Discussion continues on the Linux Kernel mailing list about the legality and morality of re-licensing BSD/GPL dual-licensed code under only the GPL. Alan Cox replied to Theo de Raadt's comments suggesting he was encouraging people to break the law, "re-read my email and then apologize. I do question the .h files where they are BSD licence and no changes were made to the work. I also point out that the dual licence on that code appears to give permission to distribute under one of those licences by choice." In response to Theo's request that code be shared both ways rather than converted to a sole GPL, "that's about the first thing I would agree on - its somewhat rude and not something I personally would usually choose todo." He then cautioned that this was a limitation of the BSD license:

"If OpenBSD wants a world where code must be returned, but you can mix it with free code in a product in some fashion and do binary only releases then OpenBSD needs to fix its licencing. Not to GPL which is clearly not the BSD intention but to something which does what BSD wants rather than an academic research licence developed thirty odd years ago for the purpose of showing that US research funds were properly spent. Perhaps its time for BSD2 licencing?"

In another email, Alan noted, "where I've reused BSD code I've aways tried to contribute it back to the BSD people or share the knowledge (and the knowledge far more than the code mattered both ways for Mac68K systems)." He pointed to drivers/net/wan/syncppp.c as an example, "which is I think the only BSD derived bit of code of mine left in the kernel - and its quite specific what it says."

Al Viro offered a couple of recent quotes from Linux creator Linus Torvalds in which he talked about dual-licensing:

"This is no different from the fact that we have some drivers that are GPLv2/BSD licensed. Within the kernel, they are GPLv2. But on their own, you can choose to use them under the BSD license, make your changes to them, and release them commercially. And correct - I cannot (and neither can anybody else) then accept those *non*GPLv2 changes back."

"Actually, normally I *do* have such a trust. It's why I have no problem with drivers that are dual-GPL/BSD, and in fact, I've told people that I don't want them to turn them into GPL-only, because that is simply not polite."

Al referred to the recent Atheros patch that converted BSD/GPL dual-licensed files to just the GPL when he added, "that's much more relevant to shooting the patch in question down (and IMO it ought to be shot down) than references to legality."

The dual-licensed code in question was originally authored by Sam Leffler, "the code in question is my code. It has my copyright (modulo bits shared with onoe-san who was consulted on the switch from dual-bsd/gpl to bsd only in freebsd)." He went on to explain, "is there a point to all this nonsense? I dual-licensed the code so folks could adopt and use it however they saw fit. As I've said before I don't care what people do with the work I give away so long as they don't claim it's their own." To avoid any uncertaininty, he reitterated his views:

"I am speaking up as the author of the code that set the dual license in place. I have the definitive say and I have said that any of my code that is dual-licensed can be made gpl only."

Other files that had been re-licensed were originally only available under the BSD license and written by Reyk Floeter, "I just returned from vacation where I was offline for about two weeks. So I totally missed the incidence and all the surrounding discussion." Regarding the files he had written he clarified, "I will not release or agree to release my code under either the GPL or any kind of a 'dual'-license." He added, "this is eating our time. Every few weeks I get a new discussion about licensing of the atheros driver etc. blah blah. Why can't they just accept the license as it is and focus on more important things?".

From: Alan Cox [email blocked]
Subject: Re: Fwd: That whole "Linux stealing our code" thing
Date:	Sat, 1 Sep 2007 21:30:52 +0100
 Windsor, Berkshire, SL4 1TE, Y Deyrnas Gyfunol. Cofrestrwyd yng Nghymru a
 Lloegr o'r rhif cofrestru 3798903

> It is illegal to modify a license unless you are the owner/author,
> because it is a legal document.  If there are multiple owners/authors,

Oh dear - Theo, go talk to a lawyer, or do a course on licencing.

The owner generally starts with the rights to control who performs acts
covered by copyright law. They pass some of those rights on to others by
contract, licence or statutory means. It is quite normal for the owner to
pass on the right to relicence or modify the licencing of a work. In many
cases the owner actually hands on all such rights to a third party
(eg an evil music company).

[Owner and author often differ as many legal systems start from the basis
that an employee produces a work for the employer rather than it being
transferred solely by contract]

The ath5k C file in question (not the headers) seems to give recipients
permission to further convey the work under a choice of two licences. It
doesn't say they must redistribute under both. So I appear to have a
right to convey the work under the GPL to a third party, who from me
receives no right to use it except under the GPL. 

The Ath5K C code is very clear about the intention of the licencing:

 * Alternatively, this software may be distributed under the terms of the
 * GNU General Public License ("GPL") version 2 as published by the Free
 * Software Foundation.

The choice appears to be delegated to the recipient very clearly and
very specifically by the licencing on the file. It does not say that I
must convey the work under both licences. It quite specifically says I may
convey the work under whichever of the two I prefer (and probably both if
I wish). Clearly if that had not been the intent it would not have
included the clause giving the choice.

This is quite different to the case Theo tries to discuss.

> In http://lkml.org/lkml/2007/8/29/183, Alan Cox managed to summarize
> what Jiri Slaby and Luis Rodriguez were trying to do by proposing a
> modification of a Dual Licenced file without the consent of all the
> authors.  Alan asks "So whats the problem ?".  Well, Alan, I must
> caution you -- your post is advising people to break the law.

