org (8.8.8/8.8.8) with ESMTP id LAA16227
          for <freebsd-smp@freebsd.org>; Wed, 4 Nov 1998 11:21:06 -0800 (PST)
          (envelope-from jc@irbs.com)
Received: (from jc@localhost)
	by irbs.irbs.com (8.8.5/8.8.5) id OAA18978;
	Wed, 4 Nov 1998 14:20:58 -0500 (EST)
Message-ID: <19981104142058.31411@irbs.com>
Date: Wed, 4 Nov 1998 14:20:58 -0500
From: John Capo <jc@irbs.com>
To: freebsd-smp@FreeBSD.ORG
Subject: Re: Dog Sloooow SMP
References: <19981103183604.49697@irbs.com>
Mime-Version: 1.0
Content-Type: text/plain; charset=us-ascii
X-Mailer: Mutt 0.89i
In-Reply-To: <19981103183604.49697@irbs.com>; from John Capo on Tue, Nov 03, 1998 at 06:36:05PM -0500
Sender: owner-freebsd-smp@FreeBSD.ORG
Precedence: bulk
X-Loop: FreeBSD.org

Turns out that SMP broke for me on 5/21/98 in identcpu.c version
1.47 when code was added to identify PIIs. Not setting cpu = CPU_II
fixes the problem.

I suspect the real breakage is in pmap.c where cpu is used.

John Capo


Quoting John Capo (jc@irbs.com):
> I have a Gigabyte 686DLX that has been running an SMP kernel from
> May 18 just fine. I upgraded to -current and realized that something
> was wrong when a buildworld was still running the next morning.
> 
> The second CPU is started and top shows both CPUs being used.  I'm
> building a system from May 18 again to make sure the hardware is
> OK.
> 

To Unsubscribe: send mail to majordomo@FreeBSD.org
with "unsubscribe freebsd-smp" in the body of the message