Matthew Netsch [Thu, 21 Dec 2017 15:47:02 +0000 (07:47 -0800)]
Fixes pixelformat utils for RGB10_A2
Some asserts were firing and the math
wasn't adding up for RGB10_A2 configs.
Components: OpenGL ES
VK-GL-CTS issue: 920
Affects:
KHR-GLES31.core.blend_equation_advanced.*
Others
Change-Id: Idf3709f71bf7de915a4390c3e8efcda902f2143f
(cherry picked from commit
bd631f4725dece3390890c9099f6ecab1ae88560)
Yabin Zheng [Fri, 31 Aug 2018 07:35:16 +0000 (15:35 +0800)]
Add specific RGB10A2 data for BlendHSLHue and BlendHSLSaturation
Some original input test data will cause divide by 0 issue. This issue
is not handled by spec, but the test should assure the valid data.
Affects: KHR-GLES31.core.blend_equation_advanced.*
VK-GL-CTS issue: 919
Components: OpenGL
Change-Id: I6278ed77de0a38aede7d840bac7b7ec57e869110
(cherry picked from commit
436e28f9d9d6e5c58ffba4bc72122650ce0ee903)
Slawomir Cygan [Fri, 23 Aug 2019 12:33:19 +0000 (14:33 +0200)]
Explicitly use -A Win32 platform in CMake Visual Studio 2019 builds
There is a new CMake switch "-A" for selecting a platform in
Visual Studion CMake generators.
It was set only for 64-bit builds (to x64), however it seems it does not
default to Win32 - so it must be set also for 32-bit builds.
Components: Framework
VK-GL-CTS Issue: 1950
Change-Id: I4d0e9f1a9c8dd0f3060e138f613c8e5326b6fc9f
(cherry picked from commit
a653debe1b4bc8a6d0c94c774aad4f3fcdda1595)
Slawomir Cygan [Tue, 9 Jul 2019 10:19:26 +0000 (12:19 +0200)]
Add support for building with Visual Studio 2019
Components: Framework
Change-Id: I52c92750e9e2874b86fd7bceea0070ad609909c1
(cherry picked from commit
cab8ce57b1400c6909a555973cb8393e1036e610)
Caio Marcelo de Oliveira Filho [Thu, 22 Feb 2018 01:54:57 +0000 (17:54 -0800)]
Handle gcc 7 warnings
- "-Wimplicit-fallthrough" warns on case conditions that
fallthrough. GCC does have a way to accept conditions that have a
comment about falling through inside the cases, but the codebase has
other cases where a single comment explains the fallthrough for the
whole switch. This warning is being ignored. It was not possible to
use "-Wno-error=WARNING" because older GCCs fail -- but they do
accept "-Wno-WARNING" where WARNING is unknown (since at least gcc
4.6.4).
- "-Wint-in-bool-context" warns (among other things) about result of
multiplication being used in boolean context. Code was changed to
use de::max() instead.
- "-Wmaybe-uninitialized" warned about some attributes of a struct not
being initialized depending on the case. Code was changed to
initialize the struct in question.
Components: Framework, Vulkan
Change-Id: Iabb26f01e047353c6e2a704ab27b673e85b185be
(cherry picked from commit
f1ee9253e64f7434cdba7c8ac048c0e9baccc133)
Mika Väinölä [Tue, 20 Mar 2018 11:40:57 +0000 (13:40 +0200)]
Fix GCC 7 implicit fallthrough warnings
GCC 7 with -Wextra or -Wimplicit-fallthrough enabled warns when
a switch case falls through. For cases where this is intentional
the warning can be suppressed with a comment. Breaks and returns
were also added for cases which end in DE_FATAL.
Removing a fallthrough in vktSparseResourcesShaderIntrinsics.cpp
resulted in new test cases.
Affects: *
New tests:
dEQP-VK.sparse_resources.shader_intrinsics.3d_sparse_fetch.*
Components: Framework, Vulkan, OpenGL
VK-GL-CTS issue: 1067
Change-Id: I0306f96cf3a50e85144b4dd22a9a88cd908607d3
(cherry picked from commit
884a93c6b9bc5dc8a7fea9392b37b8fd5ea04d7b)
Jesse Hall [Tue, 27 Mar 2018 23:49:03 +0000 (16:49 -0700)]
Add glu::BufferOffsetAsPointer utility
Clang 6.x adds a warning that doing arithmetic involving a NULL
pointer is a GNU extension. This affects the common GL pattern
(deUint8*)DE_NULL + offset
used to pass an offset into a bound buffer object as a client pointer
parameter.
This change replaces that pattern with a call to a new utility
function glu::BufferOffsetAsPointer(), which will hopefully be
warning-free on all compilers.
Change-Id: I8be939297b02c44091441c71ae75e45ceab30639
Components: Framework, OpenGL, AOSP
(cherry picked from commit
05d98991304fce712732124e5e12fb670ee5a799)
Piers Daniell [Mon, 23 Mar 2020 22:43:24 +0000 (16:43 -0600)]
Fix expected errors for some DSA functions
The OpenGL spec was fixed recently to correct some errors
for some DSA functions in
https://github.com/KhronosGroup/OpenGL-Registry/pull/300
This CL updates the CTS to match.
Affects:
KHR-GL46.direct_state_access.textures_buffer_errors
KHR-GL46.direct_state_access.textures_buffer_range_errors
KHR-GL46.direct_state_access.textures_parameter_setup_errors
Components: OpenGL
VK-GL-CTS issue: 2275
Change-Id: I783c4ca1924cbf6aa3af82d7ac5f49db4a94978b
Peter Deayton [Fri, 7 Jun 2019 19:42:35 +0000 (12:42 -0700)]
Use maximum conformant sample count configuration
Some multisample configurations may not support the maximum renderbuffer
size.
Use GL_NV_internalformat_sample_query when available to find the
maximum conformant sample count for a given format and use that count
for further testing.
Affects:
KHR-GL46.direct_state_access.renderbuffers_storage_multisample
Components: OpenGL
VK-GL-CTS issue: 1822
Change-Id: I348fda6c6b1ee89905d3e4c307e56669e2b205d0
Chad Versace [Fri, 3 Nov 2017 16:36:30 +0000 (09:36 -0700)]
Fix build when .git is a gitlink file
If ${CMAKE_SOURCE_DIR}/.git exists, then
framework/qphelper/CMakeLists.txt uses the git revision as the dEQP
revision. The CMake rules update the revision when the files
${CMAKE_SOURCE_DIR}/.git/HEAD and ${CMAKE_SOURCE_DIR}/.git/index
change.
However, Git does not require that .git be a directory. If
git-submodules or git-worktrees are used, then .git may be a gitlink
file instead and the real git directory is elsewhere.
Fix the build by using `git rev-parse --git-dir` to find the real git dir
and `git rev-parse HEAD` to query the commit sha1.
This bumps the CMake requirement from 2.8 to 2.8.2. But that should be
safe since 2.8.2 was released in 2010.
Components: Framework
Change-Id: I811630e5afada8e6aed611b0ef6c51e39be2200b
(cherry picked from commit
b5cd0ad3079ea00f0d4b243280430a36594eaae4)
Andres Gomez [Thu, 23 May 2019 16:34:13 +0000 (18:34 +0200)]
gl4cComputeShaderTests: correct conversion
Bug 13124 - gl45 conformance: Fix GL45-CTS.compute_shader.max to work with giant numbers
and
Bug 13582 - Checking values from glGet* incorrectly
didn't complete the job of avoiding conversion problems.
Now, we make sure that the comparison is done in the type that won't
cause a problem.
Components: OpenGL
VK-GL-CTS public issue: 149
Affects:
KHR-GLES31.core.compute_shader.dispatch-indirect
KHR-GLES31.core.compute_shader.max
KHR-GL43.compute_shader.dispatch-indirect
KHR-GL43.compute_shader.max
Change-Id: I69047d95ca4418a84f138ac9bfa7818fea45cc64
Tapani Pälli [Tue, 22 Jan 2019 12:00:23 +0000 (14:00 +0200)]
x11_egl: use correct cmake variable name for GLES1 include path
This commit reverts most of
7b4a66f99 that included a hack to search
for include directory next to the library path. This was invalid since
used variable name PKG_GLES1_INCLUDE_DIRS is false, correct name is
GLES1_INCLUDE_DIRS which is properly set by the PKG_CHECK_MODULES call.
This change has been succesfully tested on Fedora and Ubuntu distros by
setting up a custom PKG_CONFIG_PATH.
