mesa.git
10 years agomesa: Copy linked program data for GS.
Paul Berry [Sat, 23 Mar 2013 17:51:53 +0000 (10:51 -0700)]
mesa: Copy linked program data for GS.

The documentation for gl_shader_program.Geom and gl_geometry_program
says that the former is copied to the latter at link time, but this
wasn't happening.  This patch causes _mesa_ir_link_shader() to perform
the copy, and updates comment accordingly.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agomesa: Refactor copying of linked program data.
Paul Berry [Sat, 23 Mar 2013 17:51:53 +0000 (10:51 -0700)]
mesa: Refactor copying of linked program data.

This patch creates a single function to copy the the UsesClipDistance
flag from gl_shader_program.Vert to gl_vertex_program.  Previously
this logic was duplicated in the i965-specific function
brw_link_shader() and the core mesa function _mesa_ir_link_shader().

This logic will have to be expanded to support geometry shaders, and I
don't want to have to update it in two separate places.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agoglsl: support compilation of geometry shaders
Bryan Cain [Fri, 15 Feb 2013 15:46:50 +0000 (09:46 -0600)]
glsl: support compilation of geometry shaders

This commit adds all of the parsing and semantics for GLSL 150 style
geometry shaders.

v2 (Paul Berry <stereotype441@gmail.com>): Add a few missing calls to
get_pipeline_stage().  Fix some signed/unsigned comparison warnings.
Fix handling of NULL consumer in assign_varying_locations().

v3 (Bryan Cain <bryancain3@gmail.com>): fix indexing order of 2D
arrays.  Also, allow interpolation qualifiers in geometry shaders.

v4 (Paul Berry <stereotype441@gmail.com>): Eliminate
get_pipeline_stage()--it is no longer needed thanks to 030ca23 (mesa:
renumber shader indices according to their placement in pipeline).
Remove 2D stuff.  Move vertices_per_prim() to ir.h, so that it will be
accessible from outside the linker.  Remove
inject_num_vertices_visitor.  Rework for GLSL 1.50.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
v5 (Paul Berry <stereotype441@gmail.com>): Split out
do_set_program_inouts() argument refactoring to a separate patch.
Move geom_array_resizing_visitor to later in the series.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agoglsl/linker: Make separate allocations to track vertex and fragment shaders.
Paul Berry [Wed, 31 Jul 2013 05:38:43 +0000 (22:38 -0700)]
glsl/linker: Make separate allocations to track vertex and fragment shaders.

There's no reason to be clever about this.  By making separate
allocations for vertex and fragment shaders, we'll allow geometry
shaders to be added without introducing any complication.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agoglsl: add builtins for geometry shaders.
Bryan Cain [Fri, 15 Feb 2013 14:53:20 +0000 (08:53 -0600)]
glsl: add builtins for geometry shaders.

v2 (Paul Berry <stereotype441@gmail.com>): Account for rework of
builtin_variables.cpp.  Use INTERP_QUALIFIER_FLAT for gl_PrimitiveID
so that it will obey provoking vertex conventions.  Convert to GLSL
1.50 style geometry shaders.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
v3 (Paul Berry <stereotype441@gmail.com>): Be less obscure about
setting interpolation field of gl_Primitive variables.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agoglsl: add ir_emit_vertex and ir_end_primitive instruction types
Bryan Cain [Fri, 15 Feb 2013 15:26:35 +0000 (09:26 -0600)]
glsl: add ir_emit_vertex and ir_end_primitive instruction types

These correspond to the EmitVertex and EndPrimitive functions in GLSL.

v2 (Paul Berry <stereotype441@gmail.com>): Add stub implementations of
new pure visitor functions to i965's vec4_visitor and fs_visitor
classes.

v3 (Paul Berry <stereotype441@gmail.com>): Rename classes to be more
consistent with the names used in the GL spec.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agomesa: account for geometry shader texture fetches in update_texture_state
Bryan Cain [Fri, 15 Feb 2013 08:40:12 +0000 (02:40 -0600)]
mesa: account for geometry shader texture fetches in update_texture_state

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agomain: Allow for the possibility of GL 3.2 without ARB_geometry_shader4.
Paul Berry [Sun, 28 Jul 2013 16:23:11 +0000 (09:23 -0700)]
main: Allow for the possibility of GL 3.2 without ARB_geometry_shader4.

Previously, we assumed that the only way Mesa would expose geometry
shader support was via the ARB_geometry_shader4 extension.  But this
extension has some extra complications over GL 3.2 (interactions with
compatibility-only features, and link-time initialization of the
constant gl_VerticesIn).  So we want to allow for the possibility of
supporting GL 3.2 (with GLSL 1.50 style geometry shaders) even if
ctx->Extensions.ARB_geometry_shader4 is false.

This patch adds a new function, _mesa_has_geometry_shaders(), which
returns true if either ARB_geometry_shader4 is supported or the GL
version is at least 3.2 desktop.  Since compute_version() only enables
GL 3.2 functionality when GLSL 1.50 support is present, a sufficient
way for a back-end to advertise geometry shader support is to set
ctx->Const.GLSLVersion >= 150.

v2: Remove unnecessary ctx->Const.GeometryShaders150 constant.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agomain: Fix geometry shader error messages (missing right paren)
Paul Berry [Mon, 24 Jun 2013 15:50:04 +0000 (08:50 -0700)]
main: Fix geometry shader error messages (missing right paren)

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agoglsl: Add EXT_texture_array support for geometry shaders.
Paul Berry [Fri, 28 Jun 2013 20:02:23 +0000 (13:02 -0700)]
glsl: Add EXT_texture_array support for geometry shaders.

We can't just use a ".glsl" file since the Lod variants are only
available in vertex and geometry shaders, while the bias variants are
only available in the fragment shader.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
10 years agoglsl/linker: Make update_array_sizes apply to just uniforms.
Paul Berry [Mon, 10 Jun 2013 21:01:45 +0000 (14:01 -0700)]
glsl/linker: Make update_array_sizes apply to just uniforms.

Commit 586b4b5 (glsl: Also update implicit sizes of varyings at link
time) extended update_array_sizes() to apply to both uniforms and
shader ins/outs.  However, doing creates problems for geometry
shaders, because update_array_sizes() assumes that variables with
matching names in different parts of the pipeline should have the same
sizes.  With the addition of geometry shaders, this is no longer true
(e.g. both vertex and geometry shaders have a gl_ClipDistance output
variable, but there's no reason these variables should have the same
sizes).

