summaryrefslogtreecommitdiffstats
path: root/src/gallium/drivers/r300
Commit message (Collapse)AuthorAgeFilesLines
* r300g: Correct scissor setting, subtract 1 from window's width and heightCooper Yuan2009-08-272-7/+7
|
* r300g: Set the vector address in the input memory for ↵Cooper Yuan2009-08-241-16/+20
| | | | bypass_vs_clip_and_viewport case
* r300g: Force off ZTOP optimizations for now.Corbin Simpson2009-08-181-1/+2
|
* r300g: Utilize DONTBLOCK.Corbin Simpson2009-08-181-8/+19
| | | | Also ALGYRHYTHMS.
* r300g: Massively cleanup OQ.Corbin Simpson2009-08-186-29/+174
| | | | Still broken, but compiles cleaner, behaves better, etc.
* r300g: Add high_second_pipe cap for R3xx chipsets.Corbin Simpson2009-08-182-1/+11
| | | | | This name is totally subject to change if ever I need to separate R3xx for some other reason.
* Revert "r300-gallium, radeon-gallium: Nuke gb_pipes from orbit."Corbin Simpson2009-08-185-0/+26
| | | | | | | | | | This reverts commit 6a40d1e9d96f8e8c57bc3bbd6f567cacd4471f59. Turns out that we *do* need these for OQ after all. Go figure. Conflicts: src/gallium/winsys/drm/radeon/core/radeon_r300.h
* r300g: Fix up remaining VAP_CNTL_STATUS writes for big endian.Michel Dänzer2009-08-111-2/+10
|
* r300g: Emit relocations for pitch registers.Michel Dänzer2009-08-112-7/+12
| | | | Fixes CS failures with tiling enabled kernels.
* r300g: a typo of debug messageCooper Yuan2009-08-111-1/+1
|
* r300g: Knock out another fragment of invariant state.Corbin Simpson2009-08-071-8/+3
| | | | Colorbuffer setup will always happen.
* gallium: Move minify() to u_math.Corbin Simpson2009-08-071-7/+0
| | | | | minify() is usually used in mipmap size calculation. Strangely enough, we all defined it as MAX2(1, d >> 1); imagine that. :3
* r300g: Remove r300_constant_buffer::user_count.Corbin Simpson2009-08-073-6/+3
| | | | Not needed with new compiler.
* r300g: Slightly saner initialization of some texture / transfer fields.Michel Dänzer2009-08-042-6/+5
|
* r300g: Use radeon compiler for fragment programsNicolai Hähnle2009-07-3017-1312/+450
| | | | This is entirely untested on R500, and needs more testing on R300.
* r300g: Remove extraneous printfNicolai Hähnle2009-07-301-2/+0
| | | | Signed-off-by: Nicolai Hähnle <[email protected]>
* r300g: Use r300compiler for vertex shadersNicolai Hähnle2009-07-3015-748/+690
|
* r300g: Fix two trivial texture size issues.Corbin Simpson2009-07-261-2/+2
| | | | Next thing to fix: progs/tests/mipgen.
* r300g: Add some debugging, correct little bits of math in texture setup.Corbin Simpson2009-07-261-3/+12
| | | | Simple stuff still works, but not sure about some of the more complex things.
* r300g, radeon-gallium: Fix API, cleanup.Corbin Simpson2009-07-232-3/+6
| | | | Something called "validate" should return FALSE on failure, not TRUE.
* r300g: PIPE_CAP_BLEND_EQUATION_SEPARATE.Corbin Simpson2009-07-231-0/+2
|
* r300g: Actually mark shaders as translated/untranslated.Corbin Simpson2009-07-233-2/+8
| | | | | | Also trust that Gallium will not give us TGSI that miscounts shader consts. This creates a 20x speedup on glxgears, from 8 FPS to 160 FPS.
* gallium: simplify tgsi_full_immediate structKeith Whitwell2009-07-222-4/+2
| | | | | | | | | | | | | Remove the need to have a pointer in this struct by just including the immediate data inline. Having a pointer in the struct introduces complications like needing to alloc/free the data pointed to, uncertainty about who owns the data, etc. There doesn't seem to be a need for it, and it is unlikely to make much difference plus or minus to performance. Added some asserts as we now will trip up on immediates with more than four elements. There were actually already quite a few such asserts, but the >4 case could be used in the future to specify indexable immediate ranges, such as lookup tables.
* r300g: Guard R500 register writes by is_r500 check.Michel Dänzer2009-07-201-3/+5
| | | | | Flagged by the DRM command stream checker. This allows the driver to work on non-R500 cards.
