mesa.git
11 years agoglsl: change int->unsigned to silence MSVC warnings
Brian Paul [Sun, 4 Nov 2012 23:43:44 +0000 (16:43 -0700)]
glsl: change int->unsigned to silence MSVC warnings

Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
11 years agoscons: Require libdrm_radeon 2.4.40.
Vinson Lee [Tue, 6 Nov 2012 06:00:01 +0000 (22:00 -0800)]
scons: Require libdrm_radeon 2.4.40.

Signed-off-by: Vinson Lee <vlee@freedesktop.org>
11 years agor600g: add in-place DB decompression and texturing with DB tiling
Marek Olšák [Tue, 2 Oct 2012 20:02:54 +0000 (22:02 +0200)]
r600g: add in-place DB decompression and texturing with DB tiling

The decompression is done in-place and only the compressed tiles are
decompressed. Note: R6xx-R7xx can do that only with Z16 and Z32F.

The texture unit is programmed to use non-displayable tiling and depth
ordering of samples, so that it can fetch the texture in the native DB format.

The latest version of the libdrm surface allocator is required for stencil
texturing to work. The old one didn't create the mipmap tree correctly.
We need a separate mipmap tree for stencil, because the stencil mipmap
offsets are not really depth offsets/4.

There are still some known bugs, but this should save some memory and it also
improves performance a little bit in Lightsmark (especially with low
resolutions; tested with Radeon HD 5000).

The DB->CB copy is still used for transfers.

Reviewed-by: Jerome Glisse <jglisse@redhat.com>
11 years agoconfigure.ac: require libdrm_radeon 2.4.40
Marek Olšák [Tue, 6 Nov 2012 01:35:35 +0000 (02:35 +0100)]
configure.ac: require libdrm_radeon 2.4.40

11 years agovbo: fix glVertexAttribI* functions
Marek Olšák [Tue, 30 Oct 2012 13:44:22 +0000 (14:44 +0100)]
vbo: fix glVertexAttribI* functions

The functions were broken, because they converted ints to floats.
Now we can finally advertise OpenGL 3.0. ;)

In this commit, the vbo module also tracks the type for each attrib
in addition to the size. It can be one of FLOAT, INT, UNSIGNED_INT.

The little ugliness is the vertex attribs are declared as floats even though
there may be integer values. The code just copies integer values into them
without any conversion.

This implementation passes the glVertexAttribI piglit test which I am going
to commit in piglit soon. The test covers vertex arrays, immediate mode and
display lists.

NOTE: This is a candidate for the stable branches.

Reviewed-by: Brian Paul <brianp@vmware.com>
v2: cosmetic changes as suggested by Brian

11 years agometa: Remove redundant code in _mesa_meta_GenerateMipmap
Anuj Phogat [Fri, 2 Nov 2012 18:18:16 +0000 (11:18 -0700)]
meta: Remove redundant code in _mesa_meta_GenerateMipmap

Integer textures generate invalid operation in glGenerateMipmap.
So, the code related to integer textures is now redundant.

Note: This is a candidate for stable branches.
Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
11 years agomesa: Generate invalid operation in glGenerateMipMap for integer textures
Anuj Phogat [Fri, 2 Nov 2012 17:47:33 +0000 (10:47 -0700)]
mesa: Generate invalid operation in glGenerateMipMap for integer textures

Khronos has reached a conclusion and disallowed following texture formats in
glGenerateMipMap():
 (a) ASTC textures
 (b) integer internal formats (e.g., RGBA8UI, RG16I)
 (c) textures with stencil formats (e.g., STENCIL_INDEX8)
 (d) textures with packed depth/stencil formats (e.g, DEPTH24_STENCIL8)

https://cvs.khronos.org/bugzilla/show_bug.cgi?id=9471

Note: This is a candidate for stable branches.
Signed-off-by: Anuj Phogat <anuj.phogat@gmail.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
11 years agotrace: Prevent segfault when passing NULL to set_vertex_buffers.
José Fonseca [Mon, 5 Nov 2012 11:18:07 +0000 (11:18 +0000)]
trace: Prevent segfault when passing NULL to set_vertex_buffers.

State tracker now passes NULL buffer array to unbind buffers.

11 years agogalahad: Prevent segfault when passing NULL to set_vertex_buffers.
José Fonseca [Mon, 5 Nov 2012 11:05:34 +0000 (11:05 +0000)]
galahad: Prevent segfault when passing NULL to set_vertex_buffers.

State tracker now passes NULL buffer array to unbind buffers.

11 years agoutil: Make u_framebuffer.h C++ safe.
José Fonseca [Fri, 2 Nov 2012 16:56:30 +0000 (16:56 +0000)]
util: Make u_framebuffer.h C++ safe.

11 years agomesa: Use "non-gen name" more consistently as an error message in GL core.
Eric Anholt [Wed, 31 Oct 2012 22:37:59 +0000 (15:37 -0700)]
mesa: Use "non-gen name" more consistently as an error message in GL core.

I used this to help verify that my test was actually testing the paths I
wanted to.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
11 years agomesa: Fix core GL genned-name handling for glBeginQuery().
Eric Anholt [Wed, 31 Oct 2012 22:36:27 +0000 (15:36 -0700)]
mesa: Fix core GL genned-name handling for glBeginQuery().

