I think this is a timely article since we where just discussing this.
http://www.linux.com/article.pl?sid=06/09/29/164207
Executive Summary:
A buch of GPL code ended up getting distributed in binary only. Over a bunch of smally ports the license files ended up getting dropped.
Key point of the article for LinuxBIOS:
""" As Doelle points out, some of the problems might have been avoided if the GPL header had been included in each file, rather than depending on a single COPYING file containing the text of the GPL. At least this would make it more obvious to developers what the requirements are for their use of GPLed code. """
So we should probably do as Uwe says and start including the GPL header on all source files.
Hi,
On Sun, Oct 08, 2006 at 12:19:26PM -0500, Richard Smith wrote:
So we should probably do as Uwe says and start including the GPL header on all source files.
I agree. Do as Uwe says ;-)
Seriously though, here's a first patch which adds the GPL header to your som/ims/p2b code. This was an easy one as (accoring to svn logs) you are the sole author of all the code (please correct me if I'm wrong).
For other parts of the code base some more elaborate investigations might be needed.
I'd say everyone with svn commit access could/should just add the header to code where he/she is the sole author and owns all of the copyright to the code. We can then take care of the more complex cases later.
Note: The header I used in this patch is the "GPLv2 or later" version. Richard, if you want to use another license, please say so before this gets committed.
Is there a general guideline as to who licenses their code under which license? That would make it easier for me to prepare further patches.
Examples:
Stefan: GPLv2 or later? Who owns the copyright - you or coresystems GmbH or even SuSE(?)
Ron: Is all of your work on behalf of LANL? Do they own the copyright or do you? Do you have to use the BSD-like LANL licenseî(see arch/i386/include/arch/intel.h for an example) or can you just license your stuff under the GPLv2 or later? If so, do you want to do so?
Yinghai: Is all of your work on behalf of AMD? AMD owns the copyright for the code, correct? Is it ok to use the "GPL v2 or later" header as in this patch for the code?
Richard: Is GPLv2 or later fine? Are you the copyright owner or maybe a company you work for?
Others?
HTH, Uwe.
Forgot the patch.
rev E dual core, raminit, Cache_as_ram and CK804 is for Tyan. ... others is for AMD.
YH
On 10/8/06, Uwe Hermann uwe@hermann-uwe.de wrote:
Hi,
On Sun, Oct 08, 2006 at 12:19:26PM -0500, Richard Smith wrote:
So we should probably do as Uwe says and start including the GPL header on all source files.
I agree. Do as Uwe says ;-)
Seriously though, here's a first patch which adds the GPL header to your som/ims/p2b code. This was an easy one as (accoring to svn logs) you are the sole author of all the code (please correct me if I'm wrong).
For other parts of the code base some more elaborate investigations might be needed.
I'd say everyone with svn commit access could/should just add the header to code where he/she is the sole author and owns all of the copyright to the code. We can then take care of the more complex cases later.
Note: The header I used in this patch is the "GPLv2 or later" version. Richard, if you want to use another license, please say so before this gets committed.
Is there a general guideline as to who licenses their code under which license? That would make it easier for me to prepare further patches.
Examples:
Stefan: GPLv2 or later? Who owns the copyright - you or coresystems GmbH or even SuSE(?)
Ron: Is all of your work on behalf of LANL? Do they own the copyright or do you? Do you have to use the BSD-like LANL license��(see arch/i386/include/arch/intel.h for an example) or can you just license your stuff under the GPLv2 or later? If so, do you want to do so?
Yinghai: Is all of your work on behalf of AMD? AMD owns the copyright for the code, correct? Is it ok to use the "GPL v2 or later" header as in this patch for the code?
Richard: Is GPLv2 or later fine? Are you the copyright owner or maybe a company you work for?
Others?
HTH, Uwe.
Uwe Hermann http://www.hermann-uwe.de http://www.it-services-uh.de | http://www.crazy-hacks.org http://www.holsham-traders.de | http://www.unmaintained-free-software.org
-----BEGIN PGP SIGNATURE----- Version: GnuPG v1.4.5 (GNU/Linux)
iD8DBQFFKXxgXdVoV3jWIbQRAi84AJ9mDp7iA/Gx3M9oHCIoB/FNzV98IQCgkxRG zx4iepiauUN9hpFwzJa8Byg= =zf8L -----END PGP SIGNATURE-----
-- linuxbios mailing list linuxbios@linuxbios.org http://www.openbios.org/mailman/listinfo/linuxbios
Richard: Is GPLv2 or later fine? Are you the copyright owner or maybe a company you work for?
GPLv2 or later is fine for me. That said however, very little of that som_gx533 and p2b code was created by me. The som stuff is a copy of the Rumba board and the p2b is a copy of the ims stuff. The ims stuff is a mix and match from several of the other boards. I just stuck it all together and then fixed a few quirks to make the SPD code work.
gpl v2. Let's avoid "or later" until that controversy is resolved.
All LANL code has that weird bsd-like license text, but note it is really GPL. So GPL V2.
ron
Hi,
On Sun, Oct 08, 2006 at 07:53:34PM -0600, ron minnich wrote:
gpl v2. Let's avoid "or later" until that controversy is resolved.
I don't see a problem with "v2 or later". Everybody who doesn't like to use GPLv3 (when it's released) can just chose to use v2. That's perfectly legal.
