Brian Paul [Mon, 15 Feb 2010 04:54:17 +0000 (21:54 -0700)]
st/mesa: minor assorted clean-ups
Brian Paul [Mon, 15 Feb 2010 04:46:51 +0000 (21:46 -0700)]
st/mesa: change some function return types
Brian Paul [Mon, 15 Feb 2010 03:47:34 +0000 (20:47 -0700)]
st/mesa: move per-fragment shader fields to local vars
Brian Paul [Mon, 15 Feb 2010 01:50:49 +0000 (18:50 -0700)]
st/mesa: use _mesa_clone_fragment_program()
Brian Paul [Mon, 15 Feb 2010 01:46:56 +0000 (18:46 -0700)]
r600: use new program cloning functions
Brian Paul [Mon, 15 Feb 2010 01:46:51 +0000 (18:46 -0700)]
r300: use new program cloning functions
Brian Paul [Mon, 15 Feb 2010 01:46:37 +0000 (18:46 -0700)]
glsl: use new program cloning functions
Brian Paul [Mon, 15 Feb 2010 02:09:04 +0000 (19:09 -0700)]
mesa: added _mesa_clone_vertex/fragment_program()
To reduce casting elsewhere...
Vinson Lee [Mon, 15 Feb 2010 04:49:09 +0000 (20:49 -0800)]
r600: Silence uninitialized variable warnings.
Marek Olšák [Sun, 14 Feb 2010 23:04:32 +0000 (00:04 +0100)]
r300g: fix the size of constant buffers
4 more piglit tests pass, sweet.
José Fonseca [Mon, 15 Feb 2010 00:25:43 +0000 (00:25 +0000)]
python/retrace: Reconstruct blend state from binary blobs.
José Fonseca [Mon, 15 Feb 2010 00:22:59 +0000 (00:22 +0000)]
python: Reconstruct blend state objects from binary blobs.
José Fonseca [Mon, 15 Feb 2010 00:22:29 +0000 (00:22 +0000)]
trace: Dump blend state as a binary blob.
Gallium interfaces are too fluid -- XML or not is just too much work
trying to accommodate for differences in traces.
SWIG generated bindings are also full of quirks, making unmarshalling
the XML from too hard.
Finally, this is the first step for faster tracing.
José Fonseca [Mon, 15 Feb 2010 00:19:22 +0000 (00:19 +0000)]
trace: Fix constant buffer unwrapping.
Pointer unwrapping should happen before dump the pointer, so that
the real pipe driver pointers are shown, instead of the trace driver's.
José Fonseca [Sun, 14 Feb 2010 23:54:22 +0000 (23:54 +0000)]
util: Avoid dumping masked state. Other minor improvements.
José Fonseca [Sun, 14 Feb 2010 23:53:42 +0000 (23:53 +0000)]
os: A stream for debug logging.
Just a wrapper around os_log_message. Although it would probably make
more sense to be the other way around.
Also some comment fixes.
José Fonseca [Sun, 14 Feb 2010 23:25:01 +0000 (23:25 +0000)]
trace: No need to dump reference counts.
José Fonseca [Sun, 14 Feb 2010 23:24:08 +0000 (23:24 +0000)]
python: Pretty-printing of blend state.
José Fonseca [Sun, 14 Feb 2010 23:20:42 +0000 (23:20 +0000)]
util: Helper functions to dump all state objects.
José Fonseca [Sun, 14 Feb 2010 23:07:21 +0000 (23:07 +0000)]
os: Add a growable string stream.
José Fonseca [Sun, 14 Feb 2010 23:04:04 +0000 (23:04 +0000)]
os: Complement/improve stream inline helpers.
José Fonseca [Sun, 14 Feb 2010 23:03:10 +0000 (23:03 +0000)]
os: Fix bad calloc.
José Fonseca [Sun, 14 Feb 2010 16:55:38 +0000 (16:55 +0000)]
trace: Update for os_stream changes.
José Fonseca [Sun, 14 Feb 2010 16:55:25 +0000 (16:55 +0000)]
util: Update for os_stream changes.
José Fonseca [Sun, 14 Feb 2010 16:55:04 +0000 (16:55 +0000)]
os: Make streams abstract.
Also replace windows kernel stream with null implementation. It was
severely limited and no easy means to test it now.
Vinson Lee [Sun, 14 Feb 2010 21:47:58 +0000 (13:47 -0800)]
glut: Silence uninitialized variable warning.
Vinson Lee [Sun, 14 Feb 2010 21:23:21 +0000 (13:23 -0800)]
llvmpipe: Silence unused value warning.
Vinson Lee [Sun, 14 Feb 2010 21:10:08 +0000 (13:10 -0800)]
svga: Silence uninitialized variable warning.
Marek Olšák [Sun, 14 Feb 2010 20:58:00 +0000 (21:58 +0100)]
r300: add half_float_vertex support
Pauli Nieminen [Sun, 14 Feb 2010 12:16:20 +0000 (14:16 +0200)]
mesa: Don't pass paramter to __builtin_clz which would hve undefined result.
__builtin_clz with parameter 0 has undefined value. When
using -O3 optimizing this would result to too large next
power of two value.
Fix is to check if passed value is 1 and modify formula
for that case.
Marek Olšák [Fri, 12 Feb 2010 01:48:03 +0000 (02:48 +0100)]
r300g: set the correct tiling flags for renderbuffers
MACRO_SWITCH is applied to samplers but not renderbuffers. This commit
fixes incorrect rendering to large and small mipmaps where the large ones
are macrotiled and the small ones are not and both are emitted
in the same CS.
Note that this is still disabled by default (rework of texture transfers
is next).
Marek Olšák [Mon, 8 Feb 2010 02:17:24 +0000 (03:17 +0100)]
r300g: add macrotiling support to texture setup
Marek Olšák [Thu, 11 Feb 2010 23:48:44 +0000 (00:48 +0100)]
radeong: flush when changing tiling flags of BO referenced by CS
Marek Olšák [Sun, 14 Feb 2010 03:57:03 +0000 (04:57 +0100)]
r300g: accelerate blitting for all formats by faking the texture format
Brian Paul [Sun, 14 Feb 2010 17:02:42 +0000 (10:02 -0700)]
src/glu/mesa: fix mem leak (bug 26559)
Brian Paul [Sun, 14 Feb 2010 17:02:32 +0000 (10:02 -0700)]
src/glu/mini: fix mem leak (bug 26559)
Brian Paul [Sun, 14 Feb 2010 17:02:01 +0000 (10:02 -0700)]
progs/objviewer: close file (bug 26559)
Brian Paul [Sun, 14 Feb 2010 16:51:49 +0000 (09:51 -0700)]
progs/fp: close file (bug 26559)
Marek Olšák [Sun, 7 Feb 2010 10:36:00 +0000 (11:36 +0100)]
r300g: fix texture formats: A4R4G4B4_UNORM, A1R5G5B5_UNORM, R5G6B5_UNORM
José Fonseca [Sun, 14 Feb 2010 15:34:39 +0000 (15:34 +0000)]
scons: Control caching via new SCONS_CACHE_DIR environment variable.
This serves several purposes:
- disable caching in situations were is it useless or undesired
- share caches among all trees
- simplify purging the cache (when it's a single location)
- move the cache out of the tree, since that slows downs IDEs considerably
To retain previous behavior just define do
export SCONS_CACHE_DIR=$PWD/build/cache
before invoking scons.
José Fonseca [Sun, 14 Feb 2010 15:21:06 +0000 (15:21 +0000)]
llvmpipe: s/debug_dump_/util_dump_/
José Fonseca [Sun, 14 Feb 2010 15:20:57 +0000 (15:20 +0000)]
gallivm: s/debug_dump_/util_dump_/
José Fonseca [Sun, 14 Feb 2010 15:20:34 +0000 (15:20 +0000)]
util: Rename u_debug_dump.[ch] -> u_dump.[ch].
I have more plans for this than mere debugging -- it will be an helper
to provide human readible representations of all gallium state for
the python state tracker.
José Fonseca [Sun, 14 Feb 2010 13:28:29 +0000 (13:28 +0000)]
mesa: Add missing include.
José Fonseca [Sun, 14 Feb 2010 13:28:17 +0000 (13:28 +0000)]
gallium: Remove previously forgotten pf_name() prototype.
José Fonseca [Sun, 14 Feb 2010 13:27:00 +0000 (13:27 +0000)]
draw: Use util_format_name().
José Fonseca [Sun, 14 Feb 2010 13:26:51 +0000 (13:26 +0000)]
util: Use util_format_name().
José Fonseca [Sun, 14 Feb 2010 13:26:46 +0000 (13:26 +0000)]
cell: Use util_format_name().
José Fonseca [Sun, 14 Feb 2010 13:26:37 +0000 (13:26 +0000)]
llvmpipe: Use util_format_name().
José Fonseca [Sun, 14 Feb 2010 13:26:30 +0000 (13:26 +0000)]
nv: Use util_format_name().
José Fonseca [Sun, 14 Feb 2010 13:25:22 +0000 (13:25 +0000)]
r300g: Use util_format_name().
José Fonseca [Sun, 14 Feb 2010 13:25:06 +0000 (13:25 +0000)]
trace: Use util_format_name().
José Fonseca [Sun, 14 Feb 2010 13:24:58 +0000 (13:24 +0000)]
xorg: Use util_format_name().
