Re: non-free firmware in kernel modules, aggregation and unclear copyright notice.

Re: non-free firmware in kernel modules, aggregation and unclear copyright notice.

[Marco Colombo]

[I'm not subscribed, so this in not a real reply - sorry if it breaks
 threading somehow.]

Sven Luther wrote:
> The ftp-master are the ones reviewing the licencing problems, and they
are the
> ones handling the infrastructure, and putting their responsability on the
> stake. If they feel that some piece of software has dubious legal issues which
> come at a risk of having them personally come on the receiving end of a legal
> case, then they will say, no, we don't distribute this software, and that is
> the end of it.

I've been following the whole discussion (including later messages),
but I'm still missing one point. You seem to have investigated a lot 
on the subject, so I'll ask you. I don't get what real legal issues
distributors may have.

Let me explain with an example. Lets say:

A - is the Author (or rights owner) of the software (GPL'ed);
B - is an user, who got the a copy of the software from A;
C - is another user, who got a copy indirectly, that is from a  
    distributor;
D - is the distributor C got the copy from.
 
Now, IANAL at all. But it seems to me that B has the right to _use_ the
software by means of GPL. As long as A thinks B doesn't break GPL, B is
fine. All B needs to do is to fulfill GPL conditions (as a user, there's
little to do).

C also has the right to use the software, in a very similar way. As long
as A thinks C doesn't break GPL, C is fine.

D has the right to distribute the software, under GPL terms. As long as
A thinks D doesn't break GPL, D is fine.

Now. It seems to me that the relationship between D (distributor) and C
(target of the distribution) is _not_ regulated by GPL at all. GPL is a
license, the _owner_ of the rights (A) and the recipient of some rights
(C, as an user) are the only subjects. D _owns_ no rights on the
software, so can't grant any to C. There's no GPL between D and C.

So, even if C comes to think D is breaking GPL, all C can do is notify
A. The GPL D is supposedly breaking is an agreement between A and D
only. On which basis may C sue D? For breaking what agreement? It's up
to A to sue D for breaking GPL.

What is the risk for D, if D is distributing the source of the software
_exactly_ in the form A publicly provides it? It's not up to D to
produce the source, all D has to do is to provide verbatim copies of
it to anyone D distributes the software to, on request.

Does is really matter if C thinks the source being incomplete,
or missing? C can take the issue up with A (by means of the GPL that
exists between A and C), but not with D, since there's no GPL between
D and C. C is in the same position of B. If the source is incomplete,
they may ask A to comply to the GPL, but not D. D made no promises to
them.  

So, as long as they don't modify the source, distributors are safe.
No one can ask them to provide the "right" source, but A. And "right"
means "right for A", of course, when it's A asking, by definition.

What am I missing?

TIA,
.TM.

Re: non-free firmware in kernel modules, aggregation and unclear copyright notice.

[Sven Luther]

On Mon, Apr 11, 2005 at 06:12:22PM +0200, Marco Colombo wrote:
> [I'm not subscribed, so this in not a real reply - sorry if it breaks
>  threading somehow.]
> 
> Sven Luther wrote:
> > The ftp-master are the ones reviewing the licencing problems, and they
> are the
> > ones handling the infrastructure, and putting their responsability on the
> > stake. If they feel that some piece of software has dubious legal issues which
> > come at a risk of having them personally come on the receiving end of a legal
> > case, then they will say, no, we don't distribute this software, and that is
> > the end of it.
> 
> I've been following the whole discussion (including later messages),
> but I'm still missing one point. You seem to have investigated a lot 
> on the subject, so I'll ask you. I don't get what real legal issues
> distributors may have.
> 
> Let me explain with an example. Lets say:
> 
> A - is the Author (or rights owner) of the software (GPL'ed);
> B - is an user, who got the a copy of the software from A;
> C - is another user, who got a copy indirectly, that is from a  
>     distributor;
> D - is the distributor C got the copy from.
>  
> Now, IANAL at all. But it seems to me that B has the right to _use_ the
> software by means of GPL. As long as A thinks B doesn't break GPL, B is
> fine. All B needs to do is to fulfill GPL conditions (as a user, there's
> little to do).
> 
> C also has the right to use the software, in a very similar way. As long
> as A thinks C doesn't break GPL, C is fine.
> 
> D has the right to distribute the software, under GPL terms. As long as
> A thinks D doesn't break GPL, D is fine.
> 
> Now. It seems to me that the relationship between D (distributor) and C
> (target of the distribution) is _not_ regulated by GPL at all. GPL is a
> license, the _owner_ of the rights (A) and the recipient of some rights
> (C, as an user) are the only subjects. D _owns_ no rights on the
> software, so can't grant any to C. There's no GPL between D and C.

I think you are missing the point. D get's a licence from A, the GPL, and this
licence includes a licence, not on use, but on redistribution, and the act of
D distributing the copy to C is covered by it. In a sense A allows D to
distribute the software under the GPL to C. Now, D is only allowed to do this
distribution if he also distribute the source code of it, which he can't do
for the firmware. 

Notice also the fact that there are so many contributors to the linux kernel
in effect means that there is nobody with the full rights as A, but only a
multitude of people in the D case.

> So, even if C comes to think D is breaking GPL, all C can do is notify
> A. The GPL D is supposedly breaking is an agreement between A and D
> only. On which basis may C sue D? For breaking what agreement? It's up
> to A to sue D for breaking GPL.

This is indeed an interpretation. I am not sure myself if a user receiving
GPLed software in binary only fashion as is the case here can sue either D or
A to get access to that source code.

Now you could argue that any number of authors of GPLed bits of the linux
kernel could sue D for distributing their software as a derived work of the
binary-only bit, and the fact that D doesn't distribute the source code to the
binary bit voids any other right allowed him by the GPL, and thus he has no
right to do the distribution at all. The GPL is very clear on this topic.

> What is the risk for D, if D is distributing the source of the software
> _exactly_ in the form A publicly provides it? It's not up to D to
> produce the source, all D has to do is to provide verbatim copies of
> it to anyone D distributes the software to, on request.

Imagine one of those companies got bought up by some predatory company who
wishes us (linux, debian, redhat/suse, whoever) harm. They would then be able
to sue for damage or prejudice or whatever. And given what i have heard about
the uncertainities of the Alteon ownership, this seems indeed like a plausible
scenario, which could result in a SCO bis case.

This is the scenario i want to avoid by explicitly stating the relationships
of all copyright issues of those firmware blobs.

> Does is really matter if C thinks the source being incomplete,
> or missing? C can take the issue up with A (by means of the GPL that
> exists between A and C), but not with D, since there's no GPL between
> D and C. C is in the same position of B. If the source is incomplete,
> they may ask A to comply to the GPL, but not D. D made no promises to
> them.  

/me wonders if C then holds an illegal copy of the software, and can then be
prosecuted for piracy :)

Friendly,

Sven Luther

Re: non-free firmware in kernel modules, aggregation and unclear copyright notice.

[Marco Colombo]

On Mon, 2005-04-11 at 18:25 +0200, Sven Luther wrote:
> On Mon, Apr 11, 2005 at 06:12:22PM +0200, Marco Colombo wrote:
[...]
> > A - is the Author (or rights owner) of the software (GPL'ed);
> > B - is an user, who got the a copy of the software from A;
> > C - is another user, who got a copy indirectly, that is from a  
> >     distributor;
> > D - is the distributor C got the copy from.
[...]
> > Now. It seems to me that the relationship between D (distributor) and C
> > (target of the distribution) is _not_ regulated by GPL at all. GPL is a
> > license, the _owner_ of the rights (A) and the recipient of some rights
> > (C, as an user) are the only subjects. D _owns_ no rights on the
> > software, so can't grant any to C. There's no GPL between D and C.
> 
> I think you are missing the point. D get's a licence from A, the GPL, and this
> licence includes a licence, not on use, but on redistribution, and the act of
> D distributing the copy to C is covered by it. In a sense A allows D to
> distribute the software under the GPL to C. Now, D is only allowed to do this
> distribution if he also distribute the source code of it, which he can't do
> for the firmware. 

I think only a lawyer can answer here. What I'm saying is that the
license always comes from the copyright owner, that is A.
Sublicensing is not covered by GPL. Distribution is not sublicensing.
Quoting GPL itself:

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. ...

The wording is clear, the license is between A and C.
There's no license between D and C. There's no way C can enforce
anything on D (well, not on GPL basis).

> Notice also the fact that there are so many contributors to the linux kernel
> in effect means that there is nobody with the full rights as A, but only a
> multitude of people in the D case.

In this case, A is clearly the author (onwer of rights) of the firmware.
D is fine on respect of the other A's, since their source is actually 
(and clearly) there. It's the missing source case we're considering
and the number of A's is quite small, the copyright owners of firmware
images. Those A's are easily identified, and perfectly able to act.

> > So, even if C comes to think D is breaking GPL, all C can do is notify
> > A. The GPL D is supposedly breaking is an agreement between A and D
> > only. On which basis may C sue D? For breaking what agreement? It's up
> > to A to sue D for breaking GPL.
> 
> This is indeed an interpretation. I am not sure myself if a user receiving
> GPLed software in binary only fashion as is the case here can sue either D or
> A to get access to that source code.

The point is, if A states (even implicitly) D is distributing the right
source, there's nothing C can do to D. D is not breaking GPL, as long A
says so and it's A granting D the right to distribute. There's no way C
can prevent D from distributing A's software, if A is fine with it.
It's up to A to decide if GPL conditions are met by D.

Maybe mine it's only one interpretation. But I can't see any other.

> Now you could argue that any number of authors of GPLed bits of the linux
> kernel could sue D for distributing their software as a derived work of the
> binary-only bit, and the fact that D doesn't distribute the source code to the
> binary bit voids any other right allowed him by the GPL, and thus he has no
> right to do the distribution at all. The GPL is very clear on this topic.

We're not talking of that case. D _is_ actually distributing the right
source, according to A. It's C that is unsatisfied with it.

> > What is the risk for D, if D is distributing the source of the software
> > _exactly_ in the form A publicly provides it? It's not up to D to
> > produce the source, all D has to do is to provide verbatim copies of
> > it to anyone D distributes the software to, on request.
> 
> Imagine one of those companies got bought up by some predatory company who
> wishes us (linux, debian, redhat/suse, whoever) harm. They would then be able
> to sue for damage or prejudice or whatever. And given what i have heard about
> the uncertainities of the Alteon ownership, this seems indeed like a plausible
> scenario, which could result in a SCO bis case.

I'm not following. Are you saying what if A is bought? That is
different. Well GPL is quite clear:

1. You may copy and distribute verbatim copies of the Program's source
code as you receive it, ...

If D is distributing the source as received from A, D is in full
compliance. How could A sue D? If A distributed incomplete source
in the first place, it's not D's fault for sure. Do you really think
the following scenario is likely:

A to D: you must distribute the complete source, or the license will be
        terminated!
D to A: gimme the complete source, and I'll distribute it.
A to D: no, I'm not willing to give you the full source of my firmware,
        but you must distribute it anyway!

That, in court? Is this really what you're afraid of?
The outcome is, very likely A will be forced to release the full source.
(and D forced to distribute it, but all D's we're talking of here are
very happy with the full disclosure scenario, aren't they?)

> This is the scenario i want to avoid by explicitly stating the relationships
> of all copyright issues of those firmware blobs.

I don't see this scenario nowhere close to likely. Of course, A can 
always sue any B, C, or D for whatever reason. It's very unlikely
A will sue anyone in full compliance of GPL, but it's possible.
There's nothing we can do about it. But there's no reason to worry
either.

As for the matter of explicit copyright notices, I can only agree.
They won't harm for sure. From a purist standpoint, you're right. And
I _am_ a purist. B-) 

> > Does is really matter if C thinks the source being incomplete,
> > or missing? C can take the issue up with A (by means of the GPL that
> > exists between A and C), but not with D, since there's no GPL between
> > D and C. C is in the same position of B. If the source is incomplete,
> > they may ask A to comply to the GPL, but not D. D made no promises to
> > them.  
> 
> /me wonders if C then holds an illegal copy of the software, and can then be
> prosecuted for piracy :)

No, because GPL explicitly states that:

4. You may not copy, modify, sublicense, or distribute the Program
except as expressly provided under this License. Any attempt otherwise
to copy, modify, sublicense or distribute the Program is void, and will
automatically terminate your rights under this License. However, parties
who have received copies, or rights, from you under this License will
not have their licenses terminated so long as such parties remain in
full compliance.

Note also that GPL says nothing about how you get your copy. You can
get it while hanging from the ceiling ala Mission Impossible, but if
the software is GPL'ed, then your license is valid. The action may
still be illegal, but that's another matter: you _can_ use the software
(even if in jail). B-)

> 
> Friendly,
> 
> Sven Luther

Have a nice day,
.TM.

Re: non-free firmware in kernel modules, aggregation and unclear copyright notice.

[Sven Luther]

On Mon, Apr 11, 2005 at 10:54:50PM +0200, Marco Colombo wrote:
> In this case, A is clearly the author (onwer of rights) of the firmware.
> D is fine on respect of the other A's, since their source is actually 
> (and clearly) there. It's the missing source case we're considering
> and the number of A's is quite small, the copyright owners of firmware
> images. Those A's are easily identified, and perfectly able to act.

Well, i am not sure with your interpretation, but even if you where right, we
have no guarantee that A will continue being lenient, and no guarantee that A
will not start suing D or whoever for illegally distributing his stuff without
sources.

> > > So, even if C comes to think D is breaking GPL, all C can do is notify
> > > A. The GPL D is supposedly breaking is an agreement between A and D
> > > only. On which basis may C sue D? For breaking what agreement? It's up
> > > to A to sue D for breaking GPL.
> > 
> > This is indeed an interpretation. I am not sure myself if a user receiving
> > GPLed software in binary only fashion as is the case here can sue either D or
> > A to get access to that source code.
> 
> The point is, if A states (even implicitly) D is distributing the right
> source, there's nothing C can do to D. D is not breaking GPL, as long A

So, i get some random bit of GPLed software, i add a module or some code to
it, i distribute that code in binary format only, and claim that i have used
an hex editor to write it, or simply that it is the 'right' source.

I have some serious doubts that i will not get sued by all the authors of the
original GPLed work if i were to do that, and rightly so.

> says so and it's A granting D the right to distribute. There's no way C
> can prevent D from distributing A's software, if A is fine with it.
> It's up to A to decide if GPL conditions are met by D.

Even in that case, you still need explicit permission of A, and all the other
copyright holders of the rest of the GPLed work, to give you an explicit
exception to link with this non-free bit of code.

> Maybe mine it's only one interpretation. But I can't see any other.
> 
> > Now you could argue that any number of authors of GPLed bits of the linux
> > kernel could sue D for distributing their software as a derived work of the
> > binary-only bit, and the fact that D doesn't distribute the source code to the
> > binary bit voids any other right allowed him by the GPL, and thus he has no
> > right to do the distribution at all. The GPL is very clear on this topic.
> 
> We're not talking of that case. D _is_ actually distributing the right
> source, according to A. It's C that is unsatisfied with it.

No. The source code is clearly the prefered form of modification, not some
random intermediate state A may be claiming is source.

> > > What is the risk for D, if D is distributing the source of the software
> > > _exactly_ in the form A publicly provides it? It's not up to D to
> > > produce the source, all D has to do is to provide verbatim copies of
> > > it to anyone D distributes the software to, on request.
> > 
> > Imagine one of those companies got bought up by some predatory company who
> > wishes us (linux, debian, redhat/suse, whoever) harm. They would then be able
> > to sue for damage or prejudice or whatever. And given what i have heard about
> > the uncertainities of the Alteon ownership, this seems indeed like a plausible
> > scenario, which could result in a SCO bis case.
> 
> I'm not following. Are you saying what if A is bought? That is
> different. Well GPL is quite clear:
> 
> 1. You may copy and distribute verbatim copies of the Program's source
> code as you receive it, ...
> 
> If D is distributing the source as received from A, D is in full
> compliance. How could A sue D? If A distributed incomplete source
> in the first place, it's not D's fault for sure. Do you really think
> the following scenario is likely:
> 
> A to D: you must distribute the complete source, or the license will be
>         terminated!
> D to A: gimme the complete source, and I'll distribute it.
> A to D: no, I'm not willing to give you the full source of my firmware,
>         but you must distribute it anyway!

The result is that the code in question has to be stopped from being
distributed by D. But the case here is different, since A is not the sole
copyright owner, so he doesn't get to set random interpretations of what is
source code. 

> That, in court? Is this really what you're afraid of?
> The outcome is, very likely A will be forced to release the full source.
> (and D forced to distribute it, but all D's we're talking of here are
> very happy with the full disclosure scenario, aren't they?)

Imagine A refusing to give away the source code, and D is ordered to remove
the incriminated code it is illegally distributing from all its servers, and
recall all those thousands of CD and DVD isos containing the code it
distributed, and being fined for each day it doesn't do so ? 

> > This is the scenario i want to avoid by explicitly stating the relationships
> > of all copyright issues of those firmware blobs.
> 
> I don't see this scenario nowhere close to likely. Of course, A can 
> always sue any B, C, or D for whatever reason. It's very unlikely
> A will sue anyone in full compliance of GPL, but it's possible.
> There's nothing we can do about it. But there's no reason to worry
> either.

So, we don't take the risk and don't distribute it. If A is not ready to put a
couple of lines of disclaimer on his work explaining the copyright and
licencing issues, then we are better of not distributing its code, which is
what debian will do.

> As for the matter of explicit copyright notices, I can only agree.
> They won't harm for sure. From a purist standpoint, you're right. And
> I _am_ a purist. B-) 

:)

> > > Does is really matter if C thinks the source being incomplete,
> > > or missing? C can take the issue up with A (by means of the GPL that
> > > exists between A and C), but not with D, since there's no GPL between
> > > D and C. C is in the same position of B. If the source is incomplete,
> > > they may ask A to comply to the GPL, but not D. D made no promises to
> > > them.  
> > 
> > /me wonders if C then holds an illegal copy of the software, and can then be
> > prosecuted for piracy :)
> 
> No, because GPL explicitly states that:
> 
> 4. You may not copy, modify, sublicense, or distribute the Program
> except as expressly provided under this License. Any attempt otherwise
> to copy, modify, sublicense or distribute the Program is void, and will
> automatically terminate your rights under this License. However, parties
> who have received copies, or rights, from you under this License will
> not have their licenses terminated so long as such parties remain in
> full compliance.

Yeah, but who know what mad laws will be passed to repress piracy which will
make this void.

> Note also that GPL says nothing about how you get your copy. You can
> get it while hanging from the ceiling ala Mission Impossible, but if
> the software is GPL'ed, then your license is valid. The action may
> still be illegal, but that's another matter: you _can_ use the software
> (even if in jail). B-)

I am not sure. If i where to get a copy of windows, and manage to install it
without clicking on the "i agree" button, does that make it a legal copy of
windows to use ? 

Friendly,

Sven Luther

Re: non-free firmware in kernel modules, aggregation and unclear copyright notice.

