[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Kernel Firmware issue: are GPLed sourceless firmwares legal to distribute ?

On Tue, Oct 17, 2006 at 03:49:25PM -0400, Nathanael Nerode wrote:
> The answer to the question in the subject is simple: NO.
> This is a matter of copyright law.  If we do not have permission to 
> distribute, it is illegal to distribute.  GPL grants permission to 
> distribute *only* if we distribute source.  So, GPLed sourceless == NO 
> I will list the usual caveats so that nobody else brings them up.
> (1) Obviously if we have an alternate license (dual-licensing) which doesn't 
> require source we can use that license.
> (2) If the material is so trivial it is uncopyrightable we can obviously 
> distribute it.  (The classic example is CRC tables, which contain no 
> creative content beyond the CRC polynomial which is generally public 
> domain.)  Likewise if it was published prior to 1988 in the US without
> copyright notices, or is in the public domain for some other reason.
> (3) If the copyright holder for the firmware donated the firmware to Linux
> with the understanding that it would be redistributed by Debian and other 
> distributors, this may constitute an implicit license to distribute.  This 
> would be a case of dual-licensing, but an unpleasant one because we'd be
> relying on an *implied* license.  This requires tracing down the donation of
> the material to the Linux kernel and ascertaining the state of mind of the
> donor (perhaps by reading press releases).  This clearly applies only to 
> some of the firmware; other pieces have no such 'paper trail'.  Also, this
> implicit license *does not* include a license to modify, because I've never 
> seen any indication that any firmware donor intended that their
> firmware be modified.
> (4) If the hex lumps really are the preferred form for modification, then
> we have the source and this is not a case of 'sourceless firmware'.  I have
> not yet seen a case where there is any evidence that this is true.  It is,
> however, theoretically possible.  If the firmware author came forward and 
> said "Yes, that's the form in which we modify the firmware", this would be 
> the case.

Thanks Nerode for this complete reply.

It seems thay 3) is probably the best way we have to be able to distribute
sourceless de-facto GPLed firmwares, but as you say, it will be a mess.

I suppose the firmwares resolutions, both the one voted, and the one i
proposed, both allow to include those firmwares into main under these
conditions, for etch only, altough the resolution voted upon is probably much
less clear in its wording. It is regretable that Manoj losed patience suddenly
after more than a month an a half of discussing the issues. But we will see.

I think we all now await impatiently the statement of the RMs on what will
happend with the tg3 and acenic firmwares, and if we need a new vote or not.


Sven Luther

Reply to: