profile/ivi/mesa.git
12 years agoi965: Remove BRW_NEW_NR_VS_SURFACES dirty bit from brw_vs_surfaces.
Kenneth Graunke [Tue, 8 Nov 2011 19:32:52 +0000 (11:32 -0800)]
i965: Remove BRW_NEW_NR_VS_SURFACES dirty bit from brw_vs_surfaces.

brw_vs_surfaces _produces_ the BRW_NEW_NR_VS_SURFACES dirty bit, so it
makes no sense for it to subscribe to it.

Fixes an assertion failure in many piglit tests when INTEL_DEBUG is set:
  brw_state_upload.c:484: void brw_upload_state(struct brw_context *):
  Assertion `!check_state(&examined, &generated)' failed.

One such piglit test is vs-uniform-array-mat2-col-rd.shader_test.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
12 years agoi965: Add missing CACHE_NEW_WM_PROG dirty bit to WM pull constants.
Kenneth Graunke [Tue, 8 Nov 2011 19:21:00 +0000 (11:21 -0800)]
i965: Add missing CACHE_NEW_WM_PROG dirty bit to WM pull constants.

Comparing brw_upload_vs_pull_constants and brw_upload_wm_pull_constants,
it became evident that something was amiss: the VS code had both
CACHE_NEW_VS_PROG and BRW_NEW_VERTEX_PROGRAM, while the WM code was
missing the CACHE_NEW_WM_PROG flag.

Not observed to fix anything, but likely necessary.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
12 years agoi965: Use new vtable entries for surface state updating functions.
Kenneth Graunke [Tue, 1 Nov 2011 22:41:48 +0000 (15:41 -0700)]
i965: Use new vtable entries for surface state updating functions.

Now that we have vtable entries in place, we should use them.  This
allows us to drop the cut and pasted Gen7 brw_tracked_state atoms as
they now do exactly the same thing as their brw_wm_surface_state
counterparts.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
12 years agoi965: Add new vtable entries for surface state updating functions.
Kenneth Graunke [Tue, 1 Nov 2011 21:30:26 +0000 (14:30 -0700)]
i965: Add new vtable entries for surface state updating functions.

Gen7+ SURFACE_STATE is different from Gen4-6, so we need separate
per-generation functions for creating and updating it.  However, the
usage is the same, and callers just want to utilize the appropriate
functions with minimal pain.  So, put them in the vtable.

Since these take a brw_context pointer and are only used on Gen4, just
add a forward declaration.  This is the simplest (if not cleanest)
solution.  It would be nicer to have a i965-specific vtable, but that's
a refactor for another day.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
12 years agor600g: don't set the query result to 0 after a call to get_query_result
Marek Olšák [Thu, 10 Nov 2011 22:55:49 +0000 (23:55 +0100)]
r600g: don't set the query result to 0 after a call to get_query_result

get_query_result doesn't reset the result. Only begin_query does.

12 years agor600g: the type of OCCLUSION_PREDICATE result should be boolean
Marek Olšák [Thu, 10 Nov 2011 22:38:31 +0000 (23:38 +0100)]
r600g: the type of OCCLUSION_PREDICATE result should be boolean

12 years agomesa: silence uninitialized var warning
Brian Paul [Thu, 10 Nov 2011 22:56:04 +0000 (15:56 -0700)]
mesa: silence uninitialized var warning

12 years agoradeon: silence initializer warnings
Brian Paul [Thu, 10 Nov 2011 22:55:46 +0000 (15:55 -0700)]
radeon: silence initializer warnings

12 years agodraw/llvm: silence uninitialized variable warnings
Brian Paul [Thu, 10 Nov 2011 22:55:13 +0000 (15:55 -0700)]
draw/llvm: silence uninitialized variable warnings

12 years agoradeon: silence unused var warnings
Brian Paul [Thu, 10 Nov 2011 22:54:34 +0000 (15:54 -0700)]
radeon: silence unused var warnings

12 years agoegl: silence unused var warning
Brian Paul [Thu, 10 Nov 2011 22:54:15 +0000 (15:54 -0700)]
egl: silence unused var warning

12 years agomklib: fix static lib building by filtering out -L, -l options
Wayne E. Robertz [Thu, 10 Nov 2011 22:21:14 +0000 (15:21 -0700)]
mklib: fix static lib building by filtering out -L, -l options

Signed-off-by: Brian Paul <brianp@vmware.com>
12 years agor600g: implement PIPE_QUERY_OCCLUSION_PREDICATE
Marek Olšák [Thu, 10 Nov 2011 22:11:57 +0000 (23:11 +0100)]
r600g: implement PIPE_QUERY_OCCLUSION_PREDICATE

12 years agor300g: fix query result of GPU_FINISHED
Marek Olšák [Thu, 10 Nov 2011 21:58:34 +0000 (22:58 +0100)]
r300g: fix query result of GPU_FINISHED

12 years agor300g: implement PIPE_QUERY_GPU_FINISHED
Marek Olšák [Thu, 10 Nov 2011 21:25:21 +0000 (22:25 +0100)]
r300g: implement PIPE_QUERY_GPU_FINISHED

12 years agor300g: implement PIPE_QUERY_OCCLUSION_PREDICATE
Marek Olšák [Thu, 10 Nov 2011 20:18:10 +0000 (21:18 +0100)]
r300g: implement PIPE_QUERY_OCCLUSION_PREDICATE

12 years agollvmpipe/u_format: add support for EXT_texture_shared_exponent + EXT_packed_float
Dave Airlie [Mon, 7 Nov 2011 19:22:01 +0000 (19:22 +0000)]
llvmpipe/u_format: add support for EXT_texture_shared_exponent + EXT_packed_float

These two are fairly unique types so add specific cases for decoding them.

Passes piglit fbo-clear-format and fbo-generatemipmap-format tests for these
two extensions.

Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agollvmpipe: add NV_conditional_render support.
Dave Airlie [Mon, 7 Nov 2011 17:31:49 +0000 (17:31 +0000)]
llvmpipe: add NV_conditional_render support.

This ports the softpipe NV_conditional_render support to llvmpipe.

This passes the nv_conditional_render-* piglit tests.

Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agor600g: remove useless texture barrier
Marek Olšák [Thu, 10 Nov 2011 17:47:20 +0000 (18:47 +0100)]
r600g: remove useless texture barrier

12 years agor600g: flush the whole context when CS is full, not just hw_context
Marek Olšák [Thu, 10 Nov 2011 17:42:10 +0000 (18:42 +0100)]
r600g: flush the whole context when CS is full, not just hw_context

Because we disable render condition in r600_flush, but not in r600_context_flush.

12 years agor600g: turn some CS overflow checks into assertions
Marek Olšák [Thu, 10 Nov 2011 16:52:36 +0000 (17:52 +0100)]
r600g: turn some CS overflow checks into assertions

They're not really exhaustive and not so useful either.

12 years agor600g: put the rest of CS overflow checks in r600_need_cs_space
Marek Olšák [Thu, 10 Nov 2011 16:41:44 +0000 (17:41 +0100)]
r600g: put the rest of CS overflow checks in r600_need_cs_space

12 years agor600g: flush caches regardless of render condition
Marek Olšák [Thu, 10 Nov 2011 14:54:17 +0000 (15:54 +0100)]
r600g: flush caches regardless of render condition

What if somebody enables render condition just before we flush...

12 years agor600g: disable render condition at the end of CS, re-enable at the beginning
Marek Olšák [Thu, 10 Nov 2011 14:50:06 +0000 (15:50 +0100)]
r600g: disable render condition at the end of CS, re-enable at the beginning

12 years agor600g: don't suspend queries if they have already been suspended by blitter
Marek Olšák [Thu, 10 Nov 2011 14:22:34 +0000 (15:22 +0100)]
r600g: don't suspend queries if they have already been suspended by blitter

And add some assertions.

12 years agor600g: properly reserve CS space for queries_suspend
Marek Olšák [Thu, 10 Nov 2011 14:00:11 +0000 (15:00 +0100)]
r600g: properly reserve CS space for queries_suspend

12 years agor600g: reserve CS space for a draw command in begin_query and render_condition
Marek Olšák [Thu, 10 Nov 2011 12:35:22 +0000 (13:35 +0100)]
r600g: reserve CS space for a draw command in begin_query and render_condition

There's no point in emitting those if you can't emit a draw command too.

12 years agor600g: consolidate checking whether CS is full
Marek Olšák [Wed, 9 Nov 2011 00:42:13 +0000 (01:42 +0100)]
r600g: consolidate checking whether CS is full

This adds a new function r600_need_cs_space. Currently, it's easy to overflow
the CS - queries are not counted in. I guess that's not the only case where
the driver may crap out.

12 years agoradeon: fix tex1d-border2d
Dave Airlie [Thu, 10 Nov 2011 13:47:12 +0000 (13:47 +0000)]
radeon: fix tex1d-border2d

Fixes pigit test with T wrap usage.

Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agoradeon: fix 3-coordinate swtcl emission
Dave Airlie [Mon, 7 Nov 2011 13:46:36 +0000 (13:46 +0000)]
radeon: fix 3-coordinate swtcl emission

This fixes mipmap generation on swtcl rv100.

Signed-off-by: Dave Airlie <airlied@redhat.com>
12 years agoswrast: Add support for glReadPixels() to integer types.
Eric Anholt [Fri, 4 Nov 2011 00:27:23 +0000 (17:27 -0700)]
swrast: Add support for glReadPixels() to integer types.

With this change, i965 passes
GL_EXT_texture_integer/fbo_integer_precision_clear

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
12 years agomesa: Add support for unpacking 32-bit integer formats to int spans.
Eric Anholt [Fri, 4 Nov 2011 00:08:16 +0000 (17:08 -0700)]
mesa: Add support for unpacking 32-bit integer formats to int spans.

This is the inverse operation to _mesa_pack_rgba_span_int.  The 16-bit
code isn't done because of lack of testing and not being sure how sign
extension/clamping should be handled between, say, 16-bit int and
32-bit int or uint.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
12 years agometa: Add support for glClear() to integer color buffers.
Eric Anholt [Wed, 2 Nov 2011 21:01:25 +0000 (14:01 -0700)]
meta: Add support for glClear() to integer color buffers.

This requires using a new fragment shader to get the integer color
output, and a new vertex shader because #version has to match between
the two.

v2: Clarify that there's no need for BindFragDataLocation.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org> (v1)
12 years agoi965: Claim to support rendering to integer FBOs.
Eric Anholt [Wed, 2 Nov 2011 20:51:38 +0000 (13:51 -0700)]
i965: Claim to support rendering to integer FBOs.

We're missing support for the software paths still, but basic
rendering is working.

v2: Override RGB_INT32/UINT32 to not be renderable, since the hardware
    can't do it but we do allow texturing from it now.  Drop the
    DataType override, since the _mesa_problem() isn't in that path
    any more.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org> (v1)
12 years agoi965/fs: Add support for user-defined out variables.
Eric Anholt [Wed, 9 Nov 2011 03:27:46 +0000 (19:27 -0800)]
i965/fs: Add support for user-defined out variables.

Before, I was tracking the ir_variable * found for gl_FragColor or
gl_FragData[].  Instead, when visiting those variables, set up an
array of per-render-target fs_regs to copy the output data from.  This
cleans up the color emit path, while making handling of multiple
user-defined out variables easier.

v2: incorporate idr's feedback about ir->location (changes by Kenneth Graunke)

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
12 years agoi965/fs: Preserve the source register type when doing color writes.
Eric Anholt [Wed, 9 Nov 2011 03:26:39 +0000 (19:26 -0800)]
i965/fs: Preserve the source register type when doing color writes.

When rendering to integer color buffers, we need to be careful to use
MRFs of the correct type when emitting color writes.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
12 years agoi965: Make brw_type_for_base_type return the element type for arrays.
Eric Anholt [Wed, 9 Nov 2011 03:26:38 +0000 (19:26 -0800)]
i965: Make brw_type_for_base_type return the element type for arrays.

Previously, brw_type_for_base_type returned UD for array variables,
similar to structures.  For structures, each field may have a different
type, so every field access must explicitly override the register's type
with that field's type.  We chose to return UD in this case since it was
the least common, so errors would be more obvious.

For arrays, it makes far more sense to return the type corresponding to
an element of the array.  This allows normal array access to work
without the hassle of explicitly overriding the register's type.

This should obsolete a bunch of type overrides throughout the code.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
12 years agoi965: Enable ChooseTexFormat for supported GL_EXT_texture_integer formats.
Eric Anholt [Wed, 5 Oct 2011 20:42:19 +0000 (13:42 -0700)]
i965: Enable ChooseTexFormat for supported GL_EXT_texture_integer formats.

v2: s/GL_TRUE/true/, and re-enable RGB_INT32 based on discussion
    yesterday about required RB formats vs texture formats.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org> (v1)
12 years agoi965: Add mapping from MESA_FORMAT to BRW_SURFACEFORMAT for integer.
Eric Anholt [Wed, 5 Oct 2011 20:38:03 +0000 (13:38 -0700)]
i965: Add mapping from MESA_FORMAT to BRW_SURFACEFORMAT for integer.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
12 years agointel: Expose GL_EXT_texture_integer when GL 3.0 override is set.
Eric Anholt [Wed, 5 Oct 2011 20:25:36 +0000 (13:25 -0700)]
intel: Expose GL_EXT_texture_integer when GL 3.0 override is set.

This will let the feature be incrementally developed, hidden behind
the flag we're all using as we work on GL 3.0 support.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
12 years agodocs: Note EXT_texture_array on i965.
Eric Anholt [Wed, 5 Oct 2011 20:26:42 +0000 (13:26 -0700)]
docs: Note EXT_texture_array on i965.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
12 years agoi965: Add support for 16-bit unorm L, A, and I textures.
Eric Anholt [Tue, 8 Nov 2011 19:05:17 +0000 (11:05 -0800)]
i965: Add support for 16-bit unorm L, A, and I textures.

While not required by any particular spec version, mplayer was asking
for L16 and hoping for actual L16 without checking.  The 8 bits
allocated led to 10-bit planar video data stored in the lower 10 bits
giving only 2 bits of precision in video.  While it was an amusing
effect, give them what they actually wanted instead.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=41461

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
12 years agointel: Don't _mesa_problem when asked for an RB of a texturing-only type.
Eric Anholt [Tue, 8 Nov 2011 19:19:45 +0000 (11:19 -0800)]
intel: Don't _mesa_problem when asked for an RB of a texturing-only type.

We want to be able to support some formats for texturing that we can't
render to, which means that some choices for RenderbufferStorage end
up being incomplete (for example, L8 currently).  For these, where we
don't render to them, we don't want to have to make up an rb->DataType
that's only used for GetRow()/PutRow().