Re-read my email and then apologize. I do question the .h files where
they are BSD licence and no changes were made to the work. I also point
out that the dual licence on that code appears to give permission to
distribute under one of those licences by choice.

> - If you receive dual licensed code, you may not delete the license
>   you don't like and then distribute it.  It has to stay, because you
>   may not edit someone's else's license -- which is a three-part legal
>   document (For instance: Copyright notice, BSD, followed by GPL).

If you got BSD licenced code that doesn't give you a choice of licence
then of course the original work remains BSD and you can't go around
removing that information, the copyright holder's name and other things
protected variously by different legal systems. I would submit the ath5k
header files fit this (if they are even copyrightable works at all)

> - If you wish for everyone to remain friends, you should give code back.

That's about the first thing I would agree on - its somewhat rude and
not something I personally would usually choose todo.  However to many
there are problems as the BSD licence doesn't mean giving it back to the
community it means giving a copy to everyone who wants rip it off for
private proprietary use.

>      "Thanks for what you wrote, but this is a one-way street, you give
>      us code, and we take it, we give you you nothing back.  screw off."

The BSD licence allows this. If you in the BSD world don't want that to
happen you need to look hard at your licencing. Linux takes very little
from the BSD world this way, the big one way takers are all proprietary
software companies. Perhaps Theo should write to that nice Bill guy
instead.

> GPL fans said the great problem we would face is that companies would
> take our BSD code, modify it, and not give back.  Nope -- the great
> problem we face is that people would wrap the GPL around our code, and

You just don't realise who takes your code and what they do with it. The
proprietary people don't tell you, but the free ones you can see.

> If the Linux developers wrap GPL's around things we worked very hard
> on, it will definately not be viewed as community development.

So you'd prefer that the Linux developers worked on it and then Microsoft
took the results of all our work and didn't give anything back. At least
if the Linux work is GPL licenced its protected from further abuse. See
the viewpoint the free software people come from - you may not agree with
it but it has a logic.

If OpenBSD wants a world where code must be returned, but you can mix it
with free code in a product in some fashion and do binary only releases
then OpenBSD needs to fix its licencing. Not to GPL which is clearly not
the BSD intention but to something which does what BSD wants rather than
an academic research licence developed thirty odd years ago for the
purpose of showing that US research funds were properly spent. Perhaps
its time for BSD2 licencing ?

Alan



From: David Schwartz [email blocked]
Reply-To: [email blocked]
Subject: RE: Fwd: That whole "Linux stealing our code" thing
Date:	Sun, 2 Sep 2007 19:01:18 -0700


Alan Cox wrote:

> The ath5k C file in question (not the headers) seems to give recipients
> permission to further convey the work under a choice of two licences.

Correct.

> It doesn't say they must redistribute under both.

Correct. They need the right to redistribute the work, and they may obtain
that right from either license.

> So I appear to have a
> right to convey the work under the GPL to a third party, who from me
> receives no right to use it except under the GPL.

Here's where your train goes off the rails. They do not receive any right to
use it from you. They receive a license to use it under the GPL from the
original author. Please read GPL section 6.

"  6. Each time you redistribute the Program (or any work based on the
Program), the recipient automatically receives a license from the
original licensor to copy, distribute or modify the Program subject to
these terms and conditions.  You may not impose any further
restrictions on the recipients' exercise of the rights granted herein.
You are not responsible for enforcing compliance by third parties to
this License."

The GPL does not give you *any* right to extend anyone a license to code you
did not author. (Nor can it as such an extension would have to be done in
writing in most countries.) When you distribute a GPL'd work, the right to
use every creative element in that work is licensed to the recipients
directly from their respective authors. Under no circumstances does the GPL
ever give you the ability to license someone else's work to a third party.

>  * Alternatively, this software may be distributed under the terms of the
>  * GNU General Public License ("GPL") version 2 as published by the Free
>  * Software Foundation.
>
> The choice appears to be delegated to the recipient very clearly and
> very specifically by the licencing on the file. It does not say that I
> must convey the work under both licences. It quite specifically says I may
> convey the work under whichever of the two I prefer (and probably both if
> I wish). Clearly if that had not been the intent it would not have
> included the clause giving the choice.

Either license can grant you the right to distribute it, but how you get the
rights to distribute has *NO* effect on the recipient. They receive a lawful
copy and any rights the original author grants them under a license from
that original author. You have no power to grant or modify rights to the
original work.

This is a common misunderstanding.

Note that you may remove the text of either license from a dual-licensed
file and redistribute under the other license because neither license
requires you to retain the other license and both licenses give you the
right otherwise to modify as you wish. But the removal of a license from a
file has no effect on the grant of license. Your recipients still get a dual
license to those protectable elements in the file that were placed under a
dual license. You cannot stop the automatic grant.

DS



From:	Jeff Garzik [email blocked]
Subject: Re: Fwd: That whole "Linux stealing our code" thing
Date:	Mon, 03 Sep 2007 04:58:58 -0400

David Schwartz wrote:
> Either license can grant you the right to distribute it, but how you get the
> rights to distribute has *NO* effect on the recipient. They receive a lawful
> copy and any rights the original author grants them under a license from
> that original author. You have no power to grant or modify rights to the
> original work.

Secondary parties have the power to grant or modify rights, if delegated 
to them by the original author.

Relicensing and transfer of rights happens all the time.  How do you 
think most music gets into consumer hands?

