User: Password:
|
|
Subscribe / Log in / New account

On the value of EXPORT_SYMBOL_GPL

On the value of EXPORT_SYMBOL_GPL

Posted Oct 6, 2005 21:46 UTC (Thu) by fenrus (guest, #31654)
In reply to: On the value of EXPORT_SYMBOL_GPL by nchip
Parent article: On the value of EXPORT_SYMBOL_GPL

> Many Linux distributors bundle on their liveCD's and installers both Linux kernel and binary-only modules/firmwares, so this isn't an embedded-only phenomea.

please name one.. I'm looking for someone to sue to set a precedent for binary kernel modules and this sounds like an closed-and-shut case...


(Log in to post comments)

On the value of EXPORT_SYMBOL_GPL

Posted Oct 7, 2005 9:37 UTC (Fri) by jschrod (subscriber, #1646) [Link]

You want a distribution? Easy: SUSE and VMware.

While the source of the VMware kernel modules is distributed, it is not under an open-source license, but proprietary code.

In fact, each SUSE distribution has a package that's called kernel-nongpl. In 9.1, there were 22 non-GPL kernel modules.

Please note that for many professional users, VMware is a necessity. Luckily, Linux made clear that his opinion on binary-only kernel modules is different from yours.

So, will you now go forward and sue Novell? Good companionship you'll have.

Joachim.

On the value of EXPORT_SYMBOL_GPL

Posted Oct 7, 2005 9:41 UTC (Fri) by fenrus (guest, #31654) [Link]

> Luckily, Linus made clear that his opinion on binary-only kernel modules is different from yours.

Linus never said he considered binary only modules allowed. That's an urban myth he himself has defused many times.

On the value of EXPORT_SYMBOL_GPL

Posted Oct 7, 2005 9:50 UTC (Fri) by jschrod (subscriber, #1646) [Link]

No, you're wrong.
Read the thread at http://kerneltrap.org/node/1735

Linus says that one has to look at each kernel module seperately and judge anew if it's a derived work or not.

You told, as far as I understood you, that all binary-only modules are by definition derived works. That is not Linus' opinion, as stated by me and proofed by the lkml thread above.

Joachim

On the value of EXPORT_SYMBOL_GPL

Posted Oct 8, 2005 1:13 UTC (Sat) by fenrus (guest, #31654) [Link]

you missed the point; this part of the thread isn't about derived works; it's about the "ship together" clause in the gpl which has nothing to do with "derived work" but with "being independent".

(and anyway linus didn't say binary modules are allowed, he said the derived work part is a case by case basis)

On the value of EXPORT_SYMBOL_GPL

Posted Oct 8, 2005 11:57 UTC (Sat) by jschrod (subscriber, #1646) [Link]

you missed the point; this part of the thread isn't about derived works; it's about the "ship together" clause in the gpl which has nothing to do with "derived work" but with "being independent".
Ah, you mean you never reacted on nchip who wrote ``So it boils back to the "is the binary-only module derived from Linux" argument.'' and you never threatened to sue anybody who distributed a binary-only module because it would be ``a closed-and-shut case.''? Well, go forward, follow-up your strong words and sue Novell.
(and anyway linus didn't say binary modules are allowed, he said the derived work part is a case by case basis)
Exactly my words. Go back and read my post: I never wrote that Linus allows binary-only modules. I wrote that his opinion differs from yours in so far as distribution of binary-only modules are not a ``closed-and-shut case'' as you wrote. You just agreed to that and make it sound as if it was your opinion in the first place. Go figure, since that's your style of discussion, I'm out now.

Joachim

On the value of EXPORT_SYMBOL_GPL

Posted Oct 8, 2005 12:40 UTC (Sat) by dwmw2 (subscriber, #2063) [Link]

Ah, you mean you never reacted on nchip who wrote ``So it boils back to the "is the binary-only module derived from Linux" argument.''
Fenrus did respond to this. He made the analogy about glue -- highlighting the difference between two items which are each useful in their own right but which happen to be shipped together coincidentally, and two items which are fundamentally interdependent such that each cannot function without the other.

That's basically the difference between a collective work and 'mere aggregation on a volume of a distribution medium', which nchip seemed to have missed. It seems as if nchip was taking 'mere aggregation' to include any kind of combined work, even when the parts are fundamentally interdependent.


Copyright © 2017, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds