From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1755748AbYJ3WCm (ORCPT ); Thu, 30 Oct 2008 18:02:42 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1753166AbYJ3WCg (ORCPT ); Thu, 30 Oct 2008 18:02:36 -0400 Received: from uni13mr.unity.ncsu.edu ([152.1.224.171]:34697 "EHLO uni13mr.unity.ncsu.edu" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1753022AbYJ3WCg (ORCPT ); Thu, 30 Oct 2008 18:02:36 -0400 X-Greylist: delayed 755 seconds by postgrey-1.27 at vger.kernel.org; Thu, 30 Oct 2008 18:02:35 EDT Message-ID: <1819.152.14.92.134.1225403398.squirrel@webmail.ncsu.edu> Date: Thu, 30 Oct 2008 17:49:58 -0400 (EDT) Subject: Question about page frame management in x86_64 From: "Emre Can Sezer" To: linux-kernel@vger.kernel.org Cc: ecsezer@ncsu.edu User-Agent: SquirrelMail/1.4.10a MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7BIT X-Priority: 3 (Normal) Importance: Normal X-PMX-Version: 5.3.3.310218, Antispam-Engine: 2.5.2.313940, Antispam-Data: 2008.10.30.213114 Sender: linux-kernel-owner@vger.kernel.org List-ID: X-Mailing-List: linux-kernel@vger.kernel.org Hi, For my research I'm trying to keep track of which pages belong to an LKM. I use linux-2.6.26. When the module is first loaded, the virtual memory region (ffff ffff a000 0000) maps to a set of page frames, which I track. Later on (after all the modules have loaded), these same pages are returned by get_free_pages(), before they are released. At this point, the page table entries have also changed. I observe this for all modules. Since the modules haven't exited, I can only assume that the pages are relocated. However, neither free_more_memory() is called nor balance_pgdat() reclaims any pages. They aren't freed via any of the __free_pages() family of functions either. What could be relocating these pages and which functions would be responsible for it? Thanks in advance, John PS. I'm not subscribed to the mailing list, so please CC to me. Thanks.