	Jeff



From:	David Schwartz [email blocked]
Reply-To: [email blocked]
Subject: Re: Fwd: That whole "Linux stealing our code" thing
Date:	Mon, 03 Sep 2007 02:21:10 -0700


Jeff Garzik wrote:

> Secondary parties have the power to grant or modify rights, if
> delegated 
> to them by the original author.

Yes, but this has to be done in writing and neither the BSD nor the GPL
license claim to allow this.
 
> Relicensing and transfer of rights happens all the time.  How do you 
> think most music gets into consumer hands?

Through written agreements to permit this. Neither the BSD nor the GPL
license do. You cannot revoke someone's GPL rights to code you did not
write because they flow directly from the original author to the recipient.

The GPL is quite clear, you do not grant a license to anyone else's code,
the original author does. This is why the GPL does not need to be in writing.

DS



From:	Jeff Garzik [email blocked]
Subject: Re: Fwd: That whole "Linux stealing our code" thing
Date:	Mon, 03 Sep 2007 05:47:13 -0400

David Schwartz wrote:
> Jeff Garzik wrote:
> 
>> Secondary parties have the power to grant or modify rights, if
>> delegated 
>> to them by the original author.
> 
> Yes, but this has to be done in writing and neither the BSD nor the GPL
> license claim to allow this.

Standard dual license texts do.

	Jeff

From:	David Schwartz [email blocked]
Reply-To: [email blocked]
Subject: RE: Fwd: That whole "Linux stealing our code" thing
Date:	Mon, 3 Sep 2007 02:54:28 -0700


> > Yes, but this has to be done in writing and neither the BSD nor the GPL
> > license claim to allow this.

> Standard dual license texts do.
>
> 	Jeff

No, they don't. They simply state that *you* may obtain the right to
modify/distribute the work from either license at your option. They do not
say anything about relicensing. Again, such an agreement would have to be in
writing.

What license a distributor uses to acquire the rights to modify or
distribute a work has no effect whatsoever on the rights the recipient has
to the work. The GPL is absolutely clear about this (section 6).

DS



From:	Constantine A. Murenin [email blocked]
Subject: Re: Fwd: That whole "Linux stealing our code" thing
Date:	Sat, 1 Sep 2007 20:53:13 -0400

On 01/09/07, Adrian Bunk [email blocked] wrote:
> On Sat, Sep 01, 2007 at 07:29:39PM -0400, Constantine A. Murenin wrote:
> > On 01/09/07, Adrian Bunk [email blocked] wrote:
> > > On Sat, Sep 01, 2007 at 05:27:03PM -0400, Constantine A. Murenin wrote:
> > > > On 01/09/07, Adrian Bunk [email blocked] wrote:
> > > > > On Sat, Sep 01, 2007 at 01:37:18PM -0400, Constantine A. Murenin wrote:
> > > > > > On 01/09/07, Jeff Garzik [email blocked] wrote:
> > > > > > > Constantine A. Murenin wrote:
> > > > > > > > This will hopefully help diminish certain myths about the code licensing.
> > > > > > >
> > > > > > > What myth?  The myth that Theo understands dual licensing?
> > > > > >
> > > > > > Reyk's code was never dual licensed, so it's not like it even matters
> > > > > > to the original dispute.
> > > > >
> > > > > It's no longer dual licenced in the FreeBSD tree because the FreeBSD
> > > > > people removed the GPL choice of the dual licenced code 3 months ago.
> > > >
> > > > FreeBSD doesn't have Reyk's ath(4) HAL, which OpenHAL is based on.
> > > >
> > > > FreeBSD has a driver written by Sam, and a binary-only HAL, also written by Sam.
> > > >
> > > > > So all of Theo's accusations of people breaking the law by making this
> > > > > dual licenced code GPL-only apply as well to the FreeBSD people...
> > > >
> > > > How? FreeBSD doesn't have Reyk's ath(4) HAL from OpenBSD, so there are
> > > > no possible licensing accusations and violations.
> > >
> > > OK, I begin to understand this, there seem to be three different types
> > > of files changed by Jiri's patch:
> > > 1. dual licenced files planned to make GPL-only
> > > 2. previously dual licenced files with a too recent version used planned
> > >    to make GPL-only
> > > 3. never dual licenced files planned to make GPL-only
> > >
> > > For files under 1. and 2. Reyk did contribute to dual licenced code
> > > without touching the licence, but I missed that there's also code unter 3.
> > >
> > > So there is a problem, but not with the code under 1. (unless you plan
> > > to change the semantics of the word "alternatively"), the problem is
> > > with some headers under 2. plus the code under 3.
> > >
> > > It's funny how Theo missed the part of Jiri's patch that actually is a
> > > copyright violation and instead complains about the part that is OK...
> >
> > I'm not sure how you conclude that Theo missed the relevant parts --
> > there were many messages posted to [email blocked] mailing list and
> > to The OpenBSD Journal in the last few days, and to me it appears as
> > all of the problems were discussed ad nauseam.
> >...
>
> Then it's your fault that you forwarded the wrong email - in the email
> you forwarded the only action for which Theo accused the Linux
> developers of breaking the law was for choosing one licence when using
> dual licenced code.

For the sake of the discussion, at the time I forwarded the message,
the obvious licensing problems (e.g. the original infamous patch by
Jiri) were already addressed.

Personally, these problems were so obvious -- entirely changing the
licence under Reyk's Copyright notice -- that I didn't even take them
for real when they first came across.

