Re: Dual-Licensing Linux Kernel with GPL V2 and GPL V3

From: Daniel Hazelton
Date: Wed Jun 13 2007 - 22:43:34 EST


On Wednesday 13 June 2007 22:08:27 Adrian Bunk wrote:
> On Wed, Jun 13, 2007 at 09:40:13PM -0400, Daniel Hazelton wrote:
> > On Wednesday 13 June 2007 21:24:01 Adrian Bunk wrote:
> > > On Wed, Jun 13, 2007 at 09:01:28PM -0400, Daniel Hazelton wrote:
> > > > On Wednesday 13 June 2007 20:44:19 Adrian Bunk wrote:
> > > > > On Wed, Jun 13, 2007 at 07:46:15PM -0400, Daniel Hazelton wrote:
> > > > > > On Wednesday 13 June 2007 19:15:42 Alexandre Oliva wrote:
> > > > > > > On Jun 13, 2007, Linus Torvalds <torvalds@xxxxxxxxxxxxxxxxxxxx>
> >
> > wrote:
> > > > > > > > On Wed, 13 Jun 2007, Alan Cox wrote:
> > > > > > > >> > find offensive, so I don't choose to use it. It's
> > > > > > > >> > offensive because Tivo never did anything wrong, and the
> > > > > > > >> > FSF even acknowledged that. The fact
> > > > > > > >>
> > > > > > > >> Not all of us agree with this for the benefit of future
> > > > > > > >> legal interpretation.
> > > > > > > >
> > > > > > > > Well, even the FSF lawyers did,
> > > > > > >
> > > > > > > Or rather they didn't think an attempt to enforce that in the
> > > > > > > US would prevail (or so I'm told). That's not saying what TiVo
> > > > > > > did was right, and that's not saying that what TiVo did was
> > > > > > > permitted by the license. Only courts of law can do that.
> > > > > >
> > > > > > Wrong! Anyone with half a brain can make the distinction. What
> > > > > > TiVO did is entirely legal - they fully complied with the GPLv2.
> > > > > > Note that what they *DON'T* allow people to do is run whatever
> > > > > > version of whatever software they want on their hardware. They
> > > > > > have that right - its the "Free Software Foundation" and the GPL
> > > > > > - regardless of version - is a *SOFTWARE* license. ...
> > > > >
> > > > > The GPLv2 says:
> > > > >
> > > > > "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."
> > > > >
> > > > > The question is whether this includes private keys.
> > > > > Different people have different opinions regarding this issue.
> > > > >
> > > > > If "the complete source code" includes private keys, the GPLv2
> > > > > requires them to give any costumer the private keys.
> > > > >
> > > > > Fact is that Harald Welte did in several cases successfully
> > > > > convince vendors that private keys are part of the source code if
> > > > > they are required for running the compiled binary on some hardware.
> > > >
> > > > If the hardware was designed for the end-user to change the software
> > > > running on it - including running software that it was never meant to
> > > > run (ie: a complete webserver on cell phone) - then yes, the signing
> > > > keys are a part of the source, as the software running on the device
> > > > is designed to be updated by the user using the provided system.
> > > >
> > > > If, on the other hand, the only "software updates" the user is
> > > > expected to perform are the installation of newer versions of the
> > > > existing code for "Security" or "Bug Fix" reasons then the signing
> > > > keys aren't part of the source.
> > >
> > > Are you an idiot, or do you just choose to ignore all proof that
> > > doesn't fit your preconceived beliefs?
> >
> > Nope. Merely stating a distinction. Either a device is distributed, like
> > the common PC, that is designed for the user to change and update the
> > software on, or, like the PS2 it isn't designed for that. If I find a way
> > to update my PS2 to run Linux and find that it doesn't want to start the
> > "Linux Firmware" because I'm lacking a signing key...
> >
> > In the case of a device that internally runs Linux (or any other GPL'd
> > software) and wasn't designed for the end-user to change the software
> > running on it then the signing keys aren't part of the source. OTOH, if I
> > sell a PC running Linux that requires the kernel be signed then the
> > signing keys *are* part of the source, since a PC is designed for the
> > end-user to change the software running on it.
> >
> > BTW, nice use of irony with that line. Makes me regret letting my fingers
> > get ahead of my brain.
> >
> > > The GPL doesn't give someone distributing the software the choice of
> > > how much to limit the freedom of the user.
> >
> > Never claimed it did. I just wasn't as specific as I should have been
> > when giving my examples.
> >
> > > Either private keys required to run the kernel on the hardware are
> > > always considered part of "the complete source code" or they are never
> > > part of it.
> >
> > No. It all depends on the use-case. If the hardware is designed for the
> > user to install their own, custom versions of the code on then the
> > signing keys are part of the source as defined by the GPLv2.
> >
> > If, OTOH, the hardware was never meant for the end-user to install custom
> > versions of the software on, then while the signing keys are still
> > *technically* part of the source, in practice they are not. Why? Because
> > in most of those cases the end-user isn't granted the right to install
> > and run custom binaries on the hardware. If the manufacturer provided the
> > signing keys they'd be facilitating the commission of a crime. (call it
> > "Breach of Contract")
> >...
>
> Repetition doesn't let wrong things become true.
>
> Where does the GPLv2 talk about the distinction you are trying to make
> based on distributor intentions?
>
> We are talking about the GPLv2 licence text, not about what you would
> personally prefer.

The GPLv2 doesn't have to cover this distinction to make it a reality. This
distinction is *EXACTLY* the type of distinction a lawyer will make when
arguing the point.

Yes, it's artificial. Yes, it does appear to violate the GPLv2 - *IF* you read
the text of such in a specific manner.

However, the GPL, until version 3, *NEVER* guaranteed the right to run a given
piece of software on *ANY* hardware - not the hardware it *COMES* on.

And please, I repeated myself only because your reply seemed to imply that you
didn't understand the statement I had made. Since you have now informed me,
in a backhanded way, informed me that my interpretation of your response was
wrong, I will not repeat myself again.

Also note that I have re-examined the facts, in light of new information
presented in this discussion, and have come to the conclusion that devices
like the TiVO, in keeping the signing keys private (because of
the "Facilitation of a Crime" thing I noted earlier), is violating the GPL,
but not in the manner almost everyone is arguing. The violation is, rather,
with the clause about the license being null and void in event of laws
impacting the delivery of the source. (Because, as I also stated earlier, the
signing keys are part of the source. Since, in some cases, the license on the
hardware prevents running modified binaries (a reason for the digital
signing) companies will keep said keys private - doing otherwise can (and I
can assure you that some lawyer will do this) be construed as "Facilitating
the Commission of a crime". In this case, it'd be "Breach of Contract" -
IANAL, but IIRC, licenses fall under contract law))

DRH

>
> cu
> Adrian



--
Dialup is like pissing through a pipette. Slow and excruciatingly painful.
-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at http://vger.kernel.org/majordomo-info.html
Please read the FAQ at http://www.tux.org/lkml/