José Fonseca [Sun, 14 Feb 2010 13:24:43 +0000 (13:24 +0000)]
mesa: Use util_format_name().
José Fonseca [Sun, 14 Feb 2010 13:23:32 +0000 (13:23 +0000)]
util: Add pf_name() replacement: util_format_name().
José Fonseca [Sun, 14 Feb 2010 13:15:49 +0000 (13:15 +0000)]
makefiles: Don't complaint about changing configs if the same one is reselected.
Otherwise it creates unnecessary pain when automating builds in a build
server/farm.
José Fonseca [Sun, 14 Feb 2010 12:48:11 +0000 (12:48 +0000)]
python/retrace: Interpret pipe_screen::context_create().
José Fonseca [Sun, 14 Feb 2010 12:47:40 +0000 (12:47 +0000)]
trace: Fix generated XML.
José Fonseca [Sun, 14 Feb 2010 12:06:37 +0000 (12:06 +0000)]
python: Tell swig how to index per-rendertarget blend state.
José Fonseca [Sun, 14 Feb 2010 11:26:58 +0000 (11:26 +0000)]
python: Pass p_format.h as-is to swig now that it has no inlines.
José Fonseca [Sun, 14 Feb 2010 11:15:19 +0000 (11:15 +0000)]
progs/gallium/python: New home for python statetracker scripts.
Vinson Lee [Sun, 14 Feb 2010 10:04:17 +0000 (02:04 -0800)]
glut: Silence uninitialized variable warning.
Vinson Lee [Sun, 14 Feb 2010 08:56:57 +0000 (00:56 -0800)]
r300: Silence uninitialized variable warning.
Vinson Lee [Sun, 14 Feb 2010 08:45:23 +0000 (00:45 -0800)]
r600: Silence uninitialized variable warning.
Vinson Lee [Sun, 14 Feb 2010 07:56:48 +0000 (23:56 -0800)]
glu/sgi: Initialize members of class Curvelist.
Vinson Lee [Sun, 14 Feb 2010 07:48:40 +0000 (23:48 -0800)]
llvmpipe: Silence uninitialized variable warning.
Vinson Lee [Sun, 14 Feb 2010 06:51:34 +0000 (22:51 -0800)]
softpipe: Silence uninitialized variable warnings.
Marek Olšák [Sun, 14 Feb 2010 02:16:00 +0000 (03:16 +0100)]
st/python: fix build error after gallium-screen-context merge
Marek Olšák [Sun, 14 Feb 2010 02:14:32 +0000 (03:14 +0100)]
r300g: fix scons build
Brian Paul [Sun, 14 Feb 2010 00:47:23 +0000 (17:47 -0700)]
softpipe: fix broken cubemap / mipmap selection code
This is a quick-fix for the time being...
The per-face mipmap LOD computation was invalid at cube edges. In
mip_filter_nearest/linear() we were trying to compute LOD using
texcoords that were sometimes indexes into different cube faces.
The subtraction used to compute the partial derivatives basically
gave random values, so the LOD was unpredictable. This fix simply
uses the same cube face for all four pixels in the quad. The per-
face texcoords all reference the same cube face so the partial
deriviates are computed properly.
A more elaborate fix would involve computing the LOD at the same
time as we choose the cube faces. But for now, this solution works
well and allows the piglit/cubemap test to pass.
(cherry picked from commit
1ff9cd5079b095d7050edb8dc6a7e5b8cad36e1e)
Karl Schultz [Sun, 14 Feb 2010 00:31:58 +0000 (17:31 -0700)]
mesa: Fix compiler warnings
Add explicit casts, fix constant types, fix variable types.
Fixes about 340 warnings in MSFT Visual Studio.
Brian Paul [Sat, 13 Feb 2010 21:06:25 +0000 (14:06 -0700)]
docs: document GL_ARB_fragment_coord_conventions
Brian Paul [Sat, 13 Feb 2010 21:02:54 +0000 (14:02 -0700)]
mesa: enable GL_ARB_fragment_coord_conventions for sw drivers
This extension is now implemented in the GLSL compiler so it can be
enabled by other drivers where applicable.
Brian Paul [Sat, 13 Feb 2010 20:57:31 +0000 (13:57 -0700)]
glsl: implement layout qualifiers
For GL_ARB_fragment_coord_conventions.
This only applies to gl_FragCoord and controls pixel center origin and
pixel center integer. For example:
layout (origin_upper_left, pixel_center_integer) varying vec4 gl_FragCoord;
This features introduces the idea of re-declaring variables with a changed
type. This may also apply to arrays in some cases but that's not
implemented at this time.
