From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261970AbTFFQ3E (ORCPT ); Fri, 6 Jun 2003 12:29:04 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261989AbTFFQ3E (ORCPT ); Fri, 6 Jun 2003 12:29:04 -0400 Received: from dbl.q-ag.de ([80.146.160.66]:24521 "EHLO dbl.q-ag.de") by vger.kernel.org with ESMTP id S261970AbTFFQ3C (ORCPT ); Fri, 6 Jun 2003 12:29:02 -0400 Message-ID: <3EE0C45B.4070002@colorfullife.com> Date: Fri, 06 Jun 2003 18:42:03 +0200 From: Manfred Spraul User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.3) Gecko/20030313 X-Accept-Language: en-us, en MIME-Version: 1.0 To: ameya.mitragotri@wipro.com CC: wli@holomorphy.com, linux-kernel@vger.kernel.org, davej@suse.de, indou.takao@jp.fujitsu.com, akpm@diego.com Subject: Re: [RFC][PATCH 2.5.70] dynamically tunable semmnu and semume References: <200306061141.50965.ameya.mitragotri@wipro.com> In-Reply-To: <200306061141.50965.ameya.mitragotri@wipro.com> Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org Ameya Mitragotri wrote: >Find below the patch (RFC) that makes semmnu and semume dynamically >tunable through /proc. > AFAICS the patch implements semmnu and semume, correct? The current implemention permits an unlimited number of undo structures. 1) Why do you want to introduce a limit? Have you found a potential DoS? 2) What is counted by semume? Individual undo entries, or undo arrays? Linux differs from the normal unix sysv implementation: there is a per process, per semaphore set array of undo entries. -- Manfred