Fixes piglit gl-3.1/genned-names.

NOTE: This is a candidate for the 9.0 branch.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
11 years agomesa: Fix the core GL genned-name handling for glBindBufferBase()/Range().
Eric Anholt [Wed, 31 Oct 2012 22:33:41 +0000 (15:33 -0700)]
mesa: Fix the core GL genned-name handling for glBindBufferBase()/Range().

This is part of fixing gl-3.1/genned-names.

v2: Fix a missing return value.

NOTE: This is a candidate for the 9.0 branch.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
11 years agoi965: Fix oversized initial allocation of the state cache table pointers.
Vandrus Zoltán [Wed, 31 Oct 2012 20:11:45 +0000 (13:11 -0700)]
i965: Fix oversized initial allocation of the state cache table pointers.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=55030

11 years agoi965: Force border color A to 1 when it's not present in the GL format.
Eric Anholt [Thu, 1 Nov 2012 17:37:56 +0000 (10:37 -0700)]
i965: Force border color A to 1 when it's not present in the GL format.

It's usually forced to 1 by the surface format, but sometimes we actually have
alpha present because it's the only format available.

Fixes piglit texwrap bordercolor tests for OpenGL 1.1, GL_EXT_texture_sRGB and
GL_ARB_texture_float.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
11 years agoi965: Fix uploading user vertex arrays with basevertex set.
Eric Anholt [Wed, 31 Oct 2012 20:39:26 +0000 (13:39 -0700)]
i965: Fix uploading user vertex arrays with basevertex set.

If the index buffer is full of values like "0 1 2 3", but basevertex is 4, we
need to upload at least vertex data for elements 4 5 6 7.  Whether we also
upload 0 1 2 3 is a question of whether there are VBOs present or not -- see
the code setting start_vertex_bias in brw_draw_upload.c.

Fixes piglit draw-elements*base-vertex user_varrays

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
11 years agoi965: Set dirty state for brw_draw_upload.c when num_instances changes.
Eric Anholt [Wed, 31 Oct 2012 20:36:16 +0000 (13:36 -0700)]
i965: Set dirty state for brw_draw_upload.c when num_instances changes.

Otherwise, if we had a set of prims passed in with a num_instances varying
between them, we wouldn't upload enough (or too much!) from user vertex
arrays.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
11 years agoi965: Remove the vbo_rebase_prims() path.
Eric Anholt [Wed, 31 Oct 2012 20:56:56 +0000 (13:56 -0700)]
i965: Remove the vbo_rebase_prims() path.

The brw_draw_upload.c start_vertex_bias code has support for doing the rebase
without rewriting the index buffer by applying a basevertex.  It looks like
vbo_rebase_prims() is not equipped to handle basevertex.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
11 years agoi965/fs: Fix a comment in copy propagation.
Eric Anholt [Tue, 30 Oct 2012 22:02:23 +0000 (15:02 -0700)]
i965/fs: Fix a comment in copy propagation.

We haven't been only tracking raw GRF-GRF moves since the constant propagation
merge, and also the extension for source modifiers and uniforms.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
11 years agoi965/fs: Allow copy-propagation on pull constant load values.
Eric Anholt [Tue, 30 Oct 2012 21:38:41 +0000 (14:38 -0700)]
i965/fs: Allow copy-propagation on pull constant load values.

Given that we handle similarly-regioned GRFs registers for our copy
propagation from our UNIFORM file, there's no reason not to allow it.

The shader-db impact is negligible -- +90 instructions total, 2 shaders helped
and 7 hurt (slightly increased register pressure increased spilling), but this
is to prevent regression in other shaders when fixing copy_propagation to
reduce register pressure in the shaders that are hurt here.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
11 years agoi965/fs: Do dead code elimination just after copy propagation.
Eric Anholt [Tue, 30 Oct 2012 23:10:14 +0000 (16:10 -0700)]
i965/fs: Do dead code elimination just after copy propagation.

If we put the register coalescing in between the two, then we end up with code
sequences involving dead writes that the dead code elimination doesn't know
how to remove.  In place of making dead code elimination smart (which we
should do, too), make it less important for the moment.

shader-db results:

total instructions in shared programs: 722240 -> 721275 (-0.13%)
instructions in affected programs:     50573 -> 49608 (-1.91%)

(no shaders regressed).

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
11 years agonv50,nvc0: expose ARB_map_buffer_alignment
Lucas Stach [Wed, 31 Oct 2012 15:31:12 +0000 (16:31 +0100)]
nv50,nvc0: expose ARB_map_buffer_alignment

All HW buffers (also suballocated ones) are already aligned.
Just make sure that also the initial sysram buffers have proper
alignment.

11 years agoi965/fs: Compact the virtual GRF arrays.
Kenneth Graunke [Fri, 2 Nov 2012 05:04:50 +0000 (22:04 -0700)]
i965/fs: Compact the virtual GRF arrays.

During code generation, we create tons of temporary variables, many of
which get immediately killed and are never used.  Later optimization and
analysis passes, such as compute_live_intervals, loop over all the
virtual GRFs.  By compacting them, we can save a lot of overhead.

Reduces compilation time in L4D2's largest fragment shader from 10.2
seconds to 5.2 seconds (50%).  Drops compute_live_variables() from
10-12% of another game's startup time to 8%.

