Standard template for GPL licensing:
"This program is free software; you can redistribute it and/or modify it
under the terms of the GNU General Public License as published by the
Free Software Foundation; either version 2 of the License, or (at your
option) any later version."
As the text says, the licensee can choose the GPL version at his option,
and he is likely to choose the one with better conditions. So, newer
version can never limit the licensee's right, because he is always free
to choose version 2. Therefore, successor versions can only remove
limitations.
The institution to decide, how the new versions look like, is FSF. Being
totally paranoid, assume the FSF decision makers are infected by a SCO
designed virus to make them publish a new GPL version giving SCO the
right to exploit GPL covered intellectual property. And there is a lot
of the latter. Would be a classical "Duh!" situation.
I'm about to submit a patch for a new cipher mode called LRW, adding new
code/files to the crypto tree. My question is, especially to the
maintainers: Are you going to accept code covered by the terms:
* This program is free software; you can redistribute it and/or modify
* it under the terms of the GNU General Public License as published by
* the Free Software Foundation, version 2 of the License.
^^^^^^^^^^^^^^^^^^^^^^^^^
--
Fruhwirth Clemens <[email protected]> http://clemens.endorphin.org
On Tue, Nov 16, 2004 at 03:08:35PM +0100, Fruhwirth Clemens wrote:
> I'm about to submit a patch for a new cipher mode called LRW, adding new
> code/files to the crypto tree. My question is, especially to the
> maintainers: Are you going to accept code covered by the terms:
>
> * This program is free software; you can redistribute it and/or modify
> * it under the terms of the GNU General Public License as published by
> * the Free Software Foundation, version 2 of the License.
There's already quite some code that's only licensed with GPLv2. Look
for example at arch/arm/common/dmabounce.c which does the same as you
want but with slightly different words:
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* version 2 as published by the Free Software Foundation.
If you are the author, you can choose the license you like.
Erik (IANAL, etc.)
--
+-- Erik Mouw -- http://www.harddisk-recovery.com -- +31 70 370 12 90 --
| Lab address: Delftechpark 26, 2628 XH, Delft, The Netherlands
Fruhwirth Clemens <[email protected]> wrote:
> Standard template for GPL licensing:
>
> "This program is free software; you can redistribute it and/or modify it
> under the terms of the GNU General Public License as published by the
> Free Software Foundation; either version 2 of the License, or (at your
> option) any later version."
[snip]
That's only a recommendation by the FSF; many GPL-covered projects (including
the kernel and my own) do not use it.
The file COPYING in the kernel source says:
Also note that the only valid version of the GPL as far as the kernel
is concerned is _this_ particular version of the license (ie v2, not
v2.2 or v3.x or whatever), unless explicitly otherwise stated.
Charles
--
-----------------------------------------------------------------------
Charles Cazabon <[email protected]>
GPL'ed software available at: http://www.qcc.ca/~charlesc/software/
-----------------------------------------------------------------------
On Tue, 16 Nov 2004, Fruhwirth Clemens wrote:
> I'm about to submit a patch for a new cipher mode called LRW, adding new
> code/files to the crypto tree. My question is, especially to the
> maintainers: Are you going to accept code covered by the terms:
>
> * This program is free software; you can redistribute it and/or modify
> * it under the terms of the GNU General Public License as published by
> * the Free Software Foundation, version 2 of the License.
There are several occurences in the kernel already doing exactly that.
Tim
On Tue, 2004-11-16 at 15:35 +0100, Erik Mouw wrote:
> On Tue, Nov 16, 2004 at 03:08:35PM +0100, Fruhwirth Clemens wrote:
> > I'm about to submit a patch for a new cipher mode called LRW, adding new
> > code/files to the crypto tree. My question is, especially to the
> > maintainers: Are you going to accept code covered by the terms:
> >
> > * This program is free software; you can redistribute it and/or modify
> > * it under the terms of the GNU General Public License as published by
> > * the Free Software Foundation, version 2 of the License.
>
> There's already quite some code that's only licensed with GPLv2. Look
> for example at arch/arm/common/dmabounce.c which does the same as you
> want but with slightly different words:
Thanks, I wasn't sure, what to grep for.
--
Fruhwirth Clemens <[email protected]> http://clemens.endorphin.org
On Tue, 16 Nov 2004, Fruhwirth Clemens wrote:
> I'm about to submit a patch for a new cipher mode called LRW, adding new
> code/files to the crypto tree. My question is, especially to the
> maintainers: Are you going to accept code covered by the terms:
>
> * This program is free software; you can redistribute it and/or modify
> * it under the terms of the GNU General Public License as published by
> * the Free Software Foundation, version 2 of the License.
Sure.
This is what Linus' COPYING file in the top of the kernel tree says too.
- James
--
James Morris
<[email protected]>
On Tue, 16 Nov 2004, Tim Schmielau wrote:
> On Tue, 16 Nov 2004, Fruhwirth Clemens wrote:
>
> > I'm about to submit a patch for a new cipher mode called LRW, adding new
> > code/files to the crypto tree. My question is, especially to the
> > maintainers: Are you going to accept code covered by the terms:
> >
> > * This program is free software; you can redistribute it and/or modify
> > * it under the terms of the GNU General Public License as published by
> > * the Free Software Foundation, version 2 of the License.
>
> There are several occurences in the kernel already doing exactly that.
Indeed. See the main COPYING file for the whole kerrnel. So if you don't
explicitly state otherwise, the kernel by _default_ is v2 only.
Linus
On Maw, 2004-11-16 at 14:40, Charles Cazabon wrote:
> Also note that the only valid version of the GPL as far as the kernel
> is concerned is _this_ particular version of the license (ie v2, not
> v2.2 or v3.x or whatever), unless explicitly otherwise stated.
Version 2 explicitly states
| Each version is given a distinguishing version number. If the Program
| specifies a version number of this License which applies to it and
"any
| later version", you have the option of following the terms and
conditions
| either of that version or of any later version published by the Free
| Software Foundation. If the Program does not specify a version number
of
| this License, you may choose any version ever published by the Free
Software
| Foundation.
The use of this by some kernel people is to issue "this version only"
licenses is unfortunate, ill advised and potentially harmful. Firstly it
isn't remotely clear what it means because the license itself never
talks about this case, only the "or later" case. Secondly it may force
code chunks to be rewritten if the GPL is modified to fix a legal
problem in future and the original author or their estate or company [*]
cannot be traced. Thirdly it may actually be meaningless anyway - the
GPL doesn't talk about "this version only" in any of its text so it may
be an "additional restriction" and thus a void clause.
Think very hard before you use such a statement and if you do please
ensure it has some kind of "unlocking" clause so that if you can't be
contacted someone you trust (eg Linus) is authorised to make that
decision for you. It would be good if other people who've used this
would also execute such a change with Linus.
Alan
[*] This isn't that silly a situation the kernel community is large
enough that more than one of its contributors is sadly deceased. The
company case is even worse. Tracing company owned code from a defunct
company is nigh on impossible and if you do trace it to an official
receiver or the equivalent (such as a bankruptcy court) they may be
legally obliged to extort as much money as possible from the person
wanting to relicense it.
On Maw, 2004-11-16 at 15:46, Linus Torvalds wrote:
> Indeed. See the main COPYING file for the whole kerrnel. So if you don't
> explicitly state otherwise, the kernel by _default_ is v2 only.
Bollocks it is.
All the code I submitted I've submitted under the GPL, the real GPL not
the Linus one. All the code I submitted prior to your change and all the
code I and everyone else submitted prior to that change you have no
authorisation or right to claim ownership and relicensing rights over
beyond those originally granted. As such you have no right to restrict
them or to claim they are GPLv2 only.
Alan
On Wed, 17 Nov 2004, Alan Cox wrote:
>
> All the code I submitted I've submitted under the GPL, the real GPL not
> the Linus one.
Sorry, but the "or any later version" is not even _part_ of the GPL, much
less a "real one". It's part of the _suggested_ _header_ of a file, not
the license. Look at it yourself: it comes clearly after the "END OF TERMS
AND CONDITIONS", and it's in the "How to apply these terms to your new
program".