VK-GL-CTS issue: 1573
Change-Id: I158d816d28f5bbf30288f6226bd5c9e30cf85285
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Andres Gomez [Wed, 27 Feb 2019 13:03:56 +0000 (15:03 +0200)]
GL_ARB_enhanced_layouts: use separable programs
Currently we are testing bogus inputs without a matching counterpart
in the output interface of the previous shader. This is needed since,
otherwise, having a matching output would mean that we would have to
introduce the same error we are trying to test in the input interface
and that could lead to a misleading linking error.
However, not having an output counterpart can also lead to a
misleading linking error.
In other tests we have avoided this by not making a static use of the
tested input. However, when checking for the limit on the number of
input locations that can be used in a specific interface, this
strategy cannot be used since non statically used variable do not
count against this limit.
From page 48 (page 52 of the PDF) of the GLSL 4.60 v.5 spec:
" There is an implementation-dependent limit on the number of
locations that can be used, and if this is exceeded it will cause
a link-time error. (Declared input variables that are not
statically used do not count against this limit.)"
Therefore, the strategy used for this test is to build individual
separable programs for each stage.
From the ARB_separate_shader_objects spec v.25:
" With separable program objects, interfaces between shader stages
may involve the outputs from one program object and the inputs
from a second program object. For such interfaces, it is not
possible to detect mismatches at link time, because the programs
are linked separately. When each such program is linked, all
inputs or outputs interfacing with another program stage are
treated as active. The linker will generate an executable that
assumes the presence of a compatible program on the other side of
the interface. If a mismatch between programs occurs, no GL error
will be generated, but some or all of the inputs on the interface
will be undefined."
Components: OpenGL
VK-GL-CTS issue: 1628
Affects:
KHR-GL44.enhanced_layouts.varying_location_limit
Change-Id: If120f3f2d397be0b53c60033768db60eb6fed968
Andres Gomez [Thu, 3 Jan 2019 12:17:18 +0000 (14:17 +0200)]
GL_ARB_enhanced_layouts: no structs as VS inputs
From page 48 (page 52 of the PDF) of the GLSL 4.60 v.5 spec:
" It is a compile-time error to declare a vertex shader input with,
or that contains, any of the following types:
* A boolean type
* An opaque type
* A structure"
Components: OpenGL
VK-GL-CTS issue: 1614
Affects:
KHR-GL44.enhanced_layouts.varying_structure_member_location
Change-Id: Ied52fd09639de179db480b6a627b65208a54250a
Andres Gomez [Wed, 23 Jan 2019 16:28:59 +0000 (18:28 +0200)]
GL_ARB_enhanced_layouts: remove dead code
Additionally, make us of the new debugging macro to ease checking that
we are testing just the correct and only cause for a failure report.
Components: OpenGL
VK-GL-CTS issue: 1617
Affects:
KHR-GL44.enhanced_layouts.xfb_duplicated_stride
Change-Id: I2bb03db22743ae12d17c3da1654d082236d2e2e6
Andres Gomez [Wed, 9 Jan 2019 15:38:12 +0000 (17:38 +0200)]
GL_ARB_enhanced_layouts: fix interstage interfaces
Multiple test were reporting misleading fails due to out-in interface
mismatching.
Additionally, removed dead code paths.
Finally, make us of the new debugging macro to ease checking that we
are testing just the correct and only cause for a failure report.
Components: OpenGL
VK-GL-CTS issue: 1608
Affects:
KHR-GL44.enhanced_layouts.xfb_block_member_buffer
KHR-GL44.enhanced_layouts.xfb_capture_unsized_array
KHR-GL44.enhanced_layouts.xfb_exceed_buffer_limit
KHR-GL44.enhanced_layouts.xfb_exceed_offset_limit
KHR-GL44.enhanced_layouts.xfb_invalid_offset_alignment
KHR-GL44.enhanced_layouts.xfb_multiple_vertex_streams
KHR-GL44.enhanced_layouts.xfb_output_overlapping
KHR-GL44.enhanced_layouts.xfb_too_small_stride
Change-Id: I0fb8f851b7cbec4c8cb4ee9fbad544c32e130a99
Andres Gomez [Wed, 23 Jan 2019 15:56:33 +0000 (17:56 +0200)]
GL_ARB_enhanced_layouts: fix interstage interfaces
This test was reporting a misleading fail due to out-in interface
mismatching.
Additionally, in the past another "fail" status was "corrected" by
excluding all non-positive cases making it, actually, not a negative
test. Now, while fixing the problems in the interface matching, we
re-introduce the negative cases back.
Also, removed dead code paths.
Finally, make us of the new debugging macro to ease checking that
we are testing just the correct and only cause for a failure report.
Components: OpenGL
VK-GL-CTS issue: 1607
Affects:
KHR-GL44.enhanced_layouts.xfb_variable_stride
Change-Id: If6564a9feb9831d101b0d5073ec3d241a4a8ac81
Andres Gomez [Mon, 7 Jan 2019 23:18:40 +0000 (01:18 +0200)]
GL_ARB_enhanced_layouts: use centroid and sample
When checking for aliasing using mixed auxiliary storages, the tests
checking the usage of centroid and sample were only for the fragment
shader inputs. These qualifiers, although ignored, can also be used in
the rest of the stages.
Components: OpenGL
VK-GL-CTS issue: 1613
Affects:
KHR-GL44.enhanced_layouts.varying_location_aliasing_with_mixed_auxiliary_storage
Change-Id: I4ec1609a63a225ccd8c2ed7b82eb5477a04b6cba
Andres Gomez [Mon, 7 Jan 2019 23:00:49 +0000 (01:00 +0200)]
GL_ARB_enhanced_layouts: mind double types
The special conditions when using the component qualifier with doubles
were not taken into account.
Additionally, the test was reporting a misleading fail due to
non-coherent usage of the "flat" interpolation qualifier. This is now
fixed.
Finally, the generator has been simplified using the refactored code.
Components: OpenGL
VK-GL-CTS issue: 1611
Affects:
KHR-GL44.enhanced_layouts.varying_location_aliasing_with_mixed_auxiliary_storage
Change-Id: I4380ba720f4ea2b2326484d6c8b3e8fc58b1b5e5
Andres Gomez [Mon, 18 Feb 2019 14:49:28 +0000 (16:49 +0200)]
GL_ARB_enhanced_layouts: mind double types
The special conditions when using the component qualifier with doubles
were not taken into account.
Additionally, the generator has been simplified using the refactored
code.
Components: OpenGL
VK-GL-CTS issue: 1616
Affects:
KHR-GL44.enhanced_layouts.input_component_aliasing
KHR-GL44.enhanced_layouts.output_component_aliasing
KHR-GL44.enhanced_layouts.varying_component_of_invalid_type
KHR-GL44.enhanced_layouts.varying_component_without_location
KHR-GL44.enhanced_layouts.varying_exceeding_components
Change-Id: I33c9735a8ef02cf5257f6783cf7856911e4af9d8
Andres Gomez [Thu, 20 Dec 2018 15:29:13 +0000 (17:29 +0200)]
GL_ARB_enhanced_layouts: use component 0 for mat
When checking types that won't support the component qualifier, in the
case of matrixes, just use the universally accepted value in case the
keyword would be accepted: 0.
Otherwise, we won't be sure whether the test is passing because the
reported failure is the expected one: that matrix won't accept the
component qualifier.
Components: OpenGL
VK-GL-CTS issue: 1615
Affects:
KHR-GL44.enhanced_layouts.varying_component_of_invalid_type
Change-Id: I5e53ecdcb4e73f9ca528f3232bfa9405782db43e
Andres Gomez [Mon, 7 Jan 2019 17:46:16 +0000 (19:46 +0200)]
GL_ARB_enhanced_layouts: mind double types
The special conditions when using the component qualifier with doubles
were not taken into account.
The tests generator is now simplified using the refactored code.
Components: OpenGL
VK-GL-CTS issue: 1610
Affects:
KHR-GL44.enhanced_layouts.varying_location_aliasing_with_mixed_interpolation
Change-Id: I44763cf9e9e3b11fe17519ab307b28cc5485706d
Andres Gomez [Fri, 4 Jan 2019 12:46:27 +0000 (14:46 +0200)]
GL_ARB_enhanced_layouts: mind aliasing bit width
Currently, when location aliasing was happening, we were only checking
if the underlying numerical type was the same to allow it.
From page 67 (page 71 of the PDF) of the GLSL 4.60 v.5 spec:
" Further, when location aliasing, the aliases sharing the location
must have the same underlying numerical type and bit
width (floating-point or integer, 32-bit versus 64-bit, etc.)"
