This will hopefully help diminish certain myths about the code licensing.
C.
---------- Forwarded message ----------

From: Theo de Raadt <deraadt@cvs.openbsd.org>

Date: 31-Aug-2007 21:40

Subject: That whole "Linux stealing our code" thing

To: misc@openbsd.org
[bcc'd to Eben Moglen so that people don't flood him]
I stopped making public statements in the recent controversy because

Eben Moglen started working behind the scenes to 'improve' what Linux

people are doing wrong with licensing, and he asked me to give him

pause, so his team could work.  Honestly, I was greatly troubled by

the situation, because even people like Alan Cox were giving other

Linux developers advice to ... break the law.  And furthermore, there

are even greater potential risks for how the various communities

interact.
For the record -- I was right and the Linux developers cannot change

the licenses in any of those ways proposed in those diffs, or that

conversation (http://lkml.org/lkml/2007/8/28/157).
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,

they must all agree.  A person who receives the file under two

licenses can use the file in either way....  but if they distribute

the file (modified or unmodified!), they must distribute it with the

existing license intact, because the licenses we all use have

statements which say that the license may not be removed.
It may seem that the licenses let one _distribute_ it under either

license, but this interpretation of the license is false -- it is

still illegal to break up, cut up, or modify someone else's legal

document, and, it cannot be replaced by another license because it may

not be removed.  Hence, a dual licensed file always remains dual

licensed, every time it is distributed.
Now I've been nice enough to give Eben and his team a few days time to

communicate inside the Linux community, to convince them that what

they have proposed/discussed is ...
[ message continues ]

This is absolute nonsense. By default, you can remove a license if you want

to. This is why both the GPL and the BSD licenses have clauses requiring you

to leave them in.
A file that is under a dual license may be used under either the GPL license

or the BSD license. Neither license requires you to retain the *other*

license in the file. So there is absolutely no reason you cannot remove one

license or the other. To argue otherwise is to argue that you need to comply

with *both* licenses in a dual-licensed file to get the rights granted by

either, and that's nonsense.
You cannot, of course, modify a license and expect your modified license to

apply to protectable elements you didn't author. And anyone who receives

modified versions of the file still has all the rights the original authors

grant them.
Let's perform a thought experiment for a moment. Suppose the BSD license

explicitly said you could remove the licensing clause if you wanted to.

Would you still argue that you couldn't remove it even though it says you

can? Well, the GPL says you can modify anything you want to, except *THAT*

license. This means you can remove any other license notifications you want.
Note that your license editing or removing has no effect on the rights

people actually get except to your code.
DS
-

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.
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
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...
[ message continues ]

Correct. They need the right to redistribute the work, and they may obtain
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
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
-

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
-

License is only a promise not to sue. Only original author is permitted to

sue in BSD/GPL case. David seems right here.
If Linus releases GPL/BSD program, you choose GPL and distribute it to

me, and I put it into proprietary evil app, I am _ok_.
You can't sue me, because you are not copyright holder.
Linus can't sue me, because he BSD licensed it.
=> I'm fine. Yep, copyright law is strange, and you may be right

outside U.S.

							Pavel

--

(english) http://www.livejournal.com/~pavelmachek

(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html

-

Yes, but this has to be done in writing and neither the BSD nor the GPL
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
-

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
-

uh, p2p? :)
-

(by the way, text in caps surrounded by *'s is meant to indicate vocal stress,

not volume)
On Sunday 02 September 2007 22:01:18 David Schwartz wrote:
However, this is not what is happening here. Jiri has made changes that he has

licensed solely under the GPLv2. This means that he now becomes the

licensor - not Mr. Floeter or Mr. Leffler. *BUT* *ONLY* of the version of the

code containing his changes.
Mr. Floeter *CAN* request that his code be removed from said fork - his code

is solely licensed (AFAICT and IIRC) under the BSD/ISC license and was only

covered by the dual-license because it was integrated into a work that

carried said dual-license. (I'm not sure how well such a revocation would

work in reality, but it is Mr. Floeters right.)...
Correct. Doesn't apply in the case of the code in question (unless the changes

that were made are so tiny as to not be copyrightable).
In this case the code is question is a modified version, which means that the

right to distribute said modified version now originates with the person

holding the copyright on the modifications. (Though their right to distribute

the code, in such a situation, is lessened quite a bit by the text of the 
No misunderstanding, really.
Alan seems to have given a bad example that doesn't apply to the situation 
Agreed. When re-distributing an un-modified copy of a work. When distributing

a modified work, the "work" has the license that the person who made the

modifications places on it. But individual files and pieces of code will

still retain their original license - this is how it works.
DRH
--

Dialup is like pissing through a pipette. Slow and excruciatingly painful.

-

No. Neither the BSD license nor the GPL license permit you to revoke rights.

Mr. Floeter's code is still available under the BSD/ISC license. The BSD

license does not require you to make derived works available under a BSD

license. *His* code is still available under a BSD/ISC license, of course,

but the changed code is not.
Read the BSD license. It does not require changes to be made available under

a compatible license. This is the main difference between the BSD and GPL

licenses.
Note that it would be an error to remove the BSD license text, as the BSD

license requires you to keep it and you still need the BSD license to grant

you distribution rights to the original work. However, the license does not

apply to protectable aspects of the code not placed under the BSD license by

their original author, and it is important to add a note to that effect.
DS
-

Doesn't matter if the BSD license or the GPL *PERMITS* it or not. The fact

remains that the person making a work available under *ANY* form of copyright

license has the right to revoke said grant of license to anyone. The GPL

codifies certain situations in which the person would not, personally, have

to revoke the license, but does not limit the original copyright holders

rights (in that regard) in any way.
The BSD/ISC license has none of the automatic conditions of the GPL, but it

also cannot remove the copyright holder(s) from exercising their rights.
(And no, I haven't spoken to a lawyer about this - I did, however, ask a

recently graduated law-school student where I could look for case-law and the

text of the actual laws. What I got was some background on US copyright law

itself and an agreement that a copyright license does not - and can not - 
Have done so. And that is the only part of the license that I actually don't 
Agreed, and I've never claimed otherwise. (Nor has anyone else. I believe the

closest that anyone has come was Alan Cox saying (and I'm going to paraphrase

it because I don't think he ever stated it well) "If you've made changes to a

file that carries a dual BSD/GPL license and your changes are GPL only, the

file can no longer be distributed under the BSD license at all. So it is safe

to remove the headers from that individual file."
There is no way that a license on a constituent file can alter or affect the

license on the whole project (if it is different). It can "muddy the waters",

but that is about as far as I can see it going.
(I realize I may have said different, originally, but you'll have to forgive

me. I was not in the best of moods (or the best state of mind) to be making a

completely rational argument when I did such.)
DRH
--

Dialup is like pissing through a pipette. Slow and excruciatingly painful.

-

Not after the licence has been given and accepted (and there might be

restrictions), unless of course the licence contained such reservation.

--

Krzysztof Halasa

-

I hate to belabor the point, but you seem to be making the mistake of "The

license applies to the copyright holder" that I've seen a lot of people make

(and kept quiet about).
The person holding the copyright has all the legal standing to revoke a

license grant at any time. Licenses such as the GPL are not signed contracts,

and that means there are limits to what effect they can have on the copyright

holder.
If the license was of the "signed contract" type, or contained text stating

that the copyright holder was giving up all rights of revocation (etc...) I

could agree with you. As it stands, no "Open Source" license that I have seen

used on a major project contains any part that does that. In fact, the GPL is

the only license I can name (offhand) that even touches on the rights of the

copyright holder - and then it is in the form of "If you do X, Y or Z all

rights granted under this license are automatically revoked".
That is an "automatic clause" - not a limitation stating that the copyright

holder can only revoke under those conditions. The person holding the

copyright has quite a few rights - more than people believe - and not even

the most generous of Open Source licenses (except those that contain text

like "granted in perpetuity" or similar) even come close to being exempt from

the holder of the copyright not being able to summarily revoke a given

persons license.
DRH
--

Dialup is like pissing through a pipette. Slow and excruciatingly painful.