Reviewed-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
11 years agodispatch_sanity test: add GL CORE 3.1 test
Jordan Justen [Wed, 24 Oct 2012 16:25:52 +0000 (09:25 -0700)]
dispatch_sanity test: add GL CORE 3.1 test

The function list was generated from glcorearb.h for GL 4.3.

Note that many GL 4.X functions are commented out, and indicate
that they need to be added to Mesa's XML.

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
11 years agodispatch_sanity test: create common context creation function
Jordan Justen [Mon, 29 Oct 2012 23:06:03 +0000 (16:06 -0700)]
dispatch_sanity test: create common context creation function

We also no longer call _swrast_CreateContext, _tnl_CreateContext
or _swsetup_CreateContext when creating the context.

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
11 years agodispatch_sanity test: allow newer functions to be set to NOP
Jordan Justen [Wed, 24 Oct 2012 01:00:11 +0000 (18:00 -0700)]
dispatch_sanity test: allow newer functions to be set to NOP

If a GL function was introduced in a later GL version than the
context we are testing, then it is okay if it is set to the
_mesa_generic_nop function.

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
11 years agodispatch_sanity test: pass ctx to validate_functions/nops
Jordan Justen [Wed, 17 Oct 2012 03:50:03 +0000 (20:50 -0700)]
dispatch_sanity test: pass ctx to validate_functions/nops

This will allow validate_functions to access ctx->Version.

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
11 years agodispatch_sanity test: add version to function list
Jordan Justen [Wed, 17 Oct 2012 03:44:40 +0000 (20:44 -0700)]
dispatch_sanity test: add version to function list

This will be used by GL CORE contexts to differentiate functions that
can be set to nop from functions that are required for a particular
context version.

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
11 years agomesa: remove unimplemented FramebufferTextureFaceARB
Jordan Justen [Mon, 29 Oct 2012 17:47:19 +0000 (10:47 -0700)]
mesa: remove unimplemented FramebufferTextureFaceARB

This function can be re-added with an actual implementation
when ARB_geometry_shader4 is supported.

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
11 years agomesa: remove unimplemented FramebufferTextureARB
Jordan Justen [Mon, 29 Oct 2012 17:45:44 +0000 (10:45 -0700)]
mesa: remove unimplemented FramebufferTextureARB

This function can be re-added with an actual implementation
when ARB_geometry_shader4 is supported.

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
11 years agomesa: disable ProgramParameteri until it is needed
Jordan Justen [Mon, 29 Oct 2012 17:37:50 +0000 (10:37 -0700)]
mesa: disable ProgramParameteri until it is needed

ProgramParameteri will be required for ARB_geometry_shader4
or GLES3. Don't enable this function until either of those
is supported.

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
11 years agoglapi: alias ProgramParameteriARB to ProgramParameteri
Jordan Justen [Fri, 26 Oct 2012 22:38:23 +0000 (15:38 -0700)]
glapi: alias ProgramParameteriARB to ProgramParameteri

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
11 years agoglapi: move include for ARB_get_program_binary.xml to gl_API.xml
Jordan Justen [Fri, 26 Oct 2012 22:36:45 +0000 (15:36 -0700)]
glapi: move include for ARB_get_program_binary.xml to gl_API.xml

These functions are part in GL 4.3. Moving this will allow
ProgramParameteriARB to alias ProgramParameteri.

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
11 years agoglapi: alias FramebufferTextureARB to FramebufferTexture
Jordan Justen [Fri, 26 Oct 2012 22:33:13 +0000 (15:33 -0700)]
glapi: alias FramebufferTextureARB to FramebufferTexture

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
11 years agomesa shaderapi: don't enable various functions for GL CORE
Jordan Justen [Wed, 24 Oct 2012 01:13:32 +0000 (18:13 -0700)]
mesa shaderapi: don't enable various functions for GL CORE

These EXT_separate_shader_objects function will no longer be
enabled for CORE profiles:
* UseShaderProgramEXT
* ActiveProgramEXT
* CreateShaderProgramEXT

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
11 years agomesa api_exec: disable StencilFuncSeparateATI for API_OPENGL_CORE
Jordan Justen [Fri, 2 Nov 2012 23:05:25 +0000 (16:05 -0700)]
mesa api_exec: disable StencilFuncSeparateATI for API_OPENGL_CORE

This was mistakenly enabled in a21116f.

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
11 years agomesa api_exec: add comment regarding GetPointerv & CORE profiles
Jordan Justen [Wed, 24 Oct 2012 16:19:56 +0000 (09:19 -0700)]
mesa api_exec: add comment regarding GetPointerv & CORE profiles

GetPointerv was de-deprecated in 893ddb.

Signed-off-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
11 years agor600g: make tgsi-to-llvm generates store.pixel* intrinsic for fs
Vincent Lejeune [Sat, 29 Sep 2012 14:49:13 +0000 (16:49 +0200)]
r600g: make tgsi-to-llvm generates store.pixel* intrinsic for fs

Reviewed-by: Tom Stellard <thomas.stellard at amd.com>
11 years agoconfigure.ac: Prevent build of radeon llvm backend with llvm < 3.2
Vincent Lejeune [Wed, 31 Oct 2012 20:02:29 +0000 (21:02 +0100)]
configure.ac: Prevent build of radeon llvm backend with llvm < 3.2

Reviewed-by: Tom Stellard <thomas.stellard at amd.com>
11 years agoandroid: Update for builtin_stubs.cpp move
Thierry Reding [Fri, 2 Nov 2012 11:34:24 +0000 (12:34 +0100)]
android: Update for builtin_stubs.cpp move

This fixes the Android build after the move of builtin_stubs.cpp into
the builtin_compiler subdirectory. This patch is untested.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
11 years agoradeonsi: Implement support for vertex shader samplers.
Michel Dänzer [Fri, 2 Nov 2012 16:11:27 +0000 (17:11 +0100)]
radeonsi: Implement support for vertex shader samplers.

Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
11 years agoglsl: Fix builtin_compiler build by -I $(top_srcdir)/include.
Johannes Obermayr [Fri, 2 Nov 2012 11:30:22 +0000 (12:30 +0100)]
glsl: Fix builtin_compiler build by -I $(top_srcdir)/include.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=56664

11 years agoscons: Update for builtin_stubs.cpp
José Fonseca [Fri, 2 Nov 2012 09:42:13 +0000 (09:42 +0000)]
scons: Update for builtin_stubs.cpp

Note this by itself is not enough to fix scons build -- it will fail
until you remove:

   rm -rf build/*/glsl/builtin_compiler

because that node was a filei before, but it will be now a directory.

This also means that bisecting across this change will require wiping
the build directory..

11 years agobuild: Don't cross-compile GLSL builtin compiler
Thierry Reding [Fri, 19 Oct 2012 12:03:01 +0000 (14:03 +0200)]
build: Don't cross-compile GLSL builtin compiler

The builtin_compiler binary is used during the build process to generate
code for the builtin GLSL functions. Since this binary needs to be run
on the build host, it must not be cross-compiled.

This patch fixes the build system to compile a second version of the
source files and the builtin_compiler binary itself for the build
system. It does so by defining the CC_FOR_BUILD and CXX_FOR_BUILD
variables, which are searched for by the configure script and point to
the location of native C and C++ compilers.

In order for this to work properly, builtin_function.cpp is removed
from BUILT_SOURCES, otherwise the build system would try to generate it
before having had a chance to descend into the builtin_compiler
subdirectory. With the builtin_compiler and glsl_compiler now being
generated at different stages, the build instructions for glsl_compiler
can be simplified a bit.

Signed-off-by: Thierry Reding <thierry.reding@avionic-design.de>
Acked-by: Kenneth Graunke <kenneth@whitecape.org>
11 years agolibgl-xlib: include glheader.h instead of GL/gl.h to fix build
Brian Paul [Thu, 1 Nov 2012 23:38:40 +0000 (17:38 -0600)]
libgl-xlib: include glheader.h instead of GL/gl.h to fix build

GL/gl.h doesn't define GLfixed but glapitable.h uses it.

11 years agoi965: Remove unused variables after removing the old VS backend.
Kenneth Graunke [Thu, 1 Nov 2012 23:12:56 +0000 (16:12 -0700)]
i965: Remove unused variables after removing the old VS backend.

Fixes compiler warnings about unused variables.

11 years agoi965: Remove unnecessary walk through Mesa IR in ProgramStringNotify().
Kenneth Graunke [Fri, 26 Oct 2012 18:52:23 +0000 (11:52 -0700)]
i965: Remove unnecessary walk through Mesa IR in ProgramStringNotify().

Variable indexing of non-uniform arrays only exists in GLSL.  Likewise,
OPCODE_CAL/OPCODE_RET only existed to try and support GLSL's function
calls.  We don't use Mesa IR for GLSL, and these features are explicitly
disallowed by ARB_vertex_program/ARB_fragment_program and never
generated by ffvertex_prog.c.

Since they'll never happen, there's no need to check for them, which
saves us from walking through all the Mesa IR instructions.

Reviewed-by: Eric Anholt <eric@anholt.net>
11 years agoi965: Remove VS constant buffer read support from brw_eu_emit.c.
Kenneth Graunke [Fri, 26 Oct 2012 18:15:07 +0000 (11:15 -0700)]
i965: Remove VS constant buffer read support from brw_eu_emit.c.

brw_vec4_emit.cpp implements this directly; only the old backend used
the brw_eu_emit.c code.

Reviewed-by: Eric Anholt <eric@anholt.net>
11 years agoi965: Update comment about clipper constants.
Kenneth Graunke [Fri, 26 Oct 2012 18:01:57 +0000 (11:01 -0700)]
i965: Update comment about clipper constants.

The old VS backend doesn't exist, but I believe these still need to be
delivered to the clipper thread.

Reviewed-by: Eric Anholt <eric@anholt.net>
11 years agoi965/vs: Remove brw_vs_compile::constant_map.
Kenneth Graunke [Fri, 26 Oct 2012 17:58:04 +0000 (10:58 -0700)]
i965/vs: Remove brw_vs_compile::constant_map.

It was only used for the old backend.

Reviewed-by: Eric Anholt <eric@anholt.net>
11 years agoi965/vs: Remove support for the old parameter layout.
Kenneth Graunke [Fri, 26 Oct 2012 17:51:02 +0000 (10:51 -0700)]
i965/vs: Remove support for the old parameter layout.