The original reason for commit 586b4b5 (avoid problems with
gl_TexCoord being 0 length) has since been addressed by commit 6f53921
(linker: Ensure that unsized arrays have a size after linking).  So go
ahead and switch update_array_sizes() back to only acting on uniforms.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agoglsl: Modify ir_set_program_inouts to handle geometry shaders.
Paul Berry [Wed, 31 Jul 2013 19:49:49 +0000 (12:49 -0700)]
glsl: Modify ir_set_program_inouts to handle geometry shaders.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
10 years agoglsl: In ir_set_program_inouts, handle indexing outside array/matrix bounds.
Paul Berry [Wed, 31 Jul 2013 18:25:13 +0000 (11:25 -0700)]
glsl: In ir_set_program_inouts, handle indexing outside array/matrix bounds.

According to GLSL, indexing into an array or matrix with an
out-of-range constant results in a compile error.  However, indexing
with an out-of-range value that isn't constant merely results in
undefined results.

Since optimization passes (e.g. loop unrolling) can convert
non-constant array indices into constant array indices, it's possible
that ir_set_program_inouts will encounter a constant array index that
is out of range; if this happens, just mark the whole array as used.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agoglsl: Fallback gracefully if ir_set_program_inouts sees unexpected indexing.
Paul Berry [Wed, 31 Jul 2013 17:57:22 +0000 (10:57 -0700)]
glsl: Fallback gracefully if ir_set_program_inouts sees unexpected indexing.

The code in ir_set_program_inouts that marks just a portion of a
variable as used (rather than the whole variable) only works on a few
kinds of indexing operations:

- Indexing into matrices
- Indexing into arrays of matrices, vectors, or scalars.

Fortunately these are the only kinds of indexing operations that we
expect to see; everything else is either handled by a
previously-executed lowering pass or prohibited by GLSL.

However, that could conceivably change in the future (the GLSL rules
might change, or we might modify the lowering passes).  To avoid
mysterious bugs in the future, let's have ir_set_program_inouts report
an assertion failure if it ever encounters an unexpected kind of
indexing operation (and in release builds, fall back to just marking
the whole variable as used).

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agoglsl: Extract marking functions from ir_set_program_inouts.
Paul Berry [Wed, 31 Jul 2013 17:15:49 +0000 (10:15 -0700)]
glsl: Extract marking functions from ir_set_program_inouts.

This patch extracts the functions mark_whole_variable() and
try_mark_partial_variable() from the ir_set_program_inouts visitor
functions.  This will make the code easier to follow when we add
geometry shader support.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agoglsl: Use count_attribute_slots() in ir_set_program_inouts.
Paul Berry [Wed, 31 Jul 2013 16:43:16 +0000 (09:43 -0700)]
glsl: Use count_attribute_slots() in ir_set_program_inouts.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agoglsl: Expand count_attribute_slots() to cover structs.
Paul Berry [Wed, 31 Jul 2013 15:24:48 +0000 (08:24 -0700)]
glsl: Expand count_attribute_slots() to cover structs.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agoMove count_attribute_slots() out of the linker and into glsl_type.
Paul Berry [Wed, 31 Jul 2013 15:15:08 +0000 (08:15 -0700)]
Move count_attribute_slots() out of the linker and into glsl_type.

Our previous justification for leaving this function out of glsl_type
was that it implemented counting rules that were specific to GLSL
1.50.  However, these counting rules also describe the number of
varying slots that Mesa will assign to a varying in the absence of
varying packing.  That's useful to be able to compute from outside of
the linker code (a future patch will use it from
ir_set_program_inouts.cpp).  So go ahead and move it to glsl_type.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agoglsl: Change do_set_program_inouts' is_fragment_shader arg to shader_type.
Paul Berry [Wed, 31 Jul 2013 03:49:56 +0000 (20:49 -0700)]
glsl: Change do_set_program_inouts' is_fragment_shader arg to shader_type.

This will allow us to add geometry shader support without having to
add another boolean argument.

Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agogallivm: obey clarified shift behavior
Roland Scheidegger [Tue, 30 Jul 2013 19:26:27 +0000 (21:26 +0200)]
gallivm: obey clarified shift behavior

llvm shifts are undefined for shift counts exceeding (or matching) bit width,
so need to apply a mask for the tgsi shift instructions.

v2: only use mask for the tgsi shift instructions, not for the build shift
helpers. None of the internal callers need this behavior, and while llvm can
optimize away the masking for constants there are legitimate cases where it
might not be able to do so even if we know that shift count must be smaller
than type width (currently all such callers do not use the build shift
helpers).

Reviewed-by: Zack Rusin <zackr@vmware.com>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
10 years agotgsi: obey clarified shift behavior
Roland Scheidegger [Tue, 30 Jul 2013 15:16:17 +0000 (17:16 +0200)]
tgsi: obey clarified shift behavior

c shifts are undefined for shift counts exceeding (or matching) bit width,
so need to apply a mask (on x86 it actually would usually probably work as
shifts do masking on int domain shifts - unless some auto-vectorizer would
come along at last as simd domain does not mask the shift count).

Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
10 years agogallium: clarify shift behavior with shift count >= 32
Roland Scheidegger [Tue, 30 Jul 2013 15:08:01 +0000 (17:08 +0200)]
gallium: clarify shift behavior with shift count >= 32

Previously, nothing was said what happens with shift counts exceeding
bit width of the values to shift. In theory 3 behaviors are possible:
1) undefined (classic c definition)
2) just shift out all bits (so result is zero, or -1 potentially for ashr)
3) mask the shift count to bit width - 1
API's either require 3) or are ok with 1). In particular, GLSL (as well as a
couple uninteresting legacy GL extensions) is happy with undefined, whereas
both OpenCL and d3d10 require 3). Consequently, most hw also implements 3).
So, for simplicity we just specify that 3) is required rather than saying
undefined and then needing state trackers to work around it.
Also while here specify shift count as a vector, not scalar. As far as I
can tell this was a doc bug, neither state trackers nor drivers used scalar
shift count.

Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
10 years agodocs: Add md5sums to 9.1.6 release notes
Carl Worth [Thu, 1 Aug 2013 22:45:04 +0000 (15:45 -0700)]
docs: Add md5sums to 9.1.6 release notes

10 years agodocs: Import 9.1.6 release notes, add news item.
Carl Worth [Thu, 1 Aug 2013 22:12:25 +0000 (15:12 -0700)]
docs: Import 9.1.6 release notes, add news item.

10 years agoi965: Delete the BATCH_LOCALS macro.
Kenneth Graunke [Mon, 8 Jul 2013 02:17:02 +0000 (19:17 -0700)]
i965: Delete the BATCH_LOCALS macro.

This hasn't done anything in a long time, and it's only used in a couple
places...which means we couldn't use it without doing a bunch of work
anyway.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
10 years agoCorrect clamping of TEXTURE_{MAX, BASE}_LEVEL
Corey Richardson [Wed, 31 Jul 2013 14:46:12 +0000 (10:46 -0400)]
Correct clamping of TEXTURE_{MAX, BASE}_LEVEL

Previously, if TEXTURE_IMMUTABLE_FORMAT was TRUE, the levels were allowed to
be set like usual, but ARB_texture_storage states:

> if TEXTURE_IMMUTABLE_FORMAT is TRUE, then level_base is clamped to the range
> [0, <levels> - 1] and level_max is then clamped to the range [level_base,
> <levels> - 1], where <levels> is the parameter passed the call to
> TexStorage* for the texture object

Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Corey Richardson <corey@octayn.net>
10 years agoDe-tab and align comments in gl_texture_object
Corey Richardson [Wed, 31 Jul 2013 14:14:22 +0000 (10:14 -0400)]
De-tab and align comments in gl_texture_object

Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Corey Richardson <corey@octayn.net>
10 years agoi965 Gen4/5: clip: Don't mangle flat varyings
Chris Forbes [Sun, 7 Jul 2013 10:22:57 +0000 (22:22 +1200)]
i965 Gen4/5: clip: Don't mangle flat varyings

This patch ensures that integers will pass through unscathed.  Doing
(useless) computations on them is risky, especially when their bit
patterns correspond to values like inf or nan.

[V1-2]: Signed-off-by: Olivier Galibert <galibert at pobox.com>
Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agoi965 Gen4/5: clip: Add support for noperspective varyings
Chris Forbes [Sun, 7 Jul 2013 09:42:39 +0000 (21:42 +1200)]
i965 Gen4/5: clip: Add support for noperspective varyings

Adds support for interpolating noperspective varyings linearly in screen
space when clipping.

Based on Olivier Galibert's patch from last year:
http://lists.freedesktop.org/archives/mesa-dev/2012-July/024341.html

At this point all -fixed and -vertex interpolation tests work.

V5: Add brw_clip_compile.has_noperspective_shading rather than another
key flag.
V6: Real bools.

[V1-2]: Signed-off-by: Olivier Galibert <galibert at pobox.com>
Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Acked-by: Paul Berry <stereotype441@gmail.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agoi965 Gen4/5: clip: correctly handle flat varyings
Chris Forbes [Sun, 7 Jul 2013 07:47:19 +0000 (19:47 +1200)]
i965 Gen4/5: clip: correctly handle flat varyings

Previously we only gave special treatment to the builtin color varyings.
This patch adds support for arbitrary flat-shaded varyings, which is
required for GLSL 1.30.

Based on Olivier Galibert's patch from last year:
http://lists.freedesktop.org/archives/mesa-dev/2012-July/024340.html

V5: Move key.do_flat_shading to brw_clip_compile.has_flat_shading
V6: Real bools.

[V1-2]: Signed-off-by: Olivier Galibert <galibert at pobox.com>
Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agoi965 Gen4/5: Generalize SF interpolation setup for GLSL1.3
Chris Forbes [Sat, 6 Jul 2013 13:53:45 +0000 (01:53 +1200)]
i965 Gen4/5: Generalize SF interpolation setup for GLSL1.3

Previously the SF only handled the builtin color varying specially.
This patch generalizes that support to cover user-defined varyings,
driven by the interpolation mode array set up alongside the VUE map.

Based on the following patches from Olivier Galibert:
- http://lists.freedesktop.org/archives/mesa-dev/2012-July/024335.html
- http://lists.freedesktop.org/archives/mesa-dev/2012-July/024339.html

With this patch, all the GLSL 1.3 interpolation tests that do not clip
(spec/glsl-1.30/execution/interpolation/*-none.shader_test) pass.

V5: Move key.do_flat_shading to brw_sf_compile.has_flat_shading; drop
vestigial hunks.
V6: Real bools.

Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agoi965: Add helper functions for interpolation map
Chris Forbes [Wed, 31 Jul 2013 10:28:13 +0000 (22:28 +1200)]
i965: Add helper functions for interpolation map

V6: real bools

Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agoi965 Gen4/5: Introduce 'interpolation map' alongside the VUE map
Chris Forbes [Sat, 6 Jul 2013 14:56:14 +0000 (02:56 +1200)]
i965 Gen4/5: Introduce 'interpolation map' alongside the VUE map

The interpolation map (in brw->interpolation_mode) is a new auxiliary
structure alongside the post-GS VUE map, which describes the
interpolation modes for each VUE slot, for use by the clip and SF
stages.

This patch introduces a new state atom to compute the interpolation map,
and adjusts the program keys for the clip and SF stages, but it is not
actually used yet.

[V1-2]: Signed-off-by: Olivier Galibert <galibert at pobox.com>

V3: Updated for vue_map changes, intel -> brw merge, etc. (Chris Forbes)
V4: Compute interpolation map as a new state atom rather than tacking it
on the front of the clip setup
V5: Rework commit message, make interpolation_mode_map a struct.

Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agoget-pick-list: Allow for non-whitespace between "CC:" and "mesa-stable"
Carl Worth [Wed, 31 Jul 2013 22:49:48 +0000 (15:49 -0700)]
get-pick-list: Allow for non-whitespace between "CC:" and "mesa-stable"

We recently proposed a new syntax for stable-patch nominations such as:

CC: "9.2 and 9.1" <mesa-stable@lists.freedesktop.org>
and this has already appeared in the wild.

So we extend the regular expression to pick this up as well.

10 years agonvc0: properly align NVE4_COMPUTE_MP_TEMP_SIZE
Samuel Pitoiset [Thu, 25 Jul 2013 08:35:33 +0000 (10:35 +0200)]
nvc0: properly align NVE4_COMPUTE_MP_TEMP_SIZE

MP_TEMP_SIZE must be aligned to 0x8000, while TEMP_SIZE on NVE4_3D
must be aligned to 0x20000, so perform both alignments to be sure
we allocate enough space (actually the bo will most likely use 128
KiB pages and not aligning to that would be a waste anyway).

Cc: "9.2" mesa-stable@lists.freedesktop.org
10 years agomesa/program: remove useless YYID
Laurent Carlier [Wed, 31 Jul 2013 13:18:52 +0000 (15:18 +0200)]
mesa/program: remove useless YYID

This fixes the build with Bison 3.0. Also works with Bison 2.7.1.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agomesa/program: Switch from the deprecated YYLEX_PARAM to %lex-param.
Kenneth Graunke [Mon, 29 Jul 2013 23:03:46 +0000 (16:03 -0700)]
mesa/program: Switch from the deprecated YYLEX_PARAM to %lex-param.

YYLEX_PARAM is no longer supported as of Bison 3.0.  Instead, the Bison
developers recommend using %lex-param.

%lex-param takes a type and variable name, similar to %parse-param,
so you can't pass an arbitrary expression like state->scanner.  But Flex
insists on passing the actual scanner object, not an arbitrary object
like state.

To solve this, the parser defines a wrapper lex() function which accepts
"state," and calls Flex's lex() function with state->scanner.

Fixes the build with Bison 3.0.  Also works with Bison 2.7.1.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=67354
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Tested-by: Laurent Carlier <lordheavym@gmail.com>
Cc: "9.2" mesa-stable@lists.freedesktop.org
10 years agomesa/program: Change the program parser's namespace.
Kenneth Graunke [Mon, 29 Jul 2013 22:55:06 +0000 (15:55 -0700)]
mesa/program: Change the program parser's namespace.

Bison 3.0 removes the YYLEX_PARAM macro.  In preparation for handling
this using %lex-param, the parser needs a wrapper function for the
actual Flex lex() function.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=67354
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Tested-by: Laurent Carlier <lordheavym@gmail.com>
Cc: "9.2" mesa-stable@lists.freedesktop.org
10 years agoglsl: Switch from the deprecated YYLEX_PARAM to %lex-param.
Kenneth Graunke [Mon, 29 Jul 2013 22:28:59 +0000 (15:28 -0700)]
glsl: Switch from the deprecated YYLEX_PARAM to %lex-param.

YYLEX_PARAM is no longer supported as of Bison 3.0.  Instead, the Bison
developers recommend using %lex-param.

%lex-param takes a type and variable name, similar to %parse-param,
so you can't pass an arbitrary expression like state->scanner.  But Flex
insists on passing the actual scanner object, not an arbitrary object
like state.

To solve this, the parser defines a wrapper lex() function which accepts
"state," and calls Flex's lex() function with state->scanner.

Fixes the build with Bison 3.0.  Also works with Bison 2.7.1.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=67354
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Tested-by: Laurent Carlier <lordheavym@gmail.com>
Cc: "9.2" mesa-stable@lists.freedesktop.org
10 years agoglsl: Change the lexer's namespace.
Kenneth Graunke [Mon, 29 Jul 2013 22:27:46 +0000 (15:27 -0700)]
glsl: Change the lexer's namespace.

Bison 3.0 removes the YYLEX_PARAM macro.  In preparation for handling
this using %lex-param, the parser needs a wrapper function for the
actual Flex lex() function.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=67354
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Tested-by: Laurent Carlier <lordheavym@gmail.com>
Cc: "9.2" mesa-stable@lists.freedesktop.org
10 years agoegl: Restore "bogus" DRI2 invalidate event code.
Eric Anholt [Fri, 21 Jun 2013 22:34:52 +0000 (15:34 -0700)]
egl: Restore "bogus" DRI2 invalidate event code.

I had removed it in commit 1e7776ca2bc59a6978d9b933d23852d47078dfa8
because it was obviously wrong -- why do we care whether the server is a
version that emits events, if we're not watching for the server's events,
anyway?  And why would you only invalidate on a server that emits
invalidate events, when the comment said to emit invalidates if the server
*doesn't*?  Only, I missed that we otherwise don't flag that our buffers
might have changed at swap time at all, so the driver was only checking
for new buffers when triggered by the Viewport hack.  Of course you don't
expect Viewport to be called after a swap.

So, this is effectively a revert of the previous commit, except that I
dropped the check for only emitting invalidates on a new server -- we
*always* need to invalidate if we're doing a SwapBuffers.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=63435
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Cc: "9.1 and 9.2" <mesa-stable@lists.freedesktop.org>
10 years agogallivm: use nearest rounding for float->unorm24 conversion
Roland Scheidegger [Wed, 31 Jul 2013 14:57:50 +0000 (16:57 +0200)]
gallivm: use nearest rounding for float->unorm24 conversion

Previously we were using truncation, which gives the correct result
only for numbers in [0.5-1.0] range (because there's no mantissa bits
to do any rounding there).
This is frequently hit (and probably only used there) when converting
fragment depth to depth format (d24s8 etc.) or otherwise dealing with
depth format.

v2: as spotted by Jose, get rid of extra type (src_type is already unsigned).

Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
10 years agomesa: fix multisampling proxy textures not being queryable
Mikko Juola [Tue, 30 Jul 2013 17:39:00 +0000 (20:39 +0300)]
mesa: fix multisampling proxy textures not being queryable

The code that checks if some texture target is valid for
glGetTexLevelParameter*() was not programmed to check for multisampling
proxy textures.  This made it impossible(?) to use the proxy textures
for their intended purpose as glGetTexLevelParameter*() would just fail
on you.

Reviewed-by: Brian Paul <brianp@vmware.com>
Cc: mesa-stable@lists.freedesktop.org
10 years agomesa: fix proxy textures becoming immutable and unusable
Mikko Juola [Tue, 30 Jul 2013 03:36:43 +0000 (06:36 +0300)]
mesa: fix proxy textures becoming immutable and unusable

glTexStorage*() functions make textures immutable.  This carries on to
proxy textures.  Error checking in texture storage functions prevents
proxy textures from working after first time because internally, they
became immutable.

This commit makes the error checking ignore the immutability flag when
working with proxy textures.

Reviewed-by: Brian Paul <brianp@vmware.com>
Cc: mesa-stable@lists.freedesktop.org
10 years agomesa: fix proxy textures not working with default texture binding
Mikko Juola [Tue, 30 Jul 2013 03:29:54 +0000 (06:29 +0300)]
mesa: fix proxy textures not working with default texture binding

When working with the glTexStorage*() functions, the error checking
checks that a non-default (i.e., non-zero) texture is currently bound.
However, this check made glTexStorage*() functions fail with proxy
textures when the default texture is bound. Proxy textures do not care
about the current texture bindings so for them this check should not
be done.

Reviewed-by: Brian Paul <brianp@vmware.com>
Cc: mesa-stable@lists.freedesktop.org
10 years agomesa: fix number of mipmaps calculation for proxy textures
Mikko Juola [Tue, 30 Jul 2013 03:19:59 +0000 (06:19 +0300)]
mesa: fix number of mipmaps calculation for proxy textures

The function _mesa_get_tex_max_num_levels() is supposed to calculate
the number of mipmap levels but it was not written to handle proxy
textures, at best returning a maximum of 1 mipmap level. Because of
this, at least glTexStorage*() calls would incorrectly fail when used
with proxy textures with more than one mipmap level.

Reviewed-by: Brian Paul <brianp@vmware.com>
Cc: mesa-stable@lists.freedesktop.org
10 years agomesa: improve free() cleanup in generate_mipmap_compressed()
Brian Paul [Fri, 26 Jul 2013 15:50:36 +0000 (09:50 -0600)]
mesa: improve free() cleanup in generate_mipmap_compressed()

Free all our temporary buffers in one place at the end of the
function.  Fixes memory leak detected by Coverity.

Note: This is a candidate for the 9.x branches
Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: José Fonseca <jfonseca@vmware.com>
10 years agogallium/util: reformat, comment util_get_offset()
Brian Paul [Thu, 25 Jul 2013 17:21:50 +0000 (11:21 -0600)]
gallium/util: reformat, comment util_get_offset()

Reviewed-by: Roland Scheidegger <sroland@vmware.com>
10 years agogallium/util: comments, var renaming in u_inlines.h
Brian Paul [Thu, 25 Jul 2013 17:17:52 +0000 (11:17 -0600)]
gallium/util: comments, var renaming in u_inlines.h

The variable 'usage' was being used for two different things.
Sometimes for PIPE_USAGE_x and other times for PIPE_TRANSFER_x.
This renames usage to access when we're talking about PIPE_TRANSFER_x
flags.  Plus, add a bunch of comments to remind us what's going on.

Also, use unsigned for PIPE_TRANSFER_x bitmask to be consistent with
other places.  And add a missing const qualifier.

Reviewed-by: Roland Scheidegger <sroland@vmware.com>
10 years agosoftpipe: use new softpipe_resource_data() accessor
Brian Paul [Thu, 25 Jul 2013 15:27:09 +0000 (09:27 -0600)]
softpipe: use new softpipe_resource_data() accessor

We should probably be using map()/unmap() when accessing resource
data, but this is a little better.

v2: assert that the resource is not a display target, per Jose.

Reviewed-by: José Fonseca <jfonseca@vmware.com>
10 years agosoftpipe: don't ignore pipe_constant_buffer::buffer_offset
Brian Paul [Thu, 25 Jul 2013 15:13:47 +0000 (09:13 -0600)]
softpipe: don't ignore pipe_constant_buffer::buffer_offset

This was never a problem since the Mesa state tracker always gives
us a user-space constant buffer with buffer_offset=0.  But if another
state tracker ever gave us a "HW" constant buffer with non-zero
buffer_offset we'd mis-render.

Also, use the correct buffer size.  And move an assertion to the
top of the function.

Reviewed-by: Marek Olšák <maraeo@gmail.com>
Reviewed-by: José Fonseca <jfonseca@vmware.com>
10 years agogallium/docs: clarify definition of PIPE_CAP_USER_CONSTANT_BUFFERS, etc
Brian Paul [Thu, 25 Jul 2013 15:15:52 +0000 (09:15 -0600)]
gallium/docs: clarify definition of PIPE_CAP_USER_CONSTANT_BUFFERS, etc

The cap means _can_ accept user-space constant buffers; it doesn't
mean _only_ accepts user-space constant buffers.

v2: also update the PIPE_CAP_USER_VERTEX_BUFFERS and
PIPE_CAP_USER_INDEX_BUFFERS descriptions as well.  Per Jose.

Reviewed-by: Marek Olšák <maraeo@gmail.com>
Reviewed-by: José Fonseca <jfonseca@vmware.com>
10 years agoi965/vs: Put lod parameter in the correct place for Gen4
Chris Forbes [Tue, 30 Jul 2013 04:03:58 +0000 (16:03 +1200)]
i965/vs: Put lod parameter in the correct place for Gen4

This was never visible before due to the bogus sampler state pointer.
Fixes remaining vertex texturing breakage on Gen4.

Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Cc: mesa-stable@lists.freedesktop.org
10 years agoi965/vs: set up sampler state pointer for Gen4/5.
Chris Forbes [Mon, 29 Jul 2013 18:12:24 +0000 (06:12 +1200)]
i965/vs: set up sampler state pointer for Gen4/5.

Fixes broken filter and lod selection for vertex texturing.
(txs/txf only worked properly because they ignore the sampler state
completely)

Signed-off-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Cc: mesa-stable@lists.freedesktop.org
10 years agost/dri: add a new driconf option disable_shader_bit_encoding for Unigine
Marek Olšák [Tue, 30 Jul 2013 20:29:27 +0000 (22:29 +0200)]
st/dri: add a new driconf option disable_shader_bit_encoding for Unigine

Now Unigine Heaven 3.0 finally works with r600g.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
10 years agost/mesa: fix opcode translation for ARB_shader_bit_encoding functions
Marek Olšák [Tue, 30 Jul 2013 20:29:25 +0000 (22:29 +0200)]
st/mesa: fix opcode translation for ARB_shader_bit_encoding functions

We treat the opcodes as MOVs, but we should at least change the type
of the expression, which later affects which TGSI opcode is chosen.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
10 years agomesa,glsl,st/dri: add a new driconf option force_glsl_version for Unigine
Marek Olšák [Tue, 30 Jul 2013 20:29:24 +0000 (22:29 +0200)]
mesa,glsl,st/dri: add a new driconf option force_glsl_version for Unigine

See documentation in mtypes.h.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
10 years agomesa: add MESA_GLSL debug flag to dump shaders on compile error
Marek Olšák [Tue, 30 Jul 2013 20:29:23 +0000 (22:29 +0200)]
mesa: add MESA_GLSL debug flag to dump shaders on compile error

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
10 years agodriconf: enable app-specific workarounds for all drivers
Marek Olšák [Tue, 30 Jul 2013 20:29:22 +0000 (22:29 +0200)]
driconf: enable app-specific workarounds for all drivers

They were only enabled for i965.

Note that drirc must be installed in /etc.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
10 years agost/dri: remove driOptionCache from dri_context in favor of dri_screen
Marek Olšák [Tue, 30 Jul 2013 20:29:21 +0000 (22:29 +0200)]
st/dri: remove driOptionCache from dri_context in favor of dri_screen

There is no reason to have this duplicated.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
10 years agost/dri: move enabling postprocessing to dri_screen
Marek Olšák [Tue, 30 Jul 2013 20:29:20 +0000 (22:29 +0200)]
st/dri: move enabling postprocessing to dri_screen

The driconf options are global.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
10 years agost/dri: remove more unused driconf options
Marek Olšák [Tue, 30 Jul 2013 20:29:19 +0000 (22:29 +0200)]
st/dri: remove more unused driconf options

vblank_mode is read by dri_util.c and falls under the "dri2" driver name,
which is not connected to the actual Mesa/Gallium driver in any way.

Reviewed-by: Brian Paul <brianp@vmware.com>
10 years agost/dri: implement the driconf option force_s3tc_enable properly
Marek Olšák [Tue, 30 Jul 2013 20:29:18 +0000 (22:29 +0200)]
st/dri: implement the driconf option force_s3tc_enable properly

Reviewed-by: Brian Paul <brianp@vmware.com>
10 years agodriconf: remove the unused option allow_large_textures
Marek Olšák [Tue, 30 Jul 2013 20:29:17 +0000 (22:29 +0200)]
driconf: remove the unused option allow_large_textures

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
10 years agost/dri: support the driconf option disable_blend_func_extended
Marek Olšák [Tue, 30 Jul 2013 20:29:16 +0000 (22:29 +0200)]
st/dri: support the driconf option disable_blend_func_extended

This is needed for Unigine.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
10 years agost/osmesa: initialize disable_glsl_line_continuations
Marek Olšák [Tue, 30 Jul 2013 20:29:15 +0000 (22:29 +0200)]
st/osmesa: initialize disable_glsl_line_continuations

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
10 years agogallium/postprocessing: convert blits to pipe->blit
Marek Olšák [Tue, 30 Jul 2013 20:29:14 +0000 (22:29 +0200)]
gallium/postprocessing: convert blits to pipe->blit

PP saves current states to cso_context and then util_blit_pixels does
the same. cso_context doesn't like that and the original state is not
correctly restored.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Brian Paul <brianp@vmware.com>
10 years agogallium/postprocessing: fix shader parsing
Marek Olšák [Tue, 30 Jul 2013 20:29:12 +0000 (22:29 +0200)]
gallium/postprocessing: fix shader parsing

tokens was converted to a pointer, which made the Elements macro return 1.

Broken by e87fc11cac696881469a57955af2ac7b4929a2c7.

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
10 years agodocs/GL3: clarify core vs compatibility extension support
Marek Olšák [Tue, 30 Jul 2013 20:29:11 +0000 (22:29 +0200)]
docs/GL3: clarify core vs compatibility extension support

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
10 years agomesa: default texture buffer format should be R8 in the core profile
Marek Olšák [Tue, 30 Jul 2013 20:29:10 +0000 (22:29 +0200)]
mesa: default texture buffer format should be R8 in the core profile

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
v2: Since we don't expose the extension in the compatibility profile,
    the "if (API == CORE) .. else .." statement is removed.

10 years agomesa: default DEPTH_TEXTURE_MODE should be RED in the core profile
Marek Olšák [Tue, 30 Jul 2013 20:29:08 +0000 (22:29 +0200)]
mesa: default DEPTH_TEXTURE_MODE should be RED in the core profile

Cc: mesa-stable@lists.freedesktop.org
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agost/mesa: expose EXT_framebuffer_multisample_blit_scaled if MSAA is supported
Marek Olšák [Tue, 30 Jul 2013 20:29:06 +0000 (22:29 +0200)]
st/mesa: expose EXT_framebuffer_multisample_blit_scaled if MSAA is supported

Surprisingly all drivers supporting MSAA can already do this (r300g and r600g
for sure) and I think Christoph wanted to have this feature for his Nouveau
drivers anyway.

10 years agost/mesa: fix sRGB renderbuffers without EXT_framebuffer_sRGB support
Marek Olšák [Tue, 30 Jul 2013 20:29:04 +0000 (22:29 +0200)]
st/mesa: fix sRGB renderbuffers without EXT_framebuffer_sRGB support

https://bugs.freedesktop.org/show_bug.cgi?id=59322

Cc: mesa-stable@lists.freedesktop.org
10 years agoRevert "r300g: Give CLIP_DISABLE another try"
Marek Olšák [Tue, 30 Jul 2013 20:29:00 +0000 (22:29 +0200)]
Revert "r300g: Give CLIP_DISABLE another try"

This reverts commit e866bd1adea2c3b4971ad68e69c644752f2ab7b6.

https://bugs.freedesktop.org/show_bug.cgi?id=57875

Cc: mesa-stable@lists.freedesktop.org
10 years agoget-pick-list.sh: Include commits mentionining "CC: mesa-stable..." in pick list
Carl Worth [Tue, 30 Jul 2013 19:36:37 +0000 (12:36 -0700)]
get-pick-list.sh: Include commits mentionining "CC: mesa-stable..." in pick list

We recently adopted a new convention that patches can be nominated for the
stable branch by including a line in the commit message as follows:

CC: mesa-stable@lists.freedesktop.org
This is a convenient syntax as "git send-email" will notice this line and
automatically copy the resulting patch email to the mesa-stable mailing list.