-

I believe you are making the mistake that the license on code has

anything to do with what the author chooses to do in the future.

Releasing something as BSD does not force the author to do anything in

the future with his code, and he/she could add and relicence as he/she

feels fit.  HOWEVER, that particular code has already been released as 
I believe you are confusing the fact that an author can decide to

release code under another license, with the existence of code under

that earlier license.  The license grant comes from THE CODE (which

bears a license), not THE AUTHOR.  I can use GPL code I get in the mail

because the license on the work says I can do so, not because I

contacted the author and got a specific grant.  If such a grant were

only verbal, your theory might hold, but that doesn't apply to any OSS

software under discussion here.
If your legal theory were true, I could sell you a book and then later

demand that you destroy it.  I could also release something as public

domain, and then later rescind that (I still hold the copyright on what

I produced), and charge money from anyone who used it.  I think its safe

to say that this does not happen in practice.  Please provide some

examples to the contrary or caselaw if you want to convince me otherwise.
Furthermore, BSD/GPL software could not really exist under your legal

theory; A programmer who wrote 30 year old core BSD code could wake up

tomorrow and decide to require all BSD derivatives to remove his code or

pay him for it (and the next day he could change the price again).  Open

source software would not exist if such a liability were true, and

companies like Sun could not be built up off of derivatives of it.

Linux 0.01 is still available under a pre-GPL license if you can find a 
There are plenty of rights, but retroactive changes to the license terms

of something you've already distributed is not one of them.
  - Jim Bruce

-

I may be mistaken, but it has always been my understanding that, unless you

specifically waive your rights, they are automatically retained. (Under the

law in the US, at least).
Hence, a copyright holder can do such, where the license has not been acquired

by money changing hands.
(And actually, my above statement isn't rendered false by your rebuttal - it

still appears that the person I replied to believes that a copyright license

applies to the person holding the copyright in the same manner it applies to

the person receiving the item under said license. Though I will admit it if I 
The license is a direct grant from the author. If the author so wished, he/she

could pull the license - either entirely or in part. About the only caveat is

that the author would have to publish and attempt to contact everyone who may

have acquired the item under that license to inform them of such a change -

this does make it difficult, hell, makes it nearly impossible, but it can be 
Actually, no. A purchase does automatically grant the rights inherent in

ownership - but that is a *PURCHASE*. Mere transfer of an item with no

exchange of money cannot convey those rights. As far as the 'public domain'

argument goes... That smells of a straw-man and is as different from a grant

of license as it is from a purchase. When you release something into the

public domain you are waiving *ALL* of your rights as copyright holder. 
He could, but AFAICT, thirty years ago BSD was still run entirely by UC

Berkely and any copyrights that might be held are held entirely by UC Berkely

and not the individuals that contributed to such. (Whats more, a 30 year old

version of BSD doesn't meet the requirements of the AT&T agreement, so its

only legal in-so-far as it massively predates that agreement (and the lawsuit

which spawned it) :)
And yes, Linus actually could revoke the license on any copy of Linux from

before he started merging code written by other people into the code-base.

After...
[ message continues ]

Perhaps that is your local laws, but I'd be surprised anyway.
Do you sign contracts when shopping, or are you (and the shop)

allowed to "revoke" the transaction after it's made (I'm not

talking about shop's return policy)?
Is what you wrote only valid WRT licences?
Which country has such laws BTW?
In Poland, I can't just go and "revoke" a "statement of will"

if I haven't explicite reserved a right to do so.

Obviously I can act contrary to the statement, and be held liable.

