Have I lost my mind?
I've always thought that NVIDIA's linux kernel support was incredibly
closed source, but I swear I just saw a download link for the kernel
module sources at http://www.nvidia.com/view.asp?PAGE=linux
was I mistaken or is this something new?
Erik
On Wed, 2001-12-05 at 15:22, Erik Elmore wrote:
> Have I lost my mind?
>
> I've always thought that NVIDIA's linux kernel support was incredibly
> closed source, but I swear I just saw a download link for the kernel
> module sources at http://www.nvidia.com/view.asp?PAGE=linux
>
> was I mistaken or is this something new?
I think the kernel stuff is mostly just hooks into the kernel.
There isn't any significant proprietary information in the
kernel patch, unless I'm much mistaken. The proprietary gold
is hidden in the driver.
Miles
On Wed, 2001-12-05 at 18:22, Erik Elmore wrote:
> Have I lost my mind?
>
> I've always thought that NVIDIA's linux kernel support was incredibly
> closed source, but I swear I just saw a download link for the kernel
> module sources at http://www.nvidia.com/view.asp?PAGE=linux
>
> was I mistaken or is this something new?
The former.
Jeffrey H. Ingber (jhingber _at_ ix.netcom.com)
>
> Erik
>
>
>
> -
> To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
> the body of a message to [email protected]
> More majordomo info at http://vger.kernel.org/majordomo-info.html
> Please read the FAQ at http://www.tux.org/lkml/
On Wed, Dec 05, 2001 at 05:22:24PM -0600, Erik Elmore wrote:
I've always thought that NVIDIA's linux kernel support was incredibly
closed source, but I swear I just saw a download link for the kernel
module sources at http://www.nvidia.com/view.asp?PAGE=linux
It is a hybrid... .a + .c files
--cw
At 05:22 PM 5/12/01 -0600, Erik Elmore wrote:
>Have I lost my mind?
>
>I've always thought that NVIDIA's linux kernel support was incredibly
>closed source, but I swear I just saw a download link for the kernel
>module sources at http://www.nvidia.com/view.asp?PAGE=linux
>
>was I mistaken or is this something new?
The "sources" you see there are the following:
- A Makefile
- The Nvidia binary module
- Some source code that wraps around the binary module and interfaces to
the kernel functions, agp, etc.
This gives some flexibility when it comes to using it with different
kernels, The "kernel layer" compiles against and uses information from the
kernel at /usr/src/linux (or the one you specify through parameters). This
also means they don't have to release a new binary for every kernel release
(can you imagine how horrible that would be?), or track a huge number of
kernel changes.
Most of the guts are still a closed binary module (for x86 only) that
hasn't had any peer review.
Stuart Young - [email protected]
(aka Cefiar) - [email protected]
[All opinions expressed in the above message are my]
[own and not necessarily the views of my employer..]
> > Have I lost my mind?
> >
> > I've always thought that NVIDIA's linux kernel support was incredibly
> > closed source, but I swear I just saw a download link for the kernel
> > module sources at http://www.nvidia.com/view.asp?PAGE=linux
> >
> > was I mistaken or is this something new?
>
> The former.
heh, the part about me bing mistaken or the part about me losing my mind?
Erik
Erik Elmore wrote:
> Have I lost my mind?
>
> I've always thought that NVIDIA's linux kernel support was incredibly
> closed source, but I swear I just saw a download link for the kernel
> module sources at http://www.nvidia.com/view.asp?PAGE=linux
>
> was I mistaken or is this something new?
You're mistaken.
NVidia releases a binary module with open-source glue-code. You can
recompile the glue but not the binary part. Same as always
// Stefan
-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1
any1 got the NVIDIA module working when eg. the nvidia framebuffer code is
enabled?
best regards
Michael Kummer
- ---
Web: http://www.packetst0rm.net / http://www.sprinter-sbg.at
Email: [email protected]
UIN: 96628703
IRC: irc.packetst0rm.net #taskforce
Addr: Lieferinger-Hauptstrasse 47 / A - 5020 Salzburg
Mobil: +43 664 3333995
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://quantumlab.net/pine_privacy_guard/
iD8DBQE8D4x+TOeKpVx8ESMRAmK9AJ9VZ3UaJuleq+qrPle4iml5MUSFzQCgqMWj
RWuraqCT2CpHVnDTIz/Z0zg=
=0my1
-----END PGP SIGNATURE-----
At 04:19 PM 6/12/01 +0100, Michael Kummer wrote:
>any1 got the NVIDIA module working when eg. the nvidia framebuffer code is
>enabled?
The RivaFB code and the Nvidia module stamp all over each other afaik. Have
you tried this with their latest binary module (1.0-2313)? Least that now
plays nicely with devfs. Only real way (afaik) is unload the RivaFB module
when the Nvidia module is loaded, and vice versa.
Of course, this means that while X is running the consoles won't have
FrameBuffer, but hey, them's the trade-offs. You can do this in
modules.conf (modutils stuff) using the pre-install option.
eg:
pre-install NVdriver rmmod rivafb
If the NVdriver module is unloaded when you quit X (depends if it was
dynamically loaded by X, which "should" work if modutils isn't broken [not
usual] and the appropriate stuff is placed there by the Nvidia install),
you could use a modutils post-remove option.
eg:
post-remove NVdriver modprobe rivafb
Of course, there is no guarantee that it will unload immediately (if at
all), so you may lose the FrameBuffer till NVdriver is purged/removed. I
have no idea if this would work, as I don't use the RivaFB module at all.
If I got this wrong, I'm sure Keith Owens will enlighten me (probably with
a large clue stick - since it's in relation to one of his pet hates,
Nvidia). *grin*
Stuart Young - [email protected]
(aka Cefiar) - [email protected]
[All opinions expressed in the above message are my]
[own and not necessarily the views of my employer..]