Free as in embrace, extend and extinguish

Why do the current kernel developers forbid use of the DMA-BUF API to non-GPL drivers, most notably NVIDIA's? Isn't this a violation of my freedom to have better drivers? Why does EXPORT_SYMBOL_GPL exist in the first place anyway? If a module is statically linked to the kernel at run-time, the GPL shouldn't apply as it is dynamic. Why doesn't Linux have a stable module interface? Oh, right.

Linux is shit, use a BSD or Solaris.

>>1
Why do the current kernel developers forbid use of the DMA-BUF API to non-GPL drivers, most notably NVIDIA's?
Because they want to force license compliance for all drivers, futile though it might be.
Isn't this a violation of my freedom to have better drivers?
No
Why does EXPORT_SYMBOL_GPL exist in the first place anyway? If a module is statically linked to the kernel at run-time, the GPL shouldn't apply as it is dynamic.
That's not how it works. Modules are considered derivative works and must therefore be GPL compatible
Why doesn't Linux have a stable module interface?
Why do you want a stable module interface? That brings absolutely no benefits what-so-ever. Linux isn't a microkernel in case you haven't noticed.

>>3
Why do you want a stable module interface? That brings absolutely no benefits what-so-ever.
You mean being able to use discontinued or infrequently maintained modules flawlessly isn't a serious benefit? Most importantly, a stable module interface would dramatically lowers the maintenance costs of the Linux kernel and force its developers to actually design good interfaces and document them. Such thoughtful design would also result in increased performance.

>>2
There is not nearly enough manpower for those systems to make them usable. Also, BSD and Solaris audio suck even more than Linux (quite a feat!).

>>4
You're trying to tell me Solaris isn't usable? It's a hell of a lot better than the clusterfucked Linux development.

You are free to use the DMA-BUF API in a non-GPL driver.
You just can't distribute it.

>>6
i'll show u my secret area of non-GPL drivers if u pay me enough

critizicing solaris for audio? really?

>>4
force its developers to actually design good interfaces and document them.
Funniest thing I've read all day!

>>4
Module interface
I don't think you know what that means, unless you're seriously talking about proprietary linux kernel drivers. The module interface is not "stable" because it depends on how the kernel was compiled, which is why you need to recompile/reinstall all modules when you recompile the kernel image.

If you're talking about the internal kernel interfaces, they're unstable because they keep adapting to new technologies. With the code review that goes on you best fucking believe that they're well designed. Interfaces exposed to userspace are rarely changed, regardless of how necessary it might be. Internal macros/functions change when the developers maintaining them decide to change them because: IT'S A MONOLITHIC KERNEL, IT DOESN'T NEED NOR BENEFIT FROM STABLE INTERNAL INTERFACES.

>>1
Are you allowed to do it in BSD-licensed drivers?

^SLACKWARE_SUPREMACY ^SLACKWARE_SUPREMACY ^SLACKWARE_SUPREMACY

>>4
LOL. I play sound on OSS on any machine just fine. Everyone uses the Intel HDA driver anyways.
$ uname                                                                       
FreeBSD

$ uname
Windows

$ uname
Segmentation fault

$ uname
2.11BSD

% uname
Darwin

>>13-16
$
lel, bash.

C:\quarterly report\dont click\really dont click\go away\cp\> uname

ReactOS

>>18
Konichiwa, Thad.

[m]$ uname                     
%DCL-W-IVVERB, unrecognized command verb - check validity and spelling \UNAME[/m]

>>17
I'm >>16, and that's not bash.

>>21
I'm >>12 and what is this?

It's dubs.

>>22
Both of your posts are bad and you should feel bad.

>>14,15,16,17
LOL, thank you, you made my day!
mv -f $USER /dev/null