--

Krzysztof Halasa

-

I'll take this at face value - I might have mis-parsed your earlier statements 
US Copyright law. A copyright holder, regardless of what license he/she may

have released the work under, can still revoke the license for a specific

person or group of people. (There are some exceptions, but they do not apply 
A purchase is separate from a grant of copyright under license. If I purchase

a book I have the right to read the book and I have the right to sell the

book to someone else, but I have no other rights to it. But if I purchase

something, there are rights that go along with said purchase.
Under a license such as the GPL (I can't say the same for the BSD/ISC

license - I haven't spent enough time studying it to know for sure) no money

need change hands for access to the program *and* source. All the rights that

anyone *BESIDES* the copyright holder have to the program and/or source comes

from license. But since money has not changed hands, there is no further set

of rights or guarantees - the copyright holder has not, in general, sold a

copy of the work or granted any guarantee that the license will not be 
Ah, see - in the US the license(s) in question (and licenses in general) are

grants of rights, not a "statements of will". If there are to be any limits

on the rights of the copyright holder, under US law, they have to be

explicitly stated in the license itself. (Truthfully, in the US a license

should be read with an implicit "All rights reserved")
DRH
--

Dialup is like pissing through a pipette. Slow and excruciatingly painful.

-

Oh come on, I thought some small country in maybe central Africa,

but certainly not USA.
What you write would essentially mean GPL (and any other such licence)

is invalid in the USA.
The licence is basically a promise not to sue. It wouldn't make any
Actually (and I think it's the same in the USA), a copyrighted work

has an implicit "all rights reserved". A licence is just exception.

--

Krzysztof Halasa

-

US Law is a twisted maze - you wouldn't believe the contradictions that exist

between different sections of the US Federal Code. (And its worse as you move 
Nope. The GPL is an explicit grant of rights and is fully legal and active as 
If I was to revoke the license on something I held copyright to, I'd be forced

to make an attempt to contact everyone that may have received a copy of the

work under that license before I could ever begin filing lawsuits. This

process will take at least a month - more if the various localities where

someone might be living has laws about what constitutes an attempt to

contact. (For instance, here in Pennsylvania an attempt to contact is taking

out large format classified ad's in every newspaper in the area where the

person is known to reside - or statewide if the region is not known. The ad's

have to run for a minimum of one week)
This means that it'd take no less than five weeks - and might take as much as

six months - before I could begin filing lawsuits. (And even then I'd have to

have proof that the person in question was violating my copyright at the time 
Difference in terminology ?
A "Grant of Rights" is where you say 'Normally only I could do this, but I am

giving you the legal right to do it as well'. A "statement of will" is 'This

is what I want to have happen, in perpetuity'. In the US, a "statement of

will" can include or imply a "Grant of Rights" and vice-versa, but they are 
And? The fact remains that "All Rights Reserved" means "I am reserving all

rights I do not specifically grant or waive". ie: If a license doesn't

state 'The licenser hereby waives the right to revoke this license at any

time' then that right hasn't been lost. (A license acquired through a

purchase - as might apply to a novel - is a lot different. And contracts are

a different beast entirely)
DRH
--

Dialup is like pissing through a pipette. Slow and excruciatingly painful.

-

Sure.

That means you can't use a work without obtaining a licence first

(unless you are the copyright holder and don't need (can't have)

the licence).
I.e., as the author, you don't have to write those "copyright xxx"
Yes, it hasn't been *lost*. It just never existed :-)

(with exception of those things starting 40 or so years after the

original publication, I think we can safely skip this part).
Obviously, the copyright holder is free to distribute the work

under a different licence next time (but (s)he has to make sure it's

ok with licences (s)he uses, if it's a derived work).

--

Krzysztof Halasa

-

The OpenBSD policy page doesn't agree with you:
"...That means that having granted a permission, the copyright holder

can not retroactively say that an individual or class of individuals are

no longer granted those permissions. Likewise should the copyright

holder decide to "go commercial" he can not revoke permissions already

granted for the use of the work as distributed, though he may impose

more restrictive permissions in his future distributions of that work."
http://www.openbsd.org/policy.html
Chris

-

There are other enforceability issues as well. For instance in the

US, Copyright Law applies as soon as something is written. So,

does Copyright Law apply if I write, "You cannot read this."

Of course, it's a trivial example. Revocation of a license to

read a work is absurd. Using this theory, once somebody's

written "work" has been distributed under some license, a

different license would likely be regarded as unenforceable

by a court.
Cheers,

Dick Johnson

Penguin : Linux version 2.6.22.1 on an i686 machine (5588.30 BogoMips).

