Intel Drops Mode-Setting Rework Patch Bomb

[blinxwang]

Oh my god, why can't you guys make discrete graphics cards?
I'm sick and tired of the half-assed efforts of AMD and the arrogance and complete refusal of Nvidia to support free drivers.

[timofonic]

We're working on it We all desperately want to get there too...there's just an awful lot of work required to pull it off. We should have 3.1 fairly soon...then the next big feature is Geometry Shaders. Once those are done, we'll basically hit 3.3. Then, onwards and upwards...

If people are interested in helping out, one thing that would be extremely useful is adding test cases for Geometry Shaders (3.2) or even Hull and Domain Shaders (4.0) to Piglit, the open source OpenGL validation suite. Tests can be developed using the binary blob drivers, without waiting for us to implement the functionality. Writing tests requires some GL programming knowledge, and reading of the specs, but doesn't require a deep knowledge of Mesa internals, compilers, or the workings of GPU hardware. It's not particularly glamorous, but is extremely useful - writing tests is really the only way to know that our implementation actually works like it's supposed to. Right now, Piglit doesn't have any GL 3.2+ tests, so for every new feature, we're having to develop the driver -and- the test suite from scratch, which can be rather time consuming. Implementing these tests would also be a considerable help to the Radeon and Nouveau developers.

If anyone's interested, drop a note to the Piglit mailing list.

Intel Open Source Team isn?t so OPEN...

You are all putting the work on your codebase so others can't benefit of it.

The real deal is Gallium3D, so all drivers could use the same OpenGL implementation.

If not, you are just boycotting MESA to have a technical advantage of Intel over the rest. Not an ethical way to support Open Source at all.

Intel is a little better than Nvidia, but not much more

[entropy]

Intel Open Source Team isn?t so OPEN...

You are all putting the work on your codebase so others can't benefit of it.

The real deal is Gallium3D, so all drivers could use the same OpenGL implementation.

If not, you are just boycotting MESA to have a technical advantage of Intel over the rest. Not an ethical way to support Open Source at all.

Intel is a little better than Nvidia, but not much more

What do you mean with _same_ OpenGL implementation?
AFAIK just using Gallium does not keep you from writing mostly hardware-specific code.
So how would AMD, for instance, significantly benefit from intel using Gallium?

[timofonic]

What do you mean with _same_ OpenGL implementation?
AFAIK just using Gallium does not keep you from writing mostly hardware-specific code.
So how would AMD, for instance, significantly benefit from intel using Gallium?

So yes...

FUCK YOU, INTEL!

Gallium3D must be the future, for the benefit of the majority and not just corp(s) able to get a legion of high skilled coders for their own drivers with their own OpenGL implementation.

[entropy]

Well, while it's possible you have a point here - I don't get it.
What should that schematic explain?

Does it reveal the relative work required between the different blocks?
Obviously not.

Correct me if I'm wrong. It's not that "easy" to have just a - let's say -
abstracted OpenGL 4.x state tracker and immediately every GPU hardware
with an existing Gallium driver can profit.

[timofonic]

Well, while it's possible you have a point here - I don't get it.
What should that schematic explain?

Does it reveal the relative work required between the different blocks?
Obviously not.

Correct me if I'm wrong. It's not that "easy" to have just a - let's say -
abstracted OpenGL 4.x state tracker and immediately every GPU hardware
with an existing Gallium driver can profit.

That's one of the theorical benefits, able to share state trackers between different hardware drivers. Or at least it's what has been explained about Gallium3D since it's inception.

Anyway, it would be nice to know this from some involved developer(s).

Obviously, Intel developers will put any kind of excuses like the immaturity of Gallium3D (improve it, then) or unable to "satisfy their needs" (so why AMD and Nouveau are using it?).

[89c51]

Wasn't Michael going to publish some Intel G3D vs Intel Classic benchmarks??

[AlbertP]

Intel's OpenGL 3.0 work is really not limited to classic Mesa drivers. Radeon and Nouveau, both Gallium, were also able to take advantage of it. Both drivers quickly had OpenGL 3.0 after Intel's work.
Intel made the Mesa core support OpenGL 3. Having support in Gallium also was just a small step from there as it was even working in time for the Mesa 8.0 release.

[89c51]

Intel's OpenGL 3.0 work is really not limited to classic Mesa drivers. Radeon and Nouveau, both Gallium, were also able to take advantage of it. Both drivers quickly had OpenGL 3.0 after Intel's work.
Intel made the Mesa core support OpenGL 3. Having support in Gallium also was just a small step from there as it was even working in time for the Mesa 8.0 release.

Even that small step requires work that could be allocated somewhere else.

[Gusar]

Even that small step requires work that could be allocated somewhere else.

So the small work the radeon and nouveau teams need to do after the Intel guys have added functionality to Mesa core is too much for you, but the huge amount of work Intel would require to switch their driver to gallium is fine?

I can turn your post directly against you and say: Intel switching requires work that is much better spent bringing new stuff to Mesa.