Age | Commit message (Collapse) | Author |
|
Pointed out by Chad Versace.
|
|
Changes:
- SINGLE is removed, because its value was 0
- ALPHA is removed, because it becomes redundant with this change
- RGB asks for an RGB visual
- RGBA asks for an RGBA visual
- if RGB and RGBA are unset, it asks for a visual without RGB(A)
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
v2: fix glut
|
|
Some of the ES tests check for this error code. If the installed
glxproto.h file isn't new enough this symbol may be missing and
compilation will fail.
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
|
|
Fixes piglit build since the shader_runner change to use this
function.
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
|
|
Reviewed-by: Eric Anholt <eric@anholt.net>
|
|
The commit below caused many shader_runner tests to skip on drivers that
lack GLX_ARB_create_context and EGL_KHR_create_context. This patch fixes
those regressions buy changing the way that Piglit handles requests for
compatibility contexts.
commit 3a92ffa0eb1efff8ccbe65b5c51710cea8ef0c5c
Author: Paul Berry <stereotype441@gmail.com>
Date: Fri Feb 8 13:04:13 2013 -0800
shader_runner and glslparsertest: choose GL version more accurately.
The regressing commit set test_config->supports_gl_compat_version to the
least version that is guaranteed to support the GLSL version used by the
test. If supports_gl_compat_version > 10, then a code path was taken in
Waffle that requires the driver to support GLX_ARB_create_context or
EGL_KHR_create_context.
This patch changes piglit_wfl_framework to *not* specify a GL version when
creating a compatibility waffle_context. If context creation succeeds,
Piglit checks that the actual context version satifies the requested
version and skips the test accordingly.
CC: Paul Berry <stereotype441@gmail.com>
Tested-by: Michel Dänzer <michel@daenzer.net>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Reviewed-by: Brian Paul <brianp@vmware.com>
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
This refactor makes the function body more readable.
Understanding the function no longer requires the reader to track the
value of `ok` throughout the function body. Also, some conditional
branches have been removed by using early returns.
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Some tests declare in their config block that they support running under
both core and compatibility contexts. In this case, Piglit first attempts
to use a core context; if core context creation fails, then Piglit falls
back to the compatibility context.
This patch prints a diagnostic when the fallback occurs.
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Jordan Justen <jordan.l.justen@intel.com>
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Previously, both shader_runner and glslparsertest contained switch
statements that attempted to choose a GL version for testing based on
a GLSL version requirement, and those switch statements didn't choose
a GL version very accurately (for example, they converted GLSL version
1.50 to GL version 3.1, which doesn't work).
This patch replaces the switch statements with a single common
function (in piglit-util-gl-common) that does the right thing.
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
|
|
Otherwise this prevents any interaction with the windows in non-automatic
mode, not even the Escape key or the window decoration close button works.
That's quite annoying for investigating test failures.
Thanks to Chad Versace for pointing me in the right direction for this on
IRC.
Reviewed-by: Brian Paul <brianp@vmware.com>
Signed-off-by: Michel Dänzer <michel.daenzer@amd.com>
|
|
Added GLDEBUGPROC typedef to piglit-dispatch.h.
This update is required for the ARB_texture_buffer_range extension tests
to get the definition of the TexBufferRange function.
EDITs of gl.spec: Fixed alias of GenVertexArraysAPPLE to be
GenVertexArrays instead of GenVertexArray (singular). (Fix already
present in piglit sources).
Removed GetTransformFeedbackVarying and TransformFeedbackVaryings aliases
in gl.spec from GetTransformFeedbackVaryingNV and TransformFeedbackVaryingsNV
(different signatures). (Fix already present in piglit sourcres).
Made varyings parameter to TransformFeedbackVaryingsEXT const to match the
non-EXT version and registry/specs/EXT/transform_feedback.txt.
Added constcharPointerARB to gl.tm and used this in ShaderSourceARB to
match constness of glShaderSource.