[Marco Colombo]

On Mon, 11 Apr 2005, Sven Luther wrote:

> On Mon, Apr 11, 2005 at 10:54:50PM +0200, Marco Colombo wrote:
>> In this case, A is clearly the author (onwer of rights) of the firmware.
>> D is fine on respect of the other A's, since their source is actually
>> (and clearly) there. It's the missing source case we're considering
>> and the number of A's is quite small, the copyright owners of firmware
>> images. Those A's are easily identified, and perfectly able to act.
>
> Well, i am not sure with your interpretation, but even if you where right, we
> have no guarantee that A will continue being lenient, and no guarantee that A
> will not start suing D or whoever for illegally distributing his stuff without
> sources.

Let's keep things separated. I'm saying that the only one that may
sue D is A, not C. If we agree on this, we may abandon the case of a third
party sueing D.

As for threats coming from A, IMHO D is safe as long as he distributes
what A claims the source is, even if it's a hex string. In no world
A can publicly state "this is the source" and then sue D because
"no, that's not the source" (assuming D is copying it verbatim).

>>>> So, even if C comes to think D is breaking GPL, all C can do is notify
>>>> A. The GPL D is supposedly breaking is an agreement between A and D
>>>> only. On which basis may C sue D? For breaking what agreement? It's up
>>>> to A to sue D for breaking GPL.
>>>
>>> This is indeed an interpretation. I am not sure myself if a user receiving
>>> GPLed software in binary only fashion as is the case here can sue either D or
>>> A to get access to that source code.
>>
>> The point is, if A states (even implicitly) D is distributing the right
>> source, there's nothing C can do to D. D is not breaking GPL, as long A
>
> So, i get some random bit of GPLed software, i add a module or some code to
> it, i distribute that code in binary format only, and claim that i have used
> an hex editor to write it, or simply that it is the 'right' source.
>
> I have some serious doubts that i will not get sued by all the authors of the
> original GPLed work if i were to do that, and rightly so.

No. Please don't throw irrelevant matters in. D is not modifing the
software at all. D is a mere distributor. We're not addressing issues
related to modification, since no one is going to modify the firmware
anyway. This is not a general discussion on GPL. Issues related to
modification do not belong to this thread, which already very close
to off topic on l-k.

Which reminds me. The only reason why this thread belongs here, IMHO,
it's because when it comes to GPL, it really doesn't matter what
FSF's interpretation is, or anyone else's. The authors are choosing
GPL as a license, so _thier_ interpretation is what really matters.

>> says so and it's A granting D the right to distribute. There's no way C
>> can prevent D from distributing A's software, if A is fine with it.
>> It's up to A to decide if GPL conditions are met by D.
>
> Even in that case, you still need explicit permission of A, and all the other
> copyright holders of the rest of the GPLed work, to give you an explicit
> exception to link with this non-free bit of code.

Yes, but it does not apply to our case here. There's no "all other
copyright holders". _You_ stated that the firmware is included by mere
aggregation, so there's no other holders involved. We're talking
about the firmware case. A is one or two well identified subjects.
And A wrote it is GPL'ed. Whether you agree or not, that's the licence
A chose. A placed the copyright notice.

The licence is a matter between A and D. A may sue D and D may (less
likely) sue A, if conditions are not met. I'm not sure at all GPL
is enforceable by D upon A. Let's assume it is, for sake of discussion,
anyway.

>>> Now you could argue that any number of authors of GPLed bits of the linux
>>> kernel could sue D for distributing their software as a derived work of the
>>> binary-only bit, and the fact that D doesn't distribute the source code to the
>>> binary bit voids any other right allowed him by the GPL, and thus he has no
>>> right to do the distribution at all. The GPL is very clear on this topic.
>>
>> We're not talking of that case. D _is_ actually distributing the right
>> source, according to A. It's C that is unsatisfied with it.
>
> No. The source code is clearly the prefered form of modification, not some
> random intermediate state A may be claiming is source.

In this context, it is. Only A may sue D for not distributing the source.
Whatever D distributes, D has to make A happy. If A is happy with D
distributing `dd if=/dev/random count=1` as source, no one can stop D
from doing that. Keep in mind A is the copyright holder. He grants
rights to third parties. No one but A can remove them.

[...]
>> I'm not following. Are you saying what if A is bought? That is
>> different. Well GPL is quite clear:
>>
>> 1. You may copy and distribute verbatim copies of the Program's source
>> code as you receive it, ...
>>
>> If D is distributing the source as received from A, D is in full
>> compliance. How could A sue D? If A distributed incomplete source
>> in the first place, it's not D's fault for sure. Do you really think
>> the following scenario is likely:
>>
>> A to D: you must distribute the complete source, or the license will be
>>         terminated!
>> D to A: gimme the complete source, and I'll distribute it.
>> A to D: no, I'm not willing to give you the full source of my firmware,
>>         but you must distribute it anyway!
>
> The result is that the code in question has to be stopped from being
> distributed by D. But the case here is different, since A is not the sole
> copyright owner, so he doesn't get to set random interpretations of what is
> source code.

Definitely I'm not following. How can D be ordered to stop distributing
the software if A refuses to give the source? A is in violation. And
even if GPL can't be enforced that way, by no means D is in violation
when it's A that it's preventing D from complying.

The only way this can work the way you say is if A and D had a private
agreement, and no proof can be made by D this agreement ever existed.

But A released the software in public, and GPL'ed it. If A wants to keep
part of the source obscured, he had better to stay well away from any
court.

>> That, in court? Is this really what you're afraid of?
>> The outcome is, very likely A will be forced to release the full source.
>> (and D forced to distribute it, but all D's we're talking of here are
>> very happy with the full disclosure scenario, aren't they?)
>
> Imagine A refusing to give away the source code, and D is ordered to remove
> the incriminated code it is illegally distributing from all its servers, and
> recall all those thousands of CD and DVD isos containing the code it
> distributed, and being fined for each day it doesn't do so ?

Sorry, this is nonsense. D is well willing to distribute the source.
In this case, he _is_ distributing what A publicly stated to be the source.
That's all. You say what if A changed his mind about what the source is?
Well since it's still GPL'ed (we agree A can't chance the licence on the
fly, right?) A is to provide the new, complete version of the source.
Then, if and only if D refuses to distribute _that_ source, D is in
violation. GPL requires D to distribute source _as received by A_.
If A refuses to give the source, the specific requirement is not
enforceable.

>>> This is the scenario i want to avoid by explicitly stating the relationships
>>> of all copyright issues of those firmware blobs.
>>
>> I don't see this scenario nowhere close to likely. Of course, A can
>> always sue any B, C, or D for whatever reason. It's very unlikely
>> A will sue anyone in full compliance of GPL, but it's possible.
>> There's nothing we can do about it. But there's no reason to worry
>> either.
>
> So, we don't take the risk and don't distribute it. If A is not ready to put a
> couple of lines of disclaimer on his work explaining the copyright and
> licencing issues, then we are better of not distributing its code, which is
> what debian will do.

There no such a risk. A already placed a copyright notice on his work.
Is it the wrong one? Who cares? I'd even say that once he released it
under GPL, he can't take it back.

This is a close to impossible scenario, so why worry about it?
Anyone may sue you because he doesn't like your name. Or you logo.
Or the name of your server. The risk is by no means lower than the
case you pictured. Why not stopping being debian at all, then?

[...]
> Yeah, but who know what mad laws will be passed to repress piracy which will
> make this void.

Oh, that's totally another matter. I'm not ever sure GPL is a valid
licence for software here in Italy as regards to our anti-piracy law.
Which is criminal law.

>> Note also that GPL says nothing about how you get your copy. You can
>> get it while hanging from the ceiling ala Mission Impossible, but if
>> the software is GPL'ed, then your license is valid. The action may
>> still be illegal, but that's another matter: you _can_ use the software
>> (even if in jail). B-)
>
> I am not sure. If i where to get a copy of windows, and manage to install it
> without clicking on the "i agree" button, does that make it a legal copy of
> windows to use ?

Come on, please, do not mix things up. I've said GPL'ed software. Last time
I checked, Windows was not GPL'ed (yet). :-)

> Friendly,
>
> Sven Luther

Have a nice day,
.TM.
-- 
       ____/  ____/   /
      /      /       /			Marco Colombo
     ___/  ___  /   /		      Technical Manager
    /          /   /			 ESI s.r.l.
  _____/ _____/  _/		       Colombo@ESI.it

Re: non-free firmware in kernel modules, aggregation and unclear copyright notice.

[Sven Luther]

On Tue, Apr 12, 2005 at 02:40:48AM +0200, Marco Colombo wrote:
> Which reminds me. The only reason why this thread belongs here, IMHO,
> it's because when it comes to GPL, it really doesn't matter what
> FSF's interpretation is, or anyone else's. The authors are choosing
> GPL as a license, so _thier_ interpretation is what really matters.

The main problem is that i feel that those binary firmware copyright holders
may have put it under the GPL, but i doubt they realize that this means they
have to release the source code of said firmware blobs.

Also, i believe you are wrong in the above, the only interpretation that is
important is the one the judge will take in case someone goes to suing.

And finally, if anyone could claim that a binary is the prefered form of
modification, which is most of the time obviously false, then the GPL would be
worthless. And anyway, the GPL states this (first paragraph after subclause c
in clause 3) :

  The source code for a work means the preferred form of the work for
  making modifications to it.  For an executable work, complete source
  code means all the source code for all modules it contains, plus any
  associated interface definition files, plus the scripts used to
  control compilation and installation of the executable.

So, this is not some interpretation of the GPL by the FSF, and since it is
written black on white in the actual GPL text, i don't think there is any
doubt what a judge will decice :

  judge : so, to create this piece of work, what do you use to make
  modifications ?
  A (having sworn on the bible to say the truth and only the truth) : euh,
  some C or asm code, and a compiler or assembler to compile it.
  judge : and you did voluntarily place said code and distribute it under the
  GPL ?
  A : yes, it was going into the linux kernel, so ...
  judge : so you should distribute the source code to your work also, and
  distributing it under GPL is a breach of expectation from whoever you
  distribute it to.

Or something such.