12 years agor200: remove dangling radeon.h symlink.
Paul Berry [Tue, 8 Nov 2011 15:50:56 +0000 (07:50 -0800)]
r200: remove dangling radeon.h symlink.

Commit 1401b96b (radeon: cleanup radeon shared code after r300 and
r600 classic drivers removal) removed the file
src/mesa/drivers/dri/radeon/server/radeon.h, but it left behind the
symlink which was used to share that file into the
src/mesa/drivers/dri/r200/server directory.

This patch removes the dangling symlink.

Reviewed-by: Alex Deucher <alexdeucher@gmail.com>
12 years agoglsl: Assign transform feedback varying slots in linker.
Paul Berry [Sat, 5 Nov 2011 18:17:32 +0000 (11:17 -0700)]
glsl: Assign transform feedback varying slots in linker.

This patch modifies the GLSL linker to assign additional slots for
varying variables used by transform feedback, and record the varying
slots used by transform feedback for use by the driver back-end.

This required modifying assign_varying_locations() so that it assigns
a varying location if either (a) the varying is used by the next stage
of the GL pipeline, or (b) the varying is required by transform
feedback.  In order to avoid duplicating the code to assign a single
varying location, I moved it into its own function,
assign_varying_location().

In addition, to support transform feedback in the case where there is
no fragment shader, it is now possible to call
assign_varying_locations() with a consumer of NULL.

Reviewed-by: Marek Olšák <maraeo@gmail.com>
Tested-by: Marek Olšák <maraeo@gmail.com>
12 years agoglsl: Clamp vector indices when lowering to swizzles
Ian Romanick [Mon, 7 Nov 2011 18:58:00 +0000 (10:58 -0800)]
glsl: Clamp vector indices when lowering to swizzles

This prevents other code from seeing a swizzle of the 16th component
of a vector, for example.

NOTE: This is a candidate for the 7.11 branch.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=42517
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
Tested-by: Christian Holler <choller@mozilla.com>
12 years agoscons: Disable deprecated POSIX name MSVC warnings.
José Fonseca [Wed, 9 Nov 2011 10:33:21 +0000 (10:33 +0000)]
scons: Disable deprecated POSIX name MSVC warnings.

12 years agomesa,glsl,mapi: Put extern "C" { ... } where appropriate.
José Fonseca [Wed, 9 Nov 2011 10:20:51 +0000 (10:20 +0000)]
mesa,glsl,mapi: Put extern "C" { ... } where appropriate.

Probably a several places missing, but enough to cover all headers
(in)directly included by uniform_query.cpp, and fix the MSVC build.

12 years agoscons: Don't list m_xform.c twice.
José Fonseca [Wed, 9 Nov 2011 10:17:02 +0000 (10:17 +0000)]
scons: Don't list m_xform.c twice.

12 years agoglu: Fix deprecated conversion from string constant to ‘char*’ warning.
José Fonseca [Wed, 9 Nov 2011 09:08:58 +0000 (09:08 +0000)]
glu: Fix deprecated conversion from string constant to ‘char*’ warning.

12 years agor600g: clarify meaning of one variable in shader codegen
Marek Olšák [Tue, 8 Nov 2011 22:19:03 +0000 (23:19 +0100)]
r600g: clarify meaning of one variable in shader codegen

It's easier to read now.

12 years agor600g: cosmetic changes in query code
Marek Olšák [Tue, 8 Nov 2011 22:09:33 +0000 (23:09 +0100)]
r600g: cosmetic changes in query code

Mainly updating comments and removing one use of a magic number.

12 years agor600g: use modulo for computing index into query (ring) buffers
Marek Olšák [Tue, 8 Nov 2011 22:05:36 +0000 (23:05 +0100)]
r600g: use modulo for computing index into query (ring) buffers

12 years agor600g: make r600_query_result more generic
Marek Olšák [Tue, 8 Nov 2011 21:02:03 +0000 (22:02 +0100)]
r600g: make r600_query_result more generic

We'll soon start adding new query types, maybe even querying more than
one value per query.

12 years agoRemove tgsi_sse2.
José Fonseca [Tue, 8 Nov 2011 00:10:47 +0000 (00:10 +0000)]
Remove tgsi_sse2.

tgsi_exec is simple. llvm is fast. tgsi_sse2 ends up being neither.

12 years agomesa: fix signed/unsigned integer comparison warnings
Brian Paul [Tue, 8 Nov 2011 15:23:41 +0000 (08:23 -0700)]
mesa: fix signed/unsigned integer comparison warnings

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
12 years agoglsl: remove trailing comma to silence warning
Brian Paul [Tue, 8 Nov 2011 15:23:16 +0000 (08:23 -0700)]
glsl: remove trailing comma to silence warning

12 years agomesa: Implement glGetFragDataLocation
Ian Romanick [Fri, 4 Nov 2011 23:32:02 +0000 (16:32 -0700)]
mesa: Implement glGetFragDataLocation

Fixes piglit's getfragdatalocation test.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
12 years agolinker: Use app-specified fragment data location during linking
Ian Romanick [Fri, 4 Nov 2011 23:08:52 +0000 (16:08 -0700)]
linker: Use app-specified fragment data location during linking

Fixes piglit's bindfragdata-link-error.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
12 years agomesa: Stub implementation of glBindFragDataLocation
Ian Romanick [Fri, 4 Nov 2011 22:48:41 +0000 (15:48 -0700)]
mesa: Stub implementation of glBindFragDataLocation

This just validates the input parameters so far.

