Linux Wireless Driver Violates BSD License?

bsdphx writes "After years of encouragement from the OpenBSD community for others to use Reyk Floeter's free Atheros wireless driver, it seems that the Linux world is finally listening. Unfortunately, they seem to think that they can strip the BSD license right out of it."

No, it doesn't.

[Whiney Mac Fanboy (963289)]

It appears that someone's submitted a patch to the LKML that wrongly strips the BSD atheros driver of its license - a clear violation of copyright.

However, until it's in Linus's tree (or even the MM tree), the violation is not by "linux", but the contributor, Jiri Slaby. [blogspot.com]

Anyway, thanks to the OpenBSD team for these great drivers. Thanks to the Linux team for including them (under the correct license).

Re:No, it doesn't.

[phoebe (196531)]

It appears that someone's submitted a patch to the LKML that wrongly strips the BSD atheros driver of its license - a clear violation of copyright.

The contributor being the author of the wireless module makes this article a bit short on common sense.

First check the author of the patch, its Jiri Slaby.

Subject [PATCH 4/5] Net: ath5k, license is GPLv2
From Jiri Slaby <>
Date Tue, 28 Aug 2007 12:00:50 -0400
Digg This

ath5k, license is GPLv2

The files are available only under GPLv2 since now.

Signed-off-by: Jiri Slaby

Then check the copyright notice on top of the source files, there is a copyright to ... Jiri Slaby.

+++ b/drivers/net/wireless/ath5k_base.c
@@ -4,25 +4,9 @@
* Copyright (c) 2007 Jiri Slaby
* All rights reserved.

So an author changed the license of his own code, hit the presses!

Re:No, it doesn't.

[MMC Monster (602931)]

Not just that, but it appears that the original file was dual licensed to BSD and GPLv2.

What exactly is this article about?

It's about dividing the communities....

[by Anonymous Coward]

The BSD and GPL communities are two groups with much in common and some differences. That's especially true of the OpenBSD people who are very much committed to freedom for their software (and just disagree with the GPL people about how that is best defined). This kind of article is about trying to set up atificial disagreements between those communities so that they don't cooperate well. I write copyleft software and content (GPL/GFDL/CC-SA) but I would mostly relicence it if an important project like OpenBSD or X.org asked for it. I would make that decision based on the value of the project. If I feel that a project is harmful overall then I probably wouldn't.

The trick is that we have to not be divided and work together sensibly.

Re:It's about dividing the communities....

[PinkPanther (42194)]

[OpenBSD people] just disagree with the GPL people about how that is best defined

It isn't so much a disagreement about how "free" is defined; it is more about who the target of "free" is. The BSD-style folks focus on programmers; the GPL-style folks future end-users. Both want the code to be "free" (can do whatever they desire with the code) to their target.

Re:No... It's about something a little different..

[Phroggy (441)]

A patch to existing code would be considered a derivative work.

Correct.

Derivative works all belong to the original copyright owner. No matter who wrote them.

Only partially correct. Derivative works belong to both the original copyright owner AND the author of the derivative work. I can take BSD-licensed code, hack it however I like, and release the result as a proprietary application, as long as I give proper attribution as required under the BSD license. Doing so does NOT give the authors of the BSD code the right to use my derivative code for free; they'd have to buy it from me at whatever price I set (if I choose to offer it for sale). Although I have to comply with their license in order to use their code in this way, they don't own the result.

This is what keeps someone from trying to add in a patch that has a different license from original GPL work.

By submitting a patch, you're implicitly granting a license for your code to be included with the project, under the same license(s) as the rest of the project, but you still hold the copyright on your code. Some projects have legal wording to clarify their policy on accepting patches, but this is how things are generally understood to work.

The original licensee does not allow it.

The inclusion of BSD-licensed code in GPL-licensed software is perfectly fine, as long as the conditions of the BSD license are met. In this case, it sounds like the conditions of the BSD license may not have been met. The author of most of the driver took the whole thing (including patches owned by other people and licensed under the BSD license), and tried to relicense the whole thing under GPLv2 without dotting all the i's and crossing all the t's. The idea of what he was trying to do is totally OK, but a couple of details may not have been handled correctly and might need to be fixed.