In other words, if you didn't have that "v2 or later" in your original
patches, they were _always_ just the regular GPLv2.
That said, when I clarified (and I do want to make clear that the header
on the COPYING file is a _clarification_, not a change of license) it, I
told people that if they disagreed with me, they should send in patches
saying "v2 or later" to their own code.
Just because there may have been confused people (like you) who thought
that the "later version" thing was part of the license. It's not. It has
never been. The actual _license_ part is very clear:
If the Program specifies a version number of this License which
applies to it and "any later version", you have the option of
following the terms and conditions either of that version or of
any later version published by the Free Software Foundation.
Note the "IF". Linux _never_ had the "v2 or later" clause, so that "if"
was never an issue, and the clarification on top of the COPYING file
really _is_ just a clarification.
Alan, you need to learn to read, and not make assumptions.
Linus
Alan Cox writes:
> The use of this by some kernel people is to issue "this version only"
> licenses is unfortunate, ill advised and potentially harmful. Firstly it
> isn't remotely clear what it means because the license itself never
> talks about this case, only the "or later" case. Secondly it may force
> code chunks to be rewritten if the GPL is modified to fix a legal
> problem in future and the original author or their estate or company [*]
> cannot be traced. Thirdly it may actually be meaningless anyway - the
> GPL doesn't talk about "this version only" in any of its text so it may
> be an "additional restriction" and thus a void clause.
I have no argument that restricting it to v2 is potentially harmful,
but allowing distribution under GPL v3 is also potentially harmful
since the terms are not yet known, and later v3plus changes would
restrict the whole work to v3plus.
As for the "additional restriction" theory, the FSF apparently does
not think it is a restriction; the GPL says what happens *if* the
program says "or any later version." If the intent were to always
allow newer license, that should be clearly written into the license
rather than making it conditional. The GPL FAQ[1] implies that "or
any later version" is optional. When the issue came up on
debian-legal earlier this year, Dave Turner provided an answer[2].
Michael Poole
[1]- http://www.fsf.org/licenses/gpl-faq.html#VersionTwoOrLater
[2]- http://lists.debian.org/debian-legal/2004/08/msg00821.html
On Mer, 2004-11-17 at 16:33, Michael Poole wrote:
> I have no argument that restricting it to v2 is potentially harmful,
> but allowing distribution under GPL v3 is also potentially harmful
> since the terms are not yet known, and later v3plus changes would
> restrict the whole work to v3plus.
Oh believe me I understand exactly why Linus said "v2 only". The GPL say
"similiar license" but Linus doesn't trust Richard. That's fine, I don't
trust Richard or Linus 8)
If people are going to say "v2 only" they should most definitely make
sure they assign Linus or some similar person the right to revisit that
decision.
Alan
On Mer, 2004-11-17 at 16:28, Linus Torvalds wrote:
> In other words, if you didn't have that "v2 or later" in your original
> patches, they were _always_ just the regular GPLv2.
I did have that.
> That said, when I clarified (and I do want to make clear that the header
> on the COPYING file is a _clarification_, not a change of license) it, I
> told people that if they disagreed with me, they should send in patches
> saying "v2 or later" to their own code.
Well no obligation exists, but please add "All code owned by Alan Cox
and present in this kernel is licensed GPL v2 or later" to your copying
or another appropriate file. (A comment in the code for each one would
be rather messy)
It might be a good idea to figure out how to have a list of contributors
who've said that or "v2 - or if Linus Torvalds so chooses, a later
version"
> Note the "IF". Linux _never_ had the "v2 or later" clause, so that "if"
> was never an issue, and the clarification on top of the COPYING file
> really _is_ just a clarification.
Correction noted. I went and checked 1.2.0 and indeed it says nothing
about versions in that specific top level file.