Additionally, we were repeating two times the same tests. For example
we would check if, with float and int, they would be allowed if float
was in the component 1 and int in component 3 two times. The test
generation has been now simplified.
Components: OpenGL
VK-GL-CTS issue: 1609
Affects:
KHR-GL44.enhanced_layouts.varying_location_aliasing_with_mixed_types
Change-Id: I33652ac447c0a536bebc8175f34abee45ad1cd2e
Slawomir Cygan [Fri, 1 Mar 2019 14:45:39 +0000 (15:45 +0100)]
Fix context creation in robust_buffer_access_behavior test
Move the context creation routine to iterate() method of
tests cases, to avoid overlapping many contexts where
all tests case constructors are fired at once.
Components: OpenGL
VK-GL-CTS Issue: 1629
Affects:
KHR-GL4*.robust_buffer_access_behavior.*
Change-Id: I28658b41891437ef60a5bfd5d0a19b80342e9f1f
Andres Gomez [Thu, 3 Jan 2019 09:21:48 +0000 (11:21 +0200)]
GL_ARB_enhanced_layouts: fix arrays in {G,TC,TE}S
Several negative tests were reporting a misleading fail due to
mismatching output array dimensions in the TCS, TES and GS interfaces.
Components: OpenGL
VK-GL-CTS issue: 1606
Affects:
KHR-GL44.enhanced_layouts.output_component_aliasing
KHR-GL44.enhanced_layouts.varying_block_automatic_member_locations
KHR-GL44.enhanced_layouts.varying_component_of_invalid_type
KHR-GL44.enhanced_layouts.varying_component_without_location
KHR-GL44.enhanced_layouts.varying_exceeding_components
KHR-GL44.enhanced_layouts.varying_location_aliasing_with_mixed_auxiliary_storage
KHR-GL44.enhanced_layouts.varying_location_aliasing_with_mixed_interpolation
KHR-GL44.enhanced_layouts.varying_location_aliasing_with_mixed_types
KHR-GL44.enhanced_layouts.varying_location_limit
KHR-GL44.enhanced_layouts.varying_structure_member_location
Change-Id: I7450cdd41b72b8aca0c6e4638332eaa56a1ed954
Andres Gomez [Thu, 3 Jan 2019 10:07:02 +0000 (12:07 +0200)]
GL_ARB_enhanced_layouts: no static use of inputs
Currently we are testing bogus inputs without a matching counterpart
in the output interface of the previous shader. This is needed since,
otherwise, having a matching output would mean that we would have to
introduce the same error we are trying to test in the input
interface and that could lead to a misleading linking error.
However, not having an output counterpart can also lead to a
misleading linking error. We workaround this by not making a static
use of the tested input.
From page 47 (page 53 of the PDF) of the GLSL 4.60 v.5 spec:
" Only the input variables that are statically read need to be
written by the previous stage; it is allowed to have superfluous
declarations of input variables."
Additionally, make us of the new debugging macro to ease checking that
we are testing just the correct and only cause for a failure report.
Components: OpenGL
VK-GL-CTS issue: 1605
Affects:
KHR-GL44.enhanced_layouts.varying_block_automatic_member_locations
KHR-GL44.enhanced_layouts.varying_block_member_locations
KHR-GL44.enhanced_layouts.varying_component_of_invalid_type
KHR-GL44.enhanced_layouts.varying_component_without_location
KHR-GL44.enhanced_layouts.varying_exceeding_components
KHR-GL44.enhanced_layouts.varying_location_aliasing_with_mixed_auxiliary_storage
KHR-GL44.enhanced_layouts.varying_location_aliasing_with_mixed_interpolation
KHR-GL44.enhanced_layouts.varying_location_aliasing_with_mixed_types
KHR-GL44.enhanced_layouts.varying_structure_member_location
KHR-GL44.enhanced_layouts.xfb_input
Change-Id: I7dc136b34747f40b25b7cc2b6e3f35f029f701f7
Andres Gomez [Fri, 28 Dec 2018 18:52:13 +0000 (20:52 +0200)]
GL_ARB_enhanced_layouts: add missing out varyings
The test was "passing" due to a different error than the one expected
by the test: there were missing output varyings to be able to match
the tested inputs.
Components: OpenGL
VK-GL-CTS issue: 1604
Affects:
KHR-GL44.enhanced_layouts.input_component_aliasing
Change-Id: I038bdc3a08ed1d458c8b0988227307c5f0f3f104
Andres Gomez [Thu, 20 Dec 2018 14:31:14 +0000 (16:31 +0200)]
GL_ARB_enhanced_layouts: correct usage of "flat"
Multiple negative tests were failing due to errors in the usage of the
flat interpolation qualifier. Therefore, the tests were "passing"
while the expected error was a different one than the one reported by
the test.
Additionally, make us of the new debugging macro to ease checking that
we are testing just the correct and only cause for a failure report.
Components: OpenGL
VK-GL-CTS issue: 1603
Affects:
KHR-GL44.enhanced_layouts.input_component_aliasing
KHR-GL44.enhanced_layouts.output_component_aliasing
KHR-GL44.enhanced_layouts.varying_component_of_invalid_type
KHR-GL44.enhanced_layouts.varying_component_without_location
KHR-GL44.enhanced_layouts.varying_exceeding_components
KHR-GL44.enhanced_layouts.varying_location_aliasing_with_mixed_types
KHR-GL44.enhanced_layouts.varying_location_limit
Change-Id: I63f2b11acadfb0b3e42b7e0f176421eaff4da482
Andres Gomez [Thu, 3 Jan 2019 13:19:17 +0000 (15:19 +0200)]
GL_ARB_enhanced_layouts: no def in untested stage
Currently we are trying to match the definition of the output block
with the input block. In a negative test this has the problem that we
cannot actually match without adding the same error we are trying to
test in both interfaces. Therefore, as it is right now, even if the
tested error wouldn't happen, a misleading linking error will happen.
We avoid this by removing the input definition when testing
outputs (no need to have them at all) and, in the reverse situation,
additionally, by not doing static use of the defined input variable.
Additionally, this simplifies the test a lot.
Finally, added a debugging macro to ease checking that we are testing
just the correct and only cause for a failure report.
Components: OpenGL
VK-GL-CTS issue: 1602
Affects:
KHR-GL44.enhanced_layouts.varying_block_member_locations
Change-Id: Ia2c0ba705e975817999aaa16c1ce129d616540f6
asokolow [Wed, 6 Feb 2019 09:12:28 +0000 (10:12 +0100)]
Create context with CONTEXT_ROBUST for robust buffer access tests
These tests exercise functionality that is not available in default
context provided by the framework. New context, with CONTEXT_ROBUST
(WGL|GLX_CONTEXT_ROBUST_ACCESS_BIT_ARB) flag is created for these
tests.
Component: OpenGL
VK-GL-CTS Issue: 1588
Affects:
KHR-GL4*.robust_buffer_access_behavior.*
Change-Id: Ie698a6a434426bf147a705cc487b4fbe560a79c7
asokolow [Tue, 22 Jan 2019 06:29:33 +0000 (07:29 +0100)]
Fix deleting texture in DSA test
Components: OpenGL
VK-GL-CTS Issue: 1571
Affects:
KHR-GL45.direct_state_access.framebuffers_get_attachment_parameters
Change-Id: Ic15929f20f983e154263a79f9163b04c80b808d9
asokolow [Thu, 20 Sep 2018 09:56:25 +0000 (11:56 +0200)]
Fix GL errors expected for glClearNamedBufferData and glClearNamedBufferSubData
The spec says:
"An INVALID_VALUE error is generated if type is not one of the types in
table 8.2.
An INVALID_VALUE error is generated if format is not one of the formats
in table 8.3."
However, the test was expecting INVALID_ENUM
Components: OpenGL
VK-GL-CTS Issue: 1572
Change-Id: Ibcb338061584af48026b291c4372303c083102ea
Affects: KHR-GL45.direct_state_access.buffers_errors
Yogesh Gole [Wed, 17 Oct 2018 11:21:19 +0000 (16:51 +0530)]
Terminate EGL display after use
Previously function getDefaultEglConfigList destroyed native
display but did not terminate the EGL display. It may happen
that next create call will allocate native display at same
address as previous and as per spec multiple calls made to
eglGetPlatformDisplayEXT with the same <platform> and
<native_display> will return the same EGLDisplay handle. But
returned EGLDisplay will have references to stale data.
To fix this destroy the EGL Display along with native display.
