From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1751073AbXCOVte (ORCPT ); Thu, 15 Mar 2007 17:49:34 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S1751322AbXCOVte (ORCPT ); Thu, 15 Mar 2007 17:49:34 -0400 Received: from mail.suse.de ([195.135.220.2]:60257 "EHLO mx1.suse.de" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751073AbXCOVtd (ORCPT ); Thu, 15 Mar 2007 17:49:33 -0400 Date: Thu, 15 Mar 2007 22:49:23 +0100 From: Andrea Arcangeli To: Andrew Morton Cc: Ashif Harji , dingxn@cse.ohio-state.edu, shaggy@linux.vnet.ibm.com, andi@rhlx01.fht-esslingen.de, linux-mm@kvack.org, npiggin@suse.de, jack@suse.cz, linux-kernel@vger.kernel.org Subject: Re: [PATCH] mm/filemap.c: unconditionally call mark_page_accessed Message-ID: <20070315214923.GE6687@v2.random> References: <20070312173500.GF23532@duck.suse.cz> <20070313185554.GA5105@duck.suse.cz> <1173905741.8763.36.camel@kleikamp.austin.ibm.com> <20070314213317.GA22234@rhlx01.hs-esslingen.de> <1173910138.8763.45.camel@kleikamp.austin.ibm.com> <45F8A301.90301@cse.ohio-state.edu> <20070315110735.287c8a23.akpm@linux-foundation.org> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20070315110735.287c8a23.akpm@linux-foundation.org> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Thu, Mar 15, 2007 at 11:07:35AM -0800, Andrew Morton wrote: > > On Thu, 15 Mar 2007 01:22:45 -0400 (EDT) Ashif Harji wrote: > > I still think the simple fix of removing the > > condition is the best approach, but I'm certainly open to alternatives. > > Yes, the problem of falsely activating pages when the file is read in small > hunks is worse than the problem which your patch fixes. Really? I would have expected all performance sensitive apps to read in >=PAGE_SIZE chunks. And if they don't because they split their dataset in blocks (like some database), it may not be so wrong to activate those pages that have two "hot" blocks more aggressively than those pages with a single hot block. So I've an hard time to advocate to prefer the current behavior, but certainly this can be "fixed" by caching the last_offset like others pointed out ;)