Only the old backend used it.

Reviewed-by: Eric Anholt <eric@anholt.net>
11 years agoi965/vs: Delete the old vertex shader backend.
Kenneth Graunke [Fri, 26 Oct 2012 17:45:31 +0000 (10:45 -0700)]
i965/vs: Delete the old vertex shader backend.

It's no longer used for anything.

Reviewed-by: Eric Anholt <eric@anholt.net>
11 years agoi965/vs: Replace brw_vs_emit.c with dumping code into the vec4_visitor.
Kenneth Graunke [Mon, 8 Oct 2012 17:21:30 +0000 (10:21 -0700)]
i965/vs: Replace brw_vs_emit.c with dumping code into the vec4_visitor.

Rather than having two separate backends, just create a small layer that
translates the subset of Mesa IR used for ARB_vertex_program and fixed
function programs to the Vec4 IR.  This allows us to use the same
optimization passes, code generator, register allocator as for GLSL.

v2: Incorporate Eric's review comments.
- Fix use of uninitialized src_swiz[] values in the SWIZZLE_ZERO/ONE
  case: just initialize it to 0 (.x) since the value doesn't matter
  (those channels get writemasked out anyway).
- Properly reswizzle source register's swizzles, rather than overwriting
  the swizzle.
- Port the old brw_vs_emit code for computing .x of the EXP2 opcode.
- Update comments, removing mention of NV_vertex_program, etc.
- Delete remaining #warning lines and debug comments.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
11 years agoi965/vs: Refactor min/max handling to share code.
Kenneth Graunke [Mon, 8 Oct 2012 17:45:08 +0000 (10:45 -0700)]
i965/vs: Refactor min/max handling to share code.

v2: Properly use "conditionalmod" pre-Gen6, rather than the incorrectly
copy-and-pasted "BRW_CONDITIONAL_G".

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
11 years agoi965/vs: Add support for emitting DPH opcodes.
Kenneth Graunke [Mon, 8 Oct 2012 17:26:13 +0000 (10:26 -0700)]
i965/vs: Add support for emitting DPH opcodes.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
11 years agoi965/vs: Only do INTEL_DEBUG=perf when there's a GLSL shader.
Kenneth Graunke [Tue, 9 Oct 2012 17:19:16 +0000 (10:19 -0700)]
i965/vs: Only do INTEL_DEBUG=perf when there's a GLSL shader.

This will become necessary once we start supporting ARB programs and
fixed function in this backend.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
11 years agodispatch: stop generating separate GLES1 API code.
Paul Berry [Tue, 23 Oct 2012 20:24:17 +0000 (13:24 -0700)]
dispatch: stop generating separate GLES1 API code.

This patch removes the generated files api_exec_es1.c,
api_exec_es1_dispatch.h, and api_exec_es1_remap_helper.h (and the
source files and build rules used to generate them), since they are no
longer used.  GLES1 now uses the same dispatch table layout as all the
other APIs.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
11 years agodispatch: stop using _mesa_create_exec_table_es1() for GLES1.
Paul Berry [Tue, 23 Oct 2012 21:48:39 +0000 (14:48 -0700)]
dispatch: stop using _mesa_create_exec_table_es1() for GLES1.

This patch modifies context creation code for GLES1 to use
_mesa_create_exec_table() (which is used for all other APIs) instead
of the GLES1-specific _mesa_create_exec_table_es1().

There is a slight change in functionality.  As a result of a mistake
in the code generation of _mesa_create_exec_table_es1(), it does not
include glFlushMappedBufferRangeEXT or glMapBufferRangeEXT (this is
because when support for those two functions was added in commit
762d9ac, src/mesa/main/APIspec.xml wasn't updated).  With this patch,
glFlushMappedBufferRangeEXT and glMapBufferRangeEXT are properly
included in the dispatch table.  Accordingly, dispatch_sanity.cpp is
modified to expect these two functions to be present.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
v2: Leave GLES1.1 dispatch sanity test disabled when not building
GLES1 support.

11 years agodispatch: GLES1 fixes for _mesa_create_exec_table().
Paul Berry [Fri, 19 Oct 2012 13:31:49 +0000 (06:31 -0700)]
dispatch: GLES1 fixes for _mesa_create_exec_table().

Currently, _mesa_create_exec_table() (in api_exec.c) is used for all
APIs except GLES1.  In GLES1, _mesa_create_exec_table_es1() (a code
generated function) is used instead.

In principle, this shouldn't be necessary.  It should be possible for
api_exec.c to contain the logic for populating the dispatch table for
all API's.

This patch paves the way for using _mesa_create_exec_table() instead
of _mesa_create_exec_table_es1(), by making _mesa_create_exec_table()
(and the functions it calls) expose the correct subset of desktop GL
functions for GLES1.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
11 years agodispatch: Make a header to go along with querymatrix.c.
Paul Berry [Tue, 23 Oct 2012 21:31:27 +0000 (14:31 -0700)]
dispatch: Make a header to go along with querymatrix.c.