If A is going to say that he is the only author, and that he would never sue
because of this breach of the GPL, he could just as well have put it under a
different licence, or put a small disclaimer about it, since we cannot really
act as if we believed that A would never sue us, if he doesn't explicitly say
so.

As an example, i package the unicorn driver for the bewan soft-ADSL pci and
usb modems. These being soft-ADSL modems which use a non-free binary-only ADSL
emulating library, but are otherwise GPL, i discussed the matter with
upstream, and after council from debian-legal, and possibly the FSF people
themselves, we got to use this as GPL exception :

  In addition, as a special exception, BeWAN systems gives permission
  to link the code of this program with the modem SW library
  (modem_ant_PCI.o, modem_ant_USB.o), and distribute linked combinations
  including the two. You are also given permission to redistribute the
  modem SW library (modem_ant_PCI.o, modem_ant_USB.o) with the rest of the
  code.
  You must obey the GNU General Public License in all respects for all of
  the code used other than the modem SW library.

So, really, i doubt any manufacturer distributing non-free firmware would
really have trouble in adding to their licence something like this :

  In addition, <manufacturer>, considers the firmware blob, identified as <...>, as
  a non-derivative piece of work, and thus not covered by the GPL of the rest
  of it. <manufacturer> gives permission to distribute said firmware blob as
  part of the linux kernel module driver for their hardware. The actual syntax
  of the inclusion of the code is still covered by the GPL, as is the rest of
  the driver code.

If we where to get something as nicely pu as this, provide a patch, asking
the manufacturer to sign it of, then all issues would be void, i believe. 

> >>says so and it's A granting D the right to distribute. There's no way C
> >>can prevent D from distributing A's software, if A is fine with it.
> >>It's up to A to decide if GPL conditions are met by D.
> >
> >Even in that case, you still need explicit permission of A, and all the 
> >other
> >copyright holders of the rest of the GPLed work, to give you an explicit
> >exception to link with this non-free bit of code.
> 
> Yes, but it does not apply to our case here. There's no "all other
> copyright holders". _You_ stated that the firmware is included by mere
> aggregation, so there's no other holders involved. We're talking
> about the firmware case. A is one or two well identified subjects.
> And A wrote it is GPL'ed. Whether you agree or not, that's the licence
> A chose. A placed the copyright notice.

This is where i would need legal counsel, as to whether this means C or
someone else may stop you from distributing unless you provide the source. And
the real problem is that A didn't state anything, so we are only working on
the assumption that this may be the case, and A can change its mind later, and
the costs to defend ourselves in front of a judge, even if your
interpretations are right, are enough prohibitive for debian not to distribute
said files.

> The licence is a matter between A and D. A may sue D and D may (less
> likely) sue A, if conditions are not met. I'm not sure at all GPL
> is enforceable by D upon A. Let's assume it is, for sake of discussion,
> anyway.

Ah, but the licence is transitive, and if D may sue A, then C may also sue D,
since the GPL makes no distinction between who makes the distribution, apart
from the fact that A may relicence its code. But if he distributes it as part
of the GPL ...

> >No. The source code is clearly the prefered form of modification, not some
> >random intermediate state A may be claiming is source.
> 
> In this context, it is. Only A may sue D for not distributing the source.
> Whatever D distributes, D has to make A happy. If A is happy with D
> distributing `dd if=/dev/random count=1` as source, no one can stop D
> from doing that. Keep in mind A is the copyright holder. He grants
> rights to third parties. No one but A can remove them.

Yep, so if A was giving us an explicit right to distribute his sources,
everyone would be happy, this not being the case, we have to take the
hypothesis that A will sue us at a later time.

> [...]
> >>I'm not following. Are you saying what if A is bought? That is
> >>different. Well GPL is quite clear:
> >>
> >>1. You may copy and distribute verbatim copies of the Program's source
> >>code as you receive it, ...
> >>
> >>If D is distributing the source as received from A, D is in full
> >>compliance. How could A sue D? If A distributed incomplete source
> >>in the first place, it's not D's fault for sure. Do you really think
> >>the following scenario is likely:
> >>
> >>A to D: you must distribute the complete source, or the license will be
> >>        terminated!
> >>D to A: gimme the complete source, and I'll distribute it.
> >>A to D: no, I'm not willing to give you the full source of my firmware,
> >>        but you must distribute it anyway!
> >
> >The result is that the code in question has to be stopped from being
> >distributed by D. But the case here is different, since A is not the sole
> >copyright owner, so he doesn't get to set random interpretations of what is
> >source code.
> 
> Definitely I'm not following. How can D be ordered to stop distributing
> the software if A refuses to give the source? A is in violation. And
> even if GPL can't be enforced that way, by no means D is in violation
> when it's A that it's preventing D from complying.

The GPL says :

  If, as a consequence of a court judgment or allegation of patent
  infringement or for any other reason (not limited to patent issues),
  conditions are imposed on you (whether by court order, agreement or
  otherwise) that contradict the conditions of this License, they do not
  excuse you from the conditions of this License.  If you cannot
  distribute so as to satisfy simultaneously your obligations under this
  License and any other pertinent obligations, then as a consequence you
  may not distribute the Program at all.

The conditions of this licence are clear, the source code is the prefered form
of modification, so, if for "any other reason", you are not able to distribute
the source, then you "may not distribute the Program at all". The copyright
holder can, but there is no redistribution allowed. I will investigate about
if only A can sue over this though.

> The only way this can work the way you say is if A and D had a private
> agreement, and no proof can be made by D this agreement ever existed.

Well, there is no agreement whatsoever about this, and we don't know the
interpretation of A.

> But A released the software in public, and GPL'ed it. If A wants to keep
> part of the source obscured, he had better to stay well away from any
> court.

he could claim, as some did here, that obviously the fimrware was not GPLed,
but mere aggregation, and that he nowhere gave any right to distribute them.

> >>That, in court? Is this really what you're afraid of?
> >>The outcome is, very likely A will be forced to release the full source.
> >>(and D forced to distribute it, but all D's we're talking of here are
> >>very happy with the full disclosure scenario, aren't they?)
> >
> >Imagine A refusing to give away the source code, and D is ordered to remove
> >the incriminated code it is illegally distributing from all its servers, 
> >and
> >recall all those thousands of CD and DVD isos containing the code it
> >distributed, and being fined for each day it doesn't do so ?
> 
> Sorry, this is nonsense. D is well willing to distribute the source.
> In this case, he _is_ distributing what A publicly stated to be the source.

Yep, apart from the fact that A never did publicly state such issue, but just
passed it under silence.

> That's all. You say what if A changed his mind about what the source is?
> Well since it's still GPL'ed (we agree A can't chance the licence on the
> fly, right?) A is to provide the new, complete version of the source.
> Then, if and only if D refuses to distribute _that_ source, D is in
> violation. GPL requires D to distribute source _as received by A_.
> If A refuses to give the source, the specific requirement is not
> enforceable.

No, the GPL itself voids the distribubtion and redistribution rights if the
specific requirement is not fullfilled. This doesn't make the requirement not
enforceable, but voids the whole GPL licence and makes the work not
distributable.

> >>>This is the scenario i want to avoid by explicitly stating the 
> >>>relationships
> >>>of all copyright issues of those firmware blobs.
> >>
> >>I don't see this scenario nowhere close to likely. Of course, A can
> >>always sue any B, C, or D for whatever reason. It's very unlikely
> >>A will sue anyone in full compliance of GPL, but it's possible.
> >>There's nothing we can do about it. But there's no reason to worry
> >>either.
> >
> >So, we don't take the risk and don't distribute it. If A is not ready to 
> >put a
> >couple of lines of disclaimer on his work explaining the copyright and
> >licencing issues, then we are better of not distributing its code, which is
> >what debian will do.
> 
> There no such a risk. A already placed a copyright notice on his work.
> Is it the wrong one? Who cares? I'd even say that once he released it
> under GPL, he can't take it back.

But the GPL states that we must be able to distribute the sources, clearly
defines what said sources are, and states what happens if you can't fullfill
a clause of the GPL -> no right to distribute at all.

> >
> >I am not sure. If i where to get a copy of windows, and manage to install 
> >it
> >without clicking on the "i agree" button, does that make it a legal copy of
> >windows to use ?
> 
> Come on, please, do not mix things up. I've said GPL'ed software. Last time
> I checked, Windows was not GPL'ed (yet). :-)

What has the GPL to do with it ? 

Friendly,

Sven Luther

Re: non-free firmware in kernel modules, aggregation and unclear copyright notice.

[Marco Colombo]

On Tue, 12 Apr 2005, Sven Luther wrote:

> On Tue, Apr 12, 2005 at 02:40:48AM +0200, Marco Colombo wrote:
>> Which reminds me. The only reason why this thread belongs here, IMHO,
>> it's because when it comes to GPL, it really doesn't matter what
>> FSF's interpretation is, or anyone else's. The authors are choosing
>> GPL as a license, so _thier_ interpretation is what really matters.
>
> The main problem is that i feel that those binary firmware copyright holders
> may have put it under the GPL, but i doubt they realize that this means they
> have to release the source code of said firmware blobs.

They released it not in object format, but in the C language. An
hexstring, agreed, but still C. The copyright holders can release
their work how they please. If you think GPL can place restrictions on
what they can do, please see below.

> Also, i believe you are wrong in the above, the only interpretation that is
> important is the one the judge will take in case someone goes to suing.

Agreed, let me rephrase then. The only interpretation that is 
important _to the judge_ is the one of the parties involved.
In any agreement, the parties express their will. Here, the holders
"wrote" the agreement alone, so _their_ interpretation counts.
That is, their interpretation as it was when they licenced the software.
Not as is it after later thinking (or acquisition by some bad guy).

