From owner-freebsd-hackers  Fri Nov 14 19:33:20 1997
Return-Path: <owner-freebsd-hackers>
Received: (from root@localhost)
          by hub.freebsd.org (8.8.7/8.8.7) id TAA26140
          for hackers-outgoing; Fri, 14 Nov 1997 19:33:20 -0800 (PST)
          (envelope-from owner-freebsd-hackers)
Received: from mailhost.Ipsilon.COM (mailhost.ipsilon.com [205.226.5.12])
          by hub.freebsd.org (8.8.7/8.8.7) with SMTP id TAA26134
          for <freebsd-hackers@freebsd.org>; Fri, 14 Nov 1997 19:33:18 -0800 (PST)
          (envelope-from jre@ipsilon.com)
Received: from radio.ipsilon.com (radio.Ipsilon.COM [205.226.28.3]) by mailhost.Ipsilon.COM (8.6.11/8.6.10) with SMTP id TAA05410; Fri, 14 Nov 1997 19:32:15 -0800
Message-ID: <346D17BA.1B37ADEA@ipsilon.com>
Date: Fri, 14 Nov 1997 19:32:10 -0800
From: Joe Eykholt <jre@ipsilon.com>
Organization: Ipsilon Software Engineering
X-Mailer: Mozilla 3.01Gold (X11; I; FreeBSD 2.1.0-RELEASE i386)
MIME-Version: 1.0
To: jlemon@americantv.com
CC: freebsd-hackers@freebsd.org
Subject: Re: FreeBSD Pentium Bug fix (proposed) 
References: <199711150115.RAA18627@hub.freebsd.org>
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Sender: owner-freebsd-hackers@freebsd.org
X-Loop: FreeBSD.org
Precedence: bulk

Jonathan Lemon wrote:

> My ``fix'' is to have the IDT descriptor reference a segemnt which has
> a length of 0.  This has the effect of mapping SIGILL into SIGBUS, so that
> the `cmpxchg8' crash now generates a Bus error.  (I didn't bother  returning
> the correct signal; it can probably be added if it is important) 

Cool fix!  It should work and seems much nicer than the
two-page IDT fix.

One point, though.  The segment length is at least one byte
since the limit in the descriptor is the last valid offset
in the segment, not the length.   That means that the address might
be referenced.  The granularity should be 0 for bytes. 

I think a user can map address 0 (at least on one OS) containing a 
single-byte instruction that might be run in ring 0, 
so another, guaranteed-invalid address might be better, or you might
leave the P bit off in that segment or (better) in the 
IDT entry 6 descriptor, causing a segment-not-present fault.  
(I haven't tried any of this). 

	Joe