Age | Commit message (Collapse) | Author | Files | Lines |
|
Instead of symlinking igt into the bin/ directory, edit piglit.conf and
set the path key in the igt section to the root of intel-gpu-tools
v2: - change config key 'bindir' to 'path' (chadv)
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
Signed-off-by: Dylan Baker <baker.dylan.c@gmail.com>
|
|
Instead of symlinking a directory into bin/ for xts, edit the path entry
under the xts section to point at the root xts directory.
v2: - replace config file key 'bindir' with 'path' (chadv)
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
Signed-off-by: Dylan Baker <baker.dylan.c@gmail.com>
|
|
Mesa has an optimization that converts expressions like "v.x + v.y + v.z
+ v.w" into dot(v, 1.0). And therein lies the rub: the other operand to
the dot-product is always a float... even if the vector is an ivec or
uvec. The result is not pretty:
shader_runner: ../../src/glsl/ir_builder.cpp:257: ir_expression* ir_builder::dot(ir_builder::operand, ir_builder::operand): Assertion `a.val->type == b.val->type' failed.
Aborted (core dumped)
This was fixed on Mesa master by commit 316dafa.
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Matt Turner <mattst88@gmail.com>
|
|
Tested by tests/shaders/glsl-fs-frontfacing.shader_test.
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
|
|
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
|
|
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
|
|
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
|
|
Silence clang sometimes-uninitialized warning.
accuracy.cpp:130:13: warning: variable 'test_type' is used uninitialized whenever 'if' condition is false [-Wsometimes-uninitialized]
} else if (strcmp(argv[2], "depth_resolve") == 0) {
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
accuracy.cpp:135:21: note: uninitialized use occurs here
test = create_test(test_type, num_samples, small,
^~~~~~~~~
accuracy.cpp:130:9: note: remove the 'if' if its condition is always true
} else if (strcmp(argv[2], "depth_resolve") == 0) {
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
accuracy.cpp:119:2: note: variable 'test_type' is declared here
test_type_enum test_type;
^
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
|
|
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
|
|
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
|
|
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
|
|
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
|
|
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
|
|
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
|
|
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
|
|
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
|
|
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
|
|
The GLSL 1.10 tests currently fail (we accept floating-point suffixes
regardless of language version).
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
|
|
And throw out the glslparsertests.
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
|
|
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
|
|
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
Reviewed-by: Dylan Baker <baker.dylan.c@gmail.com>
|
|
In GLES, piglit_rgbw_texture() ignored the data type parameter and
assumed it was GL_UNSIGNED_BYTE. Remove the GLES-specific assumption and
respect the parameter on GL and GLES.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Specify the data type, s/0/GL_UNSIGNED_BYTE/. The call worked because
piglit_rgbw() ignores the type parameter on GLES.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Drop the now-empty piglit-util-gl.c.
Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
With the image generation factored out to piglit_rgbw_image and
piglit_rgbw_image_ubyte, the remainder of the implementation becomes
common between GL and GLES.
This eliminates the last helper function from piglit-util-gles.c, so
drop it.
Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
This doesn't need to remain in the GL-specific utility library.
Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
This moves the GLES implementation of piglit_rgbw_texture much closer to
the GL version.
Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
In the GLES implementation, factor out the conversion of the white and
black pixel values from float to ubyte outside the loop. That then
unifies the loop between the two implementations.
Update the documentation to note that piglit_checkerboard_texture sets
GL_CLAMP_TO_EDGE on GLES, not GL_CLAMP_TO_BORDER as it does on GL.
Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
These don't need to remain in the GL-specific utility library.
Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
These don't need to remain in the GL-specific utility library.
Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
These don't need to remain in the GL-specific utility library.
Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
These don't need to remain in the GL-specific utility library.
Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Introduce a new internal helper, piglit_read_pixels_float, that always
returns floats. On GL it just calls glReadPixels with GL_FLOAT; on GLES
it uses GL_UNSIGNED_BYTE and converts.
Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
piglit_gen_ortho_projection and piglit_ortho_projection already had a
similar implementation for GL and GLES1; common it up and remove the
ifdefs.
Move piglit_frustum_projection out of GL-only code as well, and add
support for GLES1 using glFrustumf.
Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
This doesn't need to remain in the GL-specific utility library.
Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
The unified implementation handles all the GLenum values expected by
either.
Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Piglit looked for a GL_VERSION starting with "OpenGL ES ", which did not
match "OpenGL ES-CM 1.1 Mesa 10.2.1". Stop requiring the trailing
space.
Signed-off-by: Josh Triplett <josh@joshtriplett.org>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
|
|
Fix clang sometimes-uninitialized warning.
get-platform-ids.c:62:5: warning: variable 'platforms' is used uninitialized whenever 'if' condition is true [-Wsometimes-uninitialized]
if(!piglit_cl_check_error(errNo, CL_SUCCESS)) {
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
get-platform-ids.c:93:30: note: uninitialized use occurs here
errNo = clGetPlatformIDs(0, platforms, NULL);
^~~~~~~~~
get-platform-ids.c:62:2: note: remove the 'if' if its condition is always false
if(!piglit_cl_check_error(errNo, CL_SUCCESS)) {
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
get-platform-ids.c:56:27: note: initialize the variable 'platforms' to silence this warning
cl_platform_id* platforms;
^
= NULL
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
|
|
This is a copy of the render_viewport ARB_viewport_array test. I didn't
think that it was worth it to copy the scissor/depth tests, since they
basically test the same thing but making sure that scissor/depth are
applied correctly. If that works for ARB_viewport_array, it really
should work here too.
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
|
|
Currently, this test triggers an llvm error on r600g/radeonsi.
Signed-off-by: Aaron Watry <awatry@gmail.com>
CC: Tom Stellard <thomas.stellard@amd.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=76223
|
|
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Chris Forbes <chrisf@ijw.co.nz>
|
|
Fix clang unused-const-variable warning.
egl-create-surface.c:40:21: warning: unused variable 'pixmap_attribs' [-Wunused-const-variable]
static const EGLint pixmap_attribs[] = {
^
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|
|
So the usage/help info agrees with the actual command line argument.
Trivial.
|
|
Commit 1fdb953 removed some noisy debug messages with preprocessor
guards. This patch re-enables them, but only if the environment variable
PIGLIT_DEBUG is "1" or "true".
Fixes some unused variable warnings.
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
|
|
The function logs a debug message if environment variable
PIGLIT_DEBUG is "1" or "true".
Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
|
|
Commit 04c9e870 introduced some debugging prints on success that make it
clearer what's going on. However having these output for ever test seems
unnecessary. Hide them behind compilation guards.
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
|