Re: GPLv3 Position Statement

From: Linus Torvalds
Date: Thu Sep 28 2006 - 11:25:07 EST

On Thu, 28 Sep 2006, Linus Torvalds wrote:
> It should also be pointed out that even a "GPLv2 or later" project can be
> forked two different ways: you can turn it into a "GPLv3" (with perhaps a
> "or later" added too) project, but you can _equally_ turn it into a "GPLv2
> only" project.

Btw, it should be stated here: I'm not advocating either of the above. If
a license says "v2 or later", anybody who removes an explicit right
granted by the people who originally wrote and worked on the code is just
being a total a-hole.

Quite frankly, if the FSF ever relicenses any of their projects to be
"GPLv3 or later", I will hope that everybody immediately forks, and
creates a GPLv2-only copy (and yes, you have to do it immediately, or
you're screwed forever). That way the people involved can all vote with
their feet.

I think the same is true of code that is licensed "GPL or BSD dual
licensed". If I notice a patch that removes the BSD dual-license for a
file, I won't apply it to the kernel I maintain unless there is some
really pressing reason that I can't even think of off-hand (of course,
that doesn't mean it can't have happened - if it came through a
sub-maintainer I would likely never even have noticed).

Or, indeed, as in the case of the reiserfs code: it's dual-licensed "GPLv2
or any other license as per Hans Reiser".

Btw, I have always found it funny how some people have no problem at all
with "GPLv2 or later", but then complain about reiserfs: it is _exactly_
the same dual-license, it's just that a different legal entity controls
the other yet-to-be-determined license.

The only _real_ difference is that in the case of reiserfs, the "other
entity" is actually the original author of the code, so I _personally_
actually very strongly feel that the reiserfs case is _better_ than "GPLv2
or later". In the reiserfs case, the person who does the relicensing is
actually the same people who wrote the original code and maintained it.

That's how it should be (of course, I think Reiser isn't actually
actively maintaining reiserfs any more, so at some point his "moral
rights" do end up weakening, but in the absense of any big rewrites, I
don't think that has happened yet in this example - I just wanted to
point out that things aren't "black-and-white" and "original author"
only gets you so far if you then leave the project).

I know certain people don't like the reiserfs license - they've complained
to me. At the same time, I know some of those same people themselves
expressly use "GPLv2 or later". I think those people have a serious
disconnect in their logic.

To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@xxxxxxxxxxxxxxx
More majordomo info at
Please read the FAQ at