This patch creates a header querymatrix.h, to allow functions defined
in querymatrix.c to be used from other .c files.  It also switches
from the nonstandard GL_APIENTRY to GLAPIENTRY.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
v2: Don't declare _mesa_Get{Integer,Float}v in querymatrix.c.
Instead, just include main/get.h.

Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
11 years agodispatch: Add standard boilerplate and GL_APIENTRY to es1_conversion.h.
Paul Berry [Tue, 23 Oct 2012 21:23:36 +0000 (14:23 -0700)]
dispatch: Add standard boilerplate and GL_APIENTRY to es1_conversion.h.

This patch adds the usual boilerplate (copyright notice and guards
against redundant inclusion) to es1_conversion.h.  It also moves the
definition of GL_APIENTRY from es1_conversion.c.

This allows es1_conversion.h to be safely included from other .c files.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
v2: Use copyright notice from src/mesa/main/es_generator.py (the
script that used to generate this file).

11 years agodispatch: Include GLES1-only functions in dispatch table.
Paul Berry [Tue, 23 Oct 2012 20:46:04 +0000 (13:46 -0700)]
dispatch: Include GLES1-only functions in dispatch table.

Previously dispatch table-related code was generated from gl_API.xml,
so it did not include slots for GLES1-only functions (such as those
taking fixed-point arguments).

This patch generates dispatch table-related code from
gl_and_es_API.xml, so that GLES1-only functions are included.  This
paves the way for future patches that will unify the GLES1 dispatch
table with the dispatch tables for the other APIs.

The following generated files are affected:
- glapi_x86.S
- glapi_x86-64.S
- glapi_sparc.S
- glprocs.h
- glapitemp.h
- glapitable.h
- glapi_gentable.c
- dispatch.h
- remap_helper.h

Since this change affects makefiles, a full rebuild is required.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
v2: Adjust dependencies to ensure that generated files will be rebuilt
whenever any ES-related XML source files are changed.

Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
11 years agodispatch: properly handle parameter name mismatches in glapitemp.h.
Paul Berry [Tue, 30 Oct 2012 15:45:29 +0000 (08:45 -0700)]
dispatch: properly handle parameter name mismatches in glapitemp.h.

Previously, when code-generating aliased functions in glapitemp.h, we
weren't consistent about which function alias we used to obtain the
parameter names, with the risk that we would generate incorrect code
like this:

  KEYWORD1 void KEYWORD2 NAME(Foo)(GLint x)
  {
    (void) x;
    DISPATCH(Foo, (x), (F, "glFoo(%d);\n", x));
  }
  KEYWORD1 void KEYWORD2 NAME(FooEXT)(GLint y)
  {
    (void) x;
    DISPATCH(Foo, (x), (F, "glFooEXT(%d);\n", x));
  }

At the moment there are no aliased functions with mismatched parameter
names, so this isn't the problem.  But when we introduce GLES1
functions into the dispatch table, there will be
(MapBufferRange/MapBufferRangeEXT).  This patch paves the way for that
by fixing the code generation script to handle the mismatch correctly.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
11 years agodispatch: Include glheader.h in dispatch-related files.
Paul Berry [Tue, 23 Oct 2012 21:06:32 +0000 (14:06 -0700)]
dispatch: Include glheader.h in dispatch-related files.

This ensures that GLES1-only typedefs are available in these files.
In a future patch, this will allow us to expand the dispatch table to
include GLES1-only functions.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
11 years agodispatch: Update check_table.cpp to reflect recent aliasing changes.
Paul Berry [Mon, 29 Oct 2012 18:45:10 +0000 (11:45 -0700)]
dispatch: Update check_table.cpp to reflect recent aliasing changes.

In commits bad96f6 and e7dd2e5 I added the following aliases:
- ClampColor -> ClampColorARB
- VertexAttribDivisor -> VertexAttribDivisorARB

But I neglected to update check_table.cpp, causing "make check" to
fail for non-shared-glapi builds.

This patch removes the functions that are now aliased from
check_table.cpp, so that "make check" works correctly again.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
11 years agoi965/gen4: Fix assertion failures in depthstencil piglit tests.
Eric Anholt [Wed, 31 Oct 2012 19:58:49 +0000 (12:58 -0700)]
i965/gen4: Fix assertion failures in depthstencil piglit tests.

Don't forget to set depth_mt even if !hiz_mt.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
11 years agoi965: Add "alpha to coverage" to performance debug recompile messages.
Kenneth Graunke [Sat, 27 Oct 2012 07:05:31 +0000 (00:05 -0700)]
i965: Add "alpha to coverage" to performance debug recompile messages.

This was missing and got labeled "Something else".

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
11 years agoi965: Don't replicate data for zero-stride arrays when copying to VBOs.
Kenneth Graunke [Fri, 26 Oct 2012 22:48:00 +0000 (15:48 -0700)]
i965: Don't replicate data for zero-stride arrays when copying to VBOs.

When copy_array_to_vbo_array encountered an array with src_stride == 0
and dst_stride != 0, we would replicate out the single element to the
whole size (max - min + 1).  This is unnecessary: we can simply upload
one copy and set the buffer's stride to 0.

Decreases vertex upload overhead in an upcoming Steam for Linux title.
Prior to this patch, copy_array_to_vbo_array appeared very high in the
profile (Eric quoted 20%).  After the patch, it disappeared completely.