Here we extend the regular expression in the get-pick-list.sh script to also
notice this pattern, (as well as the traditional "NOTE: This patch is a
candidate..." form.

10 years agoglsl: Remove redundant writes to prog->LinkStatus
Paul Berry [Sat, 27 Jul 2013 22:37:15 +0000 (15:37 -0700)]
glsl: Remove redundant writes to prog->LinkStatus

The linker_error() function sets prog->LinkStatus to false.  There's
no reason for the caller of linker_error() to also do so.

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
10 years agoglsl: Improve error message for interstage interface block mismatch.
Paul Berry [Sat, 27 Jul 2013 22:24:46 +0000 (15:24 -0700)]
glsl: Improve error message for interstage interface block mismatch.

We're now emitting this error from a point where we have easy access
to the name of the block that failed to match, so go ahead and include
that in the error message, as we do for intrastage interface block
mismatches.

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
10 years agoglsl: Use a consistent technique for tracking link success/failure.
Paul Berry [Sat, 27 Jul 2013 18:08:31 +0000 (11:08 -0700)]
glsl: Use a consistent technique for tracking link success/failure.

This patch changes link_shaders() so that it sets prog->LinkStatus to
true when it starts, and then relies on linker_error() to set it to
false if a link failure occurs.

Previously, link_shaders() would set prog->LinkStatus to true halfway
through its execution; as a result, linker functions that executed
during the first half of link_shaders() would have to do their own
success/failure tracking; if they didn't, then calling linker_error()
would add an error message to the log, but not cause the link to fail.
Since it wasn't always obvious from looking at a linker function
whether it was called before or after link_shaders() set
prog->LinkStatus to true, this carried a high risk of bugs.

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
10 years agoglsl: Add error message for intrastage interface block mismatch.
Paul Berry [Sat, 27 Jul 2013 21:58:43 +0000 (14:58 -0700)]
glsl: Add error message for intrastage interface block mismatch.

Previously we failed to link (which is correct), but we did not output
an error message, which could have been confusing for users.

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
10 years agoglsl: Remove bogus check on return value of link_uniform_blocks().
Paul Berry [Sat, 27 Jul 2013 22:07:08 +0000 (15:07 -0700)]
glsl: Remove bogus check on return value of link_uniform_blocks().

A comment in link_intrastage_shaders(), and an if-test that followed
it, seemed to indicate that link_uniform_blocks() would return a
negative value in the event of an error.  But this is not the
case--all error checking has already been performed by
validate_intrastage_interface_blocks(), and link_uniform_blocks() can
only return unsigned values.

So get rid of the if-test and change the return type of
link_intrastage_shaders() to clarify that it can only return unsigned
values.

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
10 years agor600g/compute: Added missing address space checking of kernel parameters
Jonathan Charest [Mon, 22 Jul 2013 13:24:56 +0000 (09:24 -0400)]
r600g/compute: Added missing address space checking of kernel parameters

To have non-static buffers in local memory, it is necessary to pass them
as arguments to the kernel.

For r600, the correct lds size must be set to the SQ_LDS_ALLOC register.
The correct size is the clover size plus the size reported by the
compiler.

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
10 years agoclover: Added missing address space checking of kernel parameters v2
Jonathan Charest [Wed, 24 Jul 2013 13:29:49 +0000 (09:29 -0400)]
clover: Added missing address space checking of kernel parameters v2

Here is an updated patch with no line wrapping and respecting 80-column limit (for my changes).

v2: Tom Stellard
  - Create global arguments for constant buffers so we don't break
    r600g.

Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
10 years agomesa: Remove broken assertion about enabled texture targets.
Kenneth Graunke [Thu, 25 Jul 2013 21:47:58 +0000 (14:47 -0700)]
mesa: Remove broken assertion about enabled texture targets.

For GLSL programs, enabledTargets can have more than one bit set.  For
example, a shader that uses sampler2D and samplerCube uniforms will have
both TEXTURE_2D_BIT and TEXTURE_CUBE_BIT set.

The code that sets _ReallyEnabled already handles this, selecting the
"highest priority" texture target.  We should simply use that.

Fixes new Piglit test incomplete-textures-of-multiple-types.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=62698
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
10 years agobuild: unify mesa version by using a VERSION file
Emil Velikov [Thu, 25 Jul 2013 22:45:45 +0000 (23:45 +0100)]
build: unify mesa version by using a VERSION file

Rather than having to keep track of all the build systems and their respecitve
definition of the mesa version, use a single top file VERSION. Every build
system is responsible for reading/parsing the file and using it

v2:
* remove useless bulletpoint from the documentation, suggested by Matt
* "Androing is Linux. Use '/' in stead of '\'", spotted by Chad V
* use cleaner code to get the version in scons, suggested by Chad V

v3:
* ensure leading and trailing whitespace characters are stripped while parsing
* android: handle GNU shell commands approapriately

Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Signed-off-by: Emil Velikov <emil.l.velikov@gmail.com>
10 years agoi965: Don't create a swrast context on ES2+.
Kenneth Graunke [Thu, 25 Jul 2013 18:31:20 +0000 (11:31 -0700)]
i965: Don't create a swrast context on ES2+.

We already skip this for API_OPENGL_CORE; ES2+ is very similar.
The primary user of the swrast context is GL_SELECT and GL_FEEDBACK,
which have never existed in ES.

This saves approximately 18MB of memory in GLBenchmark 2.7 Egypt (ES2).
No regressions in es3conform on Ivybridge.

Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
10 years agoglsl: Remove shader stage checking for extension handling.
Kenneth Graunke [Mon, 29 Jul 2013 17:40:51 +0000 (10:40 -0700)]
glsl: Remove shader stage checking for extension handling.

Certain extensions only add functionality to particular shader stages.
(For example, ARB_draw_instanced only adds variables to the vertex
shader stage.)

Previously, we only allowed such extensions to be enabled in the shader
stages where they're useful.  However, I've never found any text which
mandates that behavior; in my opinion, you should be able to turn on
extensions in any shader stage, even if they have no effect.

Fixes Piglit tests glslparsertest/glsl2/draw_buffers-05.vert and
ARB_draw_instanced/preprocessor/feature-macro-enabled.frag.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=29185
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agomesa: Expose OES_surfaceless_context.
Matt Turner [Fri, 26 Jul 2013 23:36:20 +0000 (16:36 -0700)]
mesa: Expose OES_surfaceless_context.