Alan
On Wed, 17 Nov 2004, Alan Cox wrote:
>
> Well no obligation exists, but please add "All code owned by Alan Cox
> and present in this kernel is licensed GPL v2 or later" to your copying
> or another appropriate file. (A comment in the code for each one would
> be rather messy)
It really doesn't help with a single person, since any code you wrote you
can do anything with _anyway_ (ie you're not even bound by the GPL at all,
since you can re-license it under whatever license you want). So naming
people really doesn't help anybody, while naming _files_ does (because
then _others_ than you can decide to relicense based on that file).
See what I'm saying? You as the author always have full rights, which is
why the "v2 or later" really needs to be attached to _code_, not to a
person. You may know the code you wrote, but very few other people will.
Also, attaching it to the code is the only way to have places with
multiple developers (and let's face it, these days that's pretty much
every corner of the kernel) to note the "compunded" rules.
> It might be a good idea to figure out how to have a list of contributors
> who've said that or "v2 - or if Linus Torvalds so chooses, a later
> version"
Yes, modulo again the fact that I really don't think it's about people but
about files. Not just because of the argument above, either: some people
might really do licensing per file: I've done it myself (ie I allowed my
original stupid math emulator to be relicenced under the BSD license, even
though I would not do that for other things). Also, I'd be willing to do
the "v2 or later" on my code if somebody else has been working on it a lot
and feels strongly about it, even though I do not in _general_ tend to
agree with a carte blanche for the FSF to make up any random license.
For all we know, GPL v3 will really be a good license, and everybody will
agree that we want to upgrade. I just don't _know_, and I'm not willing to
just take something that important on trust.
> > Note the "IF". Linux _never_ had the "v2 or later" clause, so that "if"
> > was never an issue, and the clarification on top of the COPYING file
> > really _is_ just a clarification.
>
> Correction noted. I went and checked 1.2.0 and indeed it says nothing
> about versions in that specific top level file.
Indeed. There are a lot of files that _do_ have the boilerplate thing, but
those obviously aren't in question (although iirc last time when this
detail came up and I pointed out what the real rules wrt that "v2 or
later" were, some people actually _removed_ their thing too, so the
confusion about these things and the preferences can obviously go either
way).
Linus
> As the text says, the licensee can choose the GPL version at his option,
> and he is likely to choose the one with better conditions. So, newer
> version can never limit the licensee's right, because he is always free
> to choose version 2. Therefore, successor versions can only remove
> limitations.
Your logic is totally flawed. Successor versions can certainly add
limitations.
Consider the following hypothetical, GPL version 3 allows you to relicense
the code under the FreeBSD license. Someone relicenses Linux (with lots of
later modification) under the FreeBSD license. Now people who receive the
binaries from this new stream of Linux are not entitled to the source code.
Not that this would ever happen, of course, but if your question is, "what
possible harm could it do", the answer is that new limitations could be put
in the newer licenses and newer code could be released with only the new
license.
When Linux opted to apply the GPL to early versions of Linux, he wasn't
concerned only with protecting that code as it existed at that instant. He
was creating the framework that shapes the future development of Linux into
the future. The "at your option" clause could be used to transfer that
contorl to the FSF.
Suppose GPL version 3 has no requirement that you make the source
available. I can then ship Linux without making any source available at all
by claiming that I'm using that later version at my option.
DS
On Wed, 17 Nov 2004, Alan Cox wrote:
> On Mer, 2004-11-17 at 16:28, Linus Torvalds wrote:
> > That said, when I clarified (and I do want to make clear that the header
> > on the COPYING file is a _clarification_, not a change of license) it, I
> > told people that if they disagreed with me, they should send in patches
> > saying "v2 or later" to their own code.
>
> Well no obligation exists, but please add "All code owned by Alan Cox
> and present in this kernel is licensed GPL v2 or later" to your copying
> or another appropriate file. (A comment in the code for each one would
> be rather messy)
>
> It might be a good idea to figure out how to have a list of contributors
> who've said that or "v2 - or if Linus Torvalds so chooses, a later
> version"
>
How about adding a new field to the entry for each person in the CREDITS
file?