Brian Paul [Sat, 13 Feb 2010 20:52:37 +0000 (13:52 -0700)]
glsl: added type layout field and new type compare func
Note: because of a weird dependency checking bug, a 'make clean' may be
needed before recompiling.
Brian Paul [Sat, 13 Feb 2010 20:51:38 +0000 (13:51 -0700)]
mesa: copy frag coord layout fields too
Brian Paul [Sat, 13 Feb 2010 20:50:26 +0000 (13:50 -0700)]
glsl/cl: implement parsing for layout qualifiers
For GL_ARB_fragment_coord_conventions support.
Brian Paul [Sat, 13 Feb 2010 20:48:32 +0000 (13:48 -0700)]
glsl/pp: fix incorrect extension enable flag for warn mode
José Fonseca [Sat, 13 Feb 2010 15:10:24 +0000 (15:10 +0000)]
mesa: Export GL_EXT_texture_cube_map.
Still used by some applications.
José Fonseca [Sat, 13 Feb 2010 15:09:46 +0000 (15:09 +0000)]
wgl: Implement wglSwapMultipleBuffers.
José Fonseca [Sat, 13 Feb 2010 09:16:57 +0000 (09:16 +0000)]
svga: Fix texture border color.
Also opposite ordering.
Marek Olšák [Fri, 12 Feb 2010 04:03:27 +0000 (05:03 +0100)]
r300g: do not invalidate texture caches when only sampler state is changed
Marek Olšák [Fri, 12 Feb 2010 03:49:30 +0000 (04:49 +0100)]
r300g: cleanup the emission of RS block state
Emit as few regs as possible.
Marek Olšák [Fri, 12 Feb 2010 03:19:55 +0000 (04:19 +0100)]
r300g: cleanup the emission of framebuffer state
Some emits are not needed anymore.
Marek Olšák [Fri, 12 Feb 2010 03:01:15 +0000 (04:01 +0100)]
r300g: disable independent blend enables
Not supported by r300.
Marek Olšák [Fri, 12 Feb 2010 02:51:41 +0000 (03:51 +0100)]
r300g: emit INDEPENDENT_COLORFORMAT_ENABLE only on r5xx
Vinson Lee [Sat, 13 Feb 2010 08:05:29 +0000 (00:05 -0800)]
mesa: s/abort/sys.exit/
Python does not have the function abort.
Vinson Lee [Sat, 13 Feb 2010 07:47:45 +0000 (23:47 -0800)]
mesa: Remove unnecessary header.
Vinson Lee [Sat, 13 Feb 2010 07:37:36 +0000 (23:37 -0800)]
i965: Remove unnecessary headers.
Vinson Lee [Sat, 13 Feb 2010 07:24:22 +0000 (23:24 -0800)]
st/xlib: Remove unnecessary header.
Vinson Lee [Sat, 13 Feb 2010 07:18:08 +0000 (23:18 -0800)]
dri: Remove unnecessary headers.
Vinson Lee [Sat, 13 Feb 2010 07:04:20 +0000 (23:04 -0800)]
mesa: Remove unnecessary headers.
Vinson Lee [Sat, 13 Feb 2010 06:33:37 +0000 (22:33 -0800)]
svga: Silence uninitialized variable warning.
Corbin Simpson [Sat, 13 Feb 2010 04:43:20 +0000 (20:43 -0800)]
util: Properly init memory for blitter CSOs.
Fixes misrender on r300g.
Brian Paul [Sat, 13 Feb 2010 00:22:10 +0000 (17:22 -0700)]
st/mesa: restore draw rasterization stage after rasterpos
This fixes invalid calls to rastpos_point/line/tri() that can occur
when glRasterPos() is called while in feedback or selection mode.
(cherry picked from commit
b3c7dc6ff232eb4c5ff6b0ddcfa90bcb56f78df1)
José Fonseca [Fri, 12 Feb 2010 23:52:18 +0000 (23:52 +0000)]
scons: Undo bad merge from cherry-pick.
José Fonseca [Fri, 12 Feb 2010 23:17:04 +0000 (23:17 +0000)]
svga: A few more tweaks to blend color state emission.
José Fonseca [Fri, 12 Feb 2010 21:30:33 +0000 (21:30 +0000)]
svga: Clamp max cube texture levels to supported max 2d levels.
José Fonseca [Fri, 12 Feb 2010 17:01:48 +0000 (17:01 +0000)]
svga: Report the 2d/3d texture sizes supported by the host.
José Fonseca [Fri, 12 Feb 2010 13:59:16 +0000 (13:59 +0000)]
wgl: Go into dormant state when DLL is unloaded unclealy.
When our DLL is unloaded, even if we leave the data structures in memory
for sake of future calls, the MS CRT will destroy the heap. Instead we
make all calls no-ops by setting stw_dev to NULL.