From: Daniel Hazelton <dhazelton@enter.net>
To: Dan Williams <dcbw@redhat.com>
Subject: Re: rt73usb: support for wireless in Kohjinsha subnotebook
Date: Tue, 23 Oct 2007 18:19:09 -0400
User-Agent: KMail/1.9.7
Cc: Pavel Machek <pavel@ucw.cz>, Ivo van Doorn <ivdoorn@gmail.com>,
       Luis Correia <luis.f.correia@gmail.com>,
       kernel list <linux-kernel@vger.kernel.org>,
       linux-wireless@vger.kernel.org, rt2400-devel@lists.sourceforge.net,
       mwallis@serialmonkey.com
References: <20071020184100.GA11640@elf.ucw.cz> <200710231541.42943.dhazelton@enter.net> <1193174828.7110.15.camel@localhost.localdomain>
In-Reply-To: <1193174828.7110.15.camel@localhost.localdomain>
MIME-Version: 1.0
Content-Type: text/plain;
  charset="utf-8"
Content-Transfer-Encoding: 7bit
Content-Disposition: inline
Message-Id: <200710231819.09919.dhazelton@enter.net>
Sender: linux-kernel-owner@vger.kernel.org
Content-Length: 7301
Lines: 148

On Tuesday 23 October 2007 17:27:07 Dan Williams wrote:
> On Tue, 2007-10-23 at 15:41 -0400, Daniel Hazelton wrote:
> > On Tuesday 23 October 2007 14:54:54 Dan Williams wrote:
> > > On Tue, 2007-10-23 at 13:07 -0400, Daniel Hazelton wrote:
> > > > On Tuesday 23 October 2007 10:05:12 Dan Williams wrote:
> > > > > On Tue, 2007-10-23 at 00:00 +0200, Pavel Machek wrote:
> > > > > > Hi!
> > > > > >
> > > > > > > > > Yes, I'm quite sure. There's MODULE_LICENCE("GPL"), IIRC.
> > > > > > > >
> > > > > > > > That doesn't say much, some manufacturers add that line to
> > > > > > > > their driver just to prevent the module loader complaining
> > > > > > > > about a non-GPL driver...
> > > > > > > >
> > > > > > > > There should be a copyright notice or a license file
> > > > > > > > accompanied with the driver that clearly states the license
> > > > > > > > of the driver.
> > > > > > >
> > > > > > > Lacking an explicitly stated license it can be argued that,
> > > > > > > since the MODULE_LICENSE() macro is meant to define the actual
> > > > > > > license on the code, this code is GPL. No, it isn't an explicit
> > > > > > > definition, but lacking any other signs of the license, the
> > > > > > > implicit declaration of it being GPL is (or should be) enough
> > > > > > > to deflect charges of copyright infringement.
> > > > > >
> > > > > > Yep, I believe this driver is GPLed. They published the source
> > > > > > and there's nothing to suggest otherwise, and there's explicit:
> > > > > >
> > > > > > #define DRIVER_AUTHOR                   "Jeff
> > > > > > Lee<YY_Lee@issc.com.tw>" #define DRIVER_DESC
> > > > > >    "IS89C35 802.11bg WLAN USB Driver" MODULE_LICENSE("GPL");
> > > > >
> > > > > If there isn't an explicit COPYING or LICENSE file or something
> > > > > distributed with the driver, and if there aren't copyright/license
> > > > > headers at the top of the files in question, I have a hard time
> > > > > agreeing that MODULE_LICENSE("GPL") _definitely_ means that the
> > > > > author has GPL-ed the driver intentionally.  Of course that's the
> > > > > way it's supposed to work, but to me this doesn't pass sufficient
> > > > > muster to be definitely called GPL without additional
> > > > > clarification.
> > > > >
> > > > > Dan
> > > >
> > > > Lacking any other indication MODULE_LICENSE is supposed to mark the
> > > > license that the code is being distributed under. If companies are
> > > > intentionally
> > >
> > > Step 1: Ask the author.
> >
> > Agreed. This should have been done before this discussion even started.
> >
> > > Step 2: if the author doesn't reply, then we can have this discussion
> > >
> > > MODULE_LICENSE is just a random string that could have been added by
> > > anybody, not necessarily the author.  Unless you can determine the
> > > intent of the author explicitly, a single MODULE_LICENSE is not
> > > sufficient to concretely determine the license of the code.  It's only
> > > in one file.  There is nothing to explicitly state the overall license
> > > of the whole work unless each file has a header referring to the
> > > license or unless there is a license document distributed with the code
> > > as a whole.
> > >
> > > In the absence of any other indication, MODULE_LICENSE doesn't not
> > > concretely determine the license of the code.  You can assume it does,
> > > but that's your gun to put to your own head.
> >
> > The intent of MODULE_LICENSE is to mark the license on the code. This is
> > clearly stated in several places in Documentation/ (if my memory serves).
> >
> > > > mis-using this to get around the "internal interfaces" limitations
> > > > (where some interfaces are not available unless the module is GPL'd)
> > > > and the warning message printed in the logs when the module is not
> > > > GPL'd then they are (technically) in violation of the law.
> > > > (interfaces that are GPL only are considered so internal to the
> > > > kernel that using them makes your code GPL because of the inclusion
> > > > of GPL'd code. And no - I am not going to get into that discussion -
> > > > it's pointless)
> > >
> > > Just because the module may be loading illegally says _nothing_ about
> > > the license of the code.
> >
> > No, but it is a mis-use of MODULE_LICENSE, which is supposed to state the
> > correct license on the code.
> >
> > > > In the end, using MODULE_LICENSE for any purpose other than declaring
> > > > the chosen license for the code is deceptive. So it is easily
> > > > arguable that by
> > >
> > > "deceptive" is also not "this code code is definitely GPL".  Doesn't
> > > matter whether it's deceptive or not.  We do not know that the code is
> > > GPL.
> >
> > Deception in order to create a situation whereby you can prosecute people
> > for violation of the law is illegal. Therefore not distributing the code
> > with any indication as to the license other than MODULE_LICENSE and
> > attempting to prosecute afterwards is illegal. QED: even if the code is
> > not GPL'd (and such could be learned by contacting the author), the fact
> > that it ships without any indication of the license other than
> > MODULE_LICENSE implies that the license is what is stated and prosecution
> > on the grounds that it isn't becomes entrapment.
> >
> > > > not including any license with the code other than the MODULE_LICENSE
> > > > statement and then trying to prosecute because MODULE_LICENSE doesn't
> > > > accurately state the license on the code is entrapment and illegal.
> > >
> > > Arguable doesn't mean that it's concrete enough to pass legal muster. 
> > > I am not a lawyer, but this just doesn't pass the bar.
> >
> > I know several and have asked one that is a very good friend. He agreed
> > with my interpretation of the presented facts - that since the only
> > indication of what the license on the code might be is MODULE_LICENSE it
> > can be safely assumed that the license is what that states. Any attempt
> > to later prosecute because the license is not what is stated would
> > constitute entrapment - which is illegal.
>
> I asked Tom Callaway, who heads up all the legal/licensing issues for
> Fedora.  He said essentially "it won't be in Fedora" on just the basis
> of MODULE_LICENSE.  I can run it past Red Hat lawyers too.  It's just
> not clear with only MODULE_LICENSE, and not clear pretty much means
> Don't Touch until it becomes clear.

I have to agree with this. There are better ways to determine the license - 
including contacting the author. However, in the lack of any other source of 
information about the license MODULE_LICENSE can be used as a (somewhat) 
defensible position. It'd cost a lot of money and be a very hard fight, 
though - which is one reason why no company would (or should) take the 
chance. Nor should any person, unless they happen to have the money to fight 
the battle.

DRH

>
> Dan
>
> > DRH
> >
> > PS: note that all legal information contained here-in is only known to be
> > valid in the US.


-- 
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@vger.kernel.org
More majordomo info at  http://vger.kernel.org/majordomo-info.html
Please read the FAQ at  http://www.tux.org/lkml/