Busybox seems to be the GPL hook that they catch all these vendors on. But I can't imagine vendors making modifications, much less usable enhancements, to such a simple set of tools. The kernel is the place SFLC should be able to get them on, but the binary driver exception is so big the GPL has no power there.
Posted Dec 14, 2009 19:59 UTC (Mon) by rahvin (subscriber, #16953)
[Link]
Harold Welte has copyright in BusyBox and is the most prolific enforcer of the GPL in the world. The SFLC does Harold's enforcement actions in the US. I should also point out that once SFLC has been involved they ensure compliance with all GPL applications, not just Busybox. The beauty is that Busybox is almost a requirement in the embedded space if you are using any GPL code and make it an easy target for enforcement because it's easy to find, easy to prove and heavily enforced making it a slam dunk case.
In addition, once SFLC files suit posting the code isn't enough, they usually revoke the license and refuse to reissue until they are reimbursed for legal fees + a little extra for the next action and almost always require the companies involved to appoint a GPL compliance officer so it doesn't happen again.
Why is always busybox?
Posted Dec 14, 2009 20:13 UTC (Mon) by BrucePerens (guest, #2510)
[Link]
Nobody ever violates the license on busybox for a smart reason. Proprietary code would not be connected with it. It's just a lack of due diligence combined with a failure to respond when contacted.
I could confuse things tremendously by taking my original code-base and applying the BSD license to it. :-)
Bruce
Why is always busybox?
Posted Dec 14, 2009 20:33 UTC (Mon) by atai (subscriber, #10977)
[Link]
No, you want to apply the GPLv3 to it. That is better!
Why is always busybox?
Posted Dec 14, 2009 20:42 UTC (Mon) by BrucePerens (guest, #2510)
[Link]
I could dual-license it.
Why is always busybox?
Posted Dec 15, 2009 22:00 UTC (Tue) by davi (guest, #18853)
[Link]
You could, but I hope you do not do it.
User freedom:
* GPL version >= 3 is the future.
* AGPL >= 3 for web applications.
Why is always busybox?
Posted Dec 14, 2009 20:37 UTC (Mon) by armijn (subscriber, #3653)
[Link]
Ehm, Harald has no copyrights on BusyBox and SFLC does not enforce his copyright.
Why is always busybox?
Posted Dec 14, 2009 22:03 UTC (Mon) by JoeF (subscriber, #4486)
[Link]
Harald Welte has copyrights on ipfilter, if I remember right.
That's also pretty much everywhere in devices that connect to the network.
The busybox developers include Bruce Perens.
Why is always busybox?
Posted Dec 14, 2009 23:15 UTC (Mon) by rahvin (subscriber, #16953)
[Link]
Maybe I'm showing my ignorance, but I thought part of the ipfilter code was in busybox these days?
Why is always busybox?
Posted Dec 14, 2009 23:30 UTC (Mon) by armijn (subscriber, #3653)
[Link]
There is definitely no netfilter or iptables in BusyBox (I just grepped through the latest sources) :-)
Why a "little extra"?
Posted Dec 14, 2009 21:30 UTC (Mon) by vondo (guest, #256)
[Link]
It seems to me the time for compliance plus a little bit of funding, a slap on the wrist, is past. Any company that is paying attention should know they can't get away with this. I would really like to see a few of these companies hit hard for an amount of money that actually *hurts*. Then maybe others won't be so cavalier about it.
As it is, a company can look at this as "Let's use this code and not worry about it. If someone finds out, we can release it later. Or we can ignore it and maybe we'll have to pay some small amount in a lawsuit." That perception has to change or SFLC is going to be doing this forever.
Why a "little extra"?
Posted Dec 15, 2009 2:51 UTC (Tue) by jamesh (guest, #1159)
[Link]
The companies should be concerned. In cases like this, the copyright holders could ask the court for an injunction against the sale of the infringing devices.
Not being able to sell your device while still having to keep the existing stock around is not cheap and should be a pretty big deterrent.
Why a "little extra"? Because that's the right amount
Posted Dec 15, 2009 10:33 UTC (Tue) by tialaramex (subscriber, #21167)
[Link]
The SFLC is probably content to do this forever. Lawyers who prosecute people for fraud or theft will be doing that forever too.
Punitive costs aren't what we want as a community - the objective is to ensure that around the world, company directors think of GPL compliance as non-optional, like paying their taxes, just a cost of doing business. Winning and settling a long series of lawsuits establishes that thinking without multi-million dollar amounts.