EGL_KHR_surfaceless_context extension allows contexts to be made current
without a default winsys fbo. This extension specifies what ES 1.1 and
2.0 should do (the ES 3.0 spec already does).

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agomesa: Return GL_FRAMEBUFFER_UNDEFINED if the winsys fbo is incomplete.
Matt Turner [Fri, 26 Jul 2013 23:36:19 +0000 (16:36 -0700)]
mesa: Return GL_FRAMEBUFFER_UNDEFINED if the winsys fbo is incomplete.

Specified by ARB_framebuffer_object, GL 3.0, and ES 3.0.

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
10 years agogles3: Update gl3.h to 2013-02-12.
Matt Turner [Fri, 26 Jul 2013 23:33:11 +0000 (16:33 -0700)]
gles3: Update gl3.h to 2013-02-12.

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
10 years agogles2: Update gl2ext.h to revision 22161.
Matt Turner [Fri, 26 Jul 2013 23:33:10 +0000 (16:33 -0700)]
gles2: Update gl2ext.h to revision 22161.

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
10 years agogles2: Update gl2.h to revision 20555.
Matt Turner [Fri, 26 Jul 2013 23:33:09 +0000 (16:33 -0700)]
gles2: Update gl2.h to revision 20555.

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
10 years agogles: Update glext.h to revision 20798.
Matt Turner [Fri, 26 Jul 2013 23:33:08 +0000 (16:33 -0700)]
gles: Update glext.h to revision 20798.

Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
10 years agogallivm: (trivial) git rid of assertion in float->uint conversion code
Roland Scheidegger [Mon, 29 Jul 2013 11:23:56 +0000 (13:23 +0200)]
gallivm: (trivial) git rid of assertion in float->uint conversion code

Commit 8c3d3622d9ce2fd2a8f46084ab8153d708fa5b09 introduced a new assertion,
but since it causes lp_test_conv failures remove it again and let's hope
we don't really hit bugs caused by the potentially bogus code (it is possible
the assert() caught some cases which work correctly too).

10 years agonvc0: force use of correct firmware file
Maarten Lankhorst [Sun, 28 Jul 2013 10:06:57 +0000 (12:06 +0200)]
nvc0: force use of correct firmware file

Signed-off-by: Maarten Lankhorst <maarten.lankhorst@canonical.com>
10 years agoglsl: Less const for glsl_type convenience accessors
Ian Romanick [Fri, 26 Jul 2013 20:37:29 +0000 (13:37 -0700)]
glsl: Less const for glsl_type convenience accessors

The second 'const' says that the pointer itself is constant.  This in
unenforcible in C++, so GCC emits a warning (see) below for each of
these functions in every file that includes glsl_types.h.  It's a lot of
warning spam.

../../../src/glsl/glsl_types.h:176:58: warning: type qualifiers ignored on function return type [-Wignored-qualifiers]

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
Cc: mesa-stable@lists.freedesktop.org
10 years agoglsl: Disallow auxiliary storage qualifiers on FS outputs.
Kenneth Graunke [Sat, 27 Jul 2013 04:18:56 +0000 (21:18 -0700)]
glsl: Disallow auxiliary storage qualifiers on FS outputs.

This has always been an error; we just forgot to check for it.

Fixes Piglit's no-aux-qual-on-fs-output.frag.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=67333
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Cc: mesa-stable@lists.freedesktop.org
10 years agoglsl: Classify "layout" like other identifiers.
Kenneth Graunke [Fri, 26 Jul 2013 19:31:06 +0000 (12:31 -0700)]
glsl: Classify "layout" like other identifiers.

When "layout" isn't being lexed as LAYOUT_TOK, we should treat it like
an ordinary identifier.  This means we need to classify it to determine
whether we should return IDENTIFIER, TYPE_IDENTIFIER, or NEW_IDENTIFIER.

Fixes the WebGL conformance test "shader-with-non-reserved-words."

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=64087
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Matt Turner <mattst88@gmail.com>
Cc: mesa-stable@lists.freedesktop.org
10 years agoglsl: Be consistent about '\n', '.', and capitalization in errors/warnings.
Paul Berry [Fri, 26 Jul 2013 02:56:43 +0000 (19:56 -0700)]
glsl: Be consistent about '\n', '.', and capitalization in errors/warnings.

The majority of calls to _mesa_glsl_error(), _mesa_glsl_warning(), and
_mesa_glsl_parse_state::check_version() use a message that begins with
a lower case letter and ends without a period.  This patch makes all
messages follow that convention.

Also, error/warning messages shouldn't end in '\n', since
_mesa_glsl_msg() automatically adds '\n' at the end of the message.

Reviewed-by: Matt Turner <mattst88@gmail.com>
10 years agogallivm: fix float->SNORM conversion
Roland Scheidegger [Sat, 27 Jul 2013 01:53:00 +0000 (03:53 +0200)]
gallivm: fix float->SNORM conversion

Just like the UNORM case we need to use round to nearest, not trunc.
(There's also another problem, we're using the formula for SNORM->float
which will produce a value below -1.0 for the most negative value which
according to both OpenGL and d3d10 would need clamping. However, no actual
failures have been observed due to that hence keep cheating on that.)

Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
10 years agoutil: don't flush overflowing values to infinity in half-float conversion
Roland Scheidegger [Sat, 27 Jul 2013 01:45:30 +0000 (03:45 +0200)]
util: don't flush overflowing values to infinity in half-float conversion

I am not able to find _any_ rounding behavior specified for OpenGL for
float to half-float conversions. However, it is specified for fp11/fp10
which suggests round to next finite value but round-to-zero would also
be allowed, but finite values must not be flushed to infinity in either
case.
Hence I believe it makes sense to do the same for half-floats too.
We could probably also use round-to-zero consistently, which is in fact
required by d3d10 (but it doesn't seem to matter much).
Does not match the mesa core function doing the same though (which is
saying it was built to match intel gpus which I don't believe for a
second as it would cause failures in d3d10, moreover the PRM (for
ivy bridge, not listed in older manuals) while not specifying rounding
behavior clearly states finite numbers are never flushed to infinity).

Reviewed-by: Jose Fonseca <jfonseca@vmware.com>