Ifîwe don't use "v2 or later" for the major parts _now_, but decide (at some point in the future) to convert to "v3 or later" we'll have to go through all of this relicensing stuff again...
All LANL code has that weird bsd-like license text, but note it is really GPL. So GPL V2.
Wait, the actual license is the GPL? It doesn't say so anywhere, so that definately needs to be clarified. As I understood things until now, was that the LANL-text is in itself a license, namely a BSD-ish one. Would LANL agree to relicense their code to GPL, and/or remove the BSD-ish text? It's really confusing...
Uwe.
gpl v2. Let's avoid "or later" until that controversy is resolved.
I don't see a problem with "v2 or later". Everybody who doesn't like to use GPLv3 (when it's released) can just chose to use v2. That's perfectly legal.
FWIW, no one but the original author(s) can decide to re-license "v2 only" code to "v2 or later" (later licenses [can] include extra restrictions, which are explicitly not allowed by the v2 license).
Segher
Hi,
On Mon, Oct 09, 2006 at 06:23:34PM +0200, Segher Boessenkool wrote:
gpl v2. Let's avoid "or later" until that controversy is resolved.
I don't see a problem with "v2 or later". Everybody who doesn't like to use GPLv3 (when it's released) can just chose to use v2. That's perfectly legal.
FWIW, no one but the original author(s) can decide to re-license "v2 only" code to "v2 or later"
Correct.
I'm notsaying to change the license of other people's code (because that's simply not possible). Only the respective copyright owner can change the license.
Uwe.
I don't see a problem with "v2 or later". Everybody who doesn't like to use GPLv3 (when it's released) can just chose to use v2. That's perfectly legal.
Actually I don't think we have a choice.
LWN.net just ran an article on the busybox GPLv2 GPLv3 issue which has caused a fork of busybox. One of the things I see in that article is that GPLv2 only code cannot co-exist with v3 code.
We have included kernel code which is v2 only. Therefore our entire base must be v2 only.
On Mon, Oct 09, 2006 at 11:28:35AM -0500, Richard Smith wrote:
I don't see a problem with "v2 or later". Everybody who doesn't like to use GPLv3 (when it's released) can just chose to use v2. That's perfectly legal.
Actually I don't think we have a choice.
LWN.net just ran an article on the busybox GPLv2 GPLv3 issue which has caused a fork of busybox.
I guess you talk about this: http://lwn.net/Articles/202106/ (which I cannot read until Oct 12, unfortunately).
One of the things I see in that article is that GPLv2 only code cannot co-exist with v3 code.
I'm not too familiar with the (plans for) GPLv3, but that may indeed be the case.
However, any "GPLv2 or later" code is very well compatible with the GPLv3 (by definition). That's one reason why we might want to use "GPLv2 or later" wherever possible.
We have included kernel code which is v2 only. Therefore our entire base must be v2 only.
I don't think so. Some parts are GPLv2, some are "GPLv2 or later", that's fine. The problems only start when we try to move all the code to GPLv3 or "GPLv3 or later". In that case we'll simply fail, as we're not the copyright owners of the kernel code (which is GPLv2 only), so we can't convert all the code to GPLv3.
Still, there would be the possibility to either reimplement the GPLv2 code, or remove it altogether, or nicely ask the kernel guys to relicense it to either GPLv3 or "GPLv3 or later".
OK, one thing I didn't say so far: I am not a lawyer! :) All I say could be complete, utter crap.
Uwe.
Uwe Hermann wrote:
Wait, the actual license is the GPL? It doesn't say so anywhere, so that definately needs to be clarified. As I understood things until now, was that the LANL-text is in itself a license, namely a BSD-ish one. Would LANL agree to relicense their code to GPL, and/or remove the BSD-ish text? It's really confusing...
it's weird, but the LANL code is gpl.
so we just add it in.
ron
Hi,
On Sun, Oct 08, 2006 at 08:42:51PM -0500, Richard Smith wrote:
Richard: Is GPLv2 or later fine? Are you the copyright owner or maybe a company you work for?
GPLv2 or later is fine for me. That said however, very little of that som_gx533 and p2b code was created by me. The som stuff is a copy of the Rumba board and the p2b is a copy of the ims stuff. The ims stuff is a mix and match from several of the other boards. I just stuck it all together and then fixed a few quirks to make the SPD code work.
Hm, ok, so more investigations are needed. I think it's pretty common in the code-base that stuff gets copied and modified. We need to find the person(s) who created the original "template" code and add them as co-authors to those "Copyright 200x" lines...
Do archives of the older CVS trees (with full history) still exist? The svn history only goes back to an "import of freebios2" or so entry. Ah, I just noted that the v1 tree seems to go back to an earlier date. Does it contain _all_ history of the code or was something used before CVS back then?
Uwe.
* Uwe Hermann uwe@hermann-uwe.de [061009 17:56]:
Do archives of the older CVS trees (with full history) still exist? The svn history only goes back to an "import of freebios2" or so entry. Ah, I just noted that the v1 tree seems to go back to an earlier date. Does it contain _all_ history of the code or was something used before CVS back then?
The initial stuff was checked into CVS, which was moved to svn including history.
However it was based on very early versions of "freebios" and "OpenBIOS" (whereas I almost doubt there's something left of that)
Stefan
-
Richard Smith -
ron minnich -
Ronald G Minnich -
Segher Boessenkool -
Stefan Reinauer -
Uwe Hermann -
yhlu