If the SFLC gets enough money from each such lawsuit to fund the next, plus a "little extra" then we're doing well enough IMO.
Why a "little extra"? Because that's the right amount
Posted Dec 18, 2009 12:42 UTC (Fri) by malor (subscriber, #2973)
[Link]
Yeah, it shouldn't be a profit center, just a method to keep the capitalists in line. Their use of free software is wonderful, and we want to encourage it, not discourage it. We want them to understand that there some rules they have to follow to get access to all that great code, but that the rules aren't especially onerous and that using GPLed code isn't scary. We want them to (correctly) think that it's often a better solution than BSD or proprietary packages.
SOME of them may even turn into contributors, but getting that code into as many devices as possible is a good end in and of itself, because it usually makes devices much more hackable. It gives the free software community more control over their hardware. Even if they never contribute a line of code, they ARE contributing devices that RUN that code.
Why a "little extra"?
Posted Dec 15, 2009 16:17 UTC (Tue) by endecotp (guest, #36428)
[Link]
> maybe [manufacturers] have to pay some small amount in a lawsuit
As far as I can recall, the amount paid has not been disclosed in any of these cases. I think that's unfortunate, and I encourage SFLC to settle future cases without a clause preventing disclosure of the amount. Without knowing if it's $1000 or $1000000, we're unable to judge whether these cases make future infringement more or less likely.
Why is always busybox?
Posted Dec 14, 2009 22:19 UTC (Mon) by arjan (subscriber, #36785)
[Link]
what binary driver exemption???
Why is always busybox?
Posted Dec 14, 2009 22:36 UTC (Mon) by daney (subscriber, #24551)
[Link]
Perhaps it is the de facto binary driver exemption that allows companies to ship products using GPUs made by $LARGE_CHIPMAKER using binary driver blobs.
Why do we have EXPORT_SYMBOL_GPL() anyhow if EXPORT_SYMBOL() doesn't imply some sort of non-GPLness?
Why is always busybox?
Posted Dec 14, 2009 22:55 UTC (Mon) by jmm82 (guest, #59425)
[Link]
I have had a long discussion about this before and basically EXPORT_SYMBOL_GPL() has no legal implications and is more of a expression of one's opinion towards the exporting of symbols in the kernel than a legal binding to the symbol.
Why is always busybox?
Posted Dec 14, 2009 23:53 UTC (Mon) by rgmoore (✭ supporter ✭, #75)
[Link]
Linus has clearly expressed his opinion that kernel modules are not necessarily derivatives of the kernel. Something that was originally written for Linux and depends explicitly on it would clearly be a derivative. But Linus's opinion is that a module that was originally written for a different OS and was ported to Linux with minimal changes would not count as a derivative of Linux. Since the GPL depends on copyright provisions that apply to derivative works, it wouldn't apply to a module that doesn't count as a derivative. As far as I can tell, Linus's analysis is correct, and a suit against a company that ported its driver for another OS and released it under a non-GPL license- NVidia being an obvious example- would fail in court.
Why is always busybox?
Posted Dec 15, 2009 5:08 UTC (Tue) by drag (subscriber, #31333)
[Link]
Who decides what a derivative is is not Linus, GPL, or anybody else except
a
judge interpreting law and precedent. It's absolutely possible that one
module can be a derivative and another one is not. Most above people (and
probably you also) understand that 'Derivative' is a very specific legal
term that is included in the copyright law of the USA. Any copyright
license, like the GPL, is limited by the scope of that legal definition. It
does not matter what the copyright holder _wants_ something to be, his
power in controlling the work of others is limited by "derivative", at
least in copyright law.
With Nvidia they are shoehorning a great big blob of Windows driver code
into the Linux kernel. The portion of the code that is specifically made
for
Linux is GPL'd. So the argument goes that the non-GPL'd parts of the driver
are actually written for a different OS or at least a multitude of OSes and
since it was not developed for Linux and does not depend on Linux kernel
code then it is not derivative.
...
Presumably. Who knows how true that is. It may be possible to sue Nvidia,
but it is certainly not a slam-dunk case and, anyways, it seems that nobody
is interested in doing so.
Why is always busybox?
Posted Dec 15, 2009 17:59 UTC (Tue) by Trelane (subscriber, #56877)
[Link]
the other thing is that the GPL is explicity *not* a use license. So long as the end user is the one doing it, they cannot argue that binary-only drivers are not legal. The interface between the binary and the GPL-ed software is the crucial bit, as it must be distributed to the end-users.