> And finally, if anyone could claim that a binary is the prefered form of
> modification, which is most of the time obviously false, then the GPL would be
> worthless. And anyway, the GPL states this (first paragraph after subclause c
> in clause 3) :

I don't care about GPL being worthless. This is not the GPL advocacy
list. I'm just saying that if you distribute the source in the form
the author published it, you can't be sued by him for breaking GPL.
That's what any linux distro and its mirrors are doing.

>  The source code for a work means the preferred form of the work for
>  making modifications to it.  For an executable work, complete source
>  code means all the source code for all modules it contains, plus any
>  associated interface definition files, plus the scripts used to
>  control compilation and installation of the executable.
>
> So, this is not some interpretation of the GPL by the FSF, and since it is
> written black on white in the actual GPL text, i don't think there is any
> doubt what a judge will decice :
>
>  judge : so, to create this piece of work, what do you use to make
>  modifications ?
>  A (having sworn on the bible to say the truth and only the truth) : euh,
>  some C or asm code, and a compiler or assembler to compile it.
>  judge : and you did voluntarily place said code and distribute it under the
>  GPL ?
>  A : yes, it was going into the linux kernel, so ...
>  judge : so you should distribute the source code to your work also, and
>  distributing it under GPL is a breach of expectation from whoever you
>  distribute it to.
>
> Or something such.

Again, I'm not following. The author release the source under GPL.
You can't release a binary under GPL, it makes no sense. So there's
no "so you should distribute the source code to your work _also_".
You released a software, it the form you claimed to be the source.
You like LISP, you release it in LISP. You like C, you release it
in C. You like hexcode, you release it in hexcode. No one can ask
you to change it.

You seems to keep forgetting what GPL is. It's a licence. The 
copyright holders grant some rights to third parties, _if_ they
comply to some conditions. Conditions are all placed on the third
parties, including the source disclosure one (source of _modifications_).
There is no condition the _holders_ have to meet. It'd be a nonsense.
The GPL says: "I grant you a right if you do this." and not:
"I grant you a right if _I_ do this.". GPL doesn't backfire.

Again, IANAL, but I see little room for "interpretation" here.

> If A is going to say that he is the only author, and that he would never sue
> because of this breach of the GPL, he could just as well have put it under a
> different licence, or put a small disclaimer about it, since we cannot really
> act as if we believed that A would never sue us, if he doesn't explicitly say
> so.

No one will ever do that. If you are distributing the software I released
under GPL, be sure I _will_ sue you if you break the licence. What do you
want from me? A promise I won't sue you if you don't? That is implicit
in the existance of the licence.

Are you implying debian will stop distributing _any_ software unless
the all the copyright holders of GPL software "explicitly say" they
won't sue you?

> As an example, i package the unicorn driver for the bewan soft-ADSL pci and
> usb modems. These being soft-ADSL modems which use a non-free binary-only ADSL
> emulating library, but are otherwise GPL, i discussed the matter with
> upstream, and after council from debian-legal, and possibly the FSF people
> themselves, we got to use this as GPL exception :
>
>  In addition, as a special exception, BeWAN systems gives permission
>  to link the code of this program with the modem SW library
>  (modem_ant_PCI.o, modem_ant_USB.o), and distribute linked combinations
>  including the two. You are also given permission to redistribute the
>  modem SW library (modem_ant_PCI.o, modem_ant_USB.o) with the rest of the
>  code.
>  You must obey the GNU General Public License in all respects for all of
>  the code used other than the modem SW library.

This is different. They are not giving the source at all. The licence
for those object files _has_ to be different. _They_ want it to be
different.

> So, really, i doubt any manufacturer distributing non-free firmware would
> really have trouble in adding to their licence something like this :
>
>  In addition, <manufacturer>, considers the firmware blob, identified as <...>, as
>  a non-derivative piece of work, and thus not covered by the GPL of the rest
>  of it. <manufacturer> gives permission to distribute said firmware blob as
>  part of the linux kernel module driver for their hardware. The actual syntax
>  of the inclusion of the code is still covered by the GPL, as is the rest of
>  the driver code.

This is fine with me. It is the existance of legal threats versus 
debian I don't agree upon.

> If we where to get something as nicely pu as this, provide a patch, asking
> the manufacturer to sign it of, then all issues would be void, i believe.

I think they are void even now.

[...]
>> Yes, but it does not apply to our case here. There's no "all other
>> copyright holders". _You_ stated that the firmware is included by mere
>> aggregation, so there's no other holders involved. We're talking
>> about the firmware case. A is one or two well identified subjects.
>> And A wrote it is GPL'ed. Whether you agree or not, that's the licence
>> A chose. A placed the copyright notice.
>
> This is where i would need legal counsel, as to whether this means C or
> someone else may stop you from distributing unless you provide the source. And
> the real problem is that A didn't state anything, so we are only working on
> the assumption that this may be the case, and A can change its mind later, and
> the costs to defend ourselves in front of a judge, even if your
> interpretations are right, are enough prohibitive for debian not to distribute
> said files.

A did put a GPL notice on it. He can't change his mind later.

>> The licence is a matter between A and D. A may sue D and D may (less
>> likely) sue A, if conditions are not met. I'm not sure at all GPL
>> is enforceable by D upon A. Let's assume it is, for sake of discussion,
>> anyway.
>
> Ah, but the licence is transitive, and if D may sue A, then C may also sue D,
> since the GPL makes no distinction between who makes the distribution, apart
> from the fact that A may relicence its code. But if he distributes it as part
> of the GPL ...

Pardon me, I have no idea of what a "transitive" licence could be.
Sublicencing or relicencing is _explicitly_ not covered by GPL anyway.

Also I have no idea of what you mean "GPL makes no distinction between
who makes the distribution". GPL for sure places no restrictions on
how A can distribute his software. A needs no license for exercising
rights on the software. He is the _owner_ of rights. A cannot "break"
the GPL. A needs no GPL to distribute. Are you saying A may sue himself?

>>> No. The source code is clearly the prefered form of modification, not some
>>> random intermediate state A may be claiming is source.
>>
>> In this context, it is. Only A may sue D for not distributing the source.
>> Whatever D distributes, D has to make A happy. If A is happy with D
>> distributing `dd if=/dev/random count=1` as source, no one can stop D
>> from doing that. Keep in mind A is the copyright holder. He grants
>> rights to third parties. No one but A can remove them.
>
> Yep, so if A was giving us an explicit right to distribute his sources,
> everyone would be happy, this not being the case, we have to take the
> hypothesis that A will sue us at a later time.

A placed it under GPL. If that is not "explicit right to distribute his
source", I'm not able to think of anything that could be it.

>> [...]
>>>> I'm not following. Are you saying what if A is bought? That is
>>>> different. Well GPL is quite clear:
>>>>
>>>> 1. You may copy and distribute verbatim copies of the Program's source
>>>> code as you receive it, ...
>>>>
>>>> If D is distributing the source as received from A, D is in full
>>>> compliance. How could A sue D? If A distributed incomplete source
>>>> in the first place, it's not D's fault for sure. Do you really think
>>>> the following scenario is likely:
>>>>
>>>> A to D: you must distribute the complete source, or the license will be
>>>>        terminated!
>>>> D to A: gimme the complete source, and I'll distribute it.
>>>> A to D: no, I'm not willing to give you the full source of my firmware,
>>>>        but you must distribute it anyway!
>>>
>>> The result is that the code in question has to be stopped from being
>>> distributed by D. But the case here is different, since A is not the sole
>>> copyright owner, so he doesn't get to set random interpretations of what is
>>> source code.
>>
>> Definitely I'm not following. How can D be ordered to stop distributing
>> the software if A refuses to give the source? A is in violation. And
>> even if GPL can't be enforced that way, by no means D is in violation
>> when it's A that it's preventing D from complying.
>
> The GPL says :
>
>  If, as a consequence of a court judgment or allegation of patent
>  infringement or for any other reason (not limited to patent issues),
>  conditions are imposed on you (whether by court order, agreement or
>  otherwise) that contradict the conditions of this License, they do not
>  excuse you from the conditions of this License.  If you cannot
>  distribute so as to satisfy simultaneously your obligations under this
>  License and any other pertinent obligations, then as a consequence you
>  may not distribute the Program at all.
>
> The conditions of this licence are clear, the source code is the prefered form
> of modification, so, if for "any other reason", you are not able to distribute
> the source, then you "may not distribute the Program at all". The copyright
> holder can, but there is no redistribution allowed. I will investigate about
> if only A can sue over this though.

So you seriously think this makes sense?

A: you have to distribute the source, or the licence will be void
D: i am, i'm distributing the source you gave me
A: that is not the source, i was lying about it
D: then give me the right source, and I'll distribute it
A: No. I refuse to give you the right source, but expect you to distribute
    it anyway, or i'll terminate the licence.

You seriouly think some judge may rule in favor of A? _If_ A sued D?
Well if you do, we can stop here. You'll never buy me on this.

>> The only way this can work the way you say is if A and D had a private
>> agreement, and no proof can be made by D this agreement ever existed.
>
> Well, there is no agreement whatsoever about this, and we don't know the
> interpretation of A.
>
>> But A released the software in public, and GPL'ed it. If A wants to keep
>> part of the source obscured, he had better to stay well away from any
>> court.
>
> he could claim, as some did here, that obviously the fimrware was not GPLed,
> but mere aggregation, and that he nowhere gave any right to distribute them.

"mere aggregation" refers to an action taken by a distributor. _You_
as debian are distributing the firmware as merely aggregated to the
kernel, assuming your intrepretation is right. That has nothing to do
with .c files.

Moreover, the firmare in not in binary form, but is part of a C source
file.

I think you're going no where. A cannot put a licence statement on
the top of a .c file stating it's GPL'ed, and then say that some part
of it is not covered because it was "merely aggregated". It makes no
sense, and there's nothing in the GPL about mere aggregation of sources.