* r300g: Small compile warning fixes.Nicolai Hähnle2009-07-133-2/+5
| | | | Signed-off-by: Corbin Simpson <[email protected]>
* r300g, radeon: Whitespace fixes.Nicolai Hähnle2009-07-131-1/+1
| | | | Signed-off-by: Corbin Simpson <[email protected]>
* r300g: Use align() instead of inline maths.Corbin Simpson2009-07-131-4/+7
|
* r300g: Remove VAP_CNTL_STATUS from invariant state.Corbin Simpson2009-07-081-6/+1
| | | | Seriously.
* r300g: Disable MSPOS registers for glisse's CS security checker.Corbin Simpson2009-07-081-4/+7
| | | | These will come back in someday, when we can properly use them.
* r300-gallium: Unify sampler and texture emit.Corbin Simpson2009-07-083-39/+26
| | | | They have to cross into each other's registers.
* r300-gallium: Mipmap setup.Corbin Simpson2009-07-082-10/+19
| | | | (cherry picked from commit 88c01a15da5639dd68a6a0133724994cb66f1316)
* r300g: Add endian fix to vertex fetcher setup.Corbin Simpson2009-07-081-2/+9
| | | | As reported and initially tested by MrCooper.
* r300g: Fix recursive Draw flush.Corbin Simpson2009-07-012-1/+7
| | | | | | Also just noticed that demos/copypix walks around the overlapping blit rules. Bad, bad Mesa. :3
* r300g: Use floats for surface_copy texcoords, use correct src and dest.Corbin Simpson2009-07-011-16/+17
| | | | This makes demos/copypix better-looking. Horizontal dimensions are right now.
* r300g: YCbCr and sRGB textures.Corbin Simpson2009-06-273-1/+15
|
* r300g: Fix EXT_packed_depth_stencil functionality.Corbin Simpson2009-06-272-2/+10
| | | | Allow Z24S8 to be a true texture.
* r300g: EXT_provoking_vertex.Corbin Simpson2009-06-262-4/+6
|
* r300g: S3TC.Corbin Simpson2009-06-262-6/+36
| | | | tests/texcompress2 doesn't work, but tests/texcmp does (more or less.)
* r300g: Use real texture formats.Corbin Simpson2009-06-263-25/+42
| | | | What bugs me is that the YUV444 format somehow worked properly. :3
* r300g: Definitively forbid unusable Z buffer/stencil formats.Corbin Simpson2009-06-261-1/+12
|
* r300g: Comment out assert for now.Corbin Simpson2009-06-261-0/+2
| | | | Will fix with better constant refactoring later.
* r300g: Be more specific on surface_copy fallbacks.Corbin Simpson2009-06-261-1/+1
|
* r300g: PIPE_CAP_TGSI_CONT_SUPPORTED.Corbin Simpson2009-06-261-0/+3
|
* r300-gallium: Ensure that no dirty state goes unemitted.Corbin Simpson2009-06-261-2/+4
|
* r300-gallium: organize fragment/vertex shadersJoakim Sindholt2009-06-2620-495/+674
| | | | | | | | | | Appart from separating r3xx/r5xx fragment shaders, a more consistent naming scheme has been applied. From now on: r300 = all chips r3xx = R300/R400 only r5xx = R500 only This way r300_fragment_shader is the master struct, and the structs r3xx_fragment_shader and r5xx_fragment_shader inherits it.
* r300-gallium: HW TCL glxgears. (Read the rest of the log.)Corbin Simpson2009-06-081-0/+5
| | | | | | | | | | | Um. So, yeah. Two massive WTF moments here. The first one is that, somehow, I never actually hooked up vertex shader emission, so the only time that the VAP gets set up is during surface_copy/surface_fill. That's why acidgears was happening. The second one is that, somehow, once I actually hooked it up, glxgears just magically worked. Without any actual, real testing, I somehow accidentally made the shader compiler work. Go figure.
* r300-gallium: Make UCP and clip work again for SW TCL.Corbin Simpson2009-06-082-19/+21
| | | | | | | SW TCL: tri-clip works, tri-userclip works HW TCL: tri-clip fails, tri-userclip works That is a 200% improvement over the previous situation. Woot.
* r300-gallium: Don't emit UCP planes for SW TCL.Corbin Simpson2009-06-083-4/+12
|
* r300-gallium: Use VAP_CLIP_CNTL.Corbin Simpson2009-06-052-4/+5
| | | | Makes tri-userclip work with HW TCL.
* r300-gallium: Emit UCP.Corbin Simpson2009-06-054-20/+53
|
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-20 16:31:53 +0000