BTW, I've now once again re-read the original message that I've
forwarded, and it does contain Theo's reiteration of his response that
the original re-licensing patch had clear violations. E.g. re-read
this part of his message:

- If you receive ISC or BSD licensed code, you may not delete the
 license.  Same principle, since the notice says so.  It's the law.
 Really.

> > After the obvious copyright violations were addressed, I think the
> > problem started being an ethical one.
> >
> > As a free software user and developer, the question I have is how come
> > the Linux community feels that they can take the BSD code that was
> > reverse-engineered at OpenBSD, and put a more restrictive licence onto
> > it, such that there will be no possibility of the changes going back
> > to OpenBSD, given that the main work on the code has happened at
> > OpenBSD? (Obviously, such a scenario it is permitted by the licence,
> > but my question is an ethical one -- after all, most components of
> > OpenHAL were specifically based on the OpenBSD's ath(4) HAL code.)
> >
> > You can see that Christoph Hellwig agrees with this ethical problem,
> > as in the message below.
>
> Is it a legal problem or is it "only" an ethical problem?

I don't particularly like to repeat myself -- after the obvious
licensing issues were addressed, it has indeed become an ethical
problem: why do you think that you as the Linux community has to act
ruder to the *BSD community than the supposed corporations that we
always hear about in the BSD/GPL licensing arguments?

I really like the response that Bob Beck gave on this question:
http://lkml.org/lkml/2007/9/1/197

> If choosing one licence when using dual licenced code is not a legal
> problem then Theo repeatedly talking about it would "break the law" in
> the email you forwarded was very unethical and the worst he could do
> for his cause.

My understanding is that with dual-licensed code, you choose to comply
with all of the terms of either licence. However, you cannot simply
remove either of these licences from the code, unless you specifically
receive such right from the copyright holder (remember, with the
copyright law, unless the rights are specifically given, they are
retained). This is what Theo was trying to educate the community on. I
don't see anything unethical in explaining the legal issues.

Constantine.



From: Alan Cox [email blocked]
Subject: Re: Fwd: That whole "Linux stealing our code" thing
Date:	Sun, 2 Sep 2007 11:36:38 +0100
 Windsor, Berkshire, SL4 1TE, Y Deyrnas Gyfunol. Cofrestrwyd yng Nghymru a
 Lloegr o'r rhif cofrestru 3798903

> - If you receive ISC or BSD licensed code, you may not delete the
>  license.  Same principle, since the notice says so.  It's the law.
>  Really.

You can shout this all you like but you would be wrong. You can remove
the licence if you have permission to do so. For the ath c files there
was permission to do so.

> My understanding is that with dual-licensed code, you choose to comply
> with all of the terms of either licence. However, you cannot simply
> remove either of these licences from the code, unless you specifically
> receive such right from the copyright holder (remember, with the
> copyright law, unless the rights are specifically given, they are
> retained). This is what Theo was trying to educate the community on. I
> don't see anything unethical in explaining the legal issues.

Your understanding isn't quite right. One of many things you may get with
dual licensed code is the right to pick a licence from several choices,
you may also get the right to remove some choices from the recipient.

A work that combines GPL and BSD licensed material is not the same as a
work which says I may choose between two licences. If both licences must
always apply (which is a perfectly possible condition to put in a
licence) then putting such a "both" GPL/BSD licence piece of code into
OpenBSD would require any OpenBSD distributed containing it was GPL
licenced when conveyed, which I am *very* sure is not the intent.

Thus what you appear to be doing by putting the ath5k C code in OpenBSD is
conveying it under the BSD licence (making a choice between the two
offered) and conveying a right for parties down the chain to convey it
under one of the licences only.

And as we've already established the header files are quite different.


Doesn't mean its not somewhat rude but illegal and rude are two very
different things.



From:	Igor Sobrado [email blocked]
Subject: Re: Fwd: That whole "Linux stealing our code" thing
Date:	Sun, 02 Sep 2007 13:20:27 +0200 (CEST)

On Sun, 2 Sep 2007, Alan Cox wrote:
> You can shout this all you like but you would be wrong. You can remove
> the licence if you have permission to do so. For the ath c files there
> was permission to do so.

There was permission to do so from Reyk Floeter?  Really?

> Your understanding isn't quite right. One of many things you may get with
> dual licensed code is the right to pick a licence from several choices,
> you may also get the right to remove some choices from the recipient.

Reyk code was never dual licensed!  His code is under truly free 
licensing terms (BSD).

> A work that combines GPL and BSD licensed material is not the same as a
> work which says I may choose between two licences. If both licences must
> always apply (which is a perfectly possible condition to put in a
> licence) then putting such a "both" GPL/BSD licence piece of code into
> OpenBSD would require any OpenBSD distributed containing it was GPL
> licenced when conveyed, which I am *very* sure is not the intent.
>
> Thus what you appear to be doing by putting the ath5k C code in OpenBSD is
> conveying it under the BSD licence (making a choice between the two
> offered) and conveying a right for parties down the chain to convey it
> under one of the licences only.

I think that Theo explained this point clearly quite a few times in the 
last days.

> And as we've already established the header files are quite different.

Is a simple change in the header files a reason to vindicate the people 
that changed the licensing terms?  Obviously, it isn't.

> Doesn't mean its not somewhat rude but illegal and rude are two very
> different things.

