From mboxrd@z Thu Jan 1 00:00:00 1970 From: ebiederm@xmission.com (Eric W. Biederman) Subject: Re: [PATCH 0/9] use network namespace for iSCSI control interfaces Date: Tue, 21 Nov 2017 13:46:03 -0600 Message-ID: <87mv3fo1ok.fsf@xmission.com> References: <20171107180156.GD29597@straylight.hirudinean.org> <20171107224513.4217-1-cleech@redhat.com> <063D6719AE5E284EB5DD2968C1650D6DD00B618E@AcuExch.aculab.com> <20171115002521.GA21082@straylight.hirudinean.org> <687d0196888f4325aebc0989a8e12ced@AcuMS.aculab.com> Mime-Version: 1.0 Content-Type: text/plain Cc: 'Chris Leech' , "netdev\@vger.kernel.org" , "containers\@lists.linux-foundation.org" To: David Laight Return-path: Received: from out01.mta.xmission.com ([166.70.13.231]:42750 "EHLO out01.mta.xmission.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750794AbdKUTqX (ORCPT ); Tue, 21 Nov 2017 14:46:23 -0500 In-Reply-To: <687d0196888f4325aebc0989a8e12ced@AcuMS.aculab.com> (David Laight's message of "Tue, 21 Nov 2017 11:26:09 +0000") Sender: netdev-owner@vger.kernel.org List-ID: David Laight writes: > To make matters even more annoying the functions for holding and > releasing a namespace are GPL_ONLY :-( I am going to pick on this by itself for a moment without mentioning anything else, so as hopefully not to derail what otherwise sounds like a good technical conversation. So far every time when someone has complained to me about things being GPL_ONLY and I have looked into it, all I have seen is someone trying to come up with a way to release derivative works of the kernel without honoring the terms of the GPLv2. I read through the US Code a while back to see if I could understand what is legaly defined as a derivative work, and my impression at the time was that the FSF is quite conservative in what they consider a derivative work, and probably the scope is much wider. So when people start complaining about things being GPLv2 those are the most annoying bug reports I ever deal with, as almost invariably people just want to take from the community and don't want to work with every one else. It is especially annoying because I have never seen a case where there is a good justification for a kernel export being anything other than GPL_ONLY. That is the kernel's license after all, and if you are using kernel internal functions the chance that your code is not a derivative work is about 0. Eric