Doing the same thing in the other direction (including GPL-licensed code in a BSD-licensed project) is a no-no, by the way. The GPL has additional requirements that the BSD license doesn't satisfy.

Re:No, it doesn't.

[kestasjk (933987)]

Linux users think Abraham was ordered to sacrifice Linus as a burnt offering, GNU users think Abraham was ordered to sacrifice Stallman as a burnt offering. Linux users believe in the holy trinity of GNU/Linux/NvidiaBlobs, BSD users believe in the one and only UNIX. Linux users are fine with Tux toys and stickers, BSD users believe in no graven images. Linux users are subdivided into those that argue about interpretations of freedom and whether the establishment is corrupt, BSD users are subdivided according to which BSD is the true descendant of UNIX(pbuh).

Re:No, it doesn't.

[jimstapleton (999106)]

actually, most of them don't have him as the copyright holder.

Looks more like some yutz decided that he didn't like the BSD licence and went in and changed all the licences to GPLv2, in the files, and didn't do anything else.

Honestly, I can't complain, as long as the copyright notices are kept, and unchanged, it is acceptable (someone posted thsi further down).

Nonetheless, someone going in, and doing nothing but removing the BSD licencing on every file (or at least the first 4 or 5, I didn't look through the whole thing), and replcaing it with "this code is now under GPLv2", seems somewhat childish, more like a tantrum than anything else.

Re:No, it doesn't.

[kernelpanicked (882802)]

Wrong! But thanks for playing. I see you convinced the greater slashdot horde to give you a few mod points for your wrong answer, congrats. The correct answer is the code is copyrighted 2004-2007 Reyk Floeter 2006-2007 Nick Kossifidis. This is a clear case of someone grabbing BSD code, stripping and replacing the license with the GPL, and submitting it as a patch to the mainline kernel.

Re:No, it doesn't.

[uglydog (944971)]

But what about the lines that say

- * 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.
- *
+ * This file is released under GPLv2

Doesn't that mean the the second person is opting to distribute under GPLv2? And the copyright notices are intact.

* Copyright (c) 2004-2007 Reyk Floeter
* Copyright (c) 2006-2007 Nick Kossifidis
*

Did you even read the original patch?

[sethawoolley (1005201)]

It appears that you can't read this particular patch style. The lines with + mean added, the lines with - mean removed.

The lines without either mean that's context for the differences.

If you look at the original patch, no attribution was removed. The attribution was in the context lines.

It looks like the .c files were handled appropriately and it was merely the .h files that had the license completely ripped out. The .c files were dual-licensed and said you could choose either. They just removed the BSD license as that was "choosing" GPLv2. The .h files are just some interfaces and don't change often anyways, so the BSD license is good enough for them (they should have left those). The .c files are the actual implementation, which would change between operating systems.

Here's a link to the actual diff as provided in the original article:

http://lkml.org/lkml/2007/8/28/157 [lkml.org]

You'll also note that the dual-licensed code had the committer's copyright notice on it. In some cases it was only his notice, originally. With the data immediately available, maybe he stripped it out in a commit before this one, but they don't seem to be accusing him of that. They are mainly accusing him of ripping out the BSD license from a couple .h files since they didn't have the dual-license notice in them. If they aren't dual-licensed under both, you can relicense as GPLv2, but you have to include the BSD notice under its own terms. The GPL itself even says not just attribution, but the original notices themselves must be preserved. One additionally might say that since the GPL says to preserve the original notice, that even in the dual-license case you must preserve the BSD license in order to initially comply with the GPL, although that's a requirement of the GPL and not a dual-licensing/BSD provision. A dual-licensing (as you can see in this case) clearly says you can pick either, since the word "Alternatively" (e.g. the ath5k_reg.h license) implies if you chose the following path, you can ignore the provisions of the previous path.

In summary, it looks like a lot of this was nit-picking over how to actually do the license notice preservation, rather than preserving somebody's attribution. I imagine it'll be fixed up in very little time and few people will care about this in more than a day or two.

Re:No, it doesn't.

[Lumpy (12016)]

Exactly,

I don't even know why this is news, Until Linus accepts it, it's some random patch submitted to the tree, tons of those are rejected daily.

