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

[openip] opencores ml (fwd)





I agree with you about the concepts but when I read the text of LGPL,
mapping to IP cores is hard to do and it is possible to do several
different mapping.  It is why I propose to produce a new License text (or
use an alternative)  so these points could be clear.  I think there is
many people out there that are stopped by this today.  So with a clear
text saying what is possible and what is not, opencores.org could gain
some popularity and usefulness.

To follow your text (the link you sent us)  what part of the work can be
copyrighted :

- vhdl/verilog ? yes
- layout ?       ok
- netlist ?      ok
- chip ?         humm, not that clear for me.

So the copyright notice and "source" shall follow restriction for any
copyrighted stuff and derived work. "Work that uses the library" in the
text.  When it is not a library anymore ?

Does the author of the FAQ has think of all those details ?

>From http://www.opencores.org/OIPC/def.shtml, the goal of opencores.org is
to get "Design of the hardware [..] available".  So the goal is to build
up open design.  Does the text of LGPL point that out clearly ?

And there is a lot of possible hardware related alternative to LGPL:
http://www.opencores.org/OIPC/otherlic.shtml

Richard


---------- Forwarded message ----------
Date: Tue, 25 Feb 2003 22:09:38 +0000
From: nico <nicolas.boulay@ifrance.com>
To: rip@step.polymtl.ca
Subject: opencores ml

I try to send this to the mailling list but does not seems to work.

nicO

On 22 Feb 2003 08:16:35 +0700
Rudolf Usselmann <rudi@asics.ws> wrote:

> On Sat, 2003-02-22 at 04:21, MikeJ wrote:
> > Hi Chaps.
> >
> > I have received an email from someone who wishes to use an opencores
> > core, but has pointed out some problems with the LGPL that many of
> > us to distribute our hardware cores.
> >
> > advice ??
> > cheers.
> > MikeJ
>
>
> This is exactly why I don't use GPL or LGPL ...
> Both of them have been written with software in mind, and can
> not be applied to IP cores.
>

http://www.gnu.org/licenses/gpl-faq.html#GPLOtherThanSoftware

"You can apply the GPL to any kind of work, as long as it is clear what
constitutes the "source code" for the work. The GPL defines this as the
preferred form of the work for making changes in it."

LGPL/GPL could be aslo imply to numerical images
(http://savannah.nongnu.org/projects/cardpics). GPL/LGPL could apply if
you have a work where copyright law could apply (code/plan/...) that
it's use to create "something" (binaries...). But even hardware, i ask a
lawyer to have a confirmation (she had worked on the applicability of
GPL in the french law).

> > The terms in the LPGPL (see below) state that the end user must be
> > able to replace the core (library) by another version. In the case
> > of an ASIC production this would imply that either the synthesized
> > netlist or the post-layout files for the whole chip should be
> > provided free of charge.
> >
> > This is unacceptable. I would hope the idea behind the open cores
> > would be that any changes made to the cores _themselves_ would have
> > to be released. But releasing the whole ASIC project (of which the
> > controller core would only be a tiny part) makes use of this type of
> > core in any sort of commercial project impossible (and together with
> > it the benefits of having commercial developers improve the core and
> > release the changes will disappear).
>
> I totally agree with you. I am doing a project for an customer as
> well, and am facing the same problem. I see two options out of that:
>
> 1) Emailing the author of the IP Cores I want to use and try
>    to make a independent agreement and license with him/her
>

That's not "open" hardware any more. That's a kind of freeware or
shareware.

> 2) Write my own cores instead of using a GPL/LGPL OpenCores core.
>
> > Is there any chance of providing an exception to those terms?
> > Something like:
> >
> > This library is free software; you can distribute it and/or modify
> > it under the terms of the GNU Lesser General Public License as
> > published by the Free Software Foundation; either version 2.1 of the
> > License, or (at your option) any later version. However, as an
> > exception to section 6 of the GNU Lesser General Public License, you
> > are allowed to provide the"work that uses the Library" in a form
> > (e.g. hardware) that does not give the user the possibility to
> > replace the Library with a modified version.
>
> Isn't there a provision in the GPL and LGPL licenses that one
> can NOT make exceptions ?
>

You can't mix code with and without exception. For exemple Linux kernel
have exception to allow proprietary modules.

But you could modify the GPL:
http://www.gnu.org/licenses/gpl-faq.html#ModifyGPL

You could add exception :
http://www.gnu.org/licenses/gpl-faq.html#WritingFSWithNFLibs

The most interresting point to controll where proprietary code could
stand :
http://www.gnu.org/licenses/gpl-faq.html#LinkingOverControlledInterface

Only the copyright holder of code could put this rules.

> > If not I'm afraid I will have to scrap your core (like all
> > LGPL-licensed cores) from my evaluation for inclusion in a future
> > ASIC.

LEON cpu written by ESA (www.gaisler.com) is plan to be used as the main
cpu of SoC in all european space industry. And it's under LGPL.

>
> Not ALL cores require GPL or LGPL licenses.
>
> Any core I have written does not require any license at all.

So it's public domain ? Almost like BSD stuff ? Are you not affraid of
what happen to the "wine" project which switch to GPL because a compagny
sell wineX which is 90% wine + the library for directx compatibility,
but not as free software. So the wine developper are "abused" by this
compagny that give nothing back to the world of free software.

Or are you not affraid to have compagny using your core but make very
bad advertising on opencores.org like microsoft as done by using BSD
stuff (like ftp, telnet, ... on NT and even the BSD IP stack) but always
make FUD against free and open software. That's true that now they only
fight the GPL because they can't "embrace and extent" such product.

> Several people have send me their own license agreements to
> sign. That is to protect them selfs from law suits. Which I
> have signed, after careful review.

GPL have a great chapter on limited responsability.

>
> I know there are other people who followed my model as well.
> You should review each core independently.

All of this sound strange. What is "open" here ? Where is the copyleft ?

Regards,
nicO

>
>
> Regards,
> rudi
--
To unsubscribe from openip mailing list please visit http://www.opencores.org/mailinglists.shtml