From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S932113AbWFGJrK (ORCPT ); Wed, 7 Jun 2006 05:47:10 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S932114AbWFGJrJ (ORCPT ); Wed, 7 Jun 2006 05:47:09 -0400 Received: from ns.suse.de ([195.135.220.2]:48530 "EHLO mx1.suse.de") by vger.kernel.org with ESMTP id S932113AbWFGJrI (ORCPT ); Wed, 7 Jun 2006 05:47:08 -0400 From: Andi Kleen To: Mel Gorman Subject: Re: [PATCH 0/5] Sizing zones and holes in an architecture independent manner V7 Date: Wed, 7 Jun 2006 11:45:04 +0200 User-Agent: KMail/1.9.3 Cc: Andrew Morton , davej@codemonkey.org.uk, tony.luck@intel.com, bob.picco@hp.com, linux-kernel@vger.kernel.org, linuxppc-dev@ozlabs.org, linux-mm@kvack.org References: <20060606134710.21419.48239.sendpatchset@skynet.skynet.ie> <20060606164311.27d4af98.akpm@osdl.org> In-Reply-To: MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit Content-Disposition: inline Message-Id: <200606071145.04938.ak@suse.de> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org > Spanned pages and holes will be different on > x86_64 because I don't account the kernel image and memmap as holes. That's a significant inaccuracy and may give worse VM results. -Andi