The entire story and Slashdot submission is plain old FUD. if it was accepted and part of a new kernel tree I can see the story, but right now it's absolutely nothing but some random guy changed. Are we going to start getting stories submitted about what someone says on their blog now?

Those in glass houses...

[Phil246 (803464)]

Im suddenly reminded of this [slashdot.org], where linux gpl'd code found its way into BSD via a wireless driver.
Those in glass houses shouldnt throw stones

Re:Those in glass houses...

[moranar (632206)]

The response from the person involved is at least much more responsible and reasonable than in the earlier incident: http://lkml.org/lkml/2007/8/29/69 [lkml.org]

Jury's Still Out

[Bob(TM) (104510)]

The fact is the original patch post was on Tue, 28 Aug 2007 12:00:50. Since then, the discussions are ongoing as how best to proceed. Recently, this was posted:

Date Wed, 29 Aug 2007 08:35:05 -0200
From "Jiri Slaby"
Subject Re: [PATCH 4/5] Net: ath5k, license is GPLv2

On 8/29/07, Johannes Berg wrote:
> On Tue, 2007-08-28 at 12:00 -0400, Jiri Slaby wrote:
>
> > The files are available only under GPLv2 since now.
>
> Since the BSD people are already getting upset about (for various
> reasons among which seem to be a clear non-understanding) I'd suggest
> changing it to:

yes, please. Can somebody do it, I'm away from my box.

> + * Parts of this file were originally licenced under the BSD licence:
> + *
> > * Permission to use, copy, modify, and 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.
> > *
> > * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL
> WARRANTIES
> > * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
> > * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
> > * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
> > * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
> > * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
> > * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
> + *
> + * Further changes to this file since the moment this notice was extended
> + * are now distributed under the terms of the GPL version two as published
> + * by the Free Software Foundation
>
> johannes
>

As mentioned before, it is the LKML, not the Rosetta stone. Things change ...

Legal Weirdness

[saterdaies (842986)]

Basically, you are allowed to link GPL'd code to BSD code. So if I wrote "The boy hit the baseball" under the BSD license and you alter it to "The large boy hit the baseball well" under the GPL, the original statement is still available for use under the BSD license - even in your second statement. As long as they remove your GPL'd addition (the intertwined words "large" and "well"), they are free to use it under the BSD's terms.

The practical point is that the BSD code, when linked with GPL code, must adhere to the restrictions of both licenses. Most people just say that it has been relicensed under the GPL. That isn't exactly true. From most practical standpoints, the BSD license has so few restrictions that it doesn't matter, but technically that BSD code is still under the BSD license and it's requirements must be met.

So, that BSD code can easily be linked and intertwined with GPL code, but those few requirements of the BSD license must be met so long as there is any BSD code in the GPL'd derivative work.

Re:No honor amongst theives

[Kokuyo (549451)]

No orthography among ACs, eh?

Re:Strange

[Whiney Mac Fanboy (963289)]

The following was removed from the license:

* Permission to use, copy, modify, and 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.

That sentence is pretty clear. Not all BSD code can be relicensed.

Let me remind you however, that this was the work of an individual who posted to a public mailing list. It hasn't been accepted into Linus's or Morton's tree.

Mod parent up!

[Eivind Eklund (5161)]

Parent has insightfully noticed his own error. And the error is modded up. So mod the parent up.

The move is clearly against the BSD license. (Also, combining GPLv2ed code and BSDed code is subtly against the GPL, as the requirement to reproduce the license - as shown and violated here - is an extra requirement compared to the GPL, violating the "no additional restrictions" clause of the GPL.)

Eivind.

Re:Strange

[fruey (563914)]

http://lkml.org/lkml/2007/8/29/69 [lkml.org]

Someone pointed out the problem and a patch is likely on its way.

Re:Strange

[nomadic (141991)]

No it isn't. You can reply to your own posts all you like. Where do people get this weird notion that it's bad form?

I agree with you, nothing wrong with that.

Well I disagree you bastard.

Re:Strange

[Eivind Eklund (5161)]

The original BSD vaersion is still out there.

Just like it would if the code was taken properitary.

