Chia-I Wu [Fri, 17 Jul 2009 04:21:51 +0000 (21:21 -0700)]
egl: Add eglcompiler.h for compiler features.
Only INLINE (from mesa/main/compiler.h) is defined now. It may be used
to deal with symbol visibility and int/pointer conversion in the future.
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Signed-off-by: Brian Paul <brianp@vmware.com>
Eric Anholt [Fri, 17 Jul 2009 01:41:03 +0000 (18:41 -0700)]
texenv: Calculate whether we need to do secondary color on our own.
The _TriangleCaps bit is deprecated, not updated when we require, and
is set based on state that hasn't been updated at that point in
_mesa_update_state_locked().
Fixes incorrect clear color in glsl/twoside.c with meta_clear_tris.
Eric Anholt [Thu, 16 Jul 2009 22:57:22 +0000 (15:57 -0700)]
i965: Add missing state dependency of sf_unit on _NEW_BUFFERS.
Keith Whitwell [Thu, 16 Jul 2009 17:45:10 +0000 (18:45 +0100)]
draw: fix ppc build regression
Found by x.org tinderbox, reported by Chris Ball.
Brian Paul [Thu, 16 Jul 2009 14:44:51 +0000 (08:44 -0600)]
st/mesa: add some array index bounds assertions
Chia-I Wu [Thu, 16 Jul 2009 14:29:19 +0000 (08:29 -0600)]
egl: Add xeglthreads
Signed-off-by: Chia-I Wu <olvaffe@gmail.com>
Keith Whitwell [Thu, 16 Jul 2009 11:49:27 +0000 (12:49 +0100)]
tgsi: simplify and fix sse KIL implementation
Use sse_movmskps to extract the correct bits of the comparison result
for use in updating the killmask. Simplify some logic around
identifying the set of necessary comparisons to make.
Keith Whitwell [Thu, 16 Jul 2009 11:47:11 +0000 (12:47 +0100)]
rtasm: export sse_movmskps
Keith Whitwell [Thu, 16 Jul 2009 10:01:19 +0000 (11:01 +0100)]
tgsi: initial texturing support on sse path
Most obvious problem is drawpixels comes out blocky, but this may be
an existing issue of KIL on the sse path.
Keith Whitwell [Thu, 16 Jul 2009 09:11:46 +0000 (10:11 +0100)]
tgsi: make sse function callout mechanism more generic
Take a list of arguments rather than hardcoding TEMP_R0.
Keith Whitwell [Thu, 16 Jul 2009 06:50:34 +0000 (07:50 +0100)]
tgsi: reduce x86 reg usage in tgsi_sse generated programs
Pass the tgsi_exec_machine struct in directly and just hold a single
pointer to this struct, rather than keeping one for each of its
internal members.
Keith Whitwell [Wed, 15 Jul 2009 23:23:33 +0000 (00:23 +0100)]
tgsi: no need to separately malloc input and output arrays
Can now guarantee alignment in the initial allocation of the tgsi exec machine.
Keith Whitwell [Wed, 15 Jul 2009 23:21:17 +0000 (00:21 +0100)]
gallium: reduce recursive include of tgsi_exec.h
A lot of draw code no longer needs to see this header.
Keith Whitwell [Wed, 15 Jul 2009 22:59:55 +0000 (23:59 +0100)]
tgsi: make function call code in tgsi_sse.c less opaque
Explictly pass src and dst arguments (previously dst argument was also
being used as a src). Separate argument handling from the rest of
the function call emit.
Keith Whitwell [Wed, 15 Jul 2009 22:44:53 +0000 (23:44 +0100)]
gallium: proper constructor and destructor for tgsi_exec_machine
Centralize the creation, initialization and destruction of this struct.
Use align_malloc instead of home-brew alternatives.
