From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S266088AbUBCUpE (ORCPT ); Tue, 3 Feb 2004 15:45:04 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S266100AbUBCUpE (ORCPT ); Tue, 3 Feb 2004 15:45:04 -0500 Received: from kinesis.swishmail.com ([209.10.110.86]:36360 "EHLO kinesis.swishmail.com") by vger.kernel.org with ESMTP id S266088AbUBCUpA (ORCPT ); Tue, 3 Feb 2004 15:45:00 -0500 Message-ID: <4020096B.7000408@techsource.com> Date: Tue, 03 Feb 2004 15:49:47 -0500 From: Timothy Miller MIME-Version: 1.0 To: Linux Kernel Mailing List Subject: Auto-regulated swappiness Content-Type: text/plain; charset=us-ascii; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org I just noticed the kerneltrap article about Con's new patchset. In particular, I am curious about the auto-regulated swappiness. I've done a little searching through the archives, but I can't seem to find the lkml posts I'm thinking about. In any event, I vaguely remember two things: - auto-regulation was developed because the kernel seemed to swap too much or too little under certain circumstances. - Someone said that auto-regulating swappiness didn't make sense, because there was some constant value that should have had the desired effect. I don't remember there being a resolution to this discussion. For my own curiosity, what happens if swappiness is too high but there isn't any pressure to swap from memory usage? Do user pages get swapped out in favor of making room for potential buffer pages? What happens if it's too low and there's lots of pressure to swap? How does the auto-regulator fix this? Thanks.