summaryrefslogtreecommitdiff
AgeCommit message (Collapse)AuthorFilesLines
2012-11-28igt: auto-enumerate subtestsigtDaniel Vetter1-4/+22
This works together with the new subtest infrastructure in intel-gpu-tools. v2: An empty string crept into the subtest list, resulting in a testcase failure. Filter it out. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
2012-11-28framework: unoicode() wrap format stringDaniel Vetter1-1/+1
Otherwise python2 blows up when the program output contains a non-ascii character. The resulting exception caused a bunch of spurious fails in the igt testscases. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
2012-11-28igt: new testinfrastructure to support drm/i915 kernel testsDaniel Vetter1-0/+101
Piglit as a testrunner is _so_ much better than automake's make check. Based upon a quick patch from Kenneth Graunke. v2: Add copyright header. Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
2012-11-28framework: add returncode parameter to ExecTest.interpretResultDaniel Vetter3-5/+5
The intel-gpu-tools testrunner needs this to figure out the testresult (since 'skip' is indicated by a returncode of 77 in i-g-t). Signed-off-by: Daniel Vetter <daniel.vetter@ffwll.ch> Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
2012-11-28egl: Fix typo in previous commitMatt Turner1-1/+1
2012-11-27egl_khr_create_context: Fix build error with old eglext.hChad Versace1-0/+7
This patch defines EGL_OPENGL_ES3_BIT_KHR, for the case when eglext.h is too old. Reported-by: Ken Phillis Jr <kphillisjr@gmail.com> Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
2012-11-26glsl-1.40: Add a test for a path in i965 where I had an XXX comment.Eric Anholt1-0/+39
2012-11-26glsl-1.40: Add a new UBO test for a potential bug in the i965 driver.Eric Anholt1-0/+42
2012-11-26glsl-es-3.00: Verify that .length() returns a constant expressionIan Romanick1-0/+28
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
2012-11-26glsl-es-3.00: Verify that .length() returns a signed integerIan Romanick3-0/+81
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
2012-11-26glsl-es-3.00: Verify that no statment is allowed before the first case in a ↵Ian Romanick2-0/+31
switch Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
2012-11-26glslparsertest: Add support for GLSL ES 3.0 via GL_ARB_ES3_compatibilityIan Romanick1-0/+2
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
2012-11-26glslparsertest: Use integers for GL and GLSL versionsIan Romanick1-21/+33
I haven't observed any failures with this yet. Since we just did the same thing with shader_runner, it seems like a good idea. v2: Fix two off-by-10x errors pointed out by Chad. Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
2012-11-26time-elapsed: Spend time in the FS rather than do a lot of draws.Paul Berry1-4/+53
The time-elapsed test needs to consume a lot of GPU time so that it can validate GPU time queries against elapsed wall-clock time. Previously we did this by painting the window thousands of times. That sometimes caused false failures due to the CPU time consumed in preparing the draw calls. This patch changes the test so that it consumes time by using a slow fragment shader rather than by painting the window over and over.
2012-11-22Revert "texture-integer-glsl130: Expect 1 for alpha in R, RG, RGB cases."Kenneth Graunke1-2/+3
This reverts commit cf47699087f74a6fd94c59a74837fc6c3c6422a8. While I still believe the test is wrong, apparently this change made it worse; the test doesn't quite work like I thought. More investigation is needed. In the meantime, restore the old behavior.
2012-11-21util: Do not use setrlimit/getrlimit if RLIMIT_AS is not defined.Vinson Lee1-1/+1
OpenBSD is one platform that does not define the symbol RLIMIT_AS. Signed-off-by: Vinson Lee <vlee@freedesktop.org>
2012-11-21texture-integer-glsl130: Expect 1 for alpha in R, RG, RGB cases.Kenneth Graunke1-3/+2
Table 3.23 (on page 221) of the OpenGL 3.0 specification is: +----------------------------------------+ | Texture Base | Texture source color | | Internal Format | C_s A_s | +-----------------+----------------------+ | ALPHA | (0, 0, 0) | A | | LUMINANCE | (L, L, L) | 1 | | LUMINANCE_ALPHA | (L, L, L) | A | | INTENSITY | (I, I, I) | I | | RED | (R, 0, 0) | 1 | | RG | (R, G, 0) | 1 | | RGB | (R, G, B) | 1 | | RGBA | (R, G, B) | A | +-----------------+----------------------+ From this, we see when sampling from RED, RG, and RGB textures, alpha is supposed to be one. v2: Change commit message to quote the right part of the spec. Reviewed-by: Eric Anholt <eric@anholt.net>
2012-11-20egl-create-context-verify-gl-flavor: Test bogus versionsChad Versace1-3/+45
Attempt to create contexts with bogus versions. Tests Mesa commit 243cf7a924eaef78ce0d5150747fae6c3c4e6974. Before that commit, the Intel driver failed to validate requested OpenGL ES context versions. Reviewed-by: Ian Romanick <ian.d.romanick@intel.com> Reviewed-by: Kenneth Graunke <kenneth@whitecape.org> Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
2012-11-20egl_khr_create_context: Check eglChooseConfig's errorsChad Versace1-0/+9
If eglChooseConfig fails with EGL_BAD_ATTRIBUTE, then fail the test. Piglit only requests valid attributes, therefore the error is incorrect. Regresses test egl-create-context-verify-gl-flavor:gles3.0 on mesa-5cf853669. Mesa fails to recoginize EGL_OPENGL_ES3_BIT_KHR despite exposing the EGL_KHR_create_context extension. Reviewed-by: Ian Romanick <ian.d.romanick@intel.com> Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
2012-11-20egl-create-context-verify-gl-flavor: Test GLES3Chad Versace1-0/+17
- Request a config with the es3 bit, EGL_OPENGL_ES3_BIT_KHR. - Try to create a GLES 3.0 context with the config. - Verify that the context version is >= 3.0. Reviewed-by: Ian Romanick <ian.d.romanick@intel.com> Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
2012-11-20glx-multithread: Replace glFlush calls with glFinishPaul Berry1-2/+2
Currently, in Mesa with the i965 driver, calling glFlush() forces all pending drawing operations to be sent to the kernel; this ensures that any subsequent drawing operations will be performed after those pending drawing operations, even if they come from other contexts. The glx-multithread test was relying on this behaviour, however, the GL and GLX specs do not garantee this. This patch modifies the glx-multithread test to use glFinish() instead of glFlush(). Reviewed-by: Chad Versace <chad.versace@linux.intel.com> Reviewed-by: Eric Anholt <eric@anholt.net>
2012-11-20glx-multithread: use XInitThreadsPaul Berry1-0/+1
To use multithreading with xlib, you're supposed to call XInitThreads before any other xlib call. Without this, xlib and GLX calls won't be thread safe. (Based on a patch by Ian Romanick <ian.d.romanick@intel.com>)
2012-11-17all.tests: Remove glean/rgbTriStrip.Kenneth Graunke1-1/+0
It doesn't actually test anything, and thus can never fail. Reviewed-by: Eric Anholt <eric@anholt.net>
2012-11-16glslparsertest: Use specified GLSL version to select API versionIan Romanick1-2/+35
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Reviewed-by: Eric Anholt <eric@anholt.net>
2012-11-16shader_runner: Partially process the requirements section before creating ↵Ian Romanick1-1/+101
the context A test script can require a specific GLSL version or a specific GL version. To satisfy this requirement, the piglit framework code needs to know about the requirement before creating the context. However, the requirements section can contain other requirements, such as minimum number of uniforms. The requirements section can't be fully processed until after the context is created, but the context can't be created until after the requirements section is processed. Do a quick can over the requirements section to find the GL and GLSL version requirements. Use these to guide context creation. v2: Use 3-digit GLSL versions as suggested by Eric. Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Reviewed-by: Chad Versace <chad.versace@linux.intel.com> Reviewed-by: Eric Anholt <eric@anholt.net>
2012-11-16shader_runner: Set a VAO in a GL 3.1 contextIan Romanick1-1/+9
There's no default VAO (without GL_ARB_compatibility), so we have to create and bind one. Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Reviewed-by: Chad Versace <chad.versace@linux.intel.com> Reviewed-by: Eric Anholt <eric@anholt.net>
2012-11-16shader_runner: Use integers for GL and GLSL versionsIan Romanick1-26/+40
In floating point, 3.1 is actually 3.0999999999, and that's just awsome for comparing GL versions. v2: Use 3-digit GLSL versions as suggested by Eric. Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Reviewed-by: Eric Anholt <eric@anholt.net>
2012-11-16shader_runner: Refactor version comparisonsIan Romanick1-8/+11
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Reviewed-by: Chad Versace <chad.versace@linux.intel.com> Reviewed-by: Eric Anholt <eric@anholt.net>
2012-11-16glx: Sanity check the various GLX extension strings that an app can queryIan Romanick3-0/+191
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com> Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=56057
2012-11-16Fix discovery of OpenCL framework on MacOS.Aaron Watry1-2/+7
Tested on Mac OS 10.7 Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
2012-11-16Fix MacOS build for OpenCL testsAaron Watry3-3/+10
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>
2012-11-16Fix test macro-definitions.clAaron Watry1-2/+2
The original version defined BUILD_OPT, but used BUILD_OPT1. Clover wasn't capable of passing the test either way. Now, the test should pass on conforming implementations. Reviewed-by: Tom Stellard <thomas.stellard@amd.com>
2012-11-16teximage-errors: remove config.window_width/height linesBrian Paul1-3/+0
So the test isn't skipped (with "warn" result) on Windows.
2012-11-15Add a test for viewport triangle clipping.Stuart Abercrombie3-0/+80
Triangles should be clipped to the clip volume and therefore shouldn't end up being rasterized outside the viewport. This was failing on Sandy Bridge with guard band clipping enabled. v2: new test name, more comments and addition to all.tests. v3: Eric's comments incorporated, and the explicit window size removed. v4: minor style tweaks by Eric, and Brian's probe-related review. Reviewed-by: Eric Anholt <eric@anholt.net>
2012-11-15GL_ARB_framebuffer_object/negative-readpixels-no-rb: Test for Mesa segfault.Eric Anholt3-0/+87
v2: Drop the window size setup. Reviewed-by: Brian Paul <brianp@vmware.com>
2012-11-15glsl-max-varyings: Add a variant that also tests >MAX_VARYING_COMPONENTS.Eric Anholt2-10/+38
Given that piglit hasn't been testing this, I don't want to modify the existing glsl-max-varyings which has been very useful. I don't want a non-rendering test, though, because gl 2.0 allows an implementation to allow shaders with >MAX_VARYING_COMPONENTS, so we should actually test the rendering if that does happen. v2: Make the error/warning message clearer. Reviewed-by: Brian Paul <brianp@vmware.com>
2012-11-15glsl-max-varyings: Make this a concurrent test.Eric Anholt1-1/+1
It's a really slow one, and I'm adding another variant of it. Reviewed-by: Brian Paul <brianp@vmware.com>
2012-11-15gitignore: Ignore ninja artifactsChad Versace1-0/+5
Ninja is a wicked-fast build system backend supported by CMake.
2012-11-14util/gl: Set WAFFLE_CONTEXT_PROFILE for compatibility profilesChad Versace1-0/+10
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>
2012-11-14textureSize: Add support for running with core GL.Eric Anholt1-2/+16
This lets the textureSize tests for TBOs run on i965. Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
2012-11-14GL_ARB_texture_buffer_object: Add support for GL core to a few other tests.Eric Anholt5-0/+5
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
2012-11-14GL_ARB_texture_buffer_object/formats: Add real support for testing GL core.Eric Anholt1-27/+69
Thanks to waffle we can finally do this. Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
2012-11-14Revert "GL_ARB_ubo/negative-bindbuffer-buffer: New test for API error."Eric Anholt3-72/+0
This reverts commit ba79f8c3f85a479e97d43f41bf6d2064a1e174fe. This test has been replaced with gl-3.0/genned-names, and the pre-GL-3.0 behavior is relatively unspecified and uninteresting. Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
2012-11-14clearbuffer-display-lists: Fix initial clear color test.Eric Anholt2-3/+3
The test claimed that the default value was 0,0,0,1, but the GL 3.0 spec page 263 disagrees. The test saw 0,0,0,1 when it was written because a visual with no alpha happened to be chosen, but there's no guarantee that you don't have alpha, so fix the test, and ask for alpha in the test so that we can reliably check those bits. Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
2012-11-14gl-3.1: Add a new test for genned object name requirements.Eric Anholt3-0/+184
Reviewed-by: Chad Versace <chad.versace@linux.intel.com>
2012-11-14GTF: Ask GTF to use a minimal visual chosen by EGL.Eric Anholt1-1/+1
2012-11-14egl: Add test egl-create-context-verify-gl-flavor (v3)Chad Versace3-0/+442
This test requests various flavors (that is, api and version and profile) of contexts and verifies that the context's actual flavor is compatible with the requested flavor. The following subcases fail on mesa-a196f43 with Intel Sandybridge: - For each OpenGL context >= 3.2, context creation fails with EGL error EGL_SUCCESS. - For OpenGL ES 3.0, the context version reported by glGetString is 2.0. All other cases skip or pass. v2: - Document that we use eglGetProcAddress to avoid piglit-dispatch. - Also verify that glGetString(GL_VERSION) and glGetIntgerv(GL_MAJOR/MINOR_VERSION) return the same version. v3: - Remove incorrect test for ES3. Reviewed-by: Iam Romanick <idr@freedesktop.org> (v2) Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
2012-11-14util/egl: Change piglit_expect_egl_error to return a boolChad Versace11-24/+39
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>
2012-11-14util: Remove unused wglew.hChad Versace1-1363/+0
Reviewed-by: Brian Paul <brianp@vmware.com> Signed-off-by: Chad Versace <chad.versace@linux.intel.com>
2012-11-14util: Kill macro glewInitChad Versace26-33/+26
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>
generated by cgit v1.2.3 (git 2.39.1) at 2024-05-15 22:22:55 +0000