>>>> That, in court? Is this really what you're afraid of?
>>>> The outcome is, very likely A will be forced to release the full source.
>>>> (and D forced to distribute it, but all D's we're talking of here are
>>>> very happy with the full disclosure scenario, aren't they?)
>>>
>>> Imagine A refusing to give away the source code, and D is ordered to remove
>>> the incriminated code it is illegally distributing from all its servers,
>>> and
>>> recall all those thousands of CD and DVD isos containing the code it
>>> distributed, and being fined for each day it doesn't do so ?
>>
>> Sorry, this is nonsense. D is well willing to distribute the source.
>> In this case, he _is_ distributing what A publicly stated to be the source.
>
> Yep, apart from the fact that A never did publicly state such issue, but just
> passed it under silence.

On the top of a .c file there's a nice copyright notice and licence statement,
isn't there? A placed it there. _You_ think it may be changed. But what
if A is fine with it?

>> That's all. You say what if A changed his mind about what the source is?
>> Well since it's still GPL'ed (we agree A can't chance the licence on the
>> fly, right?) A is to provide the new, complete version of the source.
>> Then, if and only if D refuses to distribute _that_ source, D is in
>> violation. GPL requires D to distribute source _as received by A_.
>> If A refuses to give the source, the specific requirement is not
>> enforceable.
>
> No, the GPL itself voids the distribubtion and redistribution rights if the
> specific requirement is not fullfilled. This doesn't make the requirement not
> enforceable, but voids the whole GPL licence and makes the work not
> distributable.

It makes D not suable. You can't actively prevent someone from doing
something he's supposed to do and the sue him for not doing it at the
same time.

> But the GPL states that we must be able to distribute the sources, clearly
> defines what said sources are, and states what happens if you can't fullfill
> a clause of the GPL -> no right to distribute at all.

No. GPL says you must be _willing_ to distribute the sources, as received
by A. See, GPL covers the source. There's no way you can distribute
the software in binary/executable form, unless you get the source and
complile it. That's what you do here. You compile the hexstring, and
the firmware (in binary form) gets "aggregated" to the kernel binary.
If you distribute the result, you have to distribute the source _as
you received it_. That's all. If you do, you're fine.

>>> I am not sure. If i where to get a copy of windows, and manage to install
>>> it
>>> without clicking on the "i agree" button, does that make it a legal copy of
>>> windows to use ?
>>
>> Come on, please, do not mix things up. I've said GPL'ed software. Last time
>> I checked, Windows was not GPL'ed (yet). :-)
>
> What has the GPL to do with it ?

Nothing. You introduced Windows, not me. My point was that you can
break into Fort Knox, steal a Debian CD, and give it to a friend.
His licence is valid. Your acting may be a crime. GPL says nothing
about how you got your copy. So even if the distributor is breaking
GPL, people who obtained a copy from there are fine.

.TM.
-- 
       ____/  ____/   /
      /      /       /			Marco Colombo
     ___/  ___  /   /		      Technical Manager
    /          /   /			 ESI s.r.l.
  _____/ _____/  _/		       Colombo@ESI.it

Re: non-free firmware in kernel modules, aggregation and unclear copyright notice.

[Sven Luther]

On Tue, Apr 12, 2005 at 06:14:17PM +0200, Marco Colombo wrote:
> No one will ever do that. If you are distributing the software I released
> under GPL, be sure I _will_ sue you if you break the licence. What do you
> want from me? A promise I won't sue you if you don't? That is implicit
> in the existance of the licence.
> 
> Are you implying debian will stop distributing _any_ software unless
> the all the copyright holders of GPL software "explicitly say" they
> won't sue you?

Well, we won't distribute binaries placed under the GPL, definitively not. And
if there is a dubious case, we ask for clarification of the author.

> >As an example, i package the unicorn driver for the bewan soft-ADSL pci and
> >usb modems. These being soft-ADSL modems which use a non-free binary-only 
> >ADSL
> >emulating library, but are otherwise GPL, i discussed the matter with
> >upstream, and after council from debian-legal, and possibly the FSF people
> >themselves, we got to use this as GPL exception :
> >
> > In addition, as a special exception, BeWAN systems gives permission
> > to link the code of this program with the modem SW library
> > (modem_ant_PCI.o, modem_ant_USB.o), and distribute linked combinations
> > including the two. You are also given permission to redistribute the
> > modem SW library (modem_ant_PCI.o, modem_ant_USB.o) with the rest of the
> > code.
> > You must obey the GNU General Public License in all respects for all of
> > the code used other than the modem SW library.
> 
> This is different. They are not giving the source at all. The licence
> for those object files _has_ to be different. _They_ want it to be
> different.

Sure, but in this case, the binary firmware blob is also a binary without
sources. If they really did write said firmware directly as it is, then they
should say so, but this is contrary to everyone's expectation, and a dangerous
precedent to set.

> >So, really, i doubt any manufacturer distributing non-free firmware would
> >really have trouble in adding to their licence something like this :
> >
> > In addition, <manufacturer>, considers the firmware blob, identified as 
> > <...>, as
> > a non-derivative piece of work, and thus not covered by the GPL of the 
> > rest
> > of it. <manufacturer> gives permission to distribute said firmware blob as
> > part of the linux kernel module driver for their hardware. The actual 
> > syntax
> > of the inclusion of the code is still covered by the GPL, as is the rest 
> > of
> > the driver code.
> 
> This is fine with me. It is the existance of legal threats versus 
> debian I don't agree upon.

Notice that debian can't afford to be sued even if they are right, so ...

> >>Yes, but it does not apply to our case here. There's no "all other
> >>copyright holders". _You_ stated that the firmware is included by mere
> >>aggregation, so there's no other holders involved. We're talking
> >>about the firmware case. A is one or two well identified subjects.
> >>And A wrote it is GPL'ed. Whether you agree or not, that's the licence
> >>A chose. A placed the copyright notice.
> >
> >This is where i would need legal counsel, as to whether this means C or
> >someone else may stop you from distributing unless you provide the source. 
> >And
> >the real problem is that A didn't state anything, so we are only working on
> >the assumption that this may be the case, and A can change its mind later, 
> >and
> >the costs to defend ourselves in front of a judge, even if your
> >interpretations are right, are enough prohibitive for debian not to 
> >distribute
> >said files.
> 
> A did put a GPL notice on it. He can't change his mind later.

Then he should give us the source.

> >>The licence is a matter between A and D. A may sue D and D may (less
> >>likely) sue A, if conditions are not met. I'm not sure at all GPL
> >>is enforceable by D upon A. Let's assume it is, for sake of discussion,
> >>anyway.
> >
> >Ah, but the licence is transitive, and if D may sue A, then C may also sue 
> >D,
> >since the GPL makes no distinction between who makes the distribution, 
> >apart
> >from the fact that A may relicence its code. But if he distributes it as 
> >part
> >of the GPL ...
> 
> Pardon me, I have no idea of what a "transitive" licence could be.
> Sublicencing or relicencing is _explicitly_ not covered by GPL anyway.

You give away the source to someone, he has the same rights you had, except
relicencing, this is what i meant by transitive.

> Also I have no idea of what you mean "GPL makes no distinction between
> who makes the distribution". GPL for sure places no restrictions on
> how A can distribute his software. A needs no license for exercising

No, it gives A the choice to distribute its software under the GPL, or under
another licence.

> rights on the software. He is the _owner_ of rights. A cannot "break"
> the GPL. A needs no GPL to distribute. Are you saying A may sue himself?

Yes, he can break the commonly accepted expectation of a GPLed software, which
is what happens here. He is free to distribute the software under any other
licence he sees fit, which is what i am asking here.

> >>>No. The source code is clearly the prefered form of modification, not 
> >>>some
> >>>random intermediate state A may be claiming is source.
> >>
> >>In this context, it is. Only A may sue D for not distributing the source.
> >>Whatever D distributes, D has to make A happy. If A is happy with D
> >>distributing `dd if=/dev/random count=1` as source, no one can stop D
> >>from doing that. Keep in mind A is the copyright holder. He grants
> >>rights to third parties. No one but A can remove them.
> >
> >Yep, so if A was giving us an explicit right to distribute his sources,
> >everyone would be happy, this not being the case, we have to take the
> >hypothesis that A will sue us at a later time.
> 
> A placed it under GPL. If that is not "explicit right to distribute his
> source", I'm not able to think of anything that could be it.

He is not distributing sources here, he is distributing binaries, my error.

> >The GPL says :
> >
> > If, as a consequence of a court judgment or allegation of patent
> > infringement or for any other reason (not limited to patent issues),
> > conditions are imposed on you (whether by court order, agreement or
> > otherwise) that contradict the conditions of this License, they do not
> > excuse you from the conditions of this License.  If you cannot
> > distribute so as to satisfy simultaneously your obligations under this
> > License and any other pertinent obligations, then as a consequence you
> > may not distribute the Program at all.
> >
> >The conditions of this licence are clear, the source code is the prefered 
> >form
> >of modification, so, if for "any other reason", you are not able to 
> >distribute
> >the source, then you "may not distribute the Program at all". The copyright
> >holder can, but there is no redistribution allowed. I will investigate 
> >about
> >if only A can sue over this though.
> 
> So you seriously think this makes sense?
> 
> A: you have to distribute the source, or the licence will be void
> D: i am, i'm distributing the source you gave me
> A: that is not the source, i was lying about it
> D: then give me the right source, and I'll distribute it
> A: No. I refuse to give you the right source, but expect you to distribute
>    it anyway, or i'll terminate the licence.

No, the licence auto-terminate and the software becomes undistributable.

> >he could claim, as some did here, that obviously the fimrware was not 
> >GPLed,
> >but mere aggregation, and that he nowhere gave any right to distribute 
> >them.
> 
> "mere aggregation" refers to an action taken by a distributor. _You_
> as debian are distributing the firmware as merely aggregated to the
> kernel, assuming your intrepretation is right. That has nothing to do
> with .c files.

