Brian Paul
89276e2aaf
st: as with swrast, interpret texture border color according to texture format
...
Depending on the hardware driver this might not be needed, but it will
cause no harm.
2009-04-06 14:06:08 -06:00
Brian Paul
a07e68df68
st: rename a helper function
2009-04-06 14:00:03 -06:00
Brian Paul
b48eb05f1f
swrast: fix texture border color interpretation
...
The texture border color must be interpreted according to the texture's
base format. For example, for a GL_ALPHA texture, sampling the border
color should return (0,0,0,borderAlpha). This wasn't an issue here until
I removed the legacy texenv code (we always use the combiner path now).
2009-04-06 13:53:46 -06:00
José Fonseca
8332925c3c
mesa: Fix orientation adjustment for reading stencil pixels.
...
Fixes conform spcorner.c & spclear.c failure.
2009-04-06 20:38:37 +01:00
Eric Anholt
6b187cc8a5
intel: Avoid dri2 GetBuffers round-trips for internal Viewport calls.
...
This gets us the savings for driver-internal viewport calls that
dd1c68f151
was attempting, without relying
on Xlib internals or clients handling X events.
2009-04-06 11:17:25 -07:00
Eric Anholt
e7aef006e5
i965: Use GTT maps when available to upload vertex arrays and system VBOs.
...
This speeds up OA on my GM45 by 21% (more than the original CPU cost of
the upload path). We might still be able to squeeze a few more percent out
by avoiding repeatedly mapping/unmapping buffers as we upload elements into
them.
2009-04-06 10:58:52 -07:00
Eric Anholt
5cca1ceb81
intel: Clean up some a leftover from sedding of bufmgr context->screen move.
2009-04-06 10:31:13 -07:00
José Fonseca
2d56d0839e
r300-gallium: Skeleton for integrating into the python statetracker.
2009-04-06 15:51:27 +01:00
Michel Dänzer
e798bf8053
radeon: Expose a 32 bit RGBA fbconfig even when the screen depth is 16.
...
Otherwise current xserver / libGL no longer expose a 32 bit RGBA GLX visual,
and compiz fails.
Fixes http://bugs.freedesktop.org/show_bug.cgi?id=20479 .
2009-04-06 08:59:57 +02:00
Corbin Simpson
36ae0766b9
r300-gallium: vp: Moar.
2009-04-05 02:05:08 -07:00
Corbin Simpson
84d76607ec
r300-gallium: vs: Use a tab to properly set up OVM.
2009-04-05 01:32:55 -07:00
Corbin Simpson
50ee103cf0
r300-gallium: Re-translate shaders if constants change.
2009-04-05 01:32:28 -07:00
Corbin Simpson
ce7963f338
r300-gallium: Properly interface with Draw for vert shaders.
2009-04-05 01:00:25 -07:00
Corbin Simpson
484795ff14
r300-gallium: Update state handlers/setters for vertex shaders.
2009-04-05 00:15:19 -07:00
Corbin Simpson
316b244ff1
r300-gallium: vs: Moar vert shaders.
2009-04-04 23:44:37 -07:00
Corbin Simpson
7cd535b478
r300-gallium: vs: Expand instruction emission.
2009-04-04 22:57:45 -07:00
Corbin Simpson
a4a853e593
r300-gallium: Update clear() code.
...
We have a huge optimization opportunity, but for now we'll just use the util.
2009-04-04 22:30:14 -07:00
Brian Paul
7a164411ab
cell: update clear() code to catch up to gallium changes
2009-04-04 13:45:54 -06:00
Michal Krol
1bf6af141f
python/regress: Clean up driver clear() interface.
2009-04-04 21:16:37 +02:00
Michel Dänzer
eb168e26aa
gallium: Clean up driver clear() interface.
...
Only allows clearing currently bound buffers, but colour and depth/stencil in
a single call.
2009-04-04 19:01:51 +02:00
Michel Dänzer
ba14b043bc
gallium: Add utility helper for packing combined depth/stencil values.
2009-04-04 16:04:59 +02:00
Michal Krol
b0ce915186
python/regress: Remove trailing whitespace.
2009-04-04 13:42:31 +02:00
Michal Krol
4661817dac
python/regress: Initial fragment shader test suite.
2009-04-04 13:35:53 +02:00
Michal Krol
2d8bf51ffc
python/regress: Use X8R8G8B8 rendertarget in vertex shader test.
2009-04-04 13:35:51 +02:00
Corbin Simpson
024817649f
radeon: A bit of indent and line break cleanup.
2009-04-04 02:43:09 -07:00
Corbin Simpson
48688e5e8f
r300-gallium: Calculate vert shader inputs for HW TCL.
...
This is definitely not perfect.
2009-04-04 02:38:13 -07:00
Corbin Simpson
21f1cdbe07
radeon: If the CS emit fails, dump it to stderr.
2009-04-04 02:20:29 -07:00
Corbin Simpson
23639ddbae
r300-gallium: Fix bad register write.
2009-04-04 02:19:48 -07:00
Corbin Simpson
5c50218d00
r300-gallium: Move swtcl_emit to render to reflect its true purpose.
2009-04-04 02:12:30 -07:00
Corbin Simpson
b7ffe1e876
Add scons build support for radeon/r300.
2009-04-04 00:36:05 -07:00
Corbin Simpson
d6fd672bf9
radeon: Fix compile warnings, compile errors.
2009-04-04 00:36:05 -07:00
Corbin Simpson
be1dbba0a4
r300-gallium: Clean up compile warnings and strict compile errors.
2009-04-04 00:36:04 -07:00
Corbin Simpson
a7dc04fa73
r300-gallium: r500 surface_copy fragment shader.
2009-04-04 00:36:04 -07:00
Brian Paul
4d363cc16d
intel: #include texgetimage.h
2009-04-03 17:43:03 -06:00
Brian Paul
a4bec69e72
mesa: move glGetTexImage(), glGetCompresssedTexImage() code into new file
2009-04-03 17:42:22 -06:00
Brian Paul
c7eb423c49
mesa: remove the noClamp parameter to _mesa_pack_rgba_span_float()
...
It was only set to GL_TRUE in one place where it isn't really needed
(glGetTexImage(sRGB format)).
2009-04-03 17:28:35 -06:00
Brian Paul
35d88e1ac2
mesa: clamp colors to [0,1] for glGetTexImage() when format is GL_LUMINANCE
...
For luminance, we add R+G+B and it seems we should always clamp in case.
2009-04-03 17:10:31 -06:00
Brian Paul
b7cb6650c1
mesa: whitespace and comment clean-up
2009-04-03 16:51:04 -06:00
Brian Paul
a5ce781282
mesa: remove extra semicolons
2009-04-03 16:37:36 -06:00
Brian Paul
5d3b1494b6
softpipe: add additional surface formats in tile cache code
2009-04-03 15:43:13 -06:00
Brian Paul
80197a0c1b
mesa: in mesa_add_named_constant(), avoid adding duplicate constants
2009-04-03 15:42:14 -06:00
Brian Paul
866bdd0509
mesa: fix parameter counting in ARB vertex/fragment program parsing
...
Duplicated unnamed constants were getting counted more than once.
2009-04-03 15:41:10 -06:00
Brian Paul
a4173956eb
mesa: replace >= with > when testing if we've exceeded max local params
...
Now a program that uses 256 locals works as it should.
2009-04-03 13:48:43 -06:00
Brian Paul
ce461ffc5a
mesa: only clear matrix MAT_DIRTY_INVERSE flag when we actually compute the inverse
...
If _math_matrix_analyse() got called before we allocated the inverse
matrix array we could lose the flag indicating that we needed to compute
the inverse. This could happen with certain vertex shader cases.
2009-04-03 12:49:03 -06:00
Brian Paul
b8a200ac9d
mesa: for OPCODE_LIT, use _mesa_pow() instead of exp() and log()
...
Also, s/pow/_mesa_pow/
2009-04-03 11:06:27 -06:00
Michal Krol
a7b6a28b0c
python/regress: Use A8R8G8B8 rendertarget format.
2009-04-03 18:55:28 +02:00
Brian Paul
7391ba1e9d
mesa: rename some gl_light fields to be clearer
...
EyeDirection -> SpotDirection
_NormDirection -> _NormSpotDirection
2009-04-03 10:09:28 -06:00
Brian Paul
650d147289
mesa: don't normalize spot light direction until validation time
...
In glLight() we're only supposed to transform the direction by the modelview
matrix, not normalized it too.
2009-04-03 10:09:27 -06:00
Michal Krol
be4c2d9a33
python/regress: vertex shader srcmod swz test does not use TEMP[0].
2009-04-03 17:55:21 +02:00
Michal Krol
e3d5e0aead
tgsi/exec: Actually enable switch-case for FLR.
2009-04-03 17:26:01 +02:00