Affects: CTS-Configs.*
Components: OpenGL and Framework
VK-GL-CTS Issue: 1426
Change-Id: I8e4c7d693d26bd9be6d7f892d0b5162d9b2b245c
(cherry picked from commit
90b601fa8d5c6ca10aca7b3ebb50c02400cb0049)
Piers Daniell [Tue, 2 Oct 2018 17:51:25 +0000 (11:51 -0600)]
Use non-arrayed varying name for TCS blocks
This is a partial revert of CL 2625 to restore naming the
value member of the BLOCK_INOUT interface block as
"BLOCK_INOUT.value" rather than "BLOCK_INOUT[0].value".
Affects:
KHR-GL46.tessellation_shader.single.xfb_captures_data_from_correct_stage
Components: OpenGL
VK-GL-CTS issue: 1388
Change-Id: I9ef6453ec5465a0fa5561220cc9d7bfe54298416
Alexander Galazin [Tue, 9 Oct 2018 09:01:22 +0000 (11:01 +0200)]
Update KC CTS
Components: Framework
VK-GL-CTS issue: 1404
Change-Id: Ia7b13ef1b8618b92cab3d4ebe6f593dbf1512fdc
Robert Tray [Fri, 14 Sep 2018 22:32:57 +0000 (16:32 -0600)]
Add screenLayout to AndroidManifest configChanges
Specify that dEQP will handle screenLayout config changes in the
manifest so the Android framework doesn't try to stop and restart dEQP.
dEQP can't handle a restart and just exits.
Update the minSdkVersion and targetSdkVersion to match what is
used for Khronos Vulkan dEQP manifest. Android P will launch the
DeprecatedTargetSdkVersionDialog if the sdk version is not recent
enough.
Test: deqp
Cherry-pick from https://android-review.googlesource.com/c/platform/external/deqp/+/757202
VK-GL-CTS Issue: 1357
Affects: OpenGL/OpenGL ES Android build
Change-Id: Ia9befb1d65752ece3ea5778ba1301a134559334a
Alexander Galazin [Fri, 14 Sep 2018 14:50:45 +0000 (16:50 +0200)]
Update KC CTS
Components: Framework
Change-Id: I227ab78c00577ee6ea8e90292b01d7abe3ddd2e6
James Helferty [Tue, 31 Jul 2018 19:12:08 +0000 (15:12 -0400)]
Scale texture sizes to fit within available memory
The change makes use of platform-specific information to scale the
dimensions of textures used by the advanced shader image load store
tests so they fit inside available memory.
Affects:
KHR-GL*.shader_image_load_store.advanced-copyImage
KHR-GL*.shader_image_load_store.advanced-memory-order
KHR-GL*.shader_image_load_store.advanced-sso-perSample
KHR-GL*.shader_image_load_store.advanced-sso-simple
KHR-GL*.shader_image_load_store.advanced-sso-subroutine
KHR-GL*.shader_image_load_store.advanced-sync-imageAccess2
Components: OpenGL
VK-GL-CTS issue: 1314
Change-Id: I06e7e4629b0e00f0ca5c152df56c09bbdcc05247
James Helferty [Tue, 14 Aug 2018 20:56:59 +0000 (16:56 -0400)]
Free framebuffer objects after use.
Avoid leaking framebuffer objects in verifyLookupTextureData methods.
Affects:
KHR-GL4*.sparse_texture_clamp_tests.SparseTextureClampLookupColor
KHR-GL4*.sparse_texture_clamp_tests.SparseTextureClampLookupResidency
Components: OpenGL
VK-GL-CTS issue: 1312
Change-Id: I12aedd34fa80dbd9061ac174ce622a976019396d
asokolow [Tue, 12 Jun 2018 08:59:10 +0000 (10:59 +0200)]
Do not expect TCS shader with XFB variables to fail linkging in GL
The test tries to link shader with XFB variables in TCS, without TES stage.
The test expects the shader to fail linking, but this behavior does not
seem to be mandated by GL spec.
According to GL 4.6 spec (
https://www.khronos.org/registry/OpenGL/specs/gl/glspec46.core.pdf - page 387)
- It is not forbidden to use TransformFeedbackVaryings with Tessellation Control
Shader and the spec does not mandate linking failure, even without NV_gpu_shader5
extension.
Test seems to derive this behavior from OGL ES spec where XFB is not allowed
within TCS shader.
Components: OpenGL
VK-GL-CTS Issue: 1221
Affects:
KHR-GL45.tessellation_shader.single.xfb_captures_data_from_correct_stage
Change-Id: Ie7c4209e0a36d48e56df04030ed095741bda155a
Pierre Tsai [Thu, 2 Aug 2018 15:13:14 +0000 (11:13 -0400)]
bufSize of glGetInternalformativ does not follow spec
In the specification section 22.3 Internal Format Queries
(glspec46.core.pdf page 564), it is written:
No more than bufSize integers will be written into params.
The confusion may come from the man page where in one point it is said:
"No more than bufSize basic machine units will be written to this
memory." But the above description is still a number of integers.
Affects:
KHR-GL46.sparse_texture_tests.SparseDSATextureCommitment
KHR-GL46.sparse_texture_tests.SparseTextureAllocation
KHR-GL46.sparse_texture_tests.SparseTextureCommitment
KHR-GL46.sparse_texture_tests.TextureParameterQueries
KHR-GL46.sparse_texture_tests.InternalFormatQueries
Components: OpenGL
VK-GL-CTS issue: 1296
Change-Id: I05ea0cbe54539f6f8711278cb7b08ad7e5862aa9
Mark Adams [Fri, 3 Aug 2018 21:47:39 +0000 (17:47 -0400)]
Fix resource leaks in arrays_of_arrays
The compute shader variants of the subroutine tests weren't freeing
the programs and not resetting all state. This resulted in leaked
memory on each test iteration.
VK-GL-CTS issue: 1298
Components: OpenGL
Affects:
KHR-GL*.arrays_of_arrays_gl.Subroutine*
Change-Id: Ia2152476c12f2141698dadd1b208090b28756ce3
Dave Airlie [Mon, 30 Apr 2018 04:51:38 +0000 (05:51 +0100)]
Don't used a fragment output in the shader storage buffer object test
This test doesn't require a fragment shader output, and if we are
on a system which has only 8 combined shader output resources,
this stops the test from executing.
Components: OpenGL
VK-GL-CTS Issue: 985
Affects:
KHR-GL45.shader_storage_buffer_object.advanced-unsizedArrayLength-fs-std430-vec
Change-Id: Ib0883e09d84022c1dac5bc2f22787c0e91efba1c
Neil Monday [Mon, 9 Jul 2018 13:38:16 +0000 (09:38 -0400)]
TextureBarrier required for texture_view.coherency
The third paragraph in section 9.3.1 gives conditions when
texture/framebuffer texture reads/writes are undefined. I
think that this test requires a TextureBarrier as described
in the last bullet point.
Affects:
KHR-GL46.texture_view.coherency
Components: OpenGL
VK-GL-CTS issue: 1183
Change-Id: I231ffd5acd089ebc1a9e0265c2f4a97f2a5c525c
asokolow [Fri, 30 Mar 2018 09:30:41 +0000 (11:30 +0200)]
Interface mismatch in KHR-GL45.enhanced_layouts.xfb_block_stride
There is an interface mismatch between Vertex and Geometry shaders
used in this test. VS is lacking an output variable that GS is reading.
VK-GL-CTS Issue: 1223
Components: OpenGL
Affects: KHR-GL45.enhanced_layouts.xfb_block_stride
Change-Id: I597f5dd8b553a0a47fe00eebb701822e9ef8f678
asokolow [Tue, 12 Jun 2018 08:59:10 +0000 (10:59 +0200)]
Format mismatch between bound texture and image layout specifier
According to GL 4.6 spec
(https://www.khronos.org/registry/OpenGL/specs/gl/glspec46.core.pdf - page 291)
- Any image variable used for shader loads or atomic memory operations must be
declared with a format layout qualifier matching the format of its associated
image unit.
Currently test binds the texture as GL_RGAB32UI but shader expects it to be
rgba32f. According to table 8.26 this leads to format mismatch.
Components: OpenGL
VK-GL-CTS Issue: 1222
Affected:
KHR-GL45.shader_atomic_counter_ops_tests.*
Change-Id: Ib55594371b234919b912753a4d2da1f426b55a82
Neil Monday [Wed, 27 Jun 2018 20:29:07 +0000 (16:29 -0400)]
Max compiler threads only specifies a suggestion
In this test, it is expected that calling MaxShaderCompilerThreadsKHR()
with a count of 0 will compile non-parallel. However, the extension
purposefully says this is only a "request" for non-parallel compiling
or linking.
This change will remove this specific test.
Affects:
parallel_shader_compile.compilation_completion_non_parallel
Components: OpenGL
VK-GL-CTS issue: 1226
Change-Id: I49214b3b32b4991ee7b96aad3853e24075c0f137
Alexander Galazin [Fri, 1 Jun 2018 10:30:10 +0000 (12:30 +0200)]
Exclude texture_view.view_classes
Exclude KHR-GL*.texture_view.view_classes tests
as per approved waiver VK-GL-CTS 1190
Components: OpenGL
Affects: mustpass
VK-GL-CTS issue 1141, 1190
Change-Id: I593e122b7af7cd4cd96e98400c0a876321803f78
Your Name [Fri, 13 Apr 2018 20:06:03 +0000 (16:06 -0400)]
Sparse Texture tests not following spec
CTS is testing texture types that are incompatible
with Sparse Textures.
Affects:
KHR-GL46.sparse_texture_tests.SparseDSATextureCommitment
KHR-GL46.sparse_texture_tests.SparseTextureAllocation
KHR-GL46.sparse_texture_tests.SparseTextureCommitment
KHR-GL46.sparse_texture_tests.TextureParameterQueries
KHR-GL46.sparse_texture_tests.InternalFormatQueries
1.If 1D/1D array are not clear in the spec, then CTS should delete test
1D/1D array
2.GL_TEXTURE_2D shoule set depth = 1
3.sparse texture should create width/height/depth less than page size
4.Target GL_TEXTURE_BUFFER, GL_RENDERBUFFER,
GL_TEXTURE_2D_MULTISAMPLE,GL_TEXTURE_2D_MULTISAMPLE_ARRAY are not in
the spec, so all of them sohuld be delete.
Components: OpenGL
VK-GL-CTS issue: 1121
Change-Id: I144886ae57769be85f430127d8d4211f9554a1b2
Rahul Jain [Mon, 21 May 2018 11:48:01 +0000 (17:18 +0530)]
Wrong shaders in glcParallelShaderCompileTests.cpp
In glcParallelShaderCompileTests.cpp:50
it is written as
"int main() {\n"
which is invalid as per spec.
As per spec GLSL 4.60, section 6.1 last paragraph:
It is a compile-time or link-time error to declare or define a function
main with any other parameters or return type.
in glcParallelShaderCompileTests.cpp:55
it is written as "out ver4 fragColor;\n"
it should be "out vec4 fragColor;\n"
there is no data type as "ver4"
Affects:
KHR-GL46.parallel_shader_compile.compilation_completion_parallel
Components: OpenGL
VK-GL-CTS issue: 1182
Change-Id: I2f213eeaf28039c1f36b7772257090d81e1f013e
Rahul Jain [Tue, 8 May 2018 11:16:49 +0000 (16:46 +0530)]
Fixes wrong MIN_MAP_BUFFER_ALIGNMENT limit
As per GL spec 4.6 P:632, Table 23.55,
It is given that MIN_MAP_BUFFER_ALIGNMENT should be aleast 64 or higher
Currently CTS test is limiting to 64 as maximum.
Affects:
KHR-GL46.limits.min_map_buffer_alignment
KHR-GL46.map_buffer_alignment.query
KHR-GL46.map_buffer_alignment.functional
Components: OpenGL
VK-GL-CTS issue: 1158
Change-Id: I6e4feea6e91bbf0e030b0dfcdbc49c716ffd8145
BMiles [Mon, 19 Mar 2018 18:46:26 +0000 (14:46 -0400)]
This change fixes an issue in BindBufferOffsetEXT
Change hard coded 0 to i for iterating in a loop
The test needs to be iterating through the indices of the buffer.
Affects: KHR-GL4*.transform_feedback.capture_vertex_separate_test
Components: OpenGL
VK-GL-CTS issue: 1108
Change-Id: I0654bf453c05fddad2e10c9843b9130619f4844a
asokolow [Mon, 12 Mar 2018 07:45:11 +0000 (08:45 +0100)]
Fix glTransformFeedbackBufferRange defintion
There is a mismatch between TransformFeedbackBufferRange
definitions in test and OpenGL spec.
The last parameter type is GLsizeiptr accoding
to spec (and XML registry).
Components: OpenGL
VK-GL-CTS Issue: 1093
Affects:
KHR-GL45.direct_state_access.xfb_buffers
Change-Id: Iaf688016234de106b642593d9808d88c7d765edd
Michael Chock [Thu, 8 Mar 2018 19:41:05 +0000 (11:41 -0800)]
Move khr_debug tests to NoContext
As implemented, the KHR-GL4*.khr_debug.* tests can terminate the EGL
display used by other tests. As recommended, move these tests to the
KHR-NoContext.* group.
Since the test names are already changing, fix the spelling of
'receiving' in one test case.
Components: OpenGL
VK-GL-CTS Issue: 1062
Renamed tests:
KHR-GL4*.khr_debug.* -> KHR-NoContext.gl43.khr_debug.*
Change-Id: I766131f77206cad1cb34955426a81cf10b50e159
Michael Chock [Tue, 27 Feb 2018 23:22:46 +0000 (15:22 -0800)]
Fixes for sparse texture clamp tests.
When initializing the test texture, bind to an image unit, setting the
layered parameter to GL_TRUE instead of GL_FALSE. For array textures,
the compute shader used indexes all layers of the textures. For
non-array textures, the parameter has no effect.
For cube map array textures, copy vertex array data to a higher-scoped
memory location, such that the draw call is not invoked with a dangling
pointer.
Swap the Z and W coordinates of the texture size calculated by shaders,
such that Z is the number of cube faces, and W is the number of layers
(which is how the coordinates were otherwise already being used).
Set up texture coordinates for KHR-GL46.sparse_texture_clamp_tests.*
such that the fragment shader correctly computes layer and face indices,
given the corrected texture size and existing shader generation logic.
Components: OpenGL
VK-GL-CTS issue: 1042
Affects:
KHR-GL46.sparse_texture_clamp_tests.*
KHR-GL46.sparse_texture2_tests.*
Change-Id: Ifd33b34b33494ad35f0eb48a9b4aa0ade3bb2d04
Liam Middlebrook [Fri, 19 Jan 2018 23:04:34 +0000 (15:04 -0800)]
SparseTextureAlloc check for GL_NV_deep_texture3D
When running the
SparseTextureAllocationTestCase::verifyTexStorageInvalidValueErrors test
we need to check if the GL_NV_deep_texture3D extension is supported. If
so we can't compare the depth of the texture against the
GL_MAX_3D_TEXTURE_SIZE attribute that is intended for symmetric texture
limits. Instead we need to check against GL_MAX_DEEP_3D_TEXTURE_DEPTH_NV
provided the texture width and height are within the acceptable ranges
for a deep texture.
Components: OpenGL
VK-GL-CTS issue: 959
Change-Id: I0e404d6851724b0624d0802603a12aea3af220e5
Reviewed-by: ewerness
Liam Middlebrook [Wed, 17 Jan 2018 05:17:38 +0000 (21:17 -0800)]
Fix ClearOpsBufferStorageTestCase on ppc64le
The check for the modified ClearOpsBufferStorageTestCase output currently checks
if the value present in the mapped buffer from GL is equivalent to the
calculated value under the following scheme:
(data_rgba8 & (0xFF << (n_current_byte * 8))) >> (n_current_byte * 8)
The C++ Working Draft Standard N4140 [1] Section 5.8
states the following:
The behavior is undefined if the right operand is negative or greater than
or equal to the length in bits of the promoted left operand.
Similar language is also stated in the C Working Draft Standard N1256 [2]
If the value of the right operand is negative or is greater than or equal to
the width of the promoted left operand, the behavior is undefined.
This change removes the undefined behavior that is being used to set
`expected_value`.
Additionally this change makes the literal 0xFF into 0xFFu since all of the
other components surrounding it are unsigned.
References:
[1] https://timsong-cpp.github.io/cppwp/n4140/draft.pdf
[2] http://www.open-std.org/jtc1/sc22/WG14/www/docs/n1256.pdf
Components: OpenGL
VK-GL-CTS issue: 955
Change-Id: I8f1a25b184379c966355e6d0dec81b1416049279
Reviewed-By: pdaniell
Robert Morell [Wed, 10 Jan 2018 21:31:33 +0000 (13:31 -0800)]
Fix XCreateWindow BadMatch with depth 32 visuals
For the common case of a depth 24 root window, when creating a window
with a visual of depth 32, attempting to use a depth of CopyFromParent
will result in a BadMatch error.
Fix this by passing the visual's depth to XCreateWindow explicitly --
fortunately the VisualInfo structure was already queried.