Fixes piglit's bindfragdata-invalid-parameters test.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
12 years agosoftpipe: don't clamp or do logical operations on floating-point buffers.
Morgan Armand [Mon, 7 Nov 2011 19:43:28 +0000 (20:43 +0100)]
softpipe: don't clamp or do logical operations on floating-point buffers.

Signed-off-by: Brian Paul <brianp@vmware.com>
12 years agost/mesa: Fix memory leak on error path.
Vinson Lee [Sun, 6 Nov 2011 05:36:08 +0000 (22:36 -0700)]
st/mesa: Fix memory leak on error path.

Fixes Coverity resource leak defect.

Reviewed-by: José Fonseca <jfonseca@vmware.com>
12 years agost/dri: Fix memory leak on error path.
Vinson Lee [Sun, 6 Nov 2011 05:19:31 +0000 (22:19 -0700)]
st/dri: Fix memory leak on error path.

Fixes Coverity resource leak defect.

Reviewed-by: José Fonseca <jfonseca@vmware.com>
12 years agoglsl: Generate IR for switch statements
Dan McCabe [Tue, 8 Nov 2011 00:17:58 +0000 (16:17 -0800)]
glsl: Generate IR for switch statements

Up until now modifying the GLSL compiler has been pretty straightforward.
This is where things get interesting. But still pretty straightforward.

Switch statements can be thought of a series of if/then/else statements.
Case labels are compared with the value of a test expression and the case
statements are executed if the comparison is true.

There are a couple of aspects of switch statements that complicate this simple
view of the world. The primary one is that cases can fall through sequentially
to subsequent case, unless a break statement is encountered, in which case,
the switch statement exits completely.

But break handling is further complicated by the fact that a break statement
can impact the exit of a loop. Thus, we need to coordinate break processing
between switch statements and loop statements.

The code generated by a switch statement maintains three temporary state
variables:
    int test_value;
    bool is_fallthru;
    bool is_break;

test_value is initialized to the value of the test expression at the head of
the switch statement. This is the value that case labels are compared against.

is_fallthru is used to sequentially fall through to subsequent cases and is
initialized to false. When a case label matches the test expression, this
state variable is set to true. It will also be forced to false if a break
statement has been encountered. This forcing to false on break MUST be
after every case test. In practice, we defer that forcing to immediately after
the last case comparison prior to executing a case statement, but that is
an optimization.

is_break is used to indicate that a break statement has been executed and is
initialized to false. When a break statement is encountered, it is set to true.
This state variable is then used to conditionally force is_fallthru to to false
to prevent subsequent case statements from executing.

Code generation for break statements depends on whether the break statement is
inside a switch statement or inside a loop statement. If it inside a loop
statement is inside a break statement, the same code as before gets generated.
But if a switch statement is inside a loop statement, code is emitted to set
the is_break state to true.

Just as ASTs for loop statements are managed in a stack-like
manner to handle nesting, we also add a bool to capture the innermost switch
or loop condition. Note that we still need to maintain a loop AST stack to
properly handle for-loop code generation on a continue statement. Technically,
we don't (yet) need a switch AST stack, but I am using one for orthogonality
with loop statements, in anticipation of future use. Note that a simple
boolean stack would have sufficed.

We will illustrate a switch statement with its analogous conditional code that
a switch statement corresponds to by examining an example.

Consider the following switch statement:
switch (42) {
case 0:
case 1:
gl_FragColor = vec4(1.0, 2.0, 3.0, 4.0);
case 2:
case 3:
gl_FragColor = vec4(4.0, 3.0, 2.0, 1.0);
break;
case 4:
default:
gl_FragColor = vec4(0.0, 0.0, 0.0, 0.0);
}

Note that case 0 and case 1 fall through to cases 2 and 3 if they occur.

Note that case 4 and the default case must be reached explicitly, since cases
2 and 3 break at the end of their case.

Finally, note that case 4 and the default case don't break but simply fall
through to the end of the switch.