Reviewed-by: Eric Anholt <eric@anholt.net>
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
11 years agoi965: Don't bother trying to extend the current vertex buffers.
Kenneth Graunke [Sat, 27 Oct 2012 03:36:50 +0000 (20:36 -0700)]
i965: Don't bother trying to extend the current vertex buffers.

This essentially reverts the following:

  commit c625aa19cb53ed27f91bfd16fea6ea727e9a5bbd
  Author: Chris Wilson <chris@chris-wilson.co.uk>
  Date:   Fri Feb 18 10:37:43 2011 +0000

      intel: extend current vertex buffers

While working on optimizing an upcoming Steam title, I broke this code.
Eric expressed his doubts about this optimization, and noted that the
original commit offered no performance data.

I ran before and after benchmarks on Xonotic and Citybench, and found
that this code made no difference.  So, remove it to reduce complexity
and make future work simpler.

Reviewed-by: Eric Anholt <eric@anholt.net>
11 years agor600g: re-enable handling of DISCARD_RANGE, improving performance
Marek Olšák [Thu, 1 Nov 2012 01:00:37 +0000 (02:00 +0100)]
r600g: re-enable handling of DISCARD_RANGE, improving performance

It seems to work for me now. Even the graphics corruption is gone.

This also boosts performance in Reaction Quake.

11 years agor600g: fix abysmal performance in Reaction Quake
Marek Olšák [Wed, 31 Oct 2012 23:52:19 +0000 (00:52 +0100)]
r600g: fix abysmal performance in Reaction Quake

The problem was we set VRAM|GTT for relocations of STATIC resources.
Setting just VRAM increases the framerate 4 times on my machine.

I rewrote the switch statement and adjusted the domains for window
framebuffers too.

NOTE: This is a candidate for the stable branches.

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
Reviewed-by: Jerome Glisse <jglisse@redhat.com>
11 years agogallium/u_vbuf: document how it works
Marek Olšák [Sun, 28 Oct 2012 20:31:12 +0000 (21:31 +0100)]
gallium/u_vbuf: document how it works

11 years agogallium/u_vbuf: optimize looping over the list of buffers to upload
Marek Olšák [Sun, 28 Oct 2012 19:11:32 +0000 (20:11 +0100)]
gallium/u_vbuf: optimize looping over the list of buffers to upload

11 years agogallium/u_vbuf: skip processing of buffers unused by the vertex element state
Marek Olšák [Sun, 28 Oct 2012 18:52:12 +0000 (19:52 +0100)]
gallium/u_vbuf: skip processing of buffers unused by the vertex element state

11 years agoswrast: remove explicit size from texfetch_funcs array
Brian Paul [Tue, 30 Oct 2012 15:39:20 +0000 (09:39 -0600)]
swrast: remove explicit size from texfetch_funcs array

By removing the array size, the static assertion to check for missing
elements can do its job properly.  This will catch cases where a new
Mesa format is added but the swrast texfetch code isn't updated.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
11 years agollvmpipe: Obey back writemask.
José Fonseca [Tue, 30 Oct 2012 19:45:25 +0000 (19:45 +0000)]
llvmpipe: Obey back writemask.

Tested with a modified glean tstencil2 test.

NOTE: This is a candidate for stable branches.

Reviewed-by: Brian Paul <brianp@vmware.com>
11 years agor600g: avoid shader needing too many gpr to lockup the gpu v2
Jerome Glisse [Fri, 26 Oct 2012 22:59:05 +0000 (18:59 -0400)]
r600g: avoid shader needing too many gpr to lockup the gpu v2

On r6xx/r7xx shader resource management need to make sure that the
shader does not goes over the gpr register limit. Each specific
asic has a maxmimum register that can be split btw shader stage.
For each stage the shader must not use more register than the
limit programmed.

v2: Print an error message when discarding draw. Don't add another
    boolean to context structure, but rather propagate the discard
    boolean through the call chain.

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
11 years agodraw: fix assertion failure in draw_emit_vertex_attr
Marek Olšák [Tue, 30 Oct 2012 00:07:17 +0000 (01:07 +0100)]
draw: fix assertion failure in draw_emit_vertex_attr

This is a regression since b3921e1f53833420e0a0fd581f7417.

The array stores VS outputs, not FS inputs.
Now llvmpipe can do 32 varyings too.

NOTE: This is a candidate for the stable branches.

Reviewed-by: Brian Paul <brianp@vmware.com>
11 years agor600g: use SQ_VTX_SEMANTIC_CLEAR to clear the semantic registers
Marek Olšák [Mon, 29 Oct 2012 12:18:03 +0000 (13:18 +0100)]
r600g: use SQ_VTX_SEMANTIC_CLEAR to clear the semantic registers

Reviewed-by: Alex Deucher <alexander.deucher@amd.com>
11 years agomesa: remove NV_read_buffer extension enable flag
Marek Olšák [Sun, 28 Oct 2012 14:44:53 +0000 (15:44 +0100)]
mesa: remove NV_read_buffer extension enable flag

It's been enabled by default, so the flag isn't really useful.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
11 years agomesa: remove SGIS_texture_lod extension enable flag
Marek Olšák [Sun, 28 Oct 2012 13:24:53 +0000 (14:24 +0100)]
mesa: remove SGIS_texture_lod extension enable flag