Components: Framework
VK-GL-CTS issue: 938
Change-Id: I54b142b6105c7f328e76d7ea0f9b01e635380e4e
Iago Toral Quiroga [Fri, 12 Jan 2018 09:11:42 +0000 (10:11 +0100)]
Fix checks for maximum number of samples
Some tests query MAX_SAMPLES and/or MAX_INTEGER_SAMPLES, then
attempt to create FBOs passing MAX_SAMPLES+1 samples and
expecting an error to be reported.
There are a couple of bugs in these tests:
1. One uses GL_RGB10_A2UI which is an integer format and should
use MAX_INTEGER_SAMPLES instead of MAX_SAMPLES (this is actually
what the comment above the call says it is doing).
2. Formats can actually support more samples than
MAX_SAMPLES. As stated for the SAMPLES internal format query:
"The maximum value in SAMPLES is guaranteed to be at least
the lowest of the following:
(...)
* The value of MAX_SAMPLES"
The key aspect here is "at least", which means that the
maximum number reported in SAMPLES can be larger than any
of the values in the list, which includes MAX_SAMPLES.
As a consequence of 2), it is possible for these tests to fail
because MAX_SAMPLES+1 is a valid number of samples for the target
internal format and not an invalid number like the test intended.
This patch modifies the affected tests to query SAMPLES for the
internal formats used in the checks so we have the actual maximum
number of sample counts supported for each one, that way we ensure
that SAMPLES+1 is always an invalid number of samples.
Components: OpenGL
Vk-GL-CTS issue: 944
Affects:
KHR-GL45.direct_state_access.renderbuffers_storage_multisample_errors
KHR-GL46.direct_state_access.renderbuffers_storage_multisample_errors
KHR-GL45.direct_state_access.textures_storage_errors
KHR-GL46.direct_state_access.textures_storage_errors
Change-Id: Ic294fe866a3d574fe0321821713d3258027db73b
Michael Chock [Tue, 16 Jan 2018 22:57:49 +0000 (14:57 -0800)]
Loosen color target tolerances for some tests
Some tests verify the results of glReadPixels within two bits of
tolerance. However, some pixel formats (e.g., RGB 565) cannot support
such precision. Loosen hard-coded tolerances, such that tests can pass
with 5-bit color channels.
Additionally change the types of some variables from GLubyte to int. Log
messages constructed through stream operators would treat such values as
characters, rather than numeric quantities.
Components: OpenGL
VK-GL-CTS issue: 950
Affects:
KHR-GL46.indirect_parameters_tests.MultiDrawArraysIndirectCount
KHR-GL46.indirect_parameters_tests.MultiDrawElementsIndirectCount
KHR-GL46.shader_atomic_counter_ops_tests.*
KHR-GL46.shader_viewport_layer_array.*
Change-Id: I7da30842004caf1698d83128678c1bda82fff787
Piers Daniell [Thu, 21 Dec 2017 23:50:14 +0000 (16:50 -0700)]
Remove use of unsupported SURFACETYPE_OFFSCREEN_GENERIC
The khr_debug tests were for some reason trying overriding the surface
type to SURFACETYPE_OFFSCREEN_GENERIC, which isn't supported by WGL.
This change just leaves the surface type as is.
Affects:
KHR-GL46.khr_debug.*
Components: OpenGL
VK-GL-CTS Issue: 922
Change-Id: Ib72e346ed716d19629dcbcfa423b9fc4f63d1835
Iago Toral Quiroga [Wed, 10 Jan 2018 11:16:04 +0000 (12:16 +0100)]
Implementations can use more samples than requested
KHR-GL45.copy_image.samples_mismatch tests that
implementations produce an error when attempting to
copy between images with different sample counts,
however, it relies on the hardware always allocating
exactly the number of samples requested, which may
not happen, since according to the spec for
glTexImage*Multisample:
"samples represents a request for a desired minimum
number of samples. Since different implementations
may support different sample counts for multi-
sampled textures, the actual number of samples
allocated for the texture image is implementation
dependent."
Since the test doesn't account for this, it can end
up expecting an error for the case where it requested
two textures with different sample counts but the
implementation ended up rounding one of them up to the
same number of samples used by the other one, in which
case the Copy operation will not produce an error any
more.
This patch fixes this by checking the actual sample
counts allocated for each texture and adjusting the
test expectation based on that.
Components: OpenGL
VK-GL-CTS issue: 932
Affects:
KHR-GL45.copy_image.samples_mismatch
KHR-GL46.copy_image.samples_mismatch
Change-Id: Idc1ce3026cbaae09cc65dc5194dd26f9bcddea0f
Piers Daniell [Thu, 21 Dec 2017 19:49:09 +0000 (12:49 -0700)]
Remove read value check against uncommitted pages
Reading from uncommitted pages produces undefined results so checks that
the result is zero in invalid. Only data read from committed pages has
undefined results. The PixelUnpackBufferStorageTestCase subtest was the
only subtest that was checking reads from uncommitted pages. All the
other subtest already did not do this.
Affects:
KHR-GL46.sparse_buffer_tests.BufferStorageTest
Components: OpenGL
VK-GL-CTS Issue: 921
Change-Id: Ie7e61553cbad2454af03f39b7cdbd1ae8ab66d93
Piers Daniell [Wed, 20 Dec 2017 22:27:34 +0000 (15:27 -0700)]
Add some missing GL_TEXTURE_SPARSE state setting
Some parts of the SparseTextureAllocation test were not setting the
GL_TEXTURE_SPARSE_ARB state on the texture object before
negative-testing sparse texture behavior. This meant that these parts of
the test ran as non-sparse texture. In many cases this issue was masked,
except for the case where the implementation didn't support
GL_ARB_sparse_texture2. In this case it expected a glTexStorage call
with width/height/depth not a multiple of the page size to fail. It
wouldn't fail because the texture was actually non-sparse.
Another problematic part of the test was with the
GL_SPARSE_TEXTURE_FULL_ARRAY_CUBE_MIPMAPS_ARB negative testing.
The test implemented did not fully the test spec or the OpenGL
spec. The correct negative testing of
GL_SPARSE_TEXTURE_FULL_ARRAY_CUBE_MIPMAPS_ARB requires all the
conditions of GL_TEXTURE_SPARSE_ARB set, the correct target
*and* a width/depth that doesn't follow the requirement. The
test tried to test *or* of those conditions and was just wrong.
In addition it tried to use verifyTexStorageInvalidValueErrors()
to verify invalid texture targets, which would never pass.
Affects:
KHR-GL45.sparse_texture_tests.SparseTextureAllocation
KHR-GL46.sparse_texture_tests.SparseTextureAllocation
Components: OpenGL
VK-GL-CTS issue: 917
Change-Id: If25edd0aae382b7305b87ac0a810259e9ee67523
Liam Middlebrook [Wed, 17 Jan 2018 06:22:22 +0000 (22:22 -0800)]
Fix typo in log message for ClearOpsBufferStorage
A copy-paste error led for the ClearOpsBufferStorageCase inside of
SparseBufferTests to output that the
"Unmodified buffer object region has invalid contents"
for cases referring to invalid contents for the modified buffer region.
Components: OpenGL
VK-GL-CTS issue: 956
Change-Id: I562ff6fa984977ee08ca0f2783a046b1bac235bd
Reviewed-By: pdaniell
Iago Toral Quiroga [Tue, 19 Dec 2017 12:30:14 +0000 (13:30 +0100)]
Validate instancing cases properly
The vertex attrib 64bit limits test case executes
both regular and instanced draw calls and has a
validation function that takes a boolean parameter
that informs it about whether isntancing has been
used so it can verify that all instances executed
properly. However, this function is always called
with false as parameter, indicating no instancing,
even for instanced scenarios, which means that the
validation function is only validating the first
instance even for instanced cases.
Components: OpenGL
VK-GL-CTS issue: 916
Affects:
KHR-GL45.vertex_attrib_64bit.limits_test
KHR-GL46.vertex_attrib_64bit.limits_test
Change-Id: Ibaf2c8fc0b6599a17f7bdbc7c4914b42d38d5e2a
Alexander Galazin [Wed, 10 Jan 2018 10:34:00 +0000 (11:34 +0100)]
Update gl.xml revision
Change-Id: I7cb16bd139e801df2d2090ad5dd5d6267a6d19f9
Component: Framework
Ian Romanick [Fri, 15 Dec 2017 00:18:34 +0000 (16:18 -0800)]
Don't delare arrays as static
m_texture_1D and friends may have different values on each invocation of
gl4cts::GetTextureSubImage::Errors::testTwoDimmensionalTextureErrors()
on a full CTS run. However, the rules for 'static const' initializers
may prevent the values from being updated after the first run.
