From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S964963AbWFHUVB (ORCPT ); Thu, 8 Jun 2006 16:21:01 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S964981AbWFHUVB (ORCPT ); Thu, 8 Jun 2006 16:21:01 -0400 Received: from smtp.osdl.org ([65.172.181.4]:60098 "EHLO smtp.osdl.org") by vger.kernel.org with ESMTP id S964963AbWFHUVA (ORCPT ); Thu, 8 Jun 2006 16:21:00 -0400 Date: Thu, 8 Jun 2006 13:20:35 -0700 (PDT) From: Linus Torvalds To: Nate Diller cc: Peter Zijlstra , Hugh Dickins , linux-mm@kvack.org, linux-kernel@vger.kernel.org, Andrew Morton , David Howells , Christoph Lameter , Martin Bligh , Nick Piggin Subject: Re: [PATCH] mm: tracking dirty pages -v6 In-Reply-To: <5c49b0ed0606081310q5771e8d1s55acef09b405922b@mail.gmail.com> Message-ID: References: <20060525135534.20941.91650.sendpatchset@lappy> <1149770654.4408.71.camel@lappy> <5c49b0ed0606081310q5771e8d1s55acef09b405922b@mail.gmail.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Thu, 8 Jun 2006, Nate Diller wrote: > > Does this mean that processes dirtying pages via mmap are now subject > to write throttling? That could dramatically change the performance > for tasks with a working set larger than 10% of memory. Exactly. Except it's not a "working set", it's a "dirty set". Which is the whole (and only) point of the whole patch. If you want to live on the edge, you can set the dirty_balance trigger to something much higher, it's entirely configurable if I remember correctly. Linus