The fact remains that those firmware blob have no licence, and thus defacto
fall under the GPL.

> Moreover, the firmare in not in binary form, but is part of a C source
> file.

It is in binary form. Disguised binary form maybe but still binary form.

> I think you're going no where. A cannot put a licence statement on
> the top of a .c file stating it's GPL'ed, and then say that some part
> of it is not covered because it was "merely aggregated". It makes no
> sense, and there's nothing in the GPL about mere aggregation of sources.

What i want is that A aknolwedges that he is not distributing the sources of
the firmware, and thus cannot place the binary blob under the GPL but should
chose another licence.

> >>>>That, in court? Is this really what you're afraid of?
> >>>>The outcome is, very likely A will be forced to release the full source.
> >>>>(and D forced to distribute it, but all D's we're talking of here are
> >>>>very happy with the full disclosure scenario, aren't they?)
> >>>
> >>>Imagine A refusing to give away the source code, and D is ordered to 
> >>>remove
> >>>the incriminated code it is illegally distributing from all its servers,
> >>>and
> >>>recall all those thousands of CD and DVD isos containing the code it
> >>>distributed, and being fined for each day it doesn't do so ?
> >>
> >>Sorry, this is nonsense. D is well willing to distribute the source.
> >>In this case, he _is_ distributing what A publicly stated to be the 
> >>source.
> >
> >Yep, apart from the fact that A never did publicly state such issue, but 
> >just
> >passed it under silence.
> 
> On the top of a .c file there's a nice copyright notice and licence 
> statement,
> isn't there? A placed it there. _You_ think it may be changed. But what
> if A is fine with it?

Not in the tg3.c case, no.

> >But the GPL states that we must be able to distribute the sources, clearly
> >defines what said sources are, and states what happens if you can't 
> >fullfill
> >a clause of the GPL -> no right to distribute at all.
> 
> No. GPL says you must be _willing_ to distribute the sources, as received
> by A. See, GPL covers the source. There's no way you can distribute
> the software in binary/executable form, unless you get the source and
> complile it. That's what you do here. You compile the hexstring, and
> the firmware (in binary form) gets "aggregated" to the kernel binary.
> If you distribute the result, you have to distribute the source _as
> you received it_. That's all. If you do, you're fine.

And where did those hexstrings come from ? 

Friendly,

Sven Luther

Re: non-free firmware in kernel modules, aggregation and unclear copyright notice.

[Zan Lynx]

[-- Attachment #1: Type: text/plain, Size: 1194 bytes --]

On Tue, 2005-04-12 at 20:45 +0200, Sven Luther wrote:
[snip]
> > A did put a GPL notice on it. He can't change his mind later.
> Then he should give us the source.
[snip]
> The fact remains that those firmware blob have no licence, and thus defacto
> fall under the GPL.
> 
> > Moreover, the firmare in not in binary form, but is part of a C source
> > file.
> 
> It is in binary form. Disguised binary form maybe but still binary form.
[snip]
> And where did those hexstrings come from ? 

It seems to me, that to be consistent with the argument you seem to be
presenting concerning binary data in GPLd code, that you also need to be
demanding the "source" hardware design for binary register values.

Why not consider the binary firmware in the same category as binary
register programming information?  You poke these magic bytes into these
memory locations and it works.

Where do you draw the lines between "write this byte to set the input
gate here and the output gate to there" and "write this byte sequence to
send the input byte through this loop, into this buffer, add it to the
last byte entered, and output it over there"?
-- 
Zan Lynx <zlynx@acm.org>

[-- Attachment #2: This is a digitally signed message part --]
[-- Type: application/pgp-signature, Size: 189 bytes --]

Re: non-free firmware in kernel modules, aggregation and unclear copyright notice.

[Marco Colombo]

On Tue, 2005-04-12 at 20:45 +0200, Sven Luther wrote:
> On Tue, Apr 12, 2005 at 06:14:17PM +0200, Marco Colombo wrote:
> > No one will ever do that. If you are distributing the software I released
> > under GPL, be sure I _will_ sue you if you break the licence. What do you
> > want from me? A promise I won't sue you if you don't? That is implicit
> > in the existance of the licence.
> > 
> > Are you implying debian will stop distributing _any_ software unless
> > the all the copyright holders of GPL software "explicitly say" they
> > won't sue you?
> 
> Well, we won't distribute binaries placed under the GPL, definitively not. And
> if there is a dubious case, we ask for clarification of the author.

Your choice, of course.

[...] 
> > This is different. They are not giving the source at all. The licence
> > for those object files _has_ to be different. _They_ want it to be
> > different.
> 
> Sure, but in this case, the binary firmware blob is also a binary without
> sources. If they really did write said firmware directly as it is, then they
> should say so, but this is contrary to everyone's expectation, and a dangerous
> precedent to set.

You should realize that any author can publish his work in the form he
likes. He's not bound to "everyone's expectation". I see no danger in
that.

> > >So, really, i doubt any manufacturer distributing non-free firmware would
> > >really have trouble in adding to their licence something like this :
> > >
> > > In addition, <manufacturer>, considers the firmware blob, identified as 
> > > <...>, as
> > > a non-derivative piece of work, and thus not covered by the GPL of the 
> > > rest
> > > of it. <manufacturer> gives permission to distribute said firmware blob as
> > > part of the linux kernel module driver for their hardware. The actual 
> > > syntax
> > > of the inclusion of the code is still covered by the GPL, as is the rest 
> > > of
> > > the driver code.
> > 
> > This is fine with me. It is the existance of legal threats versus 
> > debian I don't agree upon.
> 
> Notice that debian can't afford to be sued even if they are right, so ...

So what? This is not the point. You can be sued any time by any one,
even if you're right. If debian can't afford it, it can't afford
existance.

> > >>Yes, but it does not apply to our case here. There's no "all other
> > >>copyright holders". _You_ stated that the firmware is included by mere
> > >>aggregation, so there's no other holders involved. We're talking
> > >>about the firmware case. A is one or two well identified subjects.
> > >>And A wrote it is GPL'ed. Whether you agree or not, that's the licence
> > >>A chose. A placed the copyright notice.
> > >
> > >This is where i would need legal counsel, as to whether this means C or
> > >someone else may stop you from distributing unless you provide the source. 
> > >And
> > >the real problem is that A didn't state anything, so we are only working on
> > >the assumption that this may be the case, and A can change its mind later, 
> > >and
> > >the costs to defend ourselves in front of a judge, even if your
> > >interpretations are right, are enough prohibitive for debian not to 
> > >distribute
> > >said files.
> > 
> > A did put a GPL notice on it. He can't change his mind later.
> 
> Then he should give us the source.

No, why? GPL cannot place restrictions or obligations on the copyright
owner. Let's stop discussing it please, you can't buy me on this
either. I have my own interpretation of what a license it, and it seems
you don't agree with it: to me, it's one way: _you_, the licensee,
get some rights if you fulfill some conditions. Conditions are all
placed on you, none on the copyright holder. In particular, the
one about making source available is placed on distributors,
verbatim copies of the source for binary distribution of the work, or
full source of the modifications for modified versions of the work.
 
And anyway, this has nothing to do with with legal threats from the
copyright holder. My point being: he cannot sue you for not
distributing the source "as provided by him" if he failed to provide
them in the first place in a different from. That is, he has to give you
the source, if he is trying to force you distributing it.

> > >>The licence is a matter between A and D. A may sue D and D may (less
> > >>likely) sue A, if conditions are not met. I'm not sure at all GPL
> > >>is enforceable by D upon A. Let's assume it is, for sake of discussion,
> > >>anyway.
> > >
> > >Ah, but the licence is transitive, and if D may sue A, then C may also sue 
> > >D,
> > >since the GPL makes no distinction between who makes the distribution, 
> > >apart
> > >from the fact that A may relicence its code. But if he distributes it as 
> > >part
> > >of the GPL ...
> > 
> > Pardon me, I have no idea of what a "transitive" licence could be.
> > Sublicencing or relicencing is _explicitly_ not covered by GPL anyway.
> 
> You give away the source to someone, he has the same rights you had, except
> relicencing, this is what i meant by transitive.

GPL explicitly says that when you, a distributor, give the source to
someone, he receives a license, another "instance" of GPL so to say,
directly from the copyright holder and not from you. If your license
is GPL, then his license is GPL, too, and yes, he has the same rights
of you. But it's not you granting those rights to him, it's the
copyright holder.
 
> > Also I have no idea of what you mean "GPL makes no distinction between
> > who makes the distribution". GPL for sure places no restrictions on
> > how A can distribute his software. A needs no license for exercising
> 
> No, it gives A the choice to distribute its software under the GPL, or under
> another licence.

I think GPL "gives" nothing to A. A can do whatever he wants with his
software, GPL has no effect on A. It would be A granting rights to A.
Nonsense. A does not need to "release" the software under any license
to himself. A has full rights on it, already.

> > rights on the software. He is the _owner_ of rights. A cannot "break"
> > the GPL. A needs no GPL to distribute. Are you saying A may sue himself?
> 
> Yes, he can break the commonly accepted expectation of a GPLed software, which
> is what happens here. He is free to distribute the software under any other
> licence he sees fit, which is what i am asking here.

I'm sorry. We totally disagree on this. GPL is a license. The commonly
accepted expectation is that it places restrictions on the licensee,
in exchange of some rights. The copyright holder is no licensee.
He needs to comply to no conditions. He already has (and owns) full
rights. And anyway, he is the only one entitled the right of enforcing
those conditions or voiding the license. How could he enforce any
conditions on himself? How could he void a license to himself, given
that he needs no license at all?
Let's drop this, please, we're going nowhere.