I discovered this while trying to track down failures in
KHR-GL46.get_texture_sub_image.errors_test that only started occuring
with the 3rd visual tested in a full CTS run. Running the test by
itself or testing that visual by itself via glcts would not reproduce
the error. I added some extra debug logging to the driver, and I found
that testOneDimmensionalTextureErrors() and
testTwoDimmensionalTextureErrors() were using different names for the
GL_TEXTURE_1D texture. I then discovered that the targets of the
textures used from testTwoDimmensionalTextureErrors() were {
GL_TEXTURE_2D, GL_TEXTURE_RECTANGE, GL_TEXTURE_2D } instead of {
GL_TEXTURE_1D, GL_TEXTURE_1D_ARRAY, GL_TEXTURE_2D }
See also https://bugs.freedesktop.org/show_bug.cgi?id=104154
Components: OpenGL
VK-GL-CTS issue: 912
Affects:
KHR-GL46.get_texture_sub_image.errors_test
Change-Id: I3ce061d9de75df46ff69a7eea4b88f6d17164f30
Ian Romanick [Thu, 14 Dec 2017 23:56:59 +0000 (15:56 -0800)]
Fix (likely) typo in glGetTextureSubImage height calculation
The old code used "test_textures[i].id == GL_TEXTURE_1D", but
test_texturess::id contains the texture ID, not the texture target. As
a result, the GL_TEXTURE_1D test always got the error that it expected,
but it got it for the wrong reason.
Components: OpenGL
VK-GL-CTS issue: 912
Affects:
KHR-GL46.get_texture_sub_image.errors_test
Change-Id: I4d3de1ccbd3b0cc554af23ceae16d590c82e81d4
Alejandro Piñeiro [Fri, 15 Dec 2017 13:37:36 +0000 (14:37 +0100)]
gl_spirv: fix layout for uniform blocks on state queries test
The shader for spirv_modules_state_queries_test was using location for
a uniform block. Although glslang is not raising an error, that is
wrong (See section 4.4.5 of the 4.50 GLSL spec).
Additionally, we should provide a way to identify the uniform block
without using a name, so we should specify the binding (see Issue 21
at ARB_gl_spirv spec).
Modules: OpenGL
VK-GL-CTS issue: 913
Affects:
KHR-GL45.gl_spirv.spirv_modules_state_queries_test
KHR-GL46.gl_spirv.spirv_modules_state_queries_test
Change-Id: I95cb03506bccf76b4188e879960813a060c662aa
James Helferty [Fri, 8 Dec 2017 21:44:32 +0000 (16:44 -0500)]
Remove GL_RGB9_E5 from SparseTexture2Allocation
SparseTexture2Allocation tests multisample textures; since GL_RGB9_E5
is not required to be supported as a multisample texture format, it
shouldn't be included in this test.
Components: OpenGL
VK-GL-CTS issue: 892
Affects:
KHR-GL4x.sparse_texture2_tests.SparseTexture2Allocation
Change-Id: I58fa72c26d8cc0c2060b8176aeec4f7d3e43787f
Alejandro Piñeiro [Wed, 13 Dec 2017 11:43:12 +0000 (12:43 +0100)]
Update gl_spirv tests fallback binaries
ARB_gl_spirv tests include off-line fallback binaries in the case of
glslang/spirv-tools not being available. During the review process of
those tests, it was agreed to remove the noise function test, but the
binaries were not updated.
Modules: OpenGL
VK-GL-CTS issue: 897
Affects:
KHR-GL45.gl_spirv.spirv_glsl_to_spirv_builtin_functions_test
KHR-GL45.gl_spirv.spirv_glsl_to_spirv_builtin_functions_test
Change-Id: I8be837a286a3c013568101d5aa1ae3d6fd4eca8f
James Helferty [Wed, 6 Dec 2017 20:07:29 +0000 (15:07 -0500)]
Avoid redundant shader creation for clamp tests
Avoid redundant work involving creation of shaders. By rearranging the
order in which we iterate loops, we can avoid a signficant amount of
shader string manipulation and redundant shader creation. (This should
reduce the run time for this test without impacting coverage.)
Component: OpenGL
VK-GL-CTS issue: 889
Affects:
KHR-GL4x.sparse_texture_clamp_tests.SparseTextureClampLookupResidency
KHR-GL4x.sparse_texture_clamp_tests.SparseTextureClampLookupColor
Change-Id: If0cb11827c45da6b13d81115d4a8a9143d31e8c6
Alejandro Piñeiro [Fri, 8 Dec 2017 07:47:03 +0000 (08:47 +0100)]
Fix expectation for state queries SPIR-V test
The test removes the names for several variables of the shader, and
then it expect a length of 1 for all of them when doing a state
query. But on the case of TRANSFORM_FEEDBACK_VARYING_MAX_LENGTH it
should return a 0.
From spec:
"If pname is TRANSFORM_FEEDBACK_VARYING_MAX_LENGTH, the length of
the longest output variable name specified to be used for transform
feedback, including a null terminator, is returned. If no outputs
are used for transform feedback, zero is returned. If no name
reflection information is available, one is returned."
In this shader, transform feedback is not activated, so instead of
applying the last sentence, it should apply the sentence "If no
outputs are used for transform feedback, zero is returned".
Modules: OpenGL
VK-GL-CTS issue: 886
Affects:
KHR-GL45.gl_spirv.spirv_modules_state_queries_test
KHR-GL46.gl_spirv.spirv_modules_state_queries_test
Change-Id: I02ad02ee252e6703411ced4f14a7d753b3ef3eda
Alexander Galazin [Fri, 8 Dec 2017 10:54:49 +0000 (11:54 +0100)]
Update KC CTS reference
Component: OpenGL
Change-Id: I112d7af5b25398407444f31ca6b37cae836ec820
Kenneth Graunke [Thu, 30 Nov 2017 08:13:35 +0000 (00:13 -0800)]
Use less nesting in many arrays of arrays tests.
Testing 8 levels of array-of-array nesting is very thorough, but a bit
overkill. It creates a huge amount of register pressure which tends to
slow down compilers, causing the test to take an unreasonable amount of
time to run. This patch reduces it to 4 levels in many subtests, which
should still be a reasonable amount of testing while running in much
less time.
This cuts the runtime of KHR-GL46.arrays_of_arrays_gl.* by 33 minutes
(43 -> 10) on my laptop's discrete GPU. With the Intel Mesa driver,
it cuts it by 17 minutes (23 -> 6).
Some tests still exercise 8 levels of nesting, so there's still some
coverage of that scenario.
Components: OpenGL
VK-GL-CTS issue: 878
Affects:
- KHR-GL4*.arrays_of_arrays_gl.InteractionFunctionCalls1
- KHR-GL4*.arrays_of_arrays_gl.InteractionFunctionCalls2
- KHR-GL4*.arrays_of_arrays_gl.InteractionArgumentAliasing1
- KHR-GL4*.arrays_of_arrays_gl.InteractionArgumentAliasing2
- KHR-GL4*.arrays_of_arrays_gl.InteractionArgumentAliasing3
- KHR-GL4*.arrays_of_arrays_gl.InteractionArgumentAliasing4
- KHR-GL4*.arrays_of_arrays_gl.InteractionArgumentAliasing5
- KHR-GL4*.arrays_of_arrays_gl.InteractionArgumentAliasing6
- KHR-GL4*.arrays_of_arrays_gl.SubroutineFunctionCalls1
- KHR-GL4*.arrays_of_arrays_gl.SubroutineFunctionCalls2
- KHR-GL4*.arrays_of_arrays_gl.SubroutineArgumentAliasing1
- KHR-GL4*.arrays_of_arrays_gl.SubroutineArgumentAliasing2
- KHR-GL4*.arrays_of_arrays_gl.SubroutineArgumentAliasing3
- KHR-GL4*.arrays_of_arrays_gl.SubroutineArgumentAliasing4
- KHR-GLES31.core.arrays_of_arrays.InteractionFunctionCalls1
- KHR-GLES31.core.arrays_of_arrays.InteractionFunctionCalls2
- KHR-GLES31.core.arrays_of_arrays.InteractionArgumentAliasing1
- KHR-GLES31.core.arrays_of_arrays.InteractionArgumentAliasing2
- KHR-GLES31.core.arrays_of_arrays.InteractionArgumentAliasing3
- KHR-GLES31.core.arrays_of_arrays.InteractionArgumentAliasing4
- KHR-GLES31.core.arrays_of_arrays.InteractionArgumentAliasing5
- KHR-GLES31.core.arrays_of_arrays.InteractionArgumentAliasing6
Change-Id: Ia26ac5fb94799a2f80c9e2e9553c066d9c13c59b
Piers Daniell [Fri, 1 Dec 2017 21:17:54 +0000 (14:17 -0700)]
Fix OpenGL function prototype calling conventions
OpenGL doesn't use the default function calling convention
on all platforms, which is why gl.h defines the macros
GL_APICALL and GL_APIENTRY for applications to use when
defining their own OpenGL function prototypes. Not using
these may cause a crash if an OpenGL function is called
with the wrong calling convention.
