Brian Paul [Thu, 4 Mar 2010 15:36:26 +0000 (08:36 -0700)]
gallivm: added lp_build_negate()
Brian Paul [Sat, 27 Feb 2010 00:44:10 +0000 (17:44 -0700)]
gallivm: clarify unsigned vs. signed integer type construction
The lp_int_type() function was creating an unsigned type. So rename
that function to lp_uint_type() and create new lp_int_type() that
creates a signed type.
Eric Anholt [Wed, 3 Mar 2010 19:51:51 +0000 (11:51 -0800)]
i915: Enable texture tiling by default.
This now passes piglit testcases, and shows performance improvements
on openarena. Blit-heavy apps may see degradation, but the
expectation is that the common case of 3D rendering will see
significant wins.
This regresses gen-teximage on my 855, and no piglit tests on my 945.
Eric Anholt [Wed, 3 Mar 2010 19:49:39 +0000 (11:49 -0800)]
i915: Fix up i830 for tiled drawing offsets.
Corresponds to
b87406e55f029d29594ae76a4b39a4fe1007fe4f.
George Sapountzis [Thu, 4 Mar 2010 14:33:39 +0000 (16:33 +0200)]
fix getprocaddress test
Ian Romanick [Thu, 4 Mar 2010 09:56:20 +0000 (01:56 -0800)]
DRI: Put back missing break-statement
The break in the __DRI_ATTRIB_RENDER_TYPE case was accidentally
removed in commit
5cf2c5851bcd29c2d53bb04ab692b4b156f5a74d. This puts
it back.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Vinson Lee [Thu, 4 Mar 2010 09:51:50 +0000 (01:51 -0800)]
Revert "mesa: Fix unsigned comparison."
This reverts commit
f9504e75f02586a8561733e0e2711c65efa2979d.
This patch is incorrect.
Vinson Lee [Thu, 4 Mar 2010 09:35:28 +0000 (01:35 -0800)]
glsl: Remove unsigned greater than or equal zero comparison.
Vinson Lee [Thu, 4 Mar 2010 09:24:44 +0000 (01:24 -0800)]
mesa: Fix unsigned comparison.
Vinson Lee [Thu, 4 Mar 2010 09:10:44 +0000 (01:10 -0800)]
glsl: Remove unsigned greater than or equal zero comparison.
Vinson Lee [Thu, 4 Mar 2010 08:58:36 +0000 (00:58 -0800)]
glsl: Return NULL on _slan_gen_asm error path on non-debug builds.
Exit the function early on the error path, instead of continuing, to
prevent a null pointer dereference later on.
Vinson Lee [Thu, 4 Mar 2010 08:53:52 +0000 (00:53 -0800)]
glsl: Add assert to check input to strcmp.
Vinson Lee [Thu, 4 Mar 2010 08:41:21 +0000 (00:41 -0800)]
mesa: Add asserts to check inputs to memcpy.
Dave Airlie [Thu, 4 Mar 2010 05:49:57 +0000 (15:49 +1000)]
r300: make blit code work on swtcl machines.
Only tested this on my rv530 with R300_NO_TCL=1, but it works there
at least. It would be nice to fix the tex_offset by chaning the vertex
program on tcl to read from offset 6 but this isn't really necessary.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Dave Airlie [Thu, 4 Mar 2010 05:08:26 +0000 (15:08 +1000)]
r300_blit: initialise whole fp structure to 0
at least on the r500 we get a lockup unless I do this, the compiler
seems to fail to compile and we just get a null fp.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Vinson Lee [Thu, 4 Mar 2010 03:03:45 +0000 (19:03 -0800)]
Remove ffb and gamma from configs.
Vinson Lee [Thu, 4 Mar 2010 02:20:01 +0000 (18:20 -0800)]
wgl: Fix mingw32 build.
This was missed in commit
fcf438e9e03f6e75bca4a49ad372fe7c4b1abbf8.
Ian Romanick [Thu, 4 Mar 2010 01:50:11 +0000 (17:50 -0800)]
Fix unmatched parenthesis introduce by previous commits
I wasn't careful enough when removing support for GCC versions earlier
than 3.3.0. I could have sworn that I compile tested before pushing,
but apparently not. FAIL.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 4 Mar 2010 00:43:13 +0000 (16:43 -0800)]
Note removal of pre-GCC 3.3.0 support in release notes.
Ian Romanick [Thu, 4 Mar 2010 00:15:42 +0000 (16:15 -0800)]
Remove support for GCC older than 3.3.0 from generated sources
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 4 Mar 2010 00:03:24 +0000 (16:03 -0800)]
Remove support for GCC older than 3.3.0 from generator scripts
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 4 Mar 2010 00:02:45 +0000 (16:02 -0800)]
Remove support for GCC older than 3.3.0
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Wed, 3 Mar 2010 23:59:37 +0000 (15:59 -0800)]
Add version check for GCC
Mesa now requires GCC 3.3.0 or later. See the following thread from
the mesa3d-dev mailing list for more details:
http://marc.info/?t=
126748568900005&r=1&w=2
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Alex Deucher [Wed, 3 Mar 2010 21:12:22 +0000 (16:12 -0500)]
radeon/r200/r300/r600: add check_blit vtbl function
Check if the native blit formats are supported, if not,
attempt to use an alternate format.
Skip 3, >4 bpp as per comments from mcencora on irc.
Signed-off-by: Alex Deucher <alexdeucher@gmail.com>
Ian Romanick [Wed, 3 Mar 2010 13:30:14 +0000 (05:30 -0800)]
Note removal of color-index rendering in release notes.
Ian Romanick [Sat, 27 Feb 2010 01:58:53 +0000 (17:58 -0800)]
Remove stray defines of HAVE_RGBA
Now that color-index support is removed from t_dd_tritmp.h and
t_dd_unfilled.h, drivers no longer need define HAVE_RGBA.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Sat, 27 Feb 2010 01:54:59 +0000 (17:54 -0800)]
mesa/tnl_dd: Remove color-index support from t_dd_unfilled.h
The code removal and the re-indent were done together for this one
because the cause of the affected code blocks is much, much smaller
than in t_dd_tritmp.h.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Sat, 27 Feb 2010 01:45:27 +0000 (17:45 -0800)]
mesa/tnl_dd: White-space / indentation fixes after previous commit
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Sat, 27 Feb 2010 01:41:15 +0000 (17:41 -0800)]
mesa/tnl_dd: Remove color-index support from t_dd_tritmp.h
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 03:29:53 +0000 (19:29 -0800)]
mesa: Remove unused RasterIndex field
With the preceeding changes, gl_current_attrib::RasterIndex is not
used. Remove it.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 03:28:04 +0000 (19:28 -0800)]
mesa: Always return default value for CURRENT_RASTER_INDEX
Since there is no color-index rendering, it is impossible to update
this value. Just return the initial setting and be happy.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 03:25:19 +0000 (19:25 -0800)]
mesa: Eliminate index parameter to _mesa_feedback_vertex
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 03:12:30 +0000 (19:12 -0800)]
mesa: Remove ClearIndex and IndexMask from device-driver interface
These are used to inform the driver of the clear value for color-index
buffers and to control write-masking of bits in color-index buffers.
No driver use or need (not even Nouveau) these interfaces.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 03:06:29 +0000 (19:06 -0800)]
swrast: Remove _swrast_logicop_ci_span
After all the recent color-index rendering removal,
_swrast_logicop_ci_span is no longer used anywhere.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 03:03:23 +0000 (19:03 -0800)]
mesa: Remove _mesa_add_color_index_renderbuffers
After all the recent color-index visual support removal,
_mesa_add_color_index_renderbuffers is no longer used anywhere.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 03:00:50 +0000 (19:00 -0800)]
mesa: Remove checks of Visual.rgbMode in Get paths
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 02:58:38 +0000 (18:58 -0800)]
mesa: Remove checks of Visual.rgbMode
This must always be true now, so there is no reason to check it. Ever.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 02:49:33 +0000 (18:49 -0800)]
mesa: Remove support for creating color-index visuals
Remove the rgbMode and indexBits parameters from _mesa_create_visual
and _mesa_initialize_visual. These values are now hardcoded to
GL_TRUE and 0.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 02:24:07 +0000 (18:24 -0800)]
tnl: Remove color-index TNL support
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 02:23:19 +0000 (18:23 -0800)]
DRI: Remove support for color-index visuals
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 02:15:40 +0000 (18:15 -0800)]
swrast_setup: Remove more color-index rendering support
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 01:48:27 +0000 (17:48 -0800)]
swrast: Remove color-index rendering support from s_spantemp.h
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 01:44:28 +0000 (17:44 -0800)]
mesa/xlib: Remove support for color-index rendering
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 00:59:27 +0000 (16:59 -0800)]
osmesa: Remove support for color-index rendering
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 00:51:40 +0000 (16:51 -0800)]
fbdev: Remove support for color-index rendering
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 00:44:39 +0000 (16:44 -0800)]
DRI/swrast: Remove color-index support from DRI swrast driver
It appears that color-index rendering wasn't actually supported
anyway. swrastFillInModes did not previously create an color-index
configs, so it doesn't seem like there would be any way to get a
color-index visual.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 00:34:21 +0000 (16:34 -0800)]
swrast: Remove SPAN_INDEX
Also adjust the bits that appear after it to fill in the gap.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 00:11:43 +0000 (16:11 -0800)]
swrast: Remove _swrast_read_index_span
After all the recent color-index rendering removal,
_swrast_read_index_span is no longer used anywhere.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 00:10:02 +0000 (16:10 -0800)]
swrast: Remove _swrast_mask_ci_span
After all the recent color-index rendering removal,
_swrast_mask_ci_span is no longer used anywhere.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 00:07:11 +0000 (16:07 -0800)]
swrast: Remove _swrast_fog_ci_span and associated code
After all the recent color-index rendering removal,
_swrast_fog_ci_span is no longer used anywhere.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 00:05:20 +0000 (16:05 -0800)]
swrast: Remove _swrast_write_index_span and associated code
After all the recent color-index rendering removal,
_swrast_write_index_span is no longer used anywhere.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Thu, 25 Feb 2010 00:02:38 +0000 (16:02 -0800)]
swrast: Remove remaining color-index state tracking infrastructure
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Wed, 24 Feb 2010 23:54:33 +0000 (15:54 -0800)]
swrast: Remove support for ReadPixels from a color-index buffer
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Wed, 24 Feb 2010 23:45:36 +0000 (15:45 -0800)]
swrast: Remove support for Clear into a color-index buffer
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Wed, 24 Feb 2010 23:39:08 +0000 (15:39 -0800)]
swrast: Remove support for rendering antialiased lines into a color-index buffer
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Wed, 24 Feb 2010 23:26:48 +0000 (15:26 -0800)]
swrast: Remove support for rendering antialiased triangles into a color-index buffer
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Wed, 24 Feb 2010 23:22:56 +0000 (15:22 -0800)]
swrast: Remove support for rendering lines into a color-index buffer
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Wed, 24 Feb 2010 23:20:11 +0000 (15:20 -0800)]
swrast: Remove support for rendering points into a color-index buffer
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Wed, 24 Feb 2010 23:17:30 +0000 (15:17 -0800)]
swrast: Remove support for rendering triangles into a color-index buffer
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Wed, 24 Feb 2010 23:15:01 +0000 (15:15 -0800)]
swrast: Remove support for Bitmap into a color-index buffer
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Wed, 24 Feb 2010 23:11:04 +0000 (15:11 -0800)]
swrast: Remove _swrast_write_zoomed_index_span and associated code
After removing support for CopyPixels and DrawPixels involving
color-index buffers, _swrast_write_zoomed_index_span is no longer
used. Removed it and all the support for COLOR_INDEX formats in zoom_span.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Wed, 24 Feb 2010 23:02:59 +0000 (15:02 -0800)]
swrast: Remove support for CopyPixels to / from a color-index buffer
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Ian Romanick [Wed, 24 Feb 2010 22:55:25 +0000 (14:55 -0800)]
swrast: Remove support for DrawPixels into a color-index buffer
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
José Fonseca [Wed, 3 Mar 2010 19:55:31 +0000 (19:55 +0000)]
llvmpipe: Store the original triangle coordinates in the debug build.
For debugging purposes only.
José Fonseca [Wed, 3 Mar 2010 18:47:29 +0000 (18:47 +0000)]
llvmpipe: Ensure context is flushed when destroying shaders.
Eric Anholt [Wed, 3 Mar 2010 02:06:24 +0000 (18:06 -0800)]
i915: Use x,y drawing offsets instead of changing buffer offsets.
This should fix rendering into mipmaps of tiled textures.
Eric Anholt [Wed, 3 Mar 2010 02:04:40 +0000 (18:04 -0800)]
i915: Tell the kernel when we actually need fence registers on our BOs.
This improves tiled texture performance of OA on my 945 from 25.3fps
to 29.0fps, whereas untiled is 28.2fps, by avoiding stalls for fence
register changes.
Eric Anholt [Tue, 2 Mar 2010 23:34:17 +0000 (15:34 -0800)]
intel: Use drm_intel_bo_alloc_tiled for region allocs.
This moves the logic for how to align pitches, heights, and sizes of
objects to one central location. Fixes rendering with texture tiling
on i915. Note that current libdrm is required for the change for
I915_TILING_NONE pitch alignment.
Eric Anholt [Tue, 2 Mar 2010 23:28:36 +0000 (15:28 -0800)]
i915: Don't do the pitch expansion for tiled buffers.
The weirdness that led to the bumping of pitch for those 512/1024
pixels is that in taking a 2x2 subspan or bilinear filtering we'd end
up hitting the same channel in 2 different pages, leading to lower
performance. With tiling, that doesn't occur, so we don't need to
waste the memory.
Eric Anholt [Tue, 2 Mar 2010 18:41:08 +0000 (10:41 -0800)]
intel: Use GTT mapping in the blit fallback path in case there's tiling.
Eric Anholt [Tue, 2 Mar 2010 18:27:51 +0000 (10:27 -0800)]
i965: Fix up Sandybridge VS sizing.
George Sapountzis [Wed, 3 Mar 2010 19:22:38 +0000 (21:22 +0200)]
glapi: minor cosmetic
George Sapountzis [Wed, 3 Mar 2010 19:22:38 +0000 (21:22 +0200)]
glapi: drop condvar
Seems unused and replaced by functionality in os module.
Alex Deucher [Wed, 3 Mar 2010 19:09:53 +0000 (14:09 -0500)]
r600: add new rs880 pci id
Kristian Høgsberg [Tue, 2 Mar 2010 20:34:18 +0000 (15:34 -0500)]
egl_dri2: Factor out DRI2 protocol bits from initialize.
Kristian Høgsberg [Tue, 2 Mar 2010 20:41:12 +0000 (15:41 -0500)]
intel: Set InternalFormat for renderbuffers created from an EGLImage
José Fonseca [Wed, 3 Mar 2010 13:47:42 +0000 (13:47 +0000)]
progs/gallium/unit: Test (un)pack_4ub.
Most of them are failing.
José Fonseca [Wed, 3 Mar 2010 13:08:23 +0000 (13:08 +0000)]
gallium: Remove PIPE_FORMAT_R8G8B8X8_SNORM.
José Fonseca [Wed, 3 Mar 2010 13:08:08 +0000 (13:08 +0000)]
util: Remove PIPE_FORMAT_R8G8B8X8_SNORM information.
José Fonseca [Wed, 3 Mar 2010 13:07:56 +0000 (13:07 +0000)]
r300: Remove PIPE_FORMAT_R8G8B8X8_SNORM support.
José Fonseca [Wed, 3 Mar 2010 13:07:43 +0000 (13:07 +0000)]
progs/gallium/unit: Remove PIPE_FORMAT_R8G8B8X8_SNORM test cases.
José Fonseca [Wed, 3 Mar 2010 12:07:39 +0000 (12:07 +0000)]
llvmpipe: Linux is no longer a dependency. Llvmpipe builds and runs fine on windows too.