Dave Airlie [Wed, 15 Jul 2009 03:25:32 +0000 (13:25 +1000)]
radeon: update span reading micro tile code
tested on r500 with zreaddraw with Z buffer in all 3 modes
Dave Airlie [Wed, 15 Jul 2009 03:24:30 +0000 (13:24 +1000)]
r300: emit z depth pitch reloc in preparation for tiling
Dave Airlie [Wed, 15 Jul 2009 02:03:34 +0000 (12:03 +1000)]
radeon: r300 fix span reading for macro tiled buffers.
this uses the correct formula for macro tiled buffers for readback
Dave Airlie [Wed, 8 Jul 2009 01:16:56 +0000 (11:16 +1000)]
radeon: for tiling you really need to use GET/PUT VALUE not PTR.
since the surfaces aren't linear you can't just use GET_PTR
Xiang, Haihao [Wed, 15 Jul 2009 02:40:16 +0000 (10:40 +0800)]
i965: the offset of any branch/jump instruction is in unit of 64bits on IGDNG
Dave Airlie [Tue, 14 Jul 2009 23:35:09 +0000 (09:35 +1000)]
intel/radeon: add common metaops code.
Move all the metaops to a dri_metaops file and port radeon/intel
to use the new common meta ops code.
Dave Airlie [Tue, 14 Jul 2009 21:10:45 +0000 (07:10 +1000)]
radeon: fix unsigned vs signed comparison in stencil code.
This function takes a GLint not a GLuint, passing in -1
breaks internally.
Eric Anholt [Tue, 14 Jul 2009 21:49:03 +0000 (14:49 -0700)]
intel: Fix ClearDepth to not be affected by DepthRange.
Fixes new piglit depthrange-clear.c test.
Eric Anholt [Tue, 14 Jul 2009 21:30:12 +0000 (14:30 -0700)]
intel: Set DepthRange in the metaops using RasterPos[2].
RasterPos[2] is already sent through the window transform, which includes
DepthRange handling. So make DepthRange for the metaops a noop.
Fixes a failure in oglconform depthrange.c
Eric Anholt [Tue, 2 Jun 2009 14:47:20 +0000 (07:47 -0700)]
Fix state flag dependencies for fixed function fragment program updates.
I started looking into why _NEW_ARRAY punishes us, and while annotating
dependencies noticed that a bunch of dependencies were missing.
Eric Anholt [Tue, 14 Jul 2009 18:51:17 +0000 (11:51 -0700)]
intel: Flag _NEW_ARRAY state when doing array object binding in clears.
This is just following bind_vertex_array()'s behavior.
Brian Paul [Tue, 14 Jul 2009 13:47:19 +0000 (07:47 -0600)]
mesa: regenerated enums.c file
Corbin Simpson [Tue, 14 Jul 2009 08:16:25 +0000 (01:16 -0700)]
dri-st: Unbreak GL_EXT_blend_equation_separate.
Since it has a dispatch table entry (for BlendEquationSeparateEXT,)
can't omit it from this list. It'll still get disabled if the cap
(PIPE_CAP_BLEND_EQUATION_SEPARATE) isn't set.
Somebody that doesn't suck at GL (read: not me) should probably add
this into progs/samples/blendeq or similar so we can test it.
Patrice Mandin [Tue, 14 Jul 2009 07:44:49 +0000 (09:44 +0200)]
gallium: Define PIPE_CAP_BLEND_EQUATION_SEPARATE, remove extension from default extension list
Michel Dänzer [Tue, 14 Jul 2009 06:25:27 +0000 (08:25 +0200)]
radeon: Differentiate 16 bpp destination formats.
Fixes those formats in fbo_firecube.
Only tested with r300, radeon and r200 compile tested only.
Michel Dänzer [Tue, 14 Jul 2009 06:00:49 +0000 (08:00 +0200)]
radeon: Invert front face winding when rendering to FBO.
Fixes fgl_glxgears and progs/demos/fbotexture after pressing 'c'.
Tested with r300, radeon and r200 compile tested only.
Nicolai Hähnle [Mon, 13 Jul 2009 22:01:51 +0000 (15:01 -0700)]
r300g: Small compile warning fixes.
Signed-off-by: Corbin Simpson <MostAwesomeDude@gmail.com>
Nicolai Hähnle [Mon, 13 Jul 2009 21:58:45 +0000 (14:58 -0700)]
r300g, radeon: Whitespace fixes.
Signed-off-by: Corbin Simpson <MostAwesomeDude@gmail.com>
Corbin Simpson [Mon, 13 Jul 2009 21:47:36 +0000 (14:47 -0700)]
r300g: Use align() instead of inline maths.
Maciej Cencora [Mon, 13 Jul 2009 18:29:11 +0000 (20:29 +0200)]
Merge branch 'shaders_cleanup'
Eric Anholt [Fri, 10 Jul 2009 01:02:42 +0000 (18:02 -0700)]
intel: Partially fix fallback detection for glCopyTexSubImage.
Really, we should be checking that the MesaTexFormat for the read buffer
and the texture match, but the previous code wasn't even doing that, so
matching the cpp should be an improvement (and potentially fix some hangs!).
The previous code also rejected GL_RGB even though blitting the alpha
channel to the ignored channel of an x8r8g8b8 texture should be fine, which
tripped up compiz's blur plugin.
Maciej Cencora [Sun, 12 Jul 2009 13:09:15 +0000 (15:09 +0200)]
r300: always assume all components are read by fragment program
Components of input attributes that are used by fragment program aren't part of vertex program key, and that may lead to situations when vertex program writes only TEX1.xy and fragment program reads TEX1.xyz, resulting in rendering errors.
Reported-by: Nicolai Hähnle <nhaehnle@gmail.com>
Nicolai Hähnle [Sun, 12 Jul 2009 02:03:25 +0000 (02:03 +0000)]
r300: Fix fogcoord rewriting
We only care about the actual fogcoord itself now, reducing the rewriting
done for the vertex program.
The rewriting of source operand swizzles in the fragment program takes
care that fogcoord.yzw = 001.
This should fix fogcoord rewriting entirely, which had been horribly
broken in the face of dot-product instructions, and just broken (though
not horribly so) in the face of almost every other instruction (the W
component would be incorrect for most arithmetic instructions).
Signed-off-by: Nicolai Hähnle <nhaehnle@gmail.com>
Maciej Cencora [Sat, 11 Jul 2009 17:10:58 +0000 (19:10 +0200)]
r300: fix swizzle masking in getUsedComponents
Maciej Cencora [Sat, 11 Jul 2009 14:41:52 +0000 (16:41 +0200)]
r300: minor fix
Split initializations becase the vars are of different type.
Reported-by: Nicolai Hähnle <nhaehnle@gmail.com>
Maciej Cencora [Sat, 11 Jul 2009 14:37:11 +0000 (16:37 +0200)]
r300: move variables declarations
Maciej Cencora [Sat, 11 Jul 2009 14:34:04 +0000 (16:34 +0200)]
r300: document r300_vertex_program_cont structure
Reported-by: Nicolai Hähnle <nhaehnle@gmail.com>
Maciej Cencora [Sat, 11 Jul 2009 14:26:23 +0000 (16:26 +0200)]
r300: document r300_fragment_program_cont struct
Maciej Cencora [Sat, 11 Jul 2009 14:16:11 +0000 (16:16 +0200)]
r300: fix indentation
Reported-by: Nicolai Hähnle <nhaehnle@gmail.com>
Maciej Cencora [Sat, 11 Jul 2009 13:53:01 +0000 (15:53 +0200)]
r300: fix StrideB == 0 case when converting data format
Reported-by: Nicolai Hähnle <nhaehnle@gmail.com>
Maciej Cencora [Sat, 11 Jul 2009 13:09:20 +0000 (15:09 +0200)]
r300: hw can handle per component negations in vertex shaders
Reported-by: Nicolai Haehnle <nhaehnle@gmail.com>
Maciej Cencora [Sun, 5 Jul 2009 01:17:32 +0000 (03:17 +0200)]
r300: fix WPOS for SWTCL
Maciej Cencora [Sat, 4 Jul 2009 21:38:59 +0000 (23:38 +0200)]
r300: removed unnecessary params
We don't have check which attributes are used by fragment program - it's already done by NQSSADCE.
Maciej Cencora [Sun, 5 Jul 2009 00:34:48 +0000 (02:34 +0200)]
r300: use NQSSADCE for vertex programs
Maciej Cencora [Sat, 4 Jul 2009 14:52:48 +0000 (16:52 +0200)]
r300: operate on copy of a program when pairing instructions
We need to keep unpaired program for vertex program NQSSADCE.
Maciej Cencora [Mon, 13 Jul 2009 17:23:18 +0000 (19:23 +0200)]
r300: handle relative addressing in NQSSADCE
Maciej Cencora [Sat, 4 Jul 2009 13:22:22 +0000 (15:22 +0200)]
r300: handle ARB_vertex_program specific instructions in NQSSADCE
Maciej Cencora [Sat, 4 Jul 2009 14:50:25 +0000 (16:50 +0200)]
r300: move depth output rewrite out of NQSSADCE
Maciej Cencora [Sun, 5 Jul 2009 00:32:51 +0000 (02:32 +0200)]
r300: rewrite FOGC and HPOS attribs handling
Rewrite vertex and fragment programs so that we don't have to do any hacks on lower level.
Maciej Cencora [Sun, 5 Jul 2009 00:03:32 +0000 (02:03 +0200)]
r300: bind vertex program to fragment program
Maciej Cencora [Sat, 4 Jul 2009 12:57:42 +0000 (14:57 +0200)]
r300: recalculate used inputs and outputs after dead code removal
Maciej Cencora [Sat, 4 Jul 2009 11:17:35 +0000 (13:17 +0200)]
r300: move fragment program selection before vertex program selection
Prepare for wpos and fogc handling rewrite.
Maciej Cencora [Fri, 3 Jul 2009 18:14:24 +0000 (20:14 +0200)]
r300: implement proper IsProgramNative check for vertex programs
Maciej Cencora [Fri, 3 Jul 2009 18:06:23 +0000 (20:06 +0200)]
r300: don't modify original vertex program
Keep the original vertex program untouched because it may be needed after some state change for generating new r300 specific vertex program.
Maciej Cencora [Sat, 4 Jul 2009 11:26:49 +0000 (13:26 +0200)]
r300: cache translated fragment programs
Maciej Cencora [Sat, 4 Jul 2009 10:56:31 +0000 (12:56 +0200)]
r300: update state parameters only once per rendering operation
Maciej Cencora [Wed, 1 Jul 2009 16:43:14 +0000 (18:43 +0200)]
r300: translate non native insts earlier for easier debugging
Maciej Cencora [Mon, 29 Jun 2009 19:52:39 +0000 (21:52 +0200)]
r300: print vertex program after adding artificial output insts
Maciej Cencora [Mon, 29 Jun 2009 19:48:35 +0000 (21:48 +0200)]
r300: use mesa provided function for adding MVP code
Maciej Cencora [Mon, 29 Jun 2009 19:39:59 +0000 (21:39 +0200)]
r300: simplify insert_wpos a little
Brian Paul [Mon, 13 Jul 2009 14:47:37 +0000 (08:47 -0600)]
Merge branch 'mesa_7_5_branch'
Dave Airlie [Mon, 13 Jul 2009 15:22:17 +0000 (01:22 +1000)]
radeon: port more front fixes from intel.
Port fixes to read buffer from front.
Dave Airlie [Mon, 13 Jul 2009 13:41:32 +0000 (23:41 +1000)]
radeon/r200: fix color masking under dri2
Need to retrieve the bits from the rrb not from screen struct
Dave Airlie [Mon, 13 Jul 2009 13:09:17 +0000 (23:09 +1000)]
radeon: Use Stencil.Enabled instead of Stencil._Enabled in DrawBuffers.
