Another Netfilter GPL enforcement
[Posted March 2, 2004 by corbet]
[Editor's note: the following article was sent to us by Harald
Welte, the leader of the Netfilter project.]
Two weeks ago you might have read about the netfilter/iptables project
having resolved a GPL infringement case by a German networking gear vendor
called Allnet.
Today, the netfilter/iptables project made a similar
announcement.
The major difference: This time it's not about some unknown German company,
but a large international hardware vendor: Fujitsu Siemens Computers (http://www.fujitsu-siemens.com).
So you might ask yourself: Why is there a sudden rise in pushing for GPL
enforcement by the netfilter/iptables project? The remainder of this article
will try to give you an answer from the project's point of view.
Everything started with the (in)famous Linksys WRT54G case. First postings
about this issue are dating back to 7 June 2003 (http://lwn.net/Articles/35713/).
Shortly thereafter, an FSF-led alliance for making Linksys comply to the
GPL was formed. Since Linksys used netfilter/iptables to implement packet
filtering and NAT on their device, the netfilter core team was invited to
join that alliance.
Initially this seemed like great idea. The head of the netfilter core team
joined that alliance and provided information about how netfilter/iptables code
was used. Even one of our GPL-licensed libraries had been
linked into a proprietary,
binary-only library.
Two months after first contact between the FSF-led alliance and Linksys,
they finally published some code. This was just some random kernel source,
and utterly incomplete (see http://lwn.net/Articles/51399/
and http://lwn.net/Articles/53140/).
Despite a statement by the Cisco legal director about their
world-class leading GPL compliance, it took them four months to provide a
full source code release. To my knowledge, no compensation was paid, and
none of the previous customers had been informed about their rights and
obligations under the GPL.
The result is clear: The 'soft pressure' way of pushing for license
compliance doesn't work. In fact, it encourages vendors to violate the GPL
in the first place. They don't lose anything by not complying with the
license. First, they release an infringing product. Second, somebody has
to find out that they use GPL licensed code. Then, one of the original
authors has to push for license compliance. The vendor would then further
delay that issue for any time he wishes, and in the end he finally makes a
source code release. But by having not lost anything in this strategy, why
would he even bother complying with the GPL next time?
Meanwhile, more and more embedded networking devices like WLAN access
points or DSL routers turned out to use netfilter/iptables without
complying with the GPL license terms. Dissatisfied with the FSF way of GPL
enforcement, the netfilter core team decided to take legal action against
Allnet GmbH in Germany. We were prepared to apply for a preliminary
injunction banning them from further sale of their devices.
Luckily, Allnet replied immediately to the warning notice, and they were very
interested in resolving this issue out-of-court. Our main goal was and is to
make people comply with the the GPL license terms. It is not in our intent to
harass commercial users of netfilter/iptables, or to try to squeeze money out
of them. All we're asking for is license compliance. If those vendors were
using some proprietary licensed code, they would also have to comply with
its license.
Why should they allowed to behave different in case of the GPL?
Allnet understood and even agreed to inform all prior commercial buyers of the
respective products. In addition, they show their support for free software by
making donations to respected charities in the free software world.
Encouraged by this first settlement, we decided to go after every single
infringing use of netfilter/iptables code we are aware of. From your author's
point of view, this is the only way of raising conciousness about free software
licensing in corporate management.
Fujitsu-Siemens was next on the list, and resolved in an equally friendly way.
Many more are pending, stay tuned.
(
Log in to post comments)