> > >>>No. The source code is clearly the prefered form of modification, not 
> > >>>some
> > >>>random intermediate state A may be claiming is source.
> > >>
> > >>In this context, it is. Only A may sue D for not distributing the source.
> > >>Whatever D distributes, D has to make A happy. If A is happy with D
> > >>distributing `dd if=/dev/random count=1` as source, no one can stop D
> > >>from doing that. Keep in mind A is the copyright holder. He grants
> > >>rights to third parties. No one but A can remove them.
> > > 
> > >Yep, so if A was giving us an explicit right to distribute his sources,
> > >everyone would be happy, this not being the case, we have to take the
> > >hypothesis that A will sue us at a later time.
> > 
> > A placed it under GPL. If that is not "explicit right to distribute his
> > source", I'm not able to think of anything that could be it.
> 
> He is not distributing sources here, he is distributing binaries, my error.

Well, no. You are distributing the binary, after you compiled the 
kernel. The hexstring is the source, the binary aggregated to the
kernel image is the binary. The GPL notice is attached to the hexstring.

[...] 
> > So you seriously think this makes sense?
> > 
> > A: you have to distribute the source, or the licence will be void
> > D: i am, i'm distributing the source you gave me
> > A: that is not the source, i was lying about it
> > D: then give me the right source, and I'll distribute it
> > A: No. I refuse to give you the right source, but expect you to distribute
> >    it anyway, or i'll terminate the licence.
> 
> No, the licence auto-terminate and the software becomes undistributable.

Nothing "auto-terminates". A can ask the judge to order D to stop
distribution. If you seriously think that, in front of such claims by A,
any judge will rule of favor of A, well, there's no point in discussing.
D is not able to distribute the source "as provided" by A, if A is not
providing it. A cannot have such a condition enforced without providing
the source, in the first place.

The license terminates when A manages to convince the judge D is not
fulfilling a condition. I think no one would ever consider 
"i don't want to give you the right source, but you must distribute it
anyway" a valid argument.

> > >he could claim, as some did here, that obviously the fimrware was not 
> > >GPLed,
> > >but mere aggregation, and that he nowhere gave any right to distribute 
> > >them.
> > 
> > "mere aggregation" refers to an action taken by a distributor. _You_
> > as debian are distributing the firmware as merely aggregated to the
> > kernel, assuming your intrepretation is right. That has nothing to do
> > with .c files.
> 
> The fact remains that those firmware blob have no licence, and thus defacto
> fall under the GPL.

Precisely.

> > Moreover, the firmare in not in binary form, but is part of a C source
> > file.
> 
> It is in binary form. Disguised binary form maybe but still binary form.

Ah, but this is a claim A has to make in front of a judge, before the 
license of D can be terminated. "I gave him a source that is really
a disguised binary form, and not the real source. He's distributing
that 'source' as i provided it, but it's not the real one. I'm not
willing to give him the real one, but I'm requesting him to distribute
it.". Makes no sense. D: "I'm providing it as provided, which is the
requirement of GPL. Should I receive it in another form, I'd distribute
it in the new form."

> > I think you're going no where. A cannot put a licence statement on
> > the top of a .c file stating it's GPL'ed, and then say that some part
> > of it is not covered because it was "merely aggregated". It makes no
> > sense, and there's nothing in the GPL about mere aggregation of sources.
> 
> What i want is that A aknolwedges that he is not distributing the sources of
> the firmware, and thus cannot place the binary blob under the GPL but should
> chose another licence.

That would be fine. But it's a matter of having thing sorted out
correctly. 

> > >>>>That, in court? Is this really what you're afraid of?
> > >>>>The outcome is, very likely A will be forced to release the full source.
> > >>>>(and D forced to distribute it, but all D's we're talking of here are
> > >>>>very happy with the full disclosure scenario, aren't they?)
> > >>>
> > >>>Imagine A refusing to give away the source code, and D is ordered to 
> > >>>remove
> > >>>the incriminated code it is illegally distributing from all its servers,
> > >>>and
> > >>>recall all those thousands of CD and DVD isos containing the code it
> > >>>distributed, and being fined for each day it doesn't do so ?
> > >>
> > >>Sorry, this is nonsense. D is well willing to distribute the source.
> > >>In this case, he _is_ distributing what A publicly stated to be the 
> > >>source.
> > >
> > >Yep, apart from the fact that A never did publicly state such issue, but 
> > >just
> > >passed it under silence.
> > 
> > On the top of a .c file there's a nice copyright notice and licence 
> > statement,
> > isn't there? A placed it there. _You_ think it may be changed. But what
> > if A is fine with it?
> 
> Not in the tg3.c case, no.

Are you saying that the notice has been wrongly placed by the 
holders or by someone else? If they changed their mind, there's 
little they can do. If _someone else_ published the code and
put a wrong license notice on it, that's a different matter.
A _false_ notice is different from a _wrong_ legitimate one.
"Wrong" as in: they changed their mind at later time. 

> > >But the GPL states that we must be able to distribute the sources, clearly
> > >defines what said sources are, and states what happens if you can't 
> > >fullfill
> > >a clause of the GPL -> no right to distribute at all.
> > 
> > No. GPL says you must be _willing_ to distribute the sources, as received
> > by A. See, GPL covers the source. There's no way you can distribute
> > the software in binary/executable form, unless you get the source and
> > complile it. That's what you do here. You compile the hexstring, and
> > the firmware (in binary form) gets "aggregated" to the kernel binary.
> > If you distribute the result, you have to distribute the source _as
> > you received it_. That's all. If you do, you're fine.
> 
> And where did those hexstrings come from ? 

There's only one subject that can put such a question: the copyright
holder, since it's the only one that can request enforcement of a
GPL condition. "You provided it" it's enough an answer. It's a dead
end, the copyright holder can't go any further w/o providing the
real source.

I think we reached a dead end ourself.

For the debian case, I think:

- no one can sue debian, but the copyright holder, for breaking GPL
  in the act of distributing the firmware blob; debian received a GPL
  from the copyright holder when they got the program

- there's no way the copyright holder could sue debian for not
  distributing source (assuming debian is distributing the kernel
  tarball of course, which is not a problem), w/o providing something
  different they claim to be the "real" source.

As usual, IANAL.

.TM.
-- 
      ____/  ____/   /
     /      /       /                   Marco Colombo
    ___/  ___  /   /                  Technical Manager
   /          /   /                      ESI s.r.l.
 _____/ _____/  _/                      Colombo@ESI.it

Re: non-free firmware in kernel modules, aggregation and unclear copyright notice.

[Sven Luther]

On Wed, Apr 13, 2005 at 04:53:56PM +0200, Marco Colombo wrote:
> > > This is different. They are not giving the source at all. The licence
> > > for those object files _has_ to be different. _They_ want it to be
> > > different.
> > 
> > Sure, but in this case, the binary firmware blob is also a binary without
> > sources. If they really did write said firmware directly as it is, then they
> > should say so, but this is contrary to everyone's expectation, and a dangerous
> > precedent to set.
> 
> You should realize that any author can publish his work in the form he
> likes. He's not bound to "everyone's expectation". I see no danger in
> that.

I think there may be some limitation of using the GPL as licence in this case
though, as such behavior may limit its value, and the GPL itself is by no
means free software.

Friendly,

Sven Luther

RE: non-free firmware in kernel modules, aggregation and unclearcopyright notice.

[David Schwartz]

> It seems to me, that to be consistent with the argument you seem to be
> presenting concerning binary data in GPLd code, that you also need to be
> demanding the "source" hardware design for binary register values.
>
> Why not consider the binary firmware in the same category as binary
> register programming information?  You poke these magic bytes into these
> memory locations and it works.
>
> Where do you draw the lines between "write this byte to set the input
> gate here and the output gate to there" and "write this byte sequence to
> send the input byte through this loop, into this buffer, add it to the
> last byte entered, and output it over there"?

	You draw the line at the source code, the preferred form of the work for
the purpose of making modifications to it. See GPL section 3. Firmware is an
executable work.

	DS

Re: non-free firmware in kernel modules, aggregation and unclear copyright notice.

[Marco Colombo]

On Wed, 2005-04-13 at 21:47 +0200, Sven Luther wrote:
> On Wed, Apr 13, 2005 at 04:53:56PM +0200, Marco Colombo wrote:
> > > > This is different. They are not giving the source at all. The licence
> > > > for those object files _has_ to be different. _They_ want it to be
> > > > different.
> > > 
> > > Sure, but in this case, the binary firmware blob is also a binary without
> > > sources. If they really did write said firmware directly as it is, then they
> > > should say so, but this is contrary to everyone's expectation, and a dangerous
> > > precedent to set.
> > 
> > You should realize that any author can publish his work in the form he
> > likes. He's not bound to "everyone's expectation". I see no danger in
> > that.
> 
> I think there may be some limitation of using the GPL as licence in this case
> though, as such behavior may limit its value, and the GPL itself is by no
> means free software.

That GPL isn't the best license in this case (firmware included as
hexstring in the driver source), we already know. But fixing it is up
to the copyright holder. We or GPL face no risk.

Note that the holder does. I'd be interesting if someone produced a
derivative work, such a translation. A translation from the hex form
to some kind of textual formally defined language, such as, say,
assembler, or C. That would be covered by GPL. And would be
distributable under it. Say that the resulting binary is slightly
different. You are _required_ by GPL to provide the source in the
preferred form, this time, preferred by _you_. What if that is C?
Interesting enough. Can the hexstring be reverse-engineered into C,
if it's placed under GPL? Can the copyright holder really prevent that?

Something new to think of. :-)

Have a nice day,
.TM.
-- 
      ____/  ____/   /
     /      /       /                   Marco Colombo
    ___/  ___  /   /                  Technical Manager
   /          /   /                      ESI s.r.l.
 _____/ _____/  _/                      Colombo@ESI.it