Something like this perhaps :
--- linux-2.6.10-rc2-orig/CREDITS 2004-11-17 01:18:58.000000000 +0100
+++ linux-2.6.10-rc2/CREDITS 2004-11-17 22:04:00.000000000 +0100
@@ -2,8 +2,8 @@
contributed to the Linux project. It is sorted by name and
formatted to allow easy grepping and beautification by
scripts. The fields are: name (N), email (E), web-address
- (W), PGP key ID and fingerprint (P), description (D), and
- snail-mail address (S).
+ (W), PGP key ID and fingerprint (P), description (D),
+ license notes (L), and snail-mail address (S).
Thanks,
Linus
@@ -1608,6 +1608,8 @@
N: Jesper Juhl
E: [email protected]
D: Various small janitor fixes, cleanups etc.
+L: Contributed code is licensed under the GPL v2, or at your option any
+L: later version.
S: Lemnosvej 1, 3.tv
S: 2300 Copenhagen S
S: Denmark
That doesn't account for everyone who has ever contributed code, but it
does cover quite a few people if we can get anyone who's in there to add a
similar line and encourage people who are not represented in CREDITS to
submit an entry.
Of course there will still be some people it is impossible to reach (some
dead, some just non-communicative) - some of those missing people may have
left appropriate licensing info in source file comments, list emails or
similar, so maybe that info could be used for their entry. For everyone
else who either do not want to be listed in CREDITS or who are "missing"
and have not left any clues as to their licensing preferences, well, I
guess the COPYING file would then apply to them - or?
--
Jesper Juhl
On Nov 17, 2004, at 20:04, David Schwartz wrote:
> Your logic is totally flawed. Successor versions can certainly add
> limitations.
Your logic is equally flawed.
> Consider the following hypothetical, GPL version 3 allows you to
> relicense
> the code under the FreeBSD license. Someone relicenses Linux (with
> lots of
> later modification) under the FreeBSD license. Now people who receive
> the
> binaries from this new stream of Linux are not entitled to the source
> code.
s/relicense/distribute/g; You can't relicense code without owning the
original
copyright. You _can_ however, receive a license to distribute.
If I receive a binary from them, I receive the original "dual" license.
This means
that anyone who receives it may license it using GPL version "2" or, at
their
option, any later version. This means that when I get some binaries
from some
random company that used your mythical version 3 to distribute under the
terms of any BSD license, I _also_ receive a license to the same code
under
the terms of GPL v2 (at my option :-D). This means that since my
license from
company X is GPL v2 (at my option), I must receive sources under GPL
v2. In
practice this means that you _can't_ change much of what the GPL says in
future versions, at least as far as I can see, without causing said
future version
to be legally invalid.
> Suppose GPL version 3 has no requirement that you make the source
> available. I can then ship Linux without making any source available
> at all
> by claiming that I'm using that later version at my option.
And I can equally legally demand the sources under GPL version 2, at my
option, and you will be forced to give them to me, not at your option
:-D.
Cheers,
Kyle Moffett
-----BEGIN GEEK CODE BLOCK-----
Version: 3.12
GCM/CS/IT/U d- s++: a17 C++++>$ UB/L/X/*++++(+)>$ P+++(++++)>$
L++++(+++) E W++(+) N+++(++) o? K? w--- O? M++ V? PS+() PE+(-) Y+
PGP+++ t+(+++) 5 X R? tv-(--) b++++(++) DI+ D+ G e->++++$ h!*()>++$ r
!y?(-)
------END GEEK CODE BLOCK------
On Wednesday 17 November 2004 09:12 pm, Kyle Moffett wrote:
> > ??????Suppose GPL version 3 has no requirement that you make the source
> > available. I can then ship Linux without making any source available
> > at all
> > by claiming that I'm using that later version at my option.
>
> And I can equally legally demand the sources under GPL version 2, at my
> option, and you will be forced to give them to me, not at your option
> :-D.
Yes, but the additions I made will be GPLv3 only (no sources). So I will
gladly provide you with the sources off kernel.org and you will never see
the sources for a driver I wrote. See?
--
Dmitry
On Nov 17, 2004, at 21:50, Dmitry Torokhov wrote:
> Yes, but the additions I made will be GPLv3 only (no sources). So I
> will
> gladly provide you with the sources off kernel.org and you will never
> see
> the sources for a driver I wrote. See?
