lkml.org 
[lkml]   [1996]   [Apr]   [23]   [last100]   RSS Feed
Views: [wrap][no wrap]   [headers]  [forward] 
 
Messages in this thread
/
Date
From
SubjectRe: Must modules be GPL'ed?
The questions you asked are best answered by the _GNU LIBRARY GENERAL
PUBLIC LICENSE_ (commonly named COPYING.LIB) rather than the _GNU GENERAL
PUBLIC LICENSE_ (the file COPYING appearing in the majority of GNU
package).

For purposes of a program which makes system calls, the kernel ends up
acting more as a library than a program. Because of this, the COPYING.LIB
would be more approbate and probably should appear at the top of the
kernel source code tree with a header stating which files/directories it
applies too. Unfortantly, even in the Linux Kernel 1.3.94 this document
only appears buried in the linux/arch/sparc/lib directory (at least it
appears someplace).

Instead, at the root of the kernel 1.3.94 source code, there is the
classic GPL statement (the file named COPYING) with a slight modification
by Linus Torvalds to address the issue of user programs that make system
calls to the kernel. But the COPYING.LIB already addresses this type of
legal issues without any modification.

Also addressed in section 5 & 6 of COPYING.LIB is how the license effects
an user of inline functions from header files. These sections do
allow for situations where object code could be distributed for sale
without source. If the header functions from linux/include directory
where all covered by the COPYING.LIB then it clearly would be possible to
produce commerical kernel modules.

Despite this, it would appear that this document may only apply to a small
sub-section of the kernel code which isn't even used except with the Sun
Sparc port.

It would clear up questions such as this if a larger area of the kernel,
such as the header files in the linux/include directory, where clearly
stated in future releases as being covered by the GNU Library Public
Licence. I believe this would also better honor the intentions of the
GNU philosphy. But, then again, if your a strict GNU'est follower then
you probably also want see the linux/README changed to say something about
GNU/LINUX or lignux (read ftp://prep.ai.mit.edu/pub/gnu/GNUinfo/LINUX-GNU ).

There are also reasons to discourage distribution of modules in only
binary format. The most important is that the Linux kernel is not by any
means static and restructuring of the kernel may render a module
unstable/unusable. I personally would hesitate before investing in a
kernel module that does not make the source available.

- Ben Galliart (bgallia@luc.edu)

On Mon, 22 Apr 1996, Janne Peltonen wrote:

> I just read the GPL but I am still unsure whether loadable kernel
> modules qualify as work derived from the kernel (FAQ, I know..).
>
> It appears that modules are frequently thought to be independent
> entities that are using kernel services through a well defined
> interface. In that respect, modules are much similar to user mode
> programs; they just happen to reside on the other side of a protection
> boundary. However, it has been pointed out by someone that modules
> commonly contain inlined kernel functions imported from kernel header
> files.
>
> I would like to hear your opinions on the interpretation of the GPL in
> this particular case. What is the spirit of the GPL? What do the
> developers of Linux (Linus, are you reading?) think about commercial
> object only distribution of kernel modules?



\
 
 \ /
  Last update: 2005-03-22 13:37    [W:0.995 / U:0.016 seconds]
©2003-2020 Jasper Spaans|hosted at Digital Ocean and TransIP|Read the blog|Advertise on this site