From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1759338AbYAXWfp (ORCPT ); Thu, 24 Jan 2008 17:35:45 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1755161AbYAXWfd (ORCPT ); Thu, 24 Jan 2008 17:35:33 -0500 Received: from gw.goop.org ([64.81.55.164]:53130 "EHLO mail.goop.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1754994AbYAXWfb (ORCPT ); Thu, 24 Jan 2008 17:35:31 -0500 Message-ID: <479912B0.2080506@goop.org> Date: Thu, 24 Jan 2008 14:35:28 -0800 From: Jeremy Fitzhardinge User-Agent: Thunderbird 2.0.0.9 (X11/20071115) MIME-Version: 1.0 To: "H. Peter Anvin" CC: Ian Campbell , Ingo Molnar , =?UTF-8?B?TWlrYSBQZW50dGlsw6Q=?= , linux-kernel@vger.kernel.org, Thomas Gleixner , Ingo Molnar , "Eric W. Biederman" Subject: Re: [PATCH] x86: Construct 32 bit boot time page tables in native format. References: <1200758937-22386-2-git-send-email-ijc@hellion.org.uk> <47939363.8040603@kolumbus.fi> <1200950633.15491.21.camel@cthulhu.hellion.org.uk> <479510CE.7010706@zytor.com> <1200951996.15491.28.camel@cthulhu.hellion.org.uk> <479551FD.5040801@zytor.com> <1201023401.5643.52.camel@localhost.localdomain> <479634A9.1090908@zytor.com> <1201031325.15491.66.camel@cthulhu.hellion.org.uk> <47964B7B.9000905@zytor.com> <20080122203611.GA12892@elte.hu> <1201121530.30671.15.camel@cthulhu.hellion.org.uk> <4797E487.8010108@goop.org> <1201167544.30671.64.camel@cthulhu.hellion.org.uk> <47990BDC.4070008@zytor.com> In-Reply-To: <47990BDC.4070008@zytor.com> X-Enigmail-Version: 0.95.6 Content-Type: text/plain; charset=UTF-8; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org H. Peter Anvin wrote: > In other words, reusing the early page tables isn't all that > straightforward. It may easily be that it's better to build a new set > of page tables from scratch, however, it would *still* be beneficial > to have the early page tables be in the same format as the later one, > since it lets us use the fixmap area, and therefore > {bt,early}_ioremap() much sooner. Yes, and it simplifies Xen as it always starts guest domains in the appropriate pagetable mode and doesn't let the guest change it on the fly. If early_ioremap depends on non-PAE early pagetables in an otherwise PAE kernel, we'd need to go to some effort to make sure all the early_ioremap stuff is skipped (which would be possible but unpleasant for domU, but very bad in dom0). J