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
11 years agomesa: remove NV_texgen_reflection extension enable flag
Marek Olšák [Sun, 28 Oct 2012 13:24:53 +0000 (14:24 +0100)]
mesa: remove NV_texgen_reflection extension enable flag

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
11 years agomesa: remove NV_light_max_exponent extension enable flag
Marek Olšák [Sun, 28 Oct 2012 13:24:53 +0000 (14:24 +0100)]
mesa: remove NV_light_max_exponent extension enable flag

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
11 years agomesa: remove IBM_rasterpos_clip extension enable flag
Marek Olšák [Sun, 28 Oct 2012 13:24:53 +0000 (14:24 +0100)]
mesa: remove IBM_rasterpos_clip extension enable flag

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
11 years agomesa: remove IBM_multimode_draw_arrays extension enable flag
Marek Olšák [Sun, 28 Oct 2012 13:24:53 +0000 (14:24 +0100)]
mesa: remove IBM_multimode_draw_arrays extension enable flag

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
11 years agomesa: remove APPLE_packed_pixels extension enable flag
Marek Olšák [Sun, 28 Oct 2012 13:24:53 +0000 (14:24 +0100)]
mesa: remove APPLE_packed_pixels extension enable flag

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
11 years agomesa: don't always enable OES_standard_derivatives
Marek Olšák [Sun, 28 Oct 2012 14:01:52 +0000 (15:01 +0100)]
mesa: don't always enable OES_standard_derivatives

For Intel, expose it only if gen >= 4.
For Gallium, expose it only if PIPE_CAP_SM3 is advertised.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
11 years agomesa: move EXT_texture3D enabling to _mesa_init_extensions
Marek Olšák [Wed, 31 Oct 2012 01:04:16 +0000 (02:04 +0100)]
mesa: move EXT_texture3D enabling to _mesa_init_extensions

11 years agomesa: remove EXT_separate_specular_color extension enable flag
Marek Olšák [Sun, 28 Oct 2012 13:24:53 +0000 (14:24 +0100)]
mesa: remove EXT_separate_specular_color extension enable flag

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
11 years agomesa: remove EXT_rescale_normal extension enable flag
Marek Olšák [Sun, 28 Oct 2012 13:24:53 +0000 (14:24 +0100)]
mesa: remove EXT_rescale_normal extension enable flag

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
11 years agomesa: remove EXT_packed_pixels extension enable flag
Marek Olšák [Sun, 28 Oct 2012 13:24:53 +0000 (14:24 +0100)]
mesa: remove EXT_packed_pixels extension enable flag

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
11 years agomesa: remove EXT_draw_range_elements extension enable flag
Marek Olšák [Sun, 28 Oct 2012 13:24:53 +0000 (14:24 +0100)]
mesa: remove EXT_draw_range_elements extension enable flag

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
11 years agomesa: remove EXT_compiled_vertex_array extension enable flag
Marek Olšák [Sun, 28 Oct 2012 13:24:53 +0000 (14:24 +0100)]
mesa: remove EXT_compiled_vertex_array extension enable flag

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
11 years agomesa: remove ARB_window_pos extension enable flag
Marek Olšák [Sun, 28 Oct 2012 13:24:53 +0000 (14:24 +0100)]
mesa: remove ARB_window_pos extension enable flag

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
11 years agomesa: remove ARB_transpose_matrix extension enable flag
Marek Olšák [Sun, 28 Oct 2012 13:24:53 +0000 (14:24 +0100)]
mesa: remove ARB_transpose_matrix extension enable flag

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
11 years agomesa: remove ARB_copy_buffer extension enable flag
Marek Olšák [Sun, 28 Oct 2012 13:24:53 +0000 (14:24 +0100)]
mesa: remove ARB_copy_buffer extension enable flag

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
11 years agogallium: expose ARB_map_buffer_alignment on Radeon
Marek Olšák [Sun, 28 Oct 2012 16:52:48 +0000 (17:52 +0100)]
gallium: expose ARB_map_buffer_alignment on Radeon

Reviewed-by: Brian Paul <brianp@vmware.com>
v2: update relnotes-9.1
v3: use align_malloc and align_free for malloced buffers in r300g
v4: document the new CAP in the docs

11 years agomesa: implement ARB_map_buffer_alignment
Marek Olšák [Sun, 28 Oct 2012 16:51:47 +0000 (17:51 +0100)]
mesa: implement ARB_map_buffer_alignment

Reviewed-by: Brian Paul <brianp@vmware.com>
11 years agost/mesa: don't use _NEW_PROGRAM where ST_NEW_xxx_PROGRAM is sufficient
Marek Olšák [Mon, 29 Oct 2012 02:09:50 +0000 (03:09 +0100)]
st/mesa: don't use _NEW_PROGRAM where ST_NEW_xxx_PROGRAM is sufficient

Reviewed-by: Brian Paul <brianp@vmware.com>
11 years agor600g: use better sample positions for 8x MSAA
Marek Olšák [Sat, 27 Oct 2012 14:31:19 +0000 (16:31 +0200)]
r600g: use better sample positions for 8x MSAA

Taken from the intel driver. The sample positions are actually a solution
to the 8 queens puzzle.  It gives more accurate and smoother AA.