From mboxrd@z Thu Jan 1 00:00:00 1970 Received: from however ([206.171.168.138]) (authenticated user joelkatz@webmaster.com) by webmaster.com (mail1.webmaster.com [216.152.64.168]) (MDaemon.PRO.v6.8.5.R) with ESMTP id 64-md50000000029.tmp for ; Thu, 19 Feb 2004 17:06:15 -0800 From: "David Schwartz" Subject: RE: Non-GPL export of invalidate_mmap_range Date: Thu, 19 Feb 2004 17:27:28 -0800 Message-ID: MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: 7bit In-Reply-To: <20040218162858.2a230401.akpm@osdl.org> Reply-To: davids@webmaster.com Sender: owner-linux-mm@kvack.org Return-Path: To: Christoph Hellwig Cc: paulmck@us.ibm.com, arjanv@redhat.com, linux-mm@kvack.org List-ID: > Christoph Hellwig wrote: > And the "But when you distribute..." part is what the Linus doctrine rubs > out. Because it is unreasonable to say that a large piece of work such as > this is "derived" from Linux. I think you misunderstand how the Linux kernel uses the term "derive". By a "derived work", the GPL is invoking the legal copyright principle of a "derivative work". You can google this term to get a better understanding of it. The term "derived work" does not imply that the work is wholly derived. Rather, it means that some part of the protected expression of the original work is present in the work. In the specific case of Linux kernel modules, the question is whether some part of the protectable expression in the Linkx kernel is present in the module. This is a major issue for compiled modules distributed in object form because the compilation process, through header files, puts pieces of the header files in the resultant object. If the distributed work is in source code form, however, the argument becomes much different. You are not likely to find pieces of the kernel code present in the source code that's distributed. However, one possible argument is that the module is a "sequel" to the kernel. It takes the framework the kernel creates and builds on it. I can't write and sell a Star Trek novel for just this reason, it would be derived from previous such novels because it borrows their universe. Another possible argument is that the module code is so intertwined with kernel code that you can't consider the module by itself a work at all. In the present case, we have a shim that is distributed in source form. The main module works with other operating systems and doesn't contain much Linux-specific code. So the module itself is not a derived work of Linux. The shim is probably a derived work, but the shim is open source. So if there's a license issue, I don't know what it is. DS -- To unsubscribe, send a message with 'unsubscribe linux-mm' in the body to majordomo@kvack.org. For more info on Linux MM, see: http://www.linux-mm.org/ . Don't email: aart@kvack.org