From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S262965AbVG3Nuy (ORCPT ); Sat, 30 Jul 2005 09:50:54 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S262966AbVG3Nux (ORCPT ); Sat, 30 Jul 2005 09:50:53 -0400 Received: from dvhart.com ([64.146.134.43]:37306 "EHLO localhost.localdomain") by vger.kernel.org with ESMTP id S262965AbVG3Ntn (ORCPT ); Sat, 30 Jul 2005 09:49:43 -0400 Date: Sat, 30 Jul 2005 06:49:44 -0700 From: "Martin J. Bligh" Reply-To: "Martin J. Bligh" To: Andrew Morton Cc: linux-kernel@vger.kernel.org, colpatch@us.ibm.com Subject: Re: [PATCH] Fix NUMA node sizing in nr_free_zone_pages Message-ID: <259590000.1122731383@[10.10.2.4]> In-Reply-To: <20050729224043.7ce56d4e.akpm@osdl.org> References: <240970000.1122661910@[10.10.2.4]> <20050729224043.7ce56d4e.akpm@osdl.org> X-Mailer: Mulberry/2.2.1 (Linux/x86) MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Content-Disposition: inline Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org >> We are iterating over all nodes in nr_free_zone_pages(). Because the >> fallback zonelists contain all nodes in the system, and we walk all >> the zonelists, we're counting memory multiple times (once for each >> node). This caused us to make a size estimate of 32GB for an 8GB >> AMD64 box, which makes all the dirty ratio calculations, etc incorrect. >> >> There's still a further bug to fix from e820 holes causing overestimation >> as well, but this fix is separate, and good as is, and fixes one class >> of problems. Problem found by Badari, and tested by Ram Pai - thanks! > > Alas my non-NUMA EMT64 box still gets it wrong. > > nr_free_pagecache_pages() is still returning 1572864 on a 4G box. Yeah, it will do - is the e820 bug I mentioned above. Patch is half-written, will finish it off ASAP, but I'll be out today ;-( M.