From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S261771AbVBPHEs (ORCPT ); Wed, 16 Feb 2005 02:04:48 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S261818AbVBPHEs (ORCPT ); Wed, 16 Feb 2005 02:04:48 -0500 Received: from mx2.elte.hu ([157.181.151.9]:24741 "EHLO mx2.elte.hu") by vger.kernel.org with ESMTP id S261771AbVBPHEp (ORCPT ); Wed, 16 Feb 2005 02:04:45 -0500 Date: Wed, 16 Feb 2005 08:04:37 +0100 From: Ingo Molnar To: Andrew Morton Cc: Christoph Hellwig , agk@redhat.com, linux-kernel@vger.kernel.org Subject: Re: [PATCH] device-mapper: multipath Message-ID: <20050216070437.GA18912@elte.hu> References: <20050211171506.GX10195@agk.surrey.redhat.com> <20050211173143.GA11278@infradead.org> <20050211133632.2277fed9.akpm@osdl.org> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <20050211133632.2277fed9.akpm@osdl.org> User-Agent: Mutt/1.4.1i X-ELTE-SpamVersion: MailScanner 4.31.6-itk1 (ELTE 1.2) SpamAssassin 2.63 ClamAV 0.73 X-ELTE-VirusStatus: clean X-ELTE-SpamCheck: no X-ELTE-SpamCheck-Details: score=-4.9, required 5.9, autolearn=not spam, BAYES_00 -4.90 X-ELTE-SpamLevel: X-ELTE-SpamScore: -4 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org * Andrew Morton wrote: > Christoph Hellwig wrote: > > > > > +EXPORT_SYMBOL(dm_register_path_selector); > > > +EXPORT_SYMBOL(dm_unregister_path_selector); > > > > I though we agreed to only allow GPL'ed path selectors at OSDL? > > (OSDL?) > > Yup, this should be _GPL. Anything which uses these exports is a > derived work, isn't it? i'd not say it that categorically. I'd rather say that any module which uses these exports gains access to a wide range of GPL-licensed internal functionality of the kernel under the condition that the module declares that it is license-compatible with the GPL. While the use of such functionality very likely means that the module is derived work, even if it's not derived work (e.g. consider the following absurd corner-case: it is a short, trivial binary blob that was brute-force generated and blackbox tested to be a kernel module that happens to load fine) the module still has to follow the rules and must not circumvent the technological protection measure. (this fine distinction may or may not matter to you legally or otherwise, depending on your geographical coordinates and other factors.) Ingo