José Fonseca [Wed, 3 Mar 2010 12:00:25 +0000 (12:00 +0000)]
llvmpipe: Centralize LLVM initialization.
José Fonseca [Wed, 3 Mar 2010 12:00:08 +0000 (12:00 +0000)]
gallivm: Centralize initialization. Fix stack alignment issues on Windows.
José Fonseca [Wed, 3 Mar 2010 11:17:16 +0000 (11:17 +0000)]
llvmpipe: Update for lp_bld_init.h rename.
José Fonseca [Wed, 3 Mar 2010 10:54:35 +0000 (10:54 +0000)]
gallivm: Rename lp_bld_misc -> lp_bld_init.
Dave Airlie [Wed, 3 Mar 2010 06:27:21 +0000 (16:27 +1000)]
radeon/r200/r300/r600: fix blit BO buffer validation
there is no need for these to be persistent since we flush the CS
at the end of the blit, and we don't reuse these buffers at all.
on my r600 the Wine DX9 test suite doesn't crash at least anymore.
Signed-off-by: Dave Airlie <airlied@redhat.com>
Corbin Simpson [Wed, 3 Mar 2010 03:14:09 +0000 (19:14 -0800)]
r300g: Use a macro to consolidate general state handling.
My God, I've reinvented R300_STATECHANGE. Forgive me. :C
Christoph Bumiller [Tue, 2 Mar 2010 18:16:39 +0000 (19:16 +0100)]
nv50: fix RT formats after format rename fixes
Getting rid of the evil pre-processor magic, but not changing our
internal format naming convention (at least for now).
José Fonseca [Tue, 2 Mar 2010 17:26:40 +0000 (17:26 +0000)]
nouveau: Rename pipe formats.
These format uses were being hiden by pre-processor magic and were
missed in the previous search'n'replace.
José Fonseca [Tue, 2 Mar 2010 16:00:34 +0000 (16:00 +0000)]
Merge branch 'gallium-format-cleanup'
Zack Rusin [Tue, 2 Mar 2010 15:18:29 +0000 (10:18 -0500)]
llvmpipe: improve based on review from Jose and fix else clauses
else was broken in the outter most else statemants, plus the code
didn't need an inverted mask to compute the inverse of the current
condition.
Keith Whitwell [Tue, 2 Mar 2010 15:02:29 +0000 (15:02 +0000)]
llvmpipe: remove some ifdefs
No longer any need to protect expensive asserts with #ifdef DEBUG.
Keith Whitwell [Tue, 2 Mar 2010 14:41:35 +0000 (14:41 +0000)]
llvmpipe: reorganize transfer usage
Move transfer creation and mapping to the "scene" object, and out of
the rasterizer. The rasterizer operates on already-mapped
framebuffers only, and no longer needs a screen or context pointer.
The scene object has access to a pipe_context, and this reorg prepares
for moving transfer functionality from the screen to the context.
Brian Paul [Tue, 2 Mar 2010 14:50:15 +0000 (07:50 -0700)]
intel: passs correct context type to intel_prepare_render() call
Brian Paul [Tue, 2 Mar 2010 14:47:56 +0000 (07:47 -0700)]
i965: include macros.h to get CLAMP macro
Brian Paul [Tue, 2 Mar 2010 14:34:29 +0000 (07:34 -0700)]
glx: replace assertion with conditional
See fd.o bug 26832.
José Fonseca [Tue, 2 Mar 2010 12:47:16 +0000 (12:47 +0000)]
progs/gallium/unit: Add test cases for single precision floats.
José Fonseca [Tue, 2 Mar 2010 12:40:09 +0000 (12:40 +0000)]
progs/gallium/unit: Add test cases for Nx16 and Nx32 integer formats.
Michal Krol [Tue, 2 Mar 2010 12:35:30 +0000 (13:35 +0100)]
Merge branch 'gallium-no-rhw-position'
Conflicts:
src/gallium/drivers/r300/r300_blit.c
src/gallium/drivers/r300/r300_emit.c
src/gallium/drivers/r300/r300_state_derived.c