No, because this change is both rude and illegal.

Igor



From: Alan Cox [email blocked]
Subject: Re: Fwd: That whole "Linux stealing our code" thing
Date:	Sun, 2 Sep 2007 13:36:36 +0100
 Windsor, Berkshire, SL4 1TE, Y Deyrnas Gyfunol. Cofrestrwyd yng Nghymru a
 Lloegr o'r rhif cofrestru 3798903

On Sun, 02 Sep 2007 13:20:27 +0200 (CEST)
Igor Sobrado [email blocked] wrote:

> On Sun, 2 Sep 2007, Alan Cox wrote:
> > You can shout this all you like but you would be wrong. You can remove
> > the licence if you have permission to do so. For the ath c files there
> > was permission to do so.
> 
> There was permission to do so from Reyk Floeter?  Really?

The code pieces I quoted contained that choice. As far as I am concerned
that is what the discussion was about. 

Alan



From:	Igor Sobrado [email blocked]
Subject: Re: Fwd: That whole "Linux stealing our code" thing
Date:	Sun, 02 Sep 2007 14:28:28 +0200 (CEST)

On Sun, 2 Sep 2007, Adrian Bunk wrote:
> On Sun, Sep 02, 2007 at 01:20:27PM +0200, Igor Sobrado wrote:
>> Reyk code was never dual licensed!  His code is under truly free licensing
>> terms (BSD).
>
> Jiri's patch touched both files containing BSD-only code by Reyk and
> code Reyk contributed to leaving the file dual licenced.

Ok.

> You mixed two completely different things in your email:
>
> 1. Jiri's patch (that was never merged into Linux) not only removed the
>   BSD header from dual licenced files but also from not dual licenced
>   files.
>
> 2. Theo accused Alan that telling people that it was OK to choose one
>   licence for dual licenced code was "advising people to break the law".
>
> Jiri's patch was legally not OK regarding 1. - there's no discussion
> regarding this.
>
> The point 2 is what the email of Theo that was forwarded to linux-kernel
> is about and what the discussion is about. That's quite a rude action
> by Theo unless he's able to prove that this accusation is correct.

When code is multi-licensed it must be distributed under *all* these 
licensing terms concurrently.  It is easy to understand.  Removing (or 
changing) the conditions that apply to the program from the source code 
and documentation *without* an authorization from all the author(s) is 
illegal.

So, a multi-licensed file remains multi-licensed except when all authors 
agree about a change in the licensing terms.  And it is clear on the BSD 
license that a modification of the distribution terms is illegal.  It is 
the first clause on the BSD license:

  * 1. Redistributions of source code must retain the above copyright
  *    notice, this list of conditions and the following disclaimer,
  *    without modification.

So, removing (or changing) the list of conditions on the BSD license is 
not allowed.

Igor.



From: Alan Cox [email blocked]
Subject: Re: Fwd: That whole "Linux stealing our code" thing
Date:	Sun, 2 Sep 2007 13:46:12 +0100
 Windsor, Berkshire, SL4 1TE, Y Deyrnas Gyfunol. Cofrestrwyd yng Nghymru a
 Lloegr o'r rhif cofrestru 3798903

> So, a multi-licensed file remains multi-licensed except when all authors 
> agree about a change in the licensing terms.  And it is clear on the BSD 

Not strictly true. They can either agree to a change and issue one or
they can convey to other parties the right to change the terms. The GPL
for example does this for version selection.

A multi-licensed work (note work not file - don't assume a file is a
boundary of a work) which conveys the choice of licence (as some bits of
ath5k did) allows a receiving party to choose the licence it wishes.
Failing that OpenBSD would have turned itself GPL by adding that file as
according to your argument "it must be distributed under *all* these
licensing terms concurrently".

Alan



From:	Igor Sobrado [email blocked]
Subject: Re: Fwd: That whole "Linux stealing our code" thing
Date:	Sun, 02 Sep 2007 15:00:46 +0200 (CEST)

On Sun, 2 Sep 2007, Alan Cox wrote:

>> So, a multi-licensed file remains multi-licensed except when all authors
>> agree about a change in the licensing terms.  And it is clear on the BSD
>
> Not strictly true. They can either agree to a change and issue one or
> they can convey to other parties the right to change the terms. The GPL
> for example does this for version selection.

So, under a dual-licensed BSD/GPL code the latter license allows a 
developer to remove the GPL license itself and release a single-licensed 
BSD code if other parties want to do it?

> A multi-licensed work (note work not file - don't assume a file is a
> boundary of a work) which conveys the choice of licence (as some bits of
> ath5k did) allows a receiving party to choose the licence it wishes.
> Failing that OpenBSD would have turned itself GPL by adding that file as
> according to your argument "it must be distributed under *all* these
> licensing terms concurrently".

I would assume a file as a boundary of a work in the case that file is 
under different licensing terms to the rest of the software package.  On a 
lot of software packages different modules are covered under different 
licensing terms.

We can choose what license terms we will honor; however, we do not have 
the ability to remove the licensing terms we do not like.

Igor.