It's a bit rich to deny people to keep their own changes proprietary, wouldn't you say?

Except that's what the GPL tries to do. It's removing freedom.

And that's what many of us BSDers are against. We want our software to keep freedom. Including the freedom of future developers to keep their own changes private, or get paid for them. Thereby, we also allow the end users the freedom to buy those changes - a freedom they wouldn't have if the code was GPLed, because the incentive to make the changes wouldn't be there.

As an example, we have Apples operating system, partially made on code I wrote. And I'm a very happy user of it, even though I (or rather, my employer) had to pay for the extra stuff Apple has added. The ability to do so is a freedom I have partially gotten from having released my software under the BSD license.

Eivind.

Re:Strange

[bentcd (690786)]

Except that's what the GPL tries to do. It's removing freedom.
And that's what many of us BSDers are against. We want our software to keep freedom.

Freedom isn't binary, nor can it easily be discussed without qualifiers. In this case, BSD is a case of the developer saying "I want freedom, I want it for myself, and who cares what everyone else does with it" whileas GPL is a case of the developer saying "I want freedom, I want it to rest with the end user of my software and never mind if that causes me some inconvenience in the process". They are different kinds of freedom, resting in different parts of the software ecosystem.

Re:Hmmmm

[squiggleslash (241428)]

Absolutely. This situation is unbelievable.

A few months ago, a GPL'd Linux driver was incorporated into the mainline OpenBSD kernel tree (albeit some months before release, and largely non-working.) The Linux developers concerned contacted the OpenBSD team via the regular mailing list, Ccing some relevant Linux and legal people. The email was polite, more or less friendly, and constructive, offering help to the OpenBSD people to ensure the situation was resolved with both projects having a working driver at the end of the day.

The OpenBSD team's response was to go nuclear. Theo called the Linux developers "inhuman". Many argued that the copyright violation was legitimate performing coding acrobatics to pretend that real, copyrighted, code was never being distributed under the BSD license; others argued this proved the superiority of the BSD license because if it had been the other way around, the OpenBSD team would never have objected, given the BSD license allows you to do (apparently) anything, whereas the GPL prevents use in closed systems.

Well, what a bunch of, frankly, hypocritical two-faced liars. The OpenBSD team's response to an apparent BSD license violation (which we were assured would never happen, because the BSD license is so liberal) is to directly accuse the Linux developers of copyright infringement. Rather than involve appropriate mailing lists and relevant people, the complaint is made on the public Undeadly.org website. Rather than offer help, the OpenBSD developers just spit blood. And none acknowledge that the copyright infringement hasn't even happened yet, that is to say, the proposed code is a patch that has yet to be accepted into the mainline kernels.

This is the second time the OpenBSD team have owed Linux developers an apology, and I bet it's the second time we're not going to hear one, instead hearing the same self-righteous fraudulent apologetics we hear one.

OpenBSD developers have time and time again claimed "moral superiority" over GNU and Linux due to their adoption of a license that allows code to be used in closed projects. It always was a specious argument, but it's looking all the more absurd today.

Re:Hmmmm

[MysteriousPreacher (702266)]

OpenBSD developers have time and time again claimed "moral superiority" over GNU and Linux due to their adoption of a license that allows code to be used in closed projects. It always was a specious argument, but it's looking all the more absurd today.

Claiming any particular licence (BSD/GPL) to be superior is like asserting that cars are superior to helicopters. In far too many cases, the licences are dragged in to try to justify a bad argument, and the fault lies with both camps. GNU and BSD zealots alike adopt Talibanesque positions that do nothing but harm to the community.

This story should have been a simple clear-cut case it weren't for a small rabble-rousing group. Funnily enough, Theo posted a fairly decent and non-inflamatory respones in the early discussions. This is in stark contrast to the earlier GPL case (mentioned in your post) where his reaction was indefensible.

Incidentally, the BSD licence infringement has already taken place. That happened as soon as the author distributed the code with the licence stripped from it. Doesn't matter whether or not it hit a main-stream kernel. As soon as he made it available to others, distribution kicked-in. That said, the author has a case to answer for but certainly not the entire Linux community the "OMG LINUX STOLE OUR CODE!" crowd would have us think.