For this code, the equivalent code can be expressed as:
int test_val = 42; // capture value of test expression
bool is_fallthru = false; // prevent initial fall through
bool is_break = false; // capture the execution of a break stmt

is_fallthru |= (test_val == 0); // enable fallthru on case 0
is_fallthru |= (test_val == 1); // enable fallthru on case 1
is_fallthru &= !is_break; // inhibit fallthru on previous break
if (is_fallthru) {
gl_FragColor = vec4(1.0, 2.0, 3.0, 4.0);
}

is_fallthru |= (test_val == 2); // enable fallthru on case 2
is_fallthru |= (test_val == 3); // enable fallthru on case 3
is_fallthru &= !is_break; // inhibit fallthru on previous break
if (is_fallthru) {
gl_FragColor = vec4(4.0, 3.0, 2.0, 1.0);
is_break = true; // inhibit all subsequent fallthru for break
}

is_fallthru |= (test_val == 4); // enable fallthru on case 4
is_fallthru = true; // enable fallthru for default case
is_fallthru &= !is_break; // inhibit fallthru on previous break
if (is_fallthru) {
gl_FragColor = vec4(0.0, 0.0, 0.0, 0.0);
}

The code generate for |= and &= uses the conditional assignment capabilities
of the IR.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
12 years agoglsl: Reference data structure ctors in grammar
Dan McCabe [Mon, 7 Nov 2011 23:11:04 +0000 (15:11 -0800)]
glsl: Reference data structure ctors in grammar

We now tie the grammar to the ctors of the ASTs they reference.

This requires that we actually have definitions of the ctors.

In addition, we also need to define "print" and "hir" methods for the AST
classes. The Print methods are pretty simple to flesh out. However, at this
stage of the development, we simply stub out the "hir" methods and flesh
them out later.

Also, since actual class instances get returned by the productions in the
grammar, we also need to designate the type of the productions that
reference those instances.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
12 years agoglsl: Create AST structs corresponding to new productions in grammar
Dan McCabe [Mon, 7 Nov 2011 23:09:40 +0000 (15:09 -0800)]
glsl: Create AST structs corresponding to new productions in grammar

Previously we added productions for:
switch_body
case_label_list
case_statement
case_statement_list
Now add AST structs corresponding to those productions.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
12 years agoglsl: Add productions to GLSL grammar for switch statement
Dan McCabe [Mon, 7 Nov 2011 23:07:43 +0000 (15:07 -0800)]
glsl: Add productions to GLSL grammar for switch statement

The grammar is modified to support switch statements. Rather than follow the
grammar in the appendix, which allows case labels to be placed ANYWHERE
as a regular statement, we follow the development of the grammar as
described in the body of the GLSL spec.

In this variation, the switch statement has a body which consists of a list
of case statements. A case statement is preceded by a list of case labels and
ends with a list of statements.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
12 years agoglsl: Create AST data structures for switch statement and case label
Dan McCabe [Mon, 7 Nov 2011 23:05:16 +0000 (15:05 -0800)]
glsl: Create AST data structures for switch statement and case label

Data structures for switch statement and case label are created that parallel
the structure of other AST data.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
12 years agoutil: add log2f for Android
Chia-I Wu [Sun, 6 Nov 2011 21:48:11 +0000 (14:48 -0700)]
util: add log2f for Android

It is needed for nv50's new shader backend.  With this change, both u_math.h
and imports.h in core mesa define the same function.  I have to #undef log2f
here to avoid the conflict.  Not sure if there is a better way to deal with
the situation.

Acked-by: José Fonseca <jfonseca@vmware.com>
12 years agomesa: remove prog_uniform.c from SConscript
Brian Paul [Mon, 7 Nov 2011 23:35:38 +0000 (16:35 -0700)]
mesa: remove prog_uniform.c from SConscript

12 years agoDelete code made dead by previous uniform related patches
Ian Romanick [Tue, 18 Oct 2011 23:50:10 +0000 (16:50 -0700)]
Delete code made dead by previous uniform related patches

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agomesa: Add missing check for glUniform*v count > 1 on non-array
Ian Romanick [Tue, 18 Oct 2011 23:28:21 +0000 (16:28 -0700)]
mesa: Add missing check for glUniform*v count > 1 on non-array

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agomesa: Rewrite the way uniforms are tracked and handled
Ian Romanick [Tue, 18 Oct 2011 23:01:49 +0000 (16:01 -0700)]
mesa: Rewrite the way uniforms are tracked and handled

Switch all of the code in ir_to_mesa, st_glsl_to_tgsi, glUniform*,
glGetUniform, glGetUniformLocation, and glGetActiveUniforms to use the
gl_uniform_storage structures in the gl_shader_program.

A couple of notes:

 * Like most rewrite-the-world patches, this should be reviewed by
   applying the patch and examining the modified functions.

 * This leaves a lot of dead code around in linker.cpp and
   uniform_query.cpp.  This will be deleted in the next patches.

v2: Update the comment block (previously a FINISHME) in _mesa_uniform
about generating GL_INVALID_VALUE when an out-of-range sampler index
is specified.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agoi965: Move _mesa_ir_link_shader call before device-specific linking
Ian Romanick [Mon, 24 Oct 2011 23:37:01 +0000 (16:37 -0700)]
i965: Move _mesa_ir_link_shader call before device-specific linking

_mesa_ir_link_shader needs to be called before cloning the IR tree so
that the var->location field for uniforms is set.