From: Alan Cox [email blocked]
Subject: Re: Fwd: That whole "Linux stealing our code" thing
Date:	Sun, 2 Sep 2007 14:53:51 +0100
 Windsor, Berkshire, SL4 1TE, Y Deyrnas Gyfunol. Cofrestrwyd yng Nghymru a
 Lloegr o'r rhif cofrestru 3798903

> > Not strictly true. They can either agree to a change and issue one or
> > they can convey to other parties the right to change the terms. The GPL
> > for example does this for version selection.
> 
> So, under a dual-licensed BSD/GPL code the latter license allows a 
> developer to remove the GPL license itself and release a single-licensed 
> BSD code if other parties want to do it?

If the dual licence permits you to select from two alternatives as
appears to be the case in

"* Alternatively, this software may be distributed under the terms of the
 * GNU General Public License ("GPL") version 2 as published by the Free
 * Software Foundation."

Then there is no problem in doing exactly what it says and distributing
it under the terms of the GPL v2 and the GPL v2 alone (or indeed the BSD
licence alone). Anyone who took the project code and produced a binary
only proprietary product from it would for example select the BSD licence
alone and convey almost no rights at all to their customer.

> I would assume a file as a boundary of a work in the case that file is 
> under different licensing terms to the rest of the software package.  On a 

Assuming is bad, you should consult caselaw.

> lot of software packages different modules are covered under different 
> licensing terms.
> 
> We can choose what license terms we will honor; however, we do not have 
> the ability to remove the licensing terms we do not like.

If the author has conveyed that right to you, then you may usually do so.

Alan



From:	Al Viro [email blocked]
Subject: Re: Fwd: That whole "Linux stealing our code" thing
Date:	Sun, 2 Sep 2007 19:21:15 +0100

On Sun, Sep 02, 2007 at 03:00:46PM +0200, Igor Sobrado wrote:
> >Not strictly true. They can either agree to a change and issue one or
> >they can convey to other parties the right to change the terms. The GPL
> >for example does this for version selection.
> 
> So, under a dual-licensed BSD/GPL code the latter license allows a 
> developer to remove the GPL license itself and release a single-licensed 
> BSD code if other parties want to do it?

Exactly.  That's what dual-licensing is.

[quote]
This is no different from the fact that we have some drivers that are
GPLv2/BSD licensed. Within the kernel, they are GPLv2. But on their own, 
you can choose to use them under the BSD license, make your changes to
them, and release them commercially.
And correct - I cannot (and neither can anybody else) then accept those
*non*GPLv2 changes back.
[end quote]

That's from Linus and quite recently.

FWIW, it's damn hard to codify "... and changes to this code should not
change the situation".  It's certainly a very good policy and in this
case it's the only sane policy.

[quote]
Actually, normally I *do* have such a trust. It's why I have no problem
with drivers that are dual-GPL/BSD, and in fact, I've told people that I
don't want them to turn them into GPL-only, because that is simply not   
polite.
[end quote]

Same posting from Linus.  And that's much more relevant to shooting the patch
in question down (and IMO it ought to be shot down) than references to
legality.



From:	Al Viro [email blocked]
Subject: Re: That whole "Linux stealing our code" thing
Date:	Sun, 2 Sep 2007 02:53:56 +0100

On Sat, Sep 01, 2007 at 09:42:54PM -0400, Luis R. Rodriguez wrote:
 
> We asked SFLC to work with us to make sure that everyone's copyrights
> were respected in the right places, and that the licenses various developers
> wanted for their copyrights were implemented correctly.  The patch I sent
> implements SFLC's suggestions in that regard.

You know, I'm rapidly losing any respect to both sides of that.  Eben
Moglen as source of advice in "is it OK to convert to GPL-only"?  And
you seriously rely on morality of that?  Theo's rants aside, if you
have to rely on SFLC for licensing decisions...  Ouch.



From: Jonathan Gray [email blocked]
Subject: Re: That whole "Linux stealing our code" thing
Date:	Sun, 2 Sep 2007 12:22:53 +1000

On Sat, Sep 01, 2007 at 08:36:24PM -0400, Jason Dixon wrote:
>
> On Sep 1, 2007, at 5:52 PM, Adrian Bunk wrote:
>
>> OK, I begin to understand this, there seem to be three different types
>> of files changed by Jiri's patch:
>> 1. dual licenced files planned to make GPL-only
>> 2. previously dual licenced files with a too recent version used planned
>>    to make GPL-only
>> 3. never dual licenced files planned to make GPL-only
>>
>> For files under 1. and 2. Reyk did contribute to dual licenced code
>> without touching the licence, but I missed that there's also code unter 3.
>>
>> So there is a problem, but not with the code under 1. (unless you plan
>> to change the semantics of the word "alternatively"), the problem is
>> with some headers under 2. plus the code under 3.
>
> The BSD license plainly states:
>
> "Permission to use, copy, modify, and/or distribute this software for any
> purpose with or without fee is hereby granted, provided that the above
> copyright notice and this permission notice appear in all copies."
>
> Once the grantor (Reyk) releases his code under that license, it must 
> remain.  You are free to derive work and redistribute under your license, 
> but the original copyright and license permission remains intact.  Many 
> other entities (Microsoft, Apple, Sun, etc) have used BSD code and have no 
> problem understanding this.  Why is this so difficult for the Linux brain 
> share to absorb?
>
> As a former Linux advocate and current OpenBSD user/developer, I'm appalled 
> that fellow open-source developers would see fit to cavalierly disregard 
> the rights of the original copyright holder.  You wield the GPL when it 
> suits you, and trample the courtesies of non-GPL developers just because 
> you [think you] can.  As bad as Jiri's offense was, it pales to the 
> impudence displayed by Alan Cox, one of the so-called defenders of free 
> software.