The _Enabled field isn't updated at the point that DrawBuffers is called,
and the Driver.Enable() function does the testing for stencil buffer
presence anyway.
Dave Airlie [Mon, 13 Jul 2009 13:02:54 +0000 (23:02 +1000)]
radeon/fbo: stencil bits fix from Michel in intel fbo code
Peteri Andras [Mon, 13 Jul 2009 14:03:16 +0000 (08:03 -0600)]
r128: fix two-sided lighting segfault seen in GLUT's olight demo
Ian Romanick [Mon, 13 Jul 2009 04:07:38 +0000 (21:07 -0700)]
intel: Bump driver data, add RC3 tag
Xiang, Haihao [Mon, 13 Jul 2009 02:48:43 +0000 (10:48 +0800)]
i965: add support for new chipsets
1. new PCI ids
2. fix some 3D commands on new chipset
3. fix send instruction on new chipset
4. new VUE vertex header
5. ff_sync message (added by Zou Nan Hai <nanhai.zou@intel.com>)
6. the offset in JMPI is in unit of 64bits on new chipset
7. new cube map layout
Dave Airlie [Sun, 12 Jul 2009 11:35:59 +0000 (21:35 +1000)]
r300: move fallback warnings inside fallback debugging
random output is bad
Dave Airlie [Sun, 12 Jul 2009 11:35:24 +0000 (21:35 +1000)]
r300: fix clear mask to not use sw if not necessary
Nicolai Hähnle [Sun, 12 Jul 2009 09:18:43 +0000 (11:18 +0200)]
radeon: Fix crash when rendering to incomplete texture and other formats
It is possible to bind texture images of an incomplete mipmapped texture.
Software fallbacks in this case incorrectly tried to mmap the entire texture.
Additionally, add span functions for 1555 and 4444 formats.
This fixes crashes in piglit's fbo-readpixels test; unfortunately, the test
itself still fails - this needs to be investigated.
Signed-off-by: Nicolai Hähnle <nhaehnle@gmail.com>
Dave Airlie [Sun, 12 Jul 2009 07:28:14 +0000 (17:28 +1000)]
radeon: update clear code from Intel codebase.
This updates some of the clear code from Intel gives a 5x clearspd perf
for me here. played openarena also, not sure if the viewport changes
broke anything,
Dave Airlie [Sun, 12 Jul 2009 02:13:40 +0000 (12:13 +1000)]
radeon: fbo fix firecube crashes
it might still be misrendering not sure
Michel Dänzer [Sat, 11 Jul 2009 18:25:09 +0000 (20:25 +0200)]
radeon: Fix scissor rectangle calculation when rendering to FBO.
fgl_glxgears -fbo runs, though the gears don't look right yet.
Zack Rusin [Sat, 11 Jul 2009 17:48:41 +0000 (13:48 -0400)]
gallium: compare the actual register, not all the inputs
otherwise we decrement indexes for all registers
Brian Paul [Fri, 10 Jul 2009 22:26:09 +0000 (16:26 -0600)]
tgsi: update some assertions
Brian Paul [Fri, 10 Jul 2009 21:44:48 +0000 (15:44 -0600)]
tgsi: tgsi: add semantic_names[] string for TGSI_SEMANTIC_FACE
Same story as in the tgsi_dump.c code (see prev commit).
Brian Paul [Fri, 10 Jul 2009 21:41:26 +0000 (15:41 -0600)]
tgsi: add semantic_names[] string for TGSI_SEMANTIC_FACE
Fixes TGSI dump output when front/back-face register is declared.
Also, add some assertions to make sure the semantic/interpolate string
arrays have as many elements as there are tokens in the p_shader_tokens.h
file. That should catch problems like this in the future.
Brian Paul [Fri, 10 Jul 2009 19:07:45 +0000 (13:07 -0600)]
st/mesa: implement indirect addressing for destination registers
Brian Paul [Fri, 10 Jul 2009 19:07:16 +0000 (13:07 -0600)]
tgis: implement indirect addressing for destination registers
Includes the TGSI interpreter, but not the SSE/PPC/etc code generators.
