From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261336AbVHAXYW (ORCPT ); Mon, 1 Aug 2005 19:24:22 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261332AbVHAXYW (ORCPT ); Mon, 1 Aug 2005 19:24:22 -0400 Received: from graphe.net ([209.204.138.32]:20924 "EHLO graphe.net") by vger.kernel.org with ESMTP id S261336AbVHAXYU (ORCPT ); Mon, 1 Aug 2005 19:24:20 -0400 Date: Mon, 1 Aug 2005 16:24:19 -0700 (PDT) From: Christoph Lameter X-X-Sender: christoph@graphe.net To: Paul Jackson cc: linux-kernel@vger.kernel.org, akpm@osdl.org, ak@suse.de Subject: Re: [PATCH] String conversions for memory policy In-Reply-To: <20050801160351.71ee630a.pj@sgi.com> Message-ID: References: <20050729152049.4b172d78.pj@sgi.com> <20050729230026.1aa27e14.pj@sgi.com> <20050730181418.65caed1f.pj@sgi.com> <20050730190126.6bec9186.pj@sgi.com> <20050730191228.15b71533.pj@sgi.com> <20050801160351.71ee630a.pj@sgi.com> MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Spam-Score: -5.8 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Mon, 1 Aug 2005, Paul Jackson wrote: > In your related patch, you show how to merge this display of mempolicy > into the new /proc//smap (rss size of each memory area) file. > But this smap file (or, as you renamed it, emap file) is read-only, > so far as I can tell. It enables display of information, but not > changing it. How do you propose to support changing a memory policy? That is not clear yet. We need to discuss that. First I thought of just having a patch that creates /proc//numa_policy which allows to read and write the process policy (write via notifier not directly). > that other half should not also be used to display memory policies, > instead of adapting smap aka emap. e/smap is a generic way to get information about storage use of a vma. Therefore displaying numa node information etc there makes a lot of sense. > Would it be better to have two files, the first of which has one of > the strings: Yes. I initially had something like that in mind and have a partial implementation but such an approach gets extremely complicated and difficult to handle since you need to create multiple levels of directories in /proc//xx. It is easier to obtain the complete memory information from a single file. We already have that in /proc//maps.