From http://www.mac.linux-m68k.org/docs/macpaper.php

"Always be the second operating system port to an undocumented platform.
The sterling work done by the OpenBSD/Mac team was a huge help to the
Linux project. I'm also happy to say that while half of the world may
sit on usenet advocacy groups throwing manure the relationship between
the Linux and BSD Macintosh teams has always been one of mutual
co-operation. Together we advance our detective work and knowledge of
the Macintosh platforms to the good of all Macintosh users dumped"

Alan Cox circa 1999.

http://lists.freedesktop.org/archives/xorg/2007-August/027419.html

"well I'd be quite happy to see X go GPL but I'm aware
thats not the intention of the project ;)"

Alan Cox circa 2007.

What changed? Why are you guys setting out to break all of the
work underpinning UNIX and the Internet done in the 80s at
Berkeley? The reason the protocols and infrastructure took
off in the first place is due to liberal licenses that let everyone
be involved, not wrapping things up in more restrictions and lawyers.



From: Alan Cox [email blocked]
Subject: Re: That whole "Linux stealing our code" thing
Date:	Sun, 2 Sep 2007 11:18:24 +0100
 Windsor, Berkshire, SL4 1TE, Y Deyrnas Gyfunol. Cofrestrwyd yng Nghymru a
 Lloegr o'r rhif cofrestru 3798903

> co-operation. Together we advance our detective work and knowledge of
> the Macintosh platforms to the good of all Macintosh users dumped"
> 
> Alan Cox circa 1999.
> 
> http://lists.freedesktop.org/archives/xorg/2007-August/027419.html
> 
> "well I'd be quite happy to see X go GPL but I'm aware
> thats not the intention of the project ;)"
> Alan Cox circa 2007.
> 
> What changed? 

Nothing that I am aware of. You can't take Linux/Mac68K code back into
BSD either. BSD code is being used according to the BSD licence. You
could adopt a different licence if the way your code is being used
bothers you, thats.

Where I've reused BSD code I've aways tried to contribute it back to the
BSD people or share the knowledge (and the knowledge far more than the
code mattered both ways for Mac68K systems)

I suggest you read drivers/net/wan/syncppp.c, which is I think the only
BSD derived bit of code of mine left in the kernel - and its quite
specific what it says.

Ath5k isn't my code so I don't get to pick. Having the OpenBSD maintainer
make bogus remarks about that doesn't help anyone, especially when he's
wrong and doesn't appear to know about the subject in the first place.

Alan



From:	Sam Leffler [email blocked]
Subject: Re: Fwd: That whole "Linux stealing our code" thing
Date:	Sat, 01 Sep 2007 15:03:36 -0700

Adrian Bunk wrote:
> On Sat, Sep 01, 2007 at 01:37:18PM -0400, Constantine A. Murenin wrote:
>   
>> On 01/09/07, Jeff Garzik [email blocked] wrote:
>>     
>>> Constantine A. Murenin wrote:
>>>       
>>>> This will hopefully help diminish certain myths about the code licensing.
>>>>         
>>> What myth?  The myth that Theo understands dual licensing?
>>>       
>> Reyk's code was never dual licensed, so it's not like it even matters
>> to the original dispute.
>>     
>
> It's no longer dual licenced in the FreeBSD tree because the FreeBSD 
> people removed the GPL choice of the dual licenced code 3 months ago.
>
> So all of Theo's accusations of people breaking the law by making this 
> dual licenced code GPL-only apply as well to the FreeBSD people...
>   

Sigh, why actually check the facts when you can make them up.   The code 
in question is my code.  It has my copyright (modulo bits shared with 
onoe-san who was consulted on the switch from dual-bsd/gpl to bsd only 
in freebsd).  Of course what was amusing was how after I changed the 
license on the current code in freebsd certain folks retroactively 
applied the license changes to code that was 3 years old.

But is there a point to all this nonsense?  I dual-licensed the code so 
folks could adopt and use it however they saw fit.  As I've said before 
I don't care what people do with the work I give away so long as they 
don't claim it's their own.

>   
>> That said, I don't see what exact wording you consider inaccurate.
>>     
>
> Both the FreeBSD and Linux people draw the logical conclusion that this 
> "Alternatively" means everyone can always choose to remove one of the 
> two choices alternatively offered.
>
> According to Theo, that is "breaking the law"...
>
>   

I've yet to see "FreeBSD people" speak up so again you're just spouting 
jibberish.  I am speaking up as the author of the code that set the dual 
license in place.  I have the definitive say and I have said that any of 
my code that is dual-licensed can be made gpl only.

    Sam



From:	Bob Beck [email blocked]
Subject: Re: Fwd: That whole "Linux stealing our code" thing
Date:	Sat, 1 Sep 2007 18:02:26 -0600

>As a free software user and developer, the question I have is how come
>the Linux community feels that they can take the BSD code that was
>reverse-engineered at OpenBSD, and put a more restrictive licence onto
>it, such that there will be no possibility of the changes going back
>to OpenBSD, given that the main work on the code has happened at
>OpenBSD? (Obviously, such a scenario it is permitted by the licence,
>but my question is an ethical one -- after all, most components of
>OpenHAL were specifically based on the OpenBSD's ath(4) HAL code.)
>
>You can see that Christoph Hellwig agrees with this ethical problem,
>as in the message below.
>
>C.
>
>
>>On 28/08/07, Christoph Hellwig [email blocked] wrote:
>> On Tue, Aug 28, 2007 at 12:00:50PM -0400, Jiri Slaby wrote:
>> > ath5k, license is GPLv2
>> >
>> > The files are available only under GPLv2 since now.
>>
>> Is this really a good idea?  Most of the reverse-engineering was
>> done by the OpenBSD folks, and it would certainly be helpful to
>> work together with them on new hardware revisions, etc..