Made indices pointer in MultiDrawElementsEXT const to match
MultiDrawElements and registry/specs/EXT/multi_draw_arrays.txt.
v2 (Paul Berry <stereotype441@gmail.com>): restore fixes from commits
6409a67 and e9db1f0.
|
|
A very minor change to fix build issue for Android. Needed return
statement at end of function which specifies a return type.
Signed-off-by: Tom Gall <tom.gall@linaro.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
A very minor change to fix build issue for Android. Needed return
statement at end of function which specifies a return type.
Signed-off-by: Tom Gall <tom.gall@linaro.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Android is a bit more strict when it builds. Missing was ctype.h.
It doesn't break Linux.
Signed-off-by: Tom Gall <tom.gall@linaro.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
The existing code is intentionally disabling testing for channels with
1 bit of data, (since the ambiguity in rounding makes it infeasible to
test). But the disabling of testing for 0-bit channels was
unintentional collateral damage.
In particular, for RGB-only formats it's important to test that
results always provide a value of 1.0 for alpha.
So we change the condition to disable tests only when the number of
bits is equal to 1, (as opposed to less then 2). We also just fudge in
a tolerance value corresponding to 8 bits for channels with 0 bits of
storage. (The tolerance probably doesn't matter at all here since we
should really only be doing exact comparisons of values of 0.0 or
1.0).
Reviewed-by: Marek Olšák <maraeo@gmail.com>
|
|
We'll want this in GL tests where a subtest name will often be
composed of a couple of strings.
Reviewed-by: Carl Worth <cworth@cworth.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
This is the same as we have in Mesa.
Reviewed-by: Carl Worth <cworth@cworth.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
In the newly added piglit_gen_ortho_uniform it introduced a
warning due to float *[4] being passed instead of the expected
const GLfloat * as the type of the 4th param. Change the array
of floats to const GLfloat [4][4] and provide the cast.
Fixes the GCC warning:
tests/util/piglit-util-gl-common.c:547:2: warning: passing argument 4 of
‘glUniformMatrix4fv’ from incompatible pointer type [enabled by default]
Signed-off-by: Tom Gall <tom.gall@linaro.org>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Add a number of GLES2 specific workarounds to
shader_runner_gles_workarounds.h.
Add CMakeLists.gles2 into shaders directory so that
shader_runner is build when gles2 is selected.
Add piglit-vbo.cpp to util/CMakeLists.gles2.txt.
In tests/util/piglit-vbo.cpp add one small ifdef
PIGLIT_USE_OPENGL_ES2, vertex_attrib_description::setup
so the data_type GL_INTEGER for attributes is prevented
from use and emits an error message.
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
Signed-off-by: Tom Gall <tom.gall@linaro.org>
|
|
Add minmax test based on values that can be queried as specified
in chapter 6.2 of the OpenGL ES 2.0.25 spec.
Small modification to tests/util/minmax-test.* so that functions
which use GLint64 and GLuint64 types aren't compiled when
building for OpenGL ES 2.
Signed-off-by: Tom Gall <tom.gall@linaro.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
|
|
This patch adds compressed representations of uniform grayscale blocks
to piglit, in the ETC1 and FXT1 compression formats. Although they
don't exercise the full capabilities of these compression formats,
they should be sufficient to create basic tests of compression
functionality without having to rely on the GL implementation to do
compression on the fly.
Reviewed-by: Brian Paul <brianp@vmware.com>
Acked-by: Eric Anholt <eric@anholt.net>
v2: Add S3TC.
|
|
This patch adds functions piglit_gen_ortho_uniform() and
piglit_ortho_uniform(), which can be used to set up an orthographic
projection when the legacy function glOrtho() is not available.
The functions work exactly like piglit_gen_ortho_projection() and
piglit_ortho_projection(), except that instead of programming the
legacy projection matrix, they program an arbitrary uniform matrix.