My book : http://www.AbominableFirebug.com/

_
****************************************************************

The information transmitted in this message is confidential and may be privileged.  Any review, retransmission, dissemination, or other use of this information by persons or entities other than the intended recipient is prohibited.  If you are not the intended recipient, please notify Analogic Corporation immediately - by replying to this message or by sending an email to DeliveryErrors@analogic.com - and destroy all copies of this information, including any attachments, without reading or disclosing them.
Thank you.

-

By my reading, this is supported by 17 USC 203(a)(3):
  (3) Termination of the grant may be effected at any time during a

      period of five years beginning at the end of thirty-five years

      from the date of execution of the grant; or, if the grant covers

      the right of publication of the work, the period begins at the

      end of thirty-five years from the date of publication of the

      work under the grant or at the end of forty years from the date

      of execution of the grant, whichever term ends earlier.
(from http://www.law.cornell.edu/uscode/html/uscode17/usc_sec_17_00000203----0... )
I would be interested to see what other legal basis is alleged as

grounds to rescind a license.
Michael

-

Ah, I am both right and wrong, it seems. Apparently you have to wait anywhere

form 35 to 40 years, and then you only have a five year window. Seems damned

strange to me, but oh well.
(I'd totally forgotten that part of the law - or my mind decided to play

tricks on me.)
DRH
PS: See, I will admit it when I'm shown evidence that I'm wrong :)
--

Dialup is like pissing through a pipette. Slow and excruciatingly painful.

-

OpenBSD does not want this.
OpenBSD wants a world where people do things because they are the

right thing to do.
OpenBSD lets you decide; it doesn't dictate.
someone poo-poos your decision, well, it was your decision.
someone poo-poos you because you do exactly the thing that you are so

afraid of having happen to you, well, it was your decision.
--

jakemsr@sdf.lonestar.org

SDF Public Access UNIX System - http://sdf.lonestar.org

-

What myth?  The myth that Theo understands dual licensing?
	Jeff
-

Reyk's code was never dual licensed, so it's not like it even matters

to the original dispute.
That said, I don't see what exact wording you consider inaccurate.
C.

-

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 
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.
cu

Adrian
-- 
       "Is there not promise of rain?" Ling Tan asked suddenly out

        of the darkness. There had been need of rain for many days.

       "Only a promise," Lao Er said.

                                       Pearl S. Buck - Dragon Seed
-

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 
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
-

Sorry, this has been a thinko on my side:
If noone except you and onoe-san made any contributions to this code

that were non-trivial enough for automatically giving its author a

copyright on his contributions (whatever this means in various 
cu

Adrian
-- 
       "Is there not promise of rain?" Ling Tan asked suddenly out

        of the darkness. There had been need of rain for many days.

       "Only a promise," Lao Er said.

                                       Pearl S. Buck - Dragon Seed
-

FreeBSD doesn't have Reyk's ath(4) HAL, which OpenHAL is based on.
How? FreeBSD doesn't have Reyk's ath(4) HAL from OpenBSD, so there are
FreeBSD's ath(4) code, both the driver and the HAL, is entirely

written by Sam Leffler, who can licence it in whichever way he seems

reasonable. The driver part of Sam's code is also present in OpenBSD,

but the HALs in OpenBSD and FreeBSD are entirely different.
C.

-

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 
Sam also changed the licence of a file additionally containing an
cu

Adrian
-- 
       "Is there not promise of rain?" Ling Tan asked suddenly out

        of the darkness. There had been need of rain for many days.

       "Only a promise," Lao Er said.

                                       Pearl S. Buck - Dragon Seed
-

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.
Shame on you all.
---

Jason Dixon

DixonGroup Consulting

http://www.dixongroup.net
-

"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.