I couldn't agree more. The point is, while we BSD license fans know and
expect people from private industry to take our stuff and use it, at
least private industry does not come to the table with "hey, let's
cooperate" - we know who the corporate whores are, and we act accordingly. 

However, when a linux developer comes to us and say "hey lets cooperate"
usually there is a thought of "this is a kindred spirit who understands
what our mutual goals are and won't stab us in the back".  My concern
is that this situation will change if this is not rectified. 

I think the community needs to decide, should cooperation be based on
morals and trust, or does the Linux community need to be regarded with
less trust than a Corporation, something to be avoided, as while
corporations can be counted on to act without morals, the knife is up
front and visible. They do not come to you with one hand of
cooperation extended and a knife kept behind their back.

 -Bob 



From: Theo de Raadt [email blocked]
Subject: Re: Fwd: That whole "Linux stealing our code" thing 
Date:	Sat, 01 Sep 2007 18:36:36 -0600

When companies have taken our wireless device drivers, many many of
them have given changes and fixes back.  Some maybe didn't, but that
is OK.

When Linux took our changes back, they immediately locked the door
against changes moving back, by putting a GPL license on guard.

Why does our brother Linux take a file that is 90% BSD licensed,
and refuse to let us see the 10% he adds?



From:	"Constantine A. Murenin" [email blocked]
Subject: Re: Fwd: That whole "Linux stealing our code" thing
Date:	Sat, 1 Sep 2007 21:52:31 -0400

On 01/09/07, Theo de Raadt [email blocked] wrote:
> When companies have taken our wireless device drivers, many many of
> them have given changes and fixes back.  Some maybe didn't, but that
> is OK.
>
> When Linux took our changes back, they immediately locked the door
> against changes moving back, by putting a GPL license on guard.
>
> Why does our brother Linux take a file that is 90% BSD licensed,
> and refuse to let us see the 10% he adds?

Indeed, it's upsetting that people like Luis Rodriguez push for the
lawyers to be involved to (fight?) an open source project. Why, may I
ask?

Why Luis puts the phrase "legal hell" next to entirely free software?
[0] Why is he trying to go against the BSD community, which gave him
the entire HAL framework for the driver in question?

Best regards,
Constantine.

[0] http://marc.info/?l=linux-wireless&m=118857712529898&w=2



From:	Jeff Garzik [email blocked]
Subject: Re: Fwd: That whole "Linux stealing our code" thing
Date:	Sat, 01 Sep 2007 22:00:47 -0400

Constantine A. Murenin wrote:
> Indeed, it's upsetting that people like Luis Rodriguez push for the
> lawyers to be involved to (fight?) an open source project. Why, may I
> ask?


Is it not self-evident?  Legal review is the sane course of action, when 
legal issues are the bone of contention.

That said, Linux people are far more pragmatic than FSF people, and 
often disagree with the FSF.  I would not take an FSF lawyer's word as 
Gospel.

Theo manages to confuse "Linux" and "FSF" quite often, but that's 
characteristic of his muddled thinking and inexperience.

	Jeff



From: Jonathan Gray [email blocked]
Subject: [reyk at openbsd.org: Re: That whole "Linux stealing our code" thing]
Date:	Mon, 3 Sep 2007 23:25:42 +1000

This is a lot more relevant than much of the ongoing
discussion, so perhaps people could take a moment to read it over.

----- Forwarded message from Reyk Floeter [email blocked] -----

From: Reyk Floeter [email blocked]
Date: Mon, 3 Sep 2007 13:23:04 +0200
Subject: Re: That whole "Linux stealing our code" thing

Hi!

I just returned from vacation where I was offline for about two weeks.
So I totally missed the incidence and all the surrounding discussion.
I'm just digging through many many mails in my inbox from OpenBSD
users and developers, Linux people, GNU/freesoftware people, misc *BSD
people, and obviously from some trolls.

I don't want to restart the discussion but I just want to say and
repeat a few words:

- I will not release or agree to release my code under either the GPL
or any kind of a "dual"-license.

- The ISC-style license must remain including the copyright notice and
even the warranty term.

- Thanks to the OpenBSD community and especially to Theo de Raadt for
entering into it and for defending my rights as the author of the
controversial code.

- This is eating our time. Every few weeks I get a new discussion
about licensing of the atheros driver etc. blah blah. Why can't they
just accept the license as it is and focus on more important things?

I will talk to different people to get the latest state and to think
about the next steps. I don't even know if the issue has been solved
in the linux tree. But PLEASE DON'T SPAM ME with any other mails about
this, even if you want to help/support me, I will talk to the relevant
people in private.

Thanks!
reyk

----- End forwarded message -----

• Archive of above thread
• KernelTrap interview with Alan Cox
• KernelTrap interview with Jonathan Gray
• KernelTrap interview with Theo de Raadt