Brian Paul [Fri, 10 Jul 2009 18:55:30 +0000 (12:55 -0600)]
vbo: fix vbo/dlist memory leak
Based on a patch by kristof.ralovich@gmail.com
Brian Paul [Thu, 9 Jul 2009 19:59:03 +0000 (13:59 -0600)]
demos: set 4th component of texcoord to 1.0
Avoid potential randomness in resulting texcoords.
Dave Airlie [Fri, 10 Jul 2009 17:15:58 +0000 (03:15 +1000)]
radeon: enable GL_NV_texture_rectangle under dri2.
Dave Airlie [Fri, 10 Jul 2009 17:01:52 +0000 (03:01 +1000)]
radeon: set texture in state properly.
make sure to turn off when no texture is used in hw
Dave Airlie [Fri, 10 Jul 2009 17:01:04 +0000 (03:01 +1000)]
radeon: make swtcl emit size bigger
Michel Dänzer [Fri, 10 Jul 2009 12:49:46 +0000 (14:49 +0200)]
Remove stale reference to non-Gallium nouveau driver from configure.ac.
Ian Romanick [Tue, 23 Jun 2009 01:22:51 +0000 (18:22 -0700)]
mesa: From float type modifier from values to large for singles
The values
2147483648.0 and
4294967294.0 are too larget to be stored in single
precision floats. Forcing these to be singles causes bits to be lost, which
results in errors in some pixel transfer tests.
This fixes bug #22344.
(cherry picked from commit
70e72070fce6aa1e0918dcc62c1949465cee69f7)
Brian Paul [Thu, 9 Jul 2009 14:05:56 +0000 (08:05 -0600)]
Merge branch 'mesa_7_5_branch'
Brian Paul [Thu, 9 Jul 2009 14:04:07 +0000 (08:04 -0600)]
docs: document gl_TextureMatrix[i][j] array indexing bug fix
Brian Paul [Thu, 9 Jul 2009 13:58:50 +0000 (07:58 -0600)]
glsl: do const parameter optimization for array element actual parameters
When a function parameter is const-qualified we can avoid making a copy
of the actual parameter (we basically do a search/replace when inlining).
This is now done for array element params too, resulting in better code
(fewer MOV instructions).
We should allow some other types of function arguments here but let's be
conservative for the moment.
Brian Paul [Thu, 9 Jul 2009 13:57:29 +0000 (07:57 -0600)]
glsl: fix incorrect indexing for gl_TextureMatrix[i][j]
The two indexes were mixed up when accessing a row of a matrix in an array
of matrices.
Brian Paul [Wed, 8 Jul 2009 20:14:03 +0000 (14:14 -0600)]
docs: document glMaterial/glShadeModel display list optimization
Brian Paul [Tue, 5 May 2009 22:21:20 +0000 (16:21 -0600)]
mesa: only use fallback texture when using shaders, not fixed-function (take two)
The semantics are a little different for shaders vs. fixed-function when
trying to use an incomplete texture. The fallback texture returning
(0,0,0,1) should only be used with shaders. For fixed function, the texture
unit is truly disabled/ignored.
Fixes glean fbo test regression.
(cherry picked from commit
01e16befd4809e3e93ee7855a4a5d7df9fe010d9)
(cherry picked from commit
51325f44d1e7e62b47795a79f8038d10dc5bc30b)
[anholt: squashed these two together from master, skipping the mess in between]
Zack Rusin [Tue, 7 Jul 2009 19:48:50 +0000 (12:48 -0700)]
gallium: fixup register indexes for fog/frontface/point coord
Michel Dänzer [Tue, 7 Jul 2009 12:49:52 +0000 (14:49 +0200)]
gallium: Fixes for clobbering stencil values in combined depth/stencil textures.
Also fix one case where a 32 bit depth value was incorrectly converted to a
combined depth/stencil value.