-

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

-

Jiri's patch would have wrongly not only removed the BSD statement from

dual licenced files but also from not dual licenced files.
This was a mistake in this patch (that was never merged into the tree)

neither Jiri nor Alan noticed.
The only disagreement is about the following:
Theo claimed boldly in the email that started this thread on

linux-kernel it would "break the law" to choose one licence for dual

licenced code like the following:
/*      $OpenBSD: ath.c,v 1.63 2007/05/09 16:41:14 reyk Exp $  */

/*      $NetBSD: ath.c,v 1.37 2004/08/18 21:59:39 dyoung Exp $        */
/*-

 * Copyright (c) 2002-2004 Sam Leffler, Errno Consulting

 * All rights reserved.

 *

 * Redistribution and use in source and binary forms, with or without

 * modification, are permitted provided that the following conditions

 * are met:

 * 1. Redistributions of source code must retain the above copyright

 *    notice, this list of conditions and the following disclaimer,

 *    without modification.

 * 2. Redistributions in binary form must reproduce at minimum a disclaimer

 *    similar to the "NO WARRANTY" disclaimer below ("Disclaimer") and any

 *    redistribution must be conditioned upon including a substantially

 *    similar Disclaimer requirement for further binary redistribution.

 * 3. Neither the names of the above-listed copyright holders nor the names

 *    of any contributors may be used to endorse or promote products derived

 *    from this software without specific prior written permission.

 *

 * 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.

 *

 * NO WARRANTY

 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS

 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT

 * LIMITED TO, THE IMPLIED WARRANTIES OF NONINFRINGEMENT, MERCHANTIBILITY

 * AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL

 * THE...
[ message continues ]

You know, we *could* have solved this a *hell* of a lot easier if people quit

flaming about it, and we did something *productive* instead.
Like submit a corrected patch. :)

Dear Valdis,
The idea here is that no patching was needed in the first place --

most of the files are/were BSD-licensed, because they were forked from

OpenBSD.
It is beneficial for the atmosphere of both projects to keep the

licence compatible. If Linux tries to GPL future modifications to

Reyk's code, then OpenBSD would not be able to take back the changes.

But this would not be the case if all modifications to Reyk's code are

continued to be BSD-licensed. This is what this whole issue is about.
My understanding, is that Nick Kossifidis never had a problem with

licensing his changes with a BSD-license, although Jiri Slaby always

used GPLv2.
With the last patch posted by Luis [0], Jiri actually recalled his

original patch and relicensed all of his GPLv2 work under BSD (!);

however, at the very same time, Nick changed his mind, and decided to

relicense his BSD code under GPLv2 (!). (Surprise! Yes, it appears

that both Nick and Jiri decided to switch their licensing positions,

and mutually relicense their work under each other's respective

licence. :) Is everyone ready going to go back and forth now? Does

this whole story still makes any sense to you? :)
I hope that both Nick Kossifidis and Jiri Slaby can agree on licensing

their HAL code with a BSD licence, so that the code remains

licence-compatible with OpenBSD. If there are any unresolved licensing

questions, I personally would be more than happy to answer any such

questions as much as I can, and yes -- I am not a lawyer. :)
Nick, Jiri -- since much of the work on OpenHAL is based on Reyk's

HAL, could you please be so kind as to both agree to licence you

changes in OpenHAL with the same licence as Reyk does in OpenBSD's

ath(4) HAL? This step will be very welcome in the OpenBSD community at

large.
Best regards,

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

-

Oh, silly me.  For some reason, I had it in my head that Jiri's original

patch actually included some real live *code* in addition to the parts that

changed the licensing text... ;)

Why is it so difficult to understand dual licensing?
	Jeff
-

Maybe because Reyk's code was never dual-licensed?
C.

-

Suppose you saw some other variant of *nix that had some code you wanted

to use, but there was a gaping security hole in it. Wouldn't you patch

it before you incorporated it? and would it be your fault if this fix

made the code not work with the original?
We took the code and fixed a gaping security vulnerability that appeared

within the opening comment. We DO care who does what with our code, and

we fully intend to cover our balls.
The problem is yours to fix. If you actually care, use a license that

SAYS you care. Right now there's a big /* I don't give a shit */ on top

of every BSD file. We took you at your word and assumed you didn't. Now

its too late and you suddenly care, don't you?

