From mboxrd@z Thu Jan  1 00:00:00 1970
Return-Path: <linux-kernel-owner+w=401wt.eu-S1030876AbXEDPqo@vger.kernel.org>
Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand
	id S1030876AbXEDPqo (ORCPT <rfc822;w@1wt.eu>);
	Fri, 4 May 2007 11:46:44 -0400
Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1030882AbXEDPqo
	(ORCPT <rfc822;linux-kernel-outgoing>);
	Fri, 4 May 2007 11:46:44 -0400
Received: from terminus.zytor.com ([192.83.249.54]:33685 "EHLO
	terminus.zytor.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org
	with ESMTP id S1030876AbXEDPqn (ORCPT
	<rfc822;linux-kernel@vger.kernel.org>);
	Fri, 4 May 2007 11:46:43 -0400
Message-ID: <463B551E.8030701@zytor.com>
Date: Fri, 04 May 2007 08:45:34 -0700
From: "H. Peter Anvin" <hpa@zytor.com>
User-Agent: Thunderbird 2.0.0.0 (X11/20070419)
MIME-Version: 1.0
To: Jeremy Fitzhardinge <jeremy@goop.org>
CC: "Eric W. Biederman" <ebiederm@xmission.com>,
       Rusty Russell <rusty@rustcorp.com.au>, Andi Kleen <ak@suse.de>,
       Chris Wright <chrisw@sous-sol.org>, Zachary Amsden <zach@vmware.com>,
       Andrew Morton <akpm@linux-foundation.org>,
       Linus Torvalds <torvalds@linux-foundation.org>,
       lkml - Kernel Mailing List <linux-kernel@vger.kernel.org>
Subject: Re: [RFC PATCH 3/3] boot bzImages under paravirt
References: <1178283582.23670.67.camel@localhost.localdomain>	<1178283724.23670.70.camel@localhost.localdomain>	<1178284052.23670.75.camel@localhost.localdomain> <m1ps5gslfm.fsf@ebiederm.dsl.xmission.com> <463B4E12.50703@goop.org>
In-Reply-To: <463B4E12.50703@goop.org>
X-Enigmail-Version: 0.95.0
Content-Type: text/plain; charset=ISO-8859-1
Content-Transfer-Encoding: 7bit
Sender: linux-kernel-owner@vger.kernel.org
X-Mailing-List: linux-kernel@vger.kernel.org

Jeremy Fitzhardinge wrote:
> 
> So if we reload using a constant selector, then that will break under
> Xen. But if we do a:
> 
>     mov %cs, %eax
>     mov %eax, %ds
>     // etc
> 
> sequence then it should be fine.  This will work even for loading %ss,
> since the %cs CPL will equal the RPL needed for %ss.

In 32-bit mode?  Surely you're joking, Mr. Feynman!

In protected mode, you *must* have different descriptor values for the
code segment as opposed to the data segments, at least if you ever want
to write anything to the data segments!

What's worse, reloading segments here might be highly unsafe, if the
memory previously occupied by the GDT has been overwritten.  Keep in
mind the GDT is touched on a segment *load*, not on a segment *access*;
in areas such as booting that can be a huge difference.

	-hpa