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
The previous code would always add -cl-std if the version was > 1.0. This
ignored whatever the test itself wanted (e.g. the invalid-version-declaration
test). We should probably only add " -cl-std=x.y" when the test hasn't
requested a specific version.
Note: This doesn't address the rewrite of the intended functionality that Blaz,
Tom, and I discussed on the list during the week of Dec 3, 2012.
Reviewed-by: Blaž Tomažič <blaz.tomazic@gmail.com>
v2: Remove -D __OPENCL_C_VERSION__=1%d0 and -D CL_VERSION_1_0=100
Change the order in which the -cl-std checks are done.
v3: Whitespace cleanups
|
|
Subtests allow a test to produce more than one result. This is useful
for tests where it may be convenient to combine several somewhat
unrelated tests into a single program. For example, a cl test like:
kernel void test(TYPE *out, TYPE *in) {
out[0] = some_builtin_function(in[0], in[1]);
}
That uses only one kernel, but recompiles it several times with different
definitions of the macro 'TYPE' in order to test the builtin with all
the possible types.
To take advantage of the subtest feature, programs should output one
PIGLIT line per subtest as in the following example:
PIGLIT:subtest {'testA' : 'pass'}
PIGLIT:subtest {'testB' : 'fail'}
Where testA and testB are the name of the subtests.
In the result summary, this will be displayed as:
TestName 1/2
testA pass
testB fail
v2:
- Print one line for each subtest rather than printing them all
together.
Acked-by: Chad Versace <chad.versace@linux.intel.com>
|
|
v2:
- Simplify piglit_report_result()
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
The most notable changes are the removal of GL_TEXTURE_BUFFER_FORMAT
(this was not prototed to core from the ARB extension), and the
addition of a giant pile of new extesions.
Support is added to piglit_get_gl_enum_name for
GL_SHADER_BINARY_FORMATS which was previously missing from our
enumext.spec.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Paul Berry <stereotype441@gmail.com>
|
|
Fixes uninitialized scalar field defect reported by Coverity.
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
shader_runner will need this.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Fixes uninititalized pointer read defect reported by Coverity.
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
Add the following functions:
piglit_test_min_uint
piglit_test_max_uint
piglit_test_min_int64
piglit_test_max_int64
piglit_test_min_uint64
piglit_test_max_uint64
v2: Use PRId64 for printing GLint64, for Anholt.
v3: Add functions for uint, uint64.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
If piglit skips the test because the test neglected to set one of the
config attributes 'supports_gl_*_version', then print an explanation
before skipping.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Fixes uninitialized scalar field defect reported by Coverity.
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
This bug broke piglit_get_glsl_version on GLES.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
|
|
Eventually. gles3 support will be added shader_runner, and shader_runner
requires piglit-shader.[ch].
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
|
|
In the GLES1 implementations of functions
piglit_require_{GLSL,vertex_shader,fragment_shader}, print an explanation
before skipping.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
|
|
This patch kills the wrapper functions defined in piglit-shader.h for GL.
They are rendered obsolete by piglit-dispatch.
A subset of the wrappers are kept for GLES1 and GLES2.
- All GLES1 wrappers are kept and are renamed to match the core GL
names. That is, piglit_$func is renamed to gl$func. As before, the
wrappers simply raise an assertion failure because GLES1 does not
suppport shaders.
- The subset of GLES2 wrappers that raise an assertion, because the
function is not supported on GLES2, are also kept and likewise
renamed.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
Reviewed-by: Anuj Phogat <anuj.phogat@gmail.com>
|
|
This patch just adds a new file, tests/util/CMakeLists.gles3.txt, which is
closely based off tests/util/CMakeLists.gles2.txt.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
|
|
For each preprocessor directive that examines the PIGLIT_USE_OPENGL*
feature macros, extend the directive to support building with
PIGLIT_USE_OPENGL_ES3.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
|
|
In libpiglitutil_gles3, piglit-util-gl-enum.c will conflict with
piglit-util-gles3-enum.c. Therefore, require that each
test/util/CMakeLists.$api.txt add it to the sources list if needed.