-

Since when is this *your* code?  Oh that's right, when Jiri decided

to steal it by deleting Reyk's copyright and license.  Oh wait,  
The BSD license, in effect, says that we care about good code.  We

allow anyone to use it.  The only stipulation is that the copyright

and license permissions must remain intact.
Are you trying to be hateful, or are you really this ignorant?
---

Jason Dixon

DixonGroup Consulting

http://www.dixongroup.net
-

Who's "we" and would you mind showing your contributions to the tree?

git doesn't seem to find them in 2.5.0-to-current for some reason...

-

And yet a good portion of Theo's response, in particular his accusations

of Alan Cox exhorting people to break the law, were directly related to

dual licensing.
	Jeff
-

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.
  Luis

-

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.

-

I'm not sure how you conclude that Theo missed the relevant parts --

there were many messages posted to misc@openbsd.org 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.
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.
C.
( from http://lkml.org/lkml/2007/8/28/178 )

-

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 
Is it a legal problem or is it "only" an ethical problem?
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
cu

Adrian
-- 
       "Is there not promise of rain?" Ling Tan asked suddenly out

        of the darkness. There had been need of rain for many days.

       "Only a promise," Lao Er said.

                                       Pearl S. Buck - Dragon Seed
-

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.
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:
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.

-

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
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.
-

Reyk code was never dual licensed!  His code is under truly free 
I think that Theo explained this point clearly quite a few times in the 
Is a simple change in the header files a reason to vindicate the people 
No, because this change is both rude and illegal.
Igor

-

On Sun, 02 Sep 2007 13:20:27 +0200 (CEST)
The code pieces I quoted contained that choice. As far as I am concerned

that is what the discussion was about. 
Alan

-

Jiri's patch touched both files containing BSD-only code by Reyk and 
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 
cu

Adrian
-- 
       "Is there not promise of rain?" Ling Tan asked suddenly out

        of the darkness. There had been need of rain for many days.

       "Only a promise," Lao Er said.

                                       Pearl S. Buck - Dragon Seed
-

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.

-

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

-

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 
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.

-

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.

-

IANAL, but:
Of course. If it wasn't legal, dual BSD/GPL would just be equal

to GPL. Now, dual BSD/GPL equals BSD.
OTOH I'd probable leave the original licence text, something like:
The actual licence conditions:

GPL or BSD or whatever.
Portions of this file were licenced under:

[the original licence text, not valid as a licence for current file]
WRT Atheros driver I'd probably leave the thing as is (i.e., BSD/GPL

= in fact BSD), unless something like 50+% of the code is rewritten -

it's mostly their hard work after all, isn't it? Not legal

requirement, though.

--

Krzysztof Halasa

-

Yes.  This deserves to be reinforced:
There is definite value in sharing the ath5k HAL between OpenBSD and Linux.
	Jeff
-

Of course.  Sharing knowledge and efforts can only improve both the GPL

and BSD licensed code.  It is important in all cases, but becomes critical

when support from manufacturers is limited or even non existent.  In these

cases, shared efforts are required to write successful code.
Cheers,

Igor.

-

This has been pretty interesting for me to watch as I distribute my

isp driver under a dual license (at least the portions of it which are

common with the *BSD and Solaris ports)  that is almost identical to

Sam's verbiage.
I'll admit that I hadn't thought about whether redistribution included

the ability to modify the header (and thus the text of the licensing

as I had written) or not. On balance I'd say I believe that the

arguments for, on redistribution, picking one or the other license

makes sense and honored my general intent.
This allows people who modify the code (and presumably improve it) a

"chef's choice" based on where they're serving the meal.
IANAL, but I believe that none of this keeps me from continuing to put

a dual license on stuff I leave up for distribution, or changing that

to restricting the code to Martian Triathalon winners or what have

you.

-

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

-

We have the ability if the author explicitely allowed it.
This is the licencing text we are talking about:
/*-

 * Copyright (c) 2002-2004 Sam Leffler, Errno Consulting

 * All rights reserved.

 *

 * Redistribution and use in source and binary forms, with or without

 * modification, are permitted provided that the following conditions

 * are met:

 * 1. Redistributions of source code must retain the above copyright

 *    notice, this list of conditions and the following disclaimer,

 *    without modification.

 * 2. Redistributions in binary form must reproduce at minimum a disclaimer

 *    similar to the "NO WARRANTY" disclaimer below ("Disclaimer") and any

 *    redistribution must be conditioned upon including a substantially

 *    similar Disclaimer requirement for further binary redistribution.

 * 3. Neither the names of the above-listed copyright holders nor the names

 *    of any contributors may be used to endorse or promote products derived

 *    from this software without specific prior written permission.

 *

 * 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.

 *

 * NO WARRANTY

 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS

 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT

 * LIMITED TO, THE IMPLIED WARRANTIES OF NONINFRINGEMENT, MERCHANTIBILITY

 * AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL

 * THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLARY,

 * OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF

 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS

 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER

 * IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE)

 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF

 * THE POSSIBILITY OF SUCH DAMAGES.

 *...
[ message continues ]

The plain English in the dual-license text directly contradicts this

fiction.
	Jeff
-

I urge developers to not bait into this and just leave this alone.

Those involved know what they are doing and have a strong team of

attorneys watching their backs. Any *necessary* discussions are be

done privately.
  Luis

-

Err...
I don't understand why you need a lawyer to interpret this document:
/*     $OpenBSD: ar5210.c,v 1.39 2007/04/10 17:47:55 miod Exp $        */
/*

 * Copyright (c) 2004, 2005, 2006, 2007 Reyk Floeter <reyk@openbsd.org>

 *

 * 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.

 */
If you want to add more code to it, and your contribution is

significant, simply add you name next to Reyk's. Where's the problem?
I don't know how licensing could be any simpler than this. Please,

notice, that there are no additional documents (other than the

copyright law) to read here -- _this is the complete licence_! (And

you have to read the copyright law even if you use the GNU GPL.)
C.

-

Oh, and if you look at the OpenBSD CVS you see versions 4 months old

with dozens of contributions by Reyk and with:
/*      $OpenBSD: ath.c,v 1.63 2007/05/09 16:41:14 reyk Exp $  */

/*      $NetBSD: ath.c,v 1.37 2004/08/18 21:59:39 dyoung Exp $        */
/*-

 * Copyright (c) 2002-2004 Sam Leffler, Errno Consulting

 * All rights reserved.

 *

 * Redistribution and use in source and binary forms, with or without

 * modification, are permitted provided that the following conditions

 * are met:

 * 1. Redistributions of source code must retain the above copyright

 *    notice, this list of conditions and the following disclaimer,

 *    without modification.

 * 2. Redistributions in binary form must reproduce at minimum a disclaimer

 *    similar to the "NO WARRANTY" disclaimer below ("Disclaimer") and any

 *    redistribution must be conditioned upon including a substantially

 *    similar Disclaimer requirement for further binary redistribution.

 * 3. Neither the names of the above-listed copyright holders nor the names

 *    of any contributors may be used to endorse or promote products derived

 *    from this software without specific prior written permission.

 *

 * 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.

 *

 * NO WARRANTY

 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS

 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT

 * LIMITED TO, THE IMPLIED WARRANTIES OF NONINFRINGEMENT, MERCHANTIBILITY

 * AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL

 * THE COPYRIGHT HOLDERS OR CONTRIBUTORS BE LIABLE FOR SPECIAL, EXEMPLARY,

 * OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF

 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS

 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER

 * IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDI...
[ message continues ]

Where exactly do you see Reyk's copyright in the above quote?
http://www.openbsd.org/cgi-bin/man.cgi?query=ath&sektion=4#AUTHORS
C.

-

He has automatically a copyright on his contributions if they are 
cu

Adrian
-- 
       "Is there not promise of rain?" Ling Tan asked suddenly out

        of the darkness. There had been need of rain for many days.

       "Only a promise," Lao Er said.

                                       Pearl S. Buck - Dragon Seed
-