WARNING: This change breaks several integer division related piglit
tests.  The tests break because _mesa_ir_link_shader lowers integer
division to an RCP followed by a MUL.  The fix is to factor out more
of the code from ir_to_mesa so that _mesa_ir_link_shader does not need
to be called at all by the i965 driver.  This will be the subject of
several follow-on patches.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agomesa: Add log_uniform and log_program_parameters to dump data
Ian Romanick [Tue, 18 Oct 2011 22:21:04 +0000 (15:21 -0700)]
mesa: Add log_uniform and log_program_parameters to dump data

These were both useful debugging aids while developing this code.
log_uniform will be used to keep the MESA_GLSL=uniform behavior.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agoir_to_mesa: Add _mesa_associate_uniform_storage
Ian Romanick [Tue, 18 Oct 2011 22:07:45 +0000 (15:07 -0700)]
ir_to_mesa: Add _mesa_associate_uniform_storage

Connects all of the gl_program_parameter structures with the correct
gl_uniform_storage structures.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agomesa: Add _mesa_uniform_{attach,detach_all}_driver_storage functions
Ian Romanick [Tue, 18 Oct 2011 21:38:12 +0000 (14:38 -0700)]
mesa: Add _mesa_uniform_{attach,detach_all}_driver_storage functions

These functions are used to create and destroy the connections between
a uniform and the storage used by the driver to hold its value.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agomesa: Add _mesa_propagate_uniforms_to_driver_storage
Ian Romanick [Tue, 18 Oct 2011 21:29:43 +0000 (14:29 -0700)]
mesa: Add _mesa_propagate_uniforms_to_driver_storage

This function propagates the values from the backing storage of a
gl_uniform_storage structure to the driver supplied data locations.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agolinker: Track uniform locations to new tracking structures
Ian Romanick [Tue, 18 Oct 2011 18:54:48 +0000 (11:54 -0700)]
linker: Track uniform locations to new tracking structures

This is just the infrastructure and the code.  It's not used yet.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agomesa: Add structures for "new style" uniform tracking in shader programs
Ian Romanick [Tue, 18 Oct 2011 18:53:19 +0000 (11:53 -0700)]
mesa: Add structures for "new style" uniform tracking in shader programs

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agolinker: Add helper class for parcelling out backing storage to uniforms
Ian Romanick [Tue, 18 Oct 2011 18:24:58 +0000 (11:24 -0700)]
linker: Add helper class for parcelling out backing storage to uniforms

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agolinker: Add helper class for determining uniform usage
Ian Romanick [Tue, 18 Oct 2011 18:09:14 +0000 (11:09 -0700)]
linker: Add helper class for determining uniform usage

v2: Remane class count_uniform_size based on feedback from Eric:

    "Maybe just "count_uniform_size"?  "usage" makes me think "way it's
    dereferenced" or something."

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agomesa: Move most of uniforms.c to uniform_query.cpp
Ian Romanick [Thu, 13 Oct 2011 20:45:39 +0000 (13:45 -0700)]
mesa: Move most of uniforms.c to uniform_query.cpp

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agomesa: Refactor parameter validate for GetUniform, Uniform, and UniformMatrix
Ian Romanick [Fri, 14 Oct 2011 03:18:31 +0000 (20:18 -0700)]
mesa: Refactor parameter validate for GetUniform, Uniform, and UniformMatrix

v2: Update a comment block about the different treatment of
location=-1 based on feedback from Ken.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agomesa: Move {split,merge}_location_offset to uniforms.h
Ian Romanick [Thu, 13 Oct 2011 20:42:56 +0000 (13:42 -0700)]
mesa: Move {split,merge}_location_offset to uniforms.h

Prepend _mesa_uniform_ to the names and rework the calling
convention.  The calling convention was changed for a couple reasons.

1. Having a single variable named 'location' have completely different
meanings at different places in the function is confusing.  Before
calling split_location_offset the location is the encoded value
returned by glGetUniformLocation.  After calling split_location_offset
it's the index of the uniform in the gl_uniform_list::Uniforms array.

2. In a later commit the original value of 'location' is needed after
split_location_offset has been called.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agoglsl: Add new structures for tracking uniforms in linked shaders
Ian Romanick [Wed, 12 Oct 2011 00:16:18 +0000 (17:16 -0700)]
glsl: Add new structures for tracking uniforms in linked shaders

v2: Update some comments based on feedback from Eric Anholt.

v3: Remove gl_uniform_storage::dirty field.  Make
gl_uniform_storage::initialized be bool, and make
gl_uniform_storage::sampler be uint8_t.

v4: Include stdbool.h after Tom Stellard noticed a build failure that
was introduced by the changes in v2.  Oops.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agomesa: Make get_uniform available outside compilation unit
Ian Romanick [Wed, 12 Oct 2011 00:02:06 +0000 (17:02 -0700)]
mesa: Make get_uniform available outside compilation unit

