Re: [Gnewsense-dev] [KERNEL] report on actual bugs

[Sam Geeraerts]

crap0101 schreef:
> Hi!
> Here's a report about open bug regarding the kernel:
>
> * (mach64.h) http://bugs.gnewsense.org/Bugs/00212 
>
> About the mach64.h bug, there's no news.  I and the fsf licencing team
> have tried to contact ATI, but no answer yet.
> If someone have technical/licence knowledge useful for this, please
> share it :) 

After this long it's doubtful that you'll get an answer back. The FSF
have said it's better to remove it then [1]. It might be worth it to let
ATI know that we're removing it. Maybe they respond better to actions
than to words (if they care at all).

> * (mroute.h) http://bugs.gnewsense.org/Bugs/00287
>
> Regarding mroute, I wrote also in the linux-libre mailing list, this is
> the A. Oliva answer
> http://www.fsfla.org/pipermail/linux-libre/2009-June/000654.html
> about. Like suggested, I took a look into the mroute package, but like
> said in the bug's page, the fragmentetion make difficult to find the
> lines of code used in the lnux version. This may take some time (I still
> don`t do that).
> More info about this bug:
> https://fedoraproject.org/wiki/Licensing#Bad_Licenses
> http://www.mail-archive.com/address@hidden/msg04177.html

As Alexandre said, it's not certain that this code was taken directly 
from mroute and it would be a good idea to check how it came into Linux 
in the first place. You might want to contact Linux Kernel Newbies [1] 
to get an idea of how to go about that.

> * (cfi.h) http://bugs.gnewsense.org/Bugs/00243
>
> mailing the linux-libre list, A. Oliva said that's not a bug:
> http://www.fsfla.org/pipermail/linux-libre/2009-June/000653.html
> what do you think?
>
> Also, I got response for address@hidden They said:
>
> Il giorno mar, 30/06/2009 alle 16.31 -0400, Donald R Robertson III via
> RT ha scritto: 
> > Hello,
> >
> > Thanks for bringing this to our attention.
> >
> > > address@hidden - Fri Jun 05 09:06:10 2009]:
> > >
> > > Hi, 
> > > I'm Marco.
> > >
> > > checking the Linux Kernel, I've got probably non free file "cfi.h"
> > > located in /include/linux/mtd folder of the kernel's source package.
> > > Is similar to the mach64.h [gnu.org #384122] issue but,
> > In this case it might be a little different. As far as I can
> understand,
> > we have some code that was made available by Intel through that site,
> > and that site actually does include licensing terms for the content
> > being downloaded. Unfortunately the terms are extremely restrictive,
> and
> > are absolutely non-free.
> >
> > Intel has been supportive of free software in the past (they made
> > numerous assignments to the FSF), so it might be possible to get
> > alternate permission to use this code.
> > [...]

Contrary to what I said when you first brought this issue up, I tend to 
agree with Alexandre. As I understand it, this file was created with the 
help of information that may be patented, but it's not a copy of some 
non-free code. So it's different from the mach64 case, where (possibly) 
some code was copied from one or more examples.

[1] http://lists.gnu.org/archive/html/gnewsense-users/2008-11/msg00034.html
[2] http://kernelnewbies.org/