What about section 2, subsection B of the GPL:
> b) You must cause any work that you distribute or publish, that in
> whole or in part contains or is derived from the Program or any
> part thereof, to be licensed as a whole at no charge to all third
> parties under the terms of this License.
"this License", would refer to the specific version of the license.
This means
that since the original code is dual-licensed under both versions, any
code
that is a derivative work must _also_ be dual-licensed (This assumes of
course that the other license has a similar clause). In any case, any
work
derived from a GPLv2'ed work must also be licensable under the GPLv2.
Therefore, my request for _your_ source-code under the GPLv2 is
perfectly
valid.
Cheers,
Kyle Moffett
-----BEGIN GEEK CODE BLOCK-----
Version: 3.12
GCM/CS/IT/U d- s++: a17 C++++>$ UB/L/X/*++++(+)>$ P+++(++++)>$
L++++(+++) E W++(+) N+++(++) o? K? w--- O? M++ V? PS+() PE+(-) Y+
PGP+++ t+(+++) 5 X R? tv-(--) b++++(++) DI+ D+ G e->++++$ h!*()>++$ r
!y?(-)
------END GEEK CODE BLOCK------
On Wednesday 17 November 2004 10:11 pm, Kyle Moffett wrote:
> What about section 2, subsection B of the GPL:
> > b) You must cause any work that you distribute or publish, that in
> > ? ? whole or in part contains or is derived from the Program or any
> > ? ? part thereof, to be licensed as a whole at no charge to all third
> > ? ? parties under the terms of this License.
>
> "this License", would refer to the specific version of the license. ?
> This means
> that since the original code is dual-licensed under both versions, any
> code
> that is a derivative work must _also_ be dual-licensed
No, not at all. I need only _one_ license to use the code. If original
code was dual-licensed, let's say GPL/BSD, I can chose to completely
ignore GPL part and treat the code as if it was always released BSD only.
Why do you think several components, like ACPI, are dual-licensed?
Intel chose to do that so they can take ACPI interpreter implementation
and use it somewhere else, in non-GPL environment.
Q9. Under what licensing is the source released?
A9. ACPI CA can be licensed under the GNU General Public License or via a
separate license that may be more favorable to commercial OSVs. Please
see the source code license header for specifics.
> (This assumes of course that the other license has a similar clause).
> In any case, any work
> derived from a GPLv2'ed work must also be licensable under the GPLv2.
> Therefore, my request for _your_ source-code under the GPLv2 is
> perfectly
> valid.
See above. For me it was never GPLv2, if was BSD all the way and my new
code I can chose to make BSD only.
--
Dmitry
> On Nov 17, 2004, at 21:50, Dmitry Torokhov wrote:
> > Yes, but the additions I made will be GPLv3 only (no sources). So I
> > will
> > gladly provide you with the sources off kernel.org and you will never
> > see
> > the sources for a driver I wrote. See?
>
> What about section 2, subsection B of the GPL:
> > b) You must cause any work that you distribute or publish, that in
> > whole or in part contains or is derived from the Program or any
> > part thereof, to be licensed as a whole at no charge to all third
> > parties under the terms of this License.
>
> "this License", would refer to the specific version of the license.
Precisely!
> This means
> that since the original code is dual-licensed under both versions, any
> code
> that is a derivative work must _also_ be dual-licensed (This assumes of
> course that the other license has a similar clause). In any case, any
> work
> derived from a GPLv2'ed work must also be licensable under the GPLv2.
> Therefore, my request for _your_ source-code under the GPLv2 is
> perfectly
> valid.
No, the code being dual-licensed means you get to choose which license you
wish to comply with, not that you must comply with both. So long as you
comply with either license, you have the right to do whatever the license
you comply with gives you the right to do.
Let's suppose, hypothetically, that you were right. Then let's further
suppose that GPL version 3 did not let you modify the source code at all.
Would you argue that all code with the "at your option" clause now can't be
modified? (Since modifying it would violate the GPLv3 license, and you seem
to think you can't violate either license.)
DS