Intel Sandy/Ivy Bridge Gallium3D Driver Merged

[Pontostroy]

Now that Mesa has merged the Gallium driver for Intel hardware into mainline, what does this mean for distributions?

Will users of Intel graphics hardware now use the Gallium3D drivers by default, or will the classic DRI driver still serve as the defaultdriver? Also, if both the Gallium3D and classic DRI drivers are now in mainline, will there be an easy option for users to switch 'on the fly' between Gallium and DRI, either via a text command (like, modprobe -r <intel DRI driver> and modprobe <intel Gallium3D driver>) or a graphical interface?

ln -s /usr/lib/dri/ilo_dri.so /usr/lib/dri/i965_dri.so
glxinfo |grep OpenGL
ATTENTION: default value of option vblank_mode overridden by environment.
ATTENTION: default value of option vblank_mode overridden by environment.
OpenGL vendor string: LunarG, Inc.
OpenGL renderer string: Gallium 0.4 on Intel(R) Sandybridge Desktop
OpenGL version string: 2.1 Mesa 9.2.0
OpenGL shading language version string: 1.30

ln -sf /usr/lib/dri/i965c_dri.so /usr/lib/dri/i965_dri.so
glxinfo |grep OpenGL
ATTENTION: default value of option vblank_mode overridden by environment.
ATTENTION: default value of option vblank_mode overridden by environment.
OpenGL vendor string: Intel Open Source Technology Center
OpenGL renderer string: Mesa DRI Intel(R) Sandybridge Desktop x86/MMX/SSE2
OpenGL core profile version string: 3.1 (Core Profile) Mesa 9.2.0
OpenGL core profile shading language version string: 1.40
OpenGL core profile context flags: (none)
OpenGL core profile extensions:
OpenGL version string: 3.0 Mesa 9.2.0
OpenGL shading language version string: 1.30
OpenGL context flags: (none)
OpenGL extensions:

[c117152]

Intel should use Gallium, there is simply no excuse not using it, except NIH. Even rewriting it from scratch is time worthy! It would steamline the Kernel and Xorg.

Gallium is MIT while the earlier drivers were BSD\GPL mix. It's possible there are GPLed contributions in the older drivers they can't port under the new license.

Unlikely, but possible...

Besides, what if they want to FOSS a given solution but not to allow other companies to improve upon it without sharing ? The GPL forces the players to play nice. And while Intel doesn't want to play nice all the time, they sure want the other companies to play by the rules

[Veerappan]

ln -s /usr/lib/dri/ilo_dri.so /usr/lib/dri/i965_dri.so
glxinfo |grep OpenGL
ATTENTION: default value of option vblank_mode overridden by environment.
ATTENTION: default value of option vblank_mode overridden by environment.
OpenGL vendor string: LunarG, Inc.
OpenGL renderer string: Gallium 0.4 on Intel(R) Sandybridge Desktop
OpenGL version string: 2.1 Mesa 9.2.0
OpenGL shading language version string: 1.30

ln -sf /usr/lib/dri/i965c_dri.so /usr/lib/dri/i965_dri.so
glxinfo |grep OpenGL
ATTENTION: default value of option vblank_mode overridden by environment.
ATTENTION: default value of option vblank_mode overridden by environment.
OpenGL vendor string: Intel Open Source Technology Center
OpenGL renderer string: Mesa DRI Intel(R) Sandybridge Desktop x86/MMX/SSE2
OpenGL core profile version string: 3.1 (Core Profile) Mesa 9.2.0
OpenGL core profile shading language version string: 1.40
OpenGL core profile context flags: (none)
OpenGL core profile extensions:
OpenGL version string: 3.0 Mesa 9.2.0
OpenGL shading language version string: 1.30
OpenGL context flags: (none)
OpenGL extensions:

Instead of re-linking the files, you can also do the following in your xorg.conf (from Chris Wilson on the mesa-dev list):
Section "Device"
Identifier "Device0"
Driver "Intel"
Option "DRI" "ilo"
EndSection

[Kayden]

Gallium is MIT while the earlier drivers were BSD\GPL mix. It's possible there are GPLed contributions in the older drivers they can't port under the new license.

I don't know where you heard that, but the classic i965 Mesa driver is entirely under the MIT license---the exact same license used by the rest of the project. Our driver contains no GPL code whatsoever.

[Kivada]

..and it would help their competitors if Intel used its massive resources on optimizing Gallium3D infrastructure....

Intel's GPUs don't compete anyways, the silicon just doesn't have the performance. The idea of not doing it is moronic and comes from management that doesn't understand the basis of technology, that Intel is a HARDWARE company, not making sure every aspect of your hardware does exactly what it is capable of doing due to your bullshitting on drivers is only detrimental to your sales.

[Kivada]

Looking at their reactions, they're really annoyed by this "toy" driver, and did all they could to shoot it down. We'll see where it goes from there, but certainly Intel looks more interested in their competitive advantage than benefiting the entire Linux ecosystem.

Hence why you shouldn't buy Intel hardware.

[Kivada]

Gallium driver for Sandy/Ivy do not bring anything more than classic driver. Nor it provide performance gains.

Distros will stick to classic Intel driver.


GPU driver devs/enthusiasts will be able to play with ilo more easily. That is just it.

It removes the copout that it would take too long to be worth it to switch to Gallium3D.

[archibald]

I don't know where you heard that, but the classic i965 Mesa driver is entirely under the MIT license---the exact same license used by the rest of the project. Our driver contains no GPL code whatsoever.

Matthieu Herrb's "Future of X on non-Linux systems" talk says that there is a file "intel_pm.c" that's under a GPL license. I'm guessing that this is for power management and is sufficiently tied to the Linux kernel that it becomes bound by the GPL (as a derivative work)? If not, what is it/why is it under the GPL?

[ickle]

Matthieu Herrb's "Future of X on non-Linux systems" talk says that there is a file "intel_pm.c" that's under a GPL license. I'm guessing that this is for power management and is sufficiently tied to the Linux kernel that it becomes bound by the GPL (as a derivative work)? If not, what is it/why is it under the GPL?

That is part of the kernel and not part of the mesa driver, so is out of context. The respective licence for intel_pm.c is:

/*
 * Copyright ? 2012 Intel Corporation
 *
 * Permission is hereby granted, free of charge, to any person obtaining a
 * copy of this software and associated documentation files (the "Software"),
 * to deal in the Software without restriction, including without limitation
 * the rights to use, copy, modify, merge, publish, distribute, sublicense,
 * and/or sell copies of the Software, and to permit persons to whom the
 * Software is furnished to do so, subject to the following conditions:
 *
 * The above copyright notice and this permission notice (including the next
 * paragraph) shall be included in all copies or substantial portions of the
 * Software.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
 * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
 * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
 * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
 * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
 * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
 * IN THE SOFTWARE.
 */

So it is dual-licensed - GPL for inclusion into the kernel and MIT/X otherwise. If you do come across a mistake like that, please assume it was an oversight rather than malice and bring it to our attention.

[archibald]

That is part of the kernel and not part of the mesa driver, so is out of context. The respective licence for intel_pm.c is:

<snip>

So it is dual-licensed - GPL for inclusion into the kernel and MIT/X otherwise. If you do come across a mistake like that, please assume it was an oversight rather than malice and bring it to our attention.

Thank you for the clarification. I didn't assume that there was any malice involved; sorry if my message came across as differently.