v2: Add piglit-util-gl-enum.c to libpiglitutil_gl.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
|
|
1. Create a new file, piglit-util-gles3-enum.c.
2. Copy <GLES3/gl3.h> into it.
3. Massage it to resemble piglit-util-gl-enum.c.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
|
|
Only define piglit_test_oq_bits() and piglit_test_tf_bits() if the feature
macro PIGLIT_USE_OPENGL is defined. This will allow us to use
minmax-test.c for libpiglitutil_gles3.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
|
|
Allow piglit_gl_test_config::supports_gl_es_version to be >= 30. Piglit
passes the requested version directly to Waffle.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
|
|
This field allows a test to declare support for any GLES version. It
replaces fields 'supports_gl_es1' and 'supports_gl_es2'.
The valid range for the field is currently [1.0, 3.0), though it will soon
get bumped to [1.0,4.0).
One advantage of this field is that it introduces symmetry among the
'supports_gl' fields. Before this patch, two fields were integers
(supports_gl_core_version, supports_gl_compat_version) and two were
booleans (supports_gl_es1, supports_gl_es2). Now they're all integers and
behave similarly.
This patch also updates, with the substitution below, the tests that used
the old config fields.
supports_gl_es2 = true --> supports_gl_es_version = 20
supports_gl_es1 = true --> supports_gl_es_version = 10
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
|
|
OpenBSD is one platform that does not define the symbol RLIMIT_AS.
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
|
|
1) Fix creation of libpiglitutil_cl.dylib
2) Fix includes of OpenCL.h to finish fixing MacOS build
Previous error message when linking dylib was:
Linking C shared library ../../../../lib/libpiglitutil_cl.dylib
Undefined symbols for architecture x86_64:
"_piglit_cl_get_test_config", referenced from:
_piglit_cl_framework_run in piglit-framework-cl.c.o
ld: symbol(s) not found for architecture x86_64
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
|
|
If piglit_gl_test_config::supports_gl_compat_version was >= 32, then
Piglit neglected to set the WAFFLE_CONTEXT_PROFILE attribute.
Luckily this introduced no errors, because Piglit does not yet contain any
3.2 tests.
Piglit also neglected to set the profile attribute when
piglit_gl_test_config::supports_gl_core_version was >= 32, but this was
safe because the default value of WAFFLE_CONTEXT_PROFILE is
WAFFLE_CONTEXT_CORE_PROFILE for GL versions >= 3.2. However, explicitly
set it anyway for self-documentation.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Previously, piglit_expect_egl_error() would terminate the test if an
unexpected error occurred. However, there are many cases when we would
prefer to simply report the error and continue testing.
Commit 83c8ef80 made an analagous change to piglit_check_gl_error(). For
consistency with that function, this commit also renames
piglit_expect_egl_error to piglit_check_egl_error.
In all cases, I replaced
piglit_expect_egl_error(err, PIGLIT_FAIL)
with
if (!piglit_check_egl_error(err))
piglit_report_result(PIGLIT_FAIL)
or some equivalent.
Reviewed-by: Ian Romanick <idr@freedesktop.org>
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Reviewed-by: Brian Paul <brianp@vmware.com>
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
No test actually called glewInit(), because piglit-dispatch #define'd it
to piglit_dispatch_default_init().
This patch removes the macro and replaces each call to glewInit() with
piglit_dispatch_default_init().
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
On Win7 windows get resized to the min size of 116x100 pixels (which
might depend on the Aero theme, etc). This resize is detected by
the piglit framework:
Got spurious window resize in automatic run (100,100 to 116,100)
PIGLIT: {'result': 'warn'}
so we get a 'warn' result instead of actually running the test.
This sets a default window size of 150x150 pixels. Next, we'll
remove the "config.window_width/height = 100" lines from a bunch
of tests.
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|