Also rename to _mesa_get_uniform.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agomesa: Move the link check from _mesa_get_uniform_location to _mesa_GetUniformLocationARB
Ian Romanick [Tue, 11 Oct 2011 23:55:54 +0000 (16:55 -0700)]
mesa: Move the link check from _mesa_get_uniform_location to _mesa_GetUniformLocationARB

There are cases where we might want to internally query the location
of a uniform in a shader that failed linking.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agolinker: Make invalidate_variable_locations available outside the compilation unit
Ian Romanick [Tue, 11 Oct 2011 23:15:47 +0000 (16:15 -0700)]
linker: Make invalidate_variable_locations available outside the compilation unit

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agoglsl: Allow glsl_types.h to be included in C sources
Ian Romanick [Tue, 11 Oct 2011 23:11:40 +0000 (16:11 -0700)]
glsl: Allow glsl_types.h to be included in C sources

Some C code will want access to the glsl_base_type and
glsl_sampler_dim enums in the near future.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agomesa: Add string_to_uint_map::clear method to clear the map
Ian Romanick [Tue, 11 Oct 2011 22:25:00 +0000 (15:25 -0700)]
mesa: Add string_to_uint_map::clear method to clear the map

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Tested-by: Tom Stellard <thomas.stellard@amd.com>
12 years agomesa: Fix error generation for glClearBuffer{i ui}v with GL_DEPTH or GL_STENCIL
Ian Romanick [Tue, 1 Nov 2011 22:11:12 +0000 (15:11 -0700)]
mesa: Fix error generation for glClearBuffer{i ui}v with GL_DEPTH or GL_STENCIL

The spec says "Only ClearBufferiv should be used to clear
stencil buffers." and "Only ClearBufferfv should be used to clear
depth buffers."  However, on the following page it also says:

    "The result of ClearBuffer is undefined if no conversion between
    the type of the specified value and the type of the buffer being
    cleared is defined (for example, if ClearBufferiv is called for a
    fixed- or floating-point buffer, or if ClearBufferfv is called
    for a signed or unsigned integer buffer). *This is not an error.*"

Emphasis mine.

Fixes problems with piglit's clearbuffer-invalid-drawbuffer test.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
12 years agomesa: fix the selection of soft renderbuffer color formats
Brian Paul [Sat, 5 Nov 2011 18:59:12 +0000 (12:59 -0600)]
mesa: fix the selection of soft renderbuffer color formats

This fixes a regression from the recent glReadPixels changes found
with the piglit hiz tests.

Use either MESA_FORMAT_RGBA8888 or MESA_FORMAT_RGBA8888_REV for color
buffers depending on endian-ness.  Before, the gl_renderbuffer::Format
field was MESA_FORMAT_RGBA8888 but the data was really stored as
MESA_FORMAT_RGBA8888_REV when using a little endian machine.

Getting this right matters now that we can access renderbuffer data
without going through the span functions (namely glReadPixels() +
MapRenderbuffer()).

12 years agomesa: remove unneeded soft renderbuffer format-setting code
Brian Paul [Sat, 5 Nov 2011 18:54:06 +0000 (12:54 -0600)]
mesa: remove unneeded soft renderbuffer format-setting code

These vars will just get overwritten when we call _mesa_add_renderbuffer()
anyway.  We only need to set the InternalFormat field when we create the
software renderbuffer.

Reviewed-by: Eric Anholt <eric@anholt.net>
12 years agomesa: fix comment typo in intel_renderbuffer
Brian Paul [Sat, 5 Nov 2011 16:48:18 +0000 (10:48 -0600)]
mesa: fix comment typo in intel_renderbuffer

12 years agointel: update intel_texture_image comment
Brian Paul [Sat, 5 Nov 2011 16:48:18 +0000 (10:48 -0600)]
intel: update intel_texture_image comment

12 years agointel: wrap comment and fix typo
Brian Paul [Sat, 5 Nov 2011 16:48:18 +0000 (10:48 -0600)]
intel: wrap comment and fix typo

12 years agost/mesa: first implementation of Map/UnmapRenderbuffer()
Brian Paul [Sat, 5 Nov 2011 16:48:18 +0000 (10:48 -0600)]
st/mesa: first implementation of Map/UnmapRenderbuffer()

Untested, but also unused at this point.

12 years agoxlib: implement renderbuffer mapping/unmapping
Brian Paul [Sat, 5 Nov 2011 16:48:18 +0000 (10:48 -0600)]
xlib: implement renderbuffer mapping/unmapping

This fixes the glReadPixels() regression for reading from the front/back
color buffers.

Note, we only allow one mapping of an XImage/Pixmap renderbuffer
at any time.  That might need to be revisited in the future.

12 years agomesa: check for immutable texture in _mesa_test_texobj_completeness()
Brian Paul [Fri, 4 Nov 2011 01:56:51 +0000 (19:56 -0600)]
mesa: check for immutable texture in _mesa_test_texobj_completeness()

One of the points of GL_ARB_texture_storage is to make it impossible
to have malformed mipmap stacks.  If we know the texture object is
immutable, we can skip a bunch of size checking.