Age | Commit message (Collapse) | Author | Files | Lines |
|
Reported-by: Matt Turner <mattst88@gmail.com>
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
Reviewed-by: Martin Peres <martin.peres@linux.intel.com>
|
|
The multisampling test we currently have is not very useful. If you
modify the vertex shader it uses that enables ARB_post_depth_coverage
such that the lines enabling said extension are commented out, you will
notice that the test still "passes" with flying colors. While this makes
sure that ARB_post_depth_coverage doesn't break anything, it also
doesn't actually make sure that the extension even works.
I've tried fixing the original test, but honestly no matter how long I
look at it I still can't even start to understand how this was really
supposed to work in the first place (why exactly are we using depth
stenciling?). Since imirkin didn't have much more luck then I did with
figuring it out, we decided to just write a new and improved test.
So, write a new depth test that actually can't pass unless
ARB_post_depth_coverage is functional and changes the behavior of
gl_SampleMaskIn[] properly.
Signed-off-by: Lyude <lyude@redhat.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
|
|
Some intel hardware supports up to 16 samples, but nvidia only goes up
to 8. So stop these tests from failing on nvidia hardware by only
testing up to the max sample rate we can support.
Signed-off-by: Lyude <lyude@redhat.com>
Reviewed-by: Ilia Mirkin <imirkin@alum.mit.edu>
|
|
This test exposes a Mesa GLSL linker bug. The test fails with the
following error message:
error: Input block `blk' is not an output of the previous stage
Section 4.3.4 (Inputs) of the GLSL 1.50 spec says:
"Only the input variables that are actually read need to be written
by the previous stage; it is allowed to have superfluous
declarations of input variables."
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101247
|
|
It's completely possible that if piglit resume is run with the -n flag
(don't retry incomplete tests), that piglit resume could be invoked with
no tests left to run. In this case it makes sense to let piglit resume
just finalize the backend and produce a unified result.
Signed-off-by: Dylan Baker <dylanx.c.baker@intel.com>
cc: Martin Peres <martin.peres@linux.intel.com>
Tested-by: Hector Edmundo Ramirez Gomez
|
|
Querying GL_QUADS_FOLLOW_PROVOKING_VERTEX_CONVENTION is illegal in
GL 3.2 and later.
Check the GL version to determine the expected error. Note that the
test says config.supports_gl_core_version = 32. With NVIDIA's driver
we get a 3.2 context but with Mesa we get a 3.3 context.
Reviewed-by: Neha Bhende<bhenden@vmware.com>
|
|
GLPROTO_INCLUDE_DIRS are not needed for WGL. Spotted by Jose.
Reviewed-by: Jose Fonseca <jfonseca@vmware.com>
|
|
Pull the innermost code out of testPrim() into a new function.
This can make things a little easier for hacking/debugging.
Also, print whether PV control is supported and whether quads follow
the PV convention.
Reviewed-by: Neha Bhende <bhenden@vmware.com>
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
|
|
The spec/arb_shading_language_420pack/execution/binding-layout test
assumed that it could bind buffers at 16-byte offsets. This may not
be true - we need to respect GL_UNIFORM_BUFFER_OFFSET_ALIGNMENT.
This prevents crashes on drivers which require an alignment larger than
16 bytes. The closed source AMD and NVIDIA drivers appear to require
an alignment of 256 bytes, and I have a patch to bump i965 to 32 bytes.
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
|
|
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
Like the corresponding glx tests which I recently added.
Both pass on native NVIDIA drivers. The multi-context-single-window
test fails with the VMware OpenGL driver.
Reviewed-by: Neha Bhende<bhenden@vmware.com>
|
|
Like the glx tests/utility code, but for wgl.
Note, one must set the PIGLIT_PLATFORM env var to "wgl" before running
Piglit. It looks like there's some Waffle work to look at before this
can be made automatic.
|
|
|
|
Use vector operations instead of scalar.
Simplify vec2 constructor usage.
Reviewed-by: Neha Bhende<bhenden@vmware.com>
|
|
To exercise rendering into multiple windows with one context.
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
|
|
Test rendering into one window with multiple contexts.
Currently fails with llvmpipe, softpipe. Passes with VMware DRI driver.
v2: we only need to probe with context[0]. Fix comments.
Reviewed-by: Charmaine Lee <charmainel@vmware.com>
|
|
So we can run glx tests with depth testing, for example.
|
|
Check for errors clearing other mipmap levels.
Test clearing with a texture view.
Update/add comments.
|
|
The GL_ARRAY_BUFFER, GL_ELEMENT_ARRAY_BUFFER, GL_UNIFORM_BUFFER, etc.
binding points can be interpreted as hints about how buffer data may
be used, but it's not a guarantee. This test exercises setting up a
VBO with various buffer targets (and DSA) before rendering with that
buffer.
Exposes an issue with VMware's driver.
Addition tests may be added to use the targetted buffer for uniforms,
PBO transfers, etc.
|
|
Fixes
/build/piglit/tests/egl/spec/egl_mesa_platform_surfaceless/egl_mesa_platform_surfaceless.c:94:30:
warning: incompatible pointer to integer conversion passing 'void *' to
parameter of type 'NativeWindowType' (aka 'unsigned long')
[-Wint-conversion]
/*native_window*/ NULL,
^~~~
/usr/lib/llvm-3.8/bin/../lib/clang/3.8.1/include/stddef.h:105:16: note:
expanded from macro 'NULL'
^~~~~~~~~~
/build/piglit/tests/egl/spec/egl_mesa_platform_surfaceless/egl_mesa_platform_surfaceless.c:125:30:
warning: incompatible pointer to integer conversion passing 'void *' to
parameter of type 'NativePixmapType' (aka 'unsigned long')
[-Wint-conversion]
/*native_window*/ NULL,
^~~~
/usr/lib/llvm-3.8/bin/../lib/clang/3.8.1/include/stddef.h:105:16: note:
expanded from macro 'NULL'
^~~~~~~~~~
2 warnings generated.
v2: change function pointer declaration (Emil)
Reviewed-by: Emil Velikov <emil.velikov@collabora.com>
Signed-off-by: Juan A. Suarez Romero <jasuarez@igalia.com>
|
|
An existing bug in mesa made it so that overloads of any functions that
might be builtins in any language version or extension would not be
overridable in ESSL 3.00 shaders. This adds tests to ensure that
builtins available in ESSL 3.00 are rejected properly, but builtins from
future versions or non-enabled (or available) extensions are allowed to
be specified.
Signed-off-by: Ilia Mirkin <imirkin@alum.mit.edu>
Reviewed-by: Timothy Arceri <tarceri@itsqueeze.com>
|
|
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
This exposes some existing bugs in gallium and also provides more
coverage for when we start packing uniforms.
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
From the GLES2 spec:
If the transpose parameter to any of the UniformMatrix* commands
is not FALSE, an INVALID_VALUE error is generated, and no uniform
values are changed.
Fixes spurious failures in draw-vertices-half-float_gles2.
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
Fixes: 2e423dd3f4ca ("egl_khr_fence_sync: Add sw_sync lib.")
Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=101641
Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Brian Paul <brianp@vmware.com>
Reviewed-by: Rafael Antognolli <rafael.antognolli@intel.com>
|
|
Reviewed-by: Brian Paul <brianp@vmware.com>
|
|
Previously the expected result for glGetPixelMap was based on an
implicit assumption that GL_MAX_PIXEL_MAP_TABLE was 256. While this is
true for drivers based on Mesa, it is not true on the Intel Windows
driver which advertises 65536.
The GL spec only requires GL_MAX_PIXEL_MAP_TABLE >= 32.
This fixes tests/spec/gl-2.1/pbo on Intel Windows driver.
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
|
|
Reviewed-by: Samuel Pitoiset <samuel.pitoiset@gmail.com>
|
|
Enable the test on the "all" test profile.
Signed-off-by: Rafael Antognolli <rafael.antognolli@intel.com>
Cc: Dylan Baker <dylan@pnwbakers.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
Reviewed-by: Dylan Baker <dylan@pnwbakers.com>
|
|
Verify that eglClientWaitSyncKHR() correctly handles zero timeout before
and after the sw sync fence is signaled.
Signed-off-by: Rafael Antognolli <rafael.antognolli@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
|
|
If <sync> is not a valid sync object for <dpy>,
EGL_NO_NATIVE_FENCE_FD_ANDROID is returned and an EGL_BAD_PARAMETER
error is generated.
Signed-off-by: Rafael Antognolli <rafael.antognolli@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
|
|
Verify that eglDupNativeFenceFDANDROID() correctly returns a new file
descriptor.
Signed-off-by: Rafael Antognolli <rafael.antognolli@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
|
|
Add a test that creates a sync file using sw_sync, and then creates an
EGL fence sync from that sync file. It also verify that the fence sync
EGL_SYNC_CONDITION_KHR is set to EGL_SYNC_NATIVE_FENCE_SIGNALED_ANDROID,
and that EGL_SYNC_STATUS_KHR changes to EGL_SIGNALED_KHR once the fence
is signaled.
Signed-off-by: Rafael Antognolli <rafael.antognolli@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
|
|
Add a small library that helps manipulating software fences. They are
useful for testing EGL Android fences, since the latter can be created
out of them.
Signed-off-by: Rafael Antognolli <rafael.antognolli@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
|
|
Add a new "profile" for android native fences. It basically reuses some
of the tests that should be common to this and the original
EGL_KHR_fence_sync extension.
Signed-off-by: Rafael Antognolli <rafael.antognolli@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
|
|
Add a test profile to this test, so we can reuse part of the code later
and test the same features from EGL_ANDROID_native_fence_sync.
Signed-off-by: Rafael Antognolli <rafael.antognolli@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
|
|
|