From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Thu, 6 Sep 2001 12:39:46 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Thu, 6 Sep 2001 12:39:36 -0400 Received: from humbolt.nl.linux.org ([131.211.28.48]:262 "EHLO humbolt.nl.linux.org") by vger.kernel.org with ESMTP id ; Thu, 6 Sep 2001 12:39:19 -0400 Content-Type: text/plain; charset=US-ASCII From: Daniel Phillips To: Kurt Garloff , Rik van Riel Subject: Re: page_launder() on 2.4.9/10 issue Date: Thu, 6 Sep 2001 18:45:52 +0200 X-Mailer: KMail [version 1.3.1] Cc: Jan Harkes , Marcelo Tosatti , linux-kernel@vger.kernel.org In-Reply-To: <20010906124700Z16067-32383+3773@humbolt.nl.linux.org> <20010906151827.F21793@gum01m.etpnet.phys.tue.nl> In-Reply-To: <20010906151827.F21793@gum01m.etpnet.phys.tue.nl> MIME-Version: 1.0 Content-Transfer-Encoding: 7BIT Message-Id: <20010906163937Z16125-26183+11@humbolt.nl.linux.org> Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On September 6, 2001 03:18 pm, Kurt Garloff wrote: > On Thu, Sep 06, 2001 at 10:03:03AM -0300, Rik van Riel wrote: > > On Thu, 6 Sep 2001, Daniel Phillips wrote: > > > On September 6, 2001 02:32 pm, Rik van Riel wrote: > > > > Two words: "IO clustering". > > > > > > Yes, *after* the IO queue is fully loaded that makes sense. Leaving it > > > partly or fully idle while waiting for it to load up makes no sense at all. > > > > > > IO clustering will happen naturally after the queue loads up. > > > > Exactly, so we need to give the queue some time to load > > up, right ? > > Just use two limits: > * Time: After some time (say two seconds), we can always afford to write it > out > * Queue filling: After it has a certain size, it's worth doing a writing. Err, not quite the whole story. It is *never* right to leave the disk sitting idle while there are dirty, writable IO buffers. -- Daniel