Re: [Ksummit-discuss] [CORE TOPIC] GPL defense issues

[Dave Airlie <airlied@gmail.com>]

On 25 August 2016 at 17:03, Josh Triplett <josh@joshtriplett.org> wrote:
> On Thu, Aug 25, 2016 at 02:37:07AM -0400, Theodore Ts'o wrote:
>> On Wed, Aug 24, 2016 at 09:06:19PM -0700, Bradley M. Kuhn wrote:
>> > I work for an organization that holds copyrights in Linux, and Conservancy
>> > furthermore coordinates a coalition of developers who signed agreements
>> > asking us to enforce their copyrights.  We also have embedded device users
>> > writing us weekly asking us to please get the Linux sources for their
>> > devices.  We have a huge mandate, and we're going to enforce (always adhering
>> > to the Principles of Community-Oriented Enforcement, of course).
>>
>> Bradley,
>>
>> If the Conservancy is going to do what it's going to do, it's not
>> clear what's the purpose of having a discussion.
>>
>> Having a debate implies that there is a question on the floor, and to
>> date it's not clear what the question or questions would be.  Yes,
>> it's clear what you and Karen have proposed as the *topic*, and the
>> Conservancy's positions aren't really a secret --- but that's not a
>> debatable question; it's not a specific proposal that could be
>> examined.
>
> A few possibilities:
>
> - What's the right way to handle enforcement?  (The most general version
>   of this question has no hope of consensus, but I suspect we could
>   fairly easily reach consensus about what we don't want to see people
>   doing, for instance.)
>
> - What (if anything) do we do in response to the "wrong" way to handle
>   enforcement?  Would we ever have a response at the kernel level,
>   rather than just within a maintainer team (as happened with
>   netfilter)?  Is there a line where we'd stop taking patches from
>   someone because they've created problems in the legal arena?  Where is
>   that line?
>
> - Do any of our current development procedures make enforcement easier
>   or harder?
>
> - Could we incorporate anything into our development procedures that
>   improves the situation, analogous to the DCO?  (Not suggesting a live
>   brainstorming session, but rather suggesting that if a concrete
>   proposal exists for such a change, Kernel Summit seems like the place
>   to discuss it.)
>
> - EXPORT_SYMBOL_GPL and MODULE_LICENSE.  I know that some history exists
>   suggesting that they've helped with some cases.  Do they potentially
>   make enforcement more difficult, though?  (This also includes informal
>   or internal enforcement/compliance efforts; for instance,
>   EXPORT_SYMBOL_GPL may provide a supporting argument for such efforts,
>   but conversely EXPORT_SYMBOL may then hinder such efforts.)

Oh I really wish someone could answer that. I hate _GPL as I believe it makes
it legal to create GPL modules with MODULE_LICENSE, when really they
violate the license just as much if not more. If something isn't a derivate work
for MODULE_LICENSE, but using one _GPL work somehow makes it one,
is insane.

Although I can't attend this year, I think there should be less stop energy from
the senior developers, and more discussion in the only place it's probably
safe to discuss this, in a closed room with no press, I'm nearly sure this
is one of the reasons KS existed in the first place to have
unreportable discussions.

Everyone who stays in the room can later drink heavily and deny any knowledge
of having been there.

Dave.