| Age | Commit message (Collapse) | Author | Files | Lines |
|---|
|
This splits the FastSkipMixin into two classes. One that actually does
the fast skipping, and one that provides a mixin for Test.
This split will allow the class to be used in the traditional way, but
also in shader_test.MultiShaderTest.
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
|
|
This is both nicer since it provides better code quality (a separation
of roles), and because it's going to make writing a shader_test class
that can do N:1 tests much easier.
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
|
|
This makes the behavior of the Subtests class more like that of the
TestDict class, and fixes a bug in MultiShaderRunner, which will be
added in a subsequent patch in this series.
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
|
|
Which was broken by the last commit touching the logger, as were several
unittests that continued passing because of a dirty environment.
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
|
|
Otherwise there will be a spinner in the final call to summary that's
printed when piglit finishes.
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
|
|
This patch catches C-c, and any Exception while tests are running,
prints the status again with a newline, and then raises the exception.
This makes the output visible even when a stacktrace is printed.
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
|
|
This test passes with NVIDIA blob.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
|
|
compare-with-shader-subroutine.c: In function 'get_shadertype_from_program_interface':
compare-with-shader-subroutine.c:251:1: warning: control reaches end of non-void function [-Wreturn-type]
}
^
Fixes: 39ccf8c59bd5 ("arb_program_interface_query: add compare-with-shader-subroutine test")
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
|
|
This is a regression test for a very specific compiler bug in radeonsi.
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97887
Reviewed-by: Tapani Pälli <tapani.palli@intel.com>
|
|
tests/spec/arb_viewport_array/queries.c: In function ‘piglit_init’:
tests/spec/arb_viewport_array/queries.c:58:11: warning: unused variable ‘vald’
GLdouble vald[4] = {0.0, 0.0, 0.0, 0.0};
Which is only used if compiled with OpenGL support, but not when
compiled for GLES.
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
Reviewed-by: Vinson Lee <vlee@freedesktop.org>
|
|
tests/spec/ext_image_dma_buf_import/refcount.c:59:6: warning: unused variable ‘cpp’
int cpp = 4;
^~~
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
Reviewed-by: Vinson Lee <vlee@freedesktop.org>
|
|
This patch fixes this build error.
transcode-nv12-as-r8-gr88.c: In function ‘create_textures’:
transcode-nv12-as-r8-gr88.c:178:7: error: ‘DRM_FORMAT_R8’ undeclared (first use in this function)
DRM_FORMAT_R8, r8_pixels);
^
Fixes: 0222f5db00fd ("dmabuf: fix YUV tests for drivers other than intel")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97951
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
|
|
v3: - enforce local-size.c test (Nicolai)
- add mixed_variable_fixed_local_work_size.comp (Nicolai)
v2: - update formatting spec quotations (Nicolai)
- use shader_runner for the basic test (Ian)
- add a new test which checks various local sizes (Nicolai)
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
|
|
This allows to dispatch compute shaders with a variable local size
using the new "compute group size" command.
Signed-off-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
Reviewed-by: Nicolai Hähnle <nicolai.haehnle@amd.com>
|
|
If they're not defined by the installed libdrm already.
This moves code already defined in transcode-nv12-as-r8-gr88 to the
utils level, so that all tests that require these formats get them, and also
adds on additional define that wasn't in the mentioned test.
bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=97951
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
Tested-by: Vinson Lee <vlee@freedesktop.org>
|
|
This test was the opposite of standard piglit style, which was a
surprise.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
|
|
Mesa eliminates the loop and therefore the loop_count uniform, which
causes shader_runner to report failure when it tries to set it.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
|
|
Ok, so the basic problem with the YUV tests is that they currently
completely ignore driver/hw pitch requirements, since the code that
allocates the buffer doesn't know the pixel format, only the 'cpp'.
The yuv test creates a small 4x4 yuv eglimage. If, say, the hardware
requires the pitch to be aligned to, say, 32pixels, everything is fine
for the Y plane, but the subsampled U/V or U+V plane has half as many
pixels. (This did help me catch a bug in driver, not rejecting the
dmabuf import with invalid pitch, but that doesn't help to get the
piglit tests running.)
The best approach I could come up with to fix this is to pass the
fourcc all the way down to the code that creates the dmabuf (and copies
src data into the dmabuf). Unfortunately this makes the patch a bit
bigger than I was hoping, and not really sure a good way to split it
up.
This is tested on i965 (with the intel dma-buf backend) and freedreno
(with the gbm dma-buf backend). In the gbm case, it requires new
gbm format values for R8 and GR88, which is on mesa master as of
this morning. (So I bumped the gbm version dependency to 12.1.)
Signed-off-by: Rob Clark <robdclark@gmail.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Kristian Høgsberg <krh@bitplanet.net>
|
|
Fixes races in the X Server's make check environment, which runs
multiple "piglit-run.py -t xts" in parallel and would have the two
runs overwriting each other's logs in between execution and parsing.
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
|
|
Adapted for GL ES. Note that if GS does not write a viewport index (or
there is none), the value of gl_ViewportIndex is undefined, unlike in
desktop GL.
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Acked-by: Kenneth Graunke <kenneth@whitecape.org>
|
|
Mesa decides to create it as a float texture (since it's an unsized
internal format), but then it can't be rendered to in ES (without having
been created via a sized internal format).
To facilitate matters, just use GL_UNSIGNED_BYTE, since we really don't
care about the format here.
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Acked-by: Kenneth Graunke <kenneth@whitecape.org>
|
|
mesa 11.0 is needed for EGLDeviceEXT symbol.
egl_ext_device_query.c: In function ‘main’:
egl_ext_device_query.c:33:2: error: unknown type name ‘EGLDeviceEXT’
EGLDeviceEXT device = EGL_NO_DEVICE_EXT;
^
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
|
|
A bunch of additions to the ARB_query_buffer_object test:
1. test different result types (including 64 bit types)
2. compare the written result with equality to the result obtained by
querying on the CPU
3. add an additional mode to catch cache invalidation problems
4. additional checks that no values are written when they shouldn't
5. proper extension checks for the different query types
Also, rename the binary for the test to fit the extension-testname pattern.
Acked-by: Alejandro Piñeiro <apinheiro@igalia.com>
|
|
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
|
|
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
|
|
These tests used a mix of GL 1.x and newer functionality. This switches
all of them up to require GL 3.0 to avoid annoyance. All hardware that
can support texture views is GL 3.0-capable.
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
|
|
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
|
|
This makes some fairly simple changes to the existing tests to also work
with GL_OES_viewport_array, which presents largely identical
functionality. All of the tests continue to pass with desktop GL, and
also work with my soon-to-be-posted GL ES implementation for mesa.
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
|
|
needed to exercise radeonsi codepaths
|
|
|
|
In python2 the code as is works, although there might be odd corners. In
python 3 it fails because the code tries to match bytes and strs, which
doesn't work at all.
This patch fixes the issue by converting the output from bytes to str
(unicode) immediately, so we don't have to think about it again.
Reported-by: Dan Kegel <dank@kegel.com>
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
|
|
Ensure that it converts bytes to str and also collects correctly.
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
|
|
Since it's being used.
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
|
|
Currently fails with llvmpipe. Fragments with alpha <= 0.5 are discarded.
Passes with softpipe, NVIDIA driver.
v2: test with GL_MULTISAMPLE both enabled and disabled, fix all.py entry
Reviewed-by: Roland Scheidegger <sroland@vmware.com>
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
|
|
This tests the effect of PBO operation when GL_FRAMEBUFFER_SRGB is enabled.
Passes with Mesa, fails with NVIDIA.
VMware bug 1426768
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
See https://github.com/anholt/trace-db for a repository of traces and
their references images.
v2: Add a note about were to find some traces to put in here.
Reviewed-by: Dylan Baker <dylan@pnwbakers.com> (v1)
|
|
v2: Apply feedback from Dylan
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
|
|
For the apitrace test runner, I need to be able to store known-good
images for various drivers, which means I need some way to group
drivers. Not all Mesa drivers we care about are covered here, but
this is a sample to get things started.
v2: Apply feedback from Dylan, add classification for my Skylake.
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
|
|
These test so called complex loop unrolling in Mesa which just means
unrolling a loop with two exits where the trip count is only known
for one of the exits.
Reviewed-by: Alejandro Piñeiro <apinheiro@igalia.com>
|
|
This reduces the chance of false positives caused by system load:
On an 8-core test machine with an AMD Polaris, I've observed about a 50%
false positive rate on these tests when running concurrently. On the same
system, I have not observed a single failure in 10 runs when running
non-concurrently.
|
|
Fixes sometimes-uninitialized warning.
piglit-util-cl.c:1052:13: warning: variable 'errNo' is used uninitialized whenever 'if' condition is false [-Wsometimes-uninitialized]
} else if (desc->image_type == CL_MEM_OBJECT_IMAGE3D) {
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
piglit-util-cl.c:1062:28: note: uninitialized use occurs here
if(!piglit_cl_check_error(errNo, CL_SUCCESS)) {
^~~~~
piglit-util-cl.c:1052:9: note: remove the 'if' if its condition is always true
} else if (desc->image_type == CL_MEM_OBJECT_IMAGE3D) {
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
piglit-util-cl.c:1040:14: note: initialize the variable 'errNo' to silence this warning
cl_int errNo;
^
= 0
Fixes: 910cd6c222930 ("cl: add image and sampler utility functions")
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Acked-by: Jan Vesely <jan.vesely@rutgers.edu>
Reviewed-by: Serge Martin <edb+piglit@sigluy.net>
|
|
This reverts commit 5b037e06730fb4d219c3e5f374ad96594e10e3b1.
Sorry I pushed the wrong version.
|
|
Fixes sometimes-uninitialized warning.
piglit-util-cl.c:1052:13: warning: variable 'errNo' is used uninitialized whenever 'if' condition is false [-Wsometimes-uninitialized]
} else if (desc->image_type == CL_MEM_OBJECT_IMAGE3D) {
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
piglit-util-cl.c:1062:28: note: uninitialized use occurs here
if(!piglit_cl_check_error(errNo, CL_SUCCESS)) {
^~~~~
piglit-util-cl.c:1052:9: note: remove the 'if' if its condition is always true
} else if (desc->image_type == CL_MEM_OBJECT_IMAGE3D) {
^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
piglit-util-cl.c:1040:14: note: initialize the variable 'errNo' to silence this warning
cl_int errNo;
^
= 0
Fixes: 910cd6c222930 ("cl: add image and sampler utility functions")
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Acked-by: Jan Vesely <jan.vesely@rutgers.edu>
Reviewed-by: Serge Martin <edb+piglit@sigluy.net>
|
|
Also enables the target in travis and appveyor.
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
|
|
This commit adds support in the json backend for using an external
library that I wrote called jsonstreams. It's a pretty self-explanatory
library, and for piglit has several advantages. First, I've measured
a consistent 10-15 second speed up for running the quick profile.
Second, it *vastly* reduces the amount of memory piglit needs to write
the final JSON document out.
This is not implemented as a separate backend because the way that
piglit's backend are implemented only one backend can handle a specific
file extension. While this can be worked around by setting the
extension to something like '.junit.xml' that doesn't really make sense
here, since they are writing the same format, just using different
means.
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
|
|
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
|
|
It turns out that the backends coincidently put certain fields in, but
the new jsonstreams backend does not automatically add these fields, it
must be passed them. (This is the behavior of piglit when not testing).
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
|
|
We generate this on the fly if it's not there and add it.
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
|
|
Takes over 40 minutes and 16GB on Skylake. Most of the blend
functionality in llvmpipe is adequately covered by Mesa's in-tree tests
anyway.
Signed-off-by: Adam Jackson <ajax@redhat.com>
|
|
New eglext.h has the function pointer typedef corrected to end in PROC
like everything else. We'd like to build against both old and new
headers (at least for a while), while using the idiomatic PROC name in
the actual code.
To do this, before including any headers, tell the preprocessor to
substitute the ...PROC string for any ...NOK string it encounters.
Against old headers this will correct the typedef to the standard form;
against new headers, it will do nothing.
Signed-off-by: Adam Jackson <ajax@redhat.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
|