Affects:
KHR-GL46.transform_feedback.*
Components: OpenGL
VK-GL-CTS issue: 882
Change-Id: I8766262b32a04b8eff8e110e1f116c2f915274e9
Alexander Galazin [Fri, 1 Dec 2017 14:02:32 +0000 (15:02 +0100)]
Update KC CTS reference
Component: OpenGL
Change-Id: I5cfa8633cdc7af147351951e7c1cd89908ea3f45
Piers Daniell [Mon, 27 Nov 2017 19:26:19 +0000 (12:26 -0700)]
Add OpenGL 4.6 configuration
The CTS-Configs.gl46 entry was missing from the configs.qpa
file and results in validation failure.
Affects:
cts-runner --type=gl46
Components: OpenGL
VK-GL-CTS issue: 866
Change-Id: I16745a7ce36040f8ef60cfbb1e4ffac250b289ee
Iago Toral Quiroga [Thu, 23 Nov 2017 08:30:35 +0000 (09:30 +0100)]
Fix maximum location for varying location limit test
The varying location limit tests for enhanced layouts produces
shader code where it attempts to use a location that is 1 slot
beyond the limit available for each shader stage and expects
linker errors as a result.
However, the functions it uses to compute the maximum location
slot available in a shader stage also consider the stage it will
be linked to to cap that limit (which is required to produce valid
shader code in other tests that also use these functions).
This affects scenarios where, for example, one stage allows up to
32 output locations but the stage it is linked to only accepts
up to 16 inputs. In such scenario, these functions will return
a limit of 16 for the stage that can output 32 so the shader code
produced can link.
However, the varying location limit test intends to produce an
invalid location by adding 1 to the maximum location slot for that
stage (so it wants to produce 32 + 1 = 33 in the case above), so
in this particular case we want the connecting stage to be
ignored.
Add a boolean flag to these functions so we can select between both
behaviors when computing the maximum location for a given stage
and make the varying location limit test force the behavior where
the maximum location in the connecting shader stages is ignored.
Components: OpenGL
VK-GL-CTS issue: 863
Affects:
KHR-GL45.enhanced_layouts.varying_location_limit
KHR-GL46.enhanced_layouts.varying_location_limit
KHR-GL45.enhanced_layouts.varying_locations
KHR-GL46.enhanced_layouts.varying_locations
KHR-GL45.enhanced_layouts.varying_array_locations
KHR-GL46.enhanced_layouts.varying_array_locations
KHR-GL45.enhanced_layouts.varying_components
KHR-GL46.enhanced_layouts.varying_components
KHR-GL45.enhanced_layouts.varying_array_components
KHR-GL46.enhanced_layouts.varying_array_components
Change-Id: If766582bf494942abdad3e1dde5e1ed8149f4333
Nicolai Hähnle [Mon, 20 Nov 2017 16:45:49 +0000 (17:45 +0100)]
gl3cTransformFeedbackTests: use delete[] to match new[]
C++ requires that memory allocated by new[] is freed using delete[]
instead of just plain delete. This does not really have an effect in
practice, but it does fix an error reported by Address Sanitizer:
=================================================================
==12338==ERROR: AddressSanitizer: alloc-dealloc-mismatch (operator new [] vs operator delete) on 0x6020000634b0
#0 0x7f21af5b3580 in operator delete(void*) (/usr/lib/x86_64-linux-gnu/libasan.so.3+0xc8580)
#1 0x55d795f901ae in gl3cts::TransformFeedback::APIErrors::testExtension1() /home/nha/amd/tests/cts/external/openglcts/modules/gl/gl3cTransformFeedbackTests.cpp:837
#2 0x55d795f9d71f in gl3cts::TransformFeedback::APIErrors::iterate() /home/nha/amd/tests/cts/external/openglcts/modules/gl/gl3cTransformFeedbackTests.cpp:251
#3 0x55d795f64736 in gl3cts::TestCaseWrapper::iterate(tcu::TestCase*) /home/nha/amd/tests/cts/external/openglcts/modules/gl/gl3cTestPackages.cpp:100
#4 0x55d797d10682 in tcu::TestSessionExecutor::iterateTestCase(tcu::TestCase*) /home/nha/amd/tests/cts/framework/common/tcuTestSessionExecutor.cpp:272
#5 0x55d797d10798 in tcu::TestSessionExecutor::iterate() /home/nha/amd/tests/cts/framework/common/tcuTestSessionExecutor.cpp:135
#6 0x55d797cf0748 in tcu::App::iterate() /home/nha/amd/tests/cts/framework/common/tcuApp.cpp:173
#7 0x55d795f15967 in main /home/nha/amd/tests/cts/framework/platform/tcuMain.cpp:55
#8 0x7f21adceb3f0 in __libc_start_main (/lib/x86_64-linux-gnu/libc.so.6+0x203f0)
#9 0x55d795f5e449 in _start (/mnt/
9a5f7d49-0504-436d-83ae-
10198f821499/nha/amd/tests/cts/build/external/openglcts/modules/glcts+0xa55449)
0x6020000634b0 is located 0 bytes inside of 3-byte region [0x6020000634b0,0x6020000634b3)
allocated by thread T0 here:
#0 0x7f21af5b30a0 in operator new[](unsigned long) (/usr/lib/x86_64-linux-gnu/libasan.so.3+0xc80a0)
#1 0x55d795f900da in gl3cts::TransformFeedback::APIErrors::testExtension1() /home/nha/amd/tests/cts/external/openglcts/modules/gl/gl3cTransformFeedbackTests.cpp:826
Component: OpenGL
VK-GL-CTS issue: 858
Affects:
KHR-GL45.transform_feedback.api_errors_test
Change-Id: Ie0872535d224c27cc6864fb71aaf7eb219661a02
Iago Toral Quiroga [Wed, 15 Nov 2017 07:44:22 +0000 (08:44 +0100)]
Fix expectation for multisample SPIR-V test
The last subtest of the builtin variable decorations
SPIR-V test renders to a 2-sample multisample FBO by
clearing to black and then using a fragment shader
that uses gl_SampleMask to select only one of the 2
samples. The selected sample always outputs 100% green.
Because the sample locations can be different acrosss
implementations, we can't tell what the resolved green
component will be, only that it cannot be 0x00 (since
one green sample is selected) and it cannot
be 0xFF (since sample coverage is only 50%).
Modules: OpenGL
VK-GL-CTS issue: 850
Affects:
KHR-GL45.gl_spirv.spirv_validation_builtin_variable_decorations_test
KHR-GL46.gl_spirv.spirv_validation_builtin_variable_decorations_test
Change-Id: I7f225199954c9193ff0b5e962d4de0ca905a5a76
Scott D Phillips [Tue, 14 Nov 2017 22:06:50 +0000 (14:06 -0800)]
Remove luminance and 3D depth requirement from ARB_texture_filter_minmax
Support not required by the extension.
Component: OpenGL
VK-GL-CTS issue: 849
Affects:
* KHR-GL45.texture_filter_minmax_tests.*
Change-Id: I53ede0343aa15e7528e12bf387f337f70ac155dd
Daniel Koch [Wed, 15 Nov 2017 15:20:54 +0000 (10:20 -0500)]
Add interaction with NVX_cross_process_interop
Update registry references to include
GL_NV_gpu_multicast and GL_NVX_cross_process_interop
VK-GL-CTS issue: 851
Component: OpenGL
Affects:
KHR-GL45.direct_state_access.buffers_errors
KHR-GL46.direct_state_access.buffers_errors
Change-Id: I42cdb7ad37f9c2ed60547aa817d263cd777f0f6f
Neil Roberts [Wed, 15 Nov 2017 21:45:39 +0000 (22:45 +0100)]
Fix strict aliasing violations in gl3cCommonBugsTests
In two places in PerVertexValidationTest::getTestIterationProperties
it first stores a value in a pointer to an enum and then later tries
to update the enum via a reference casted to an int reference. I
believe this violates the strict aliasing rules and so is undefined
behaviour. At least with gcc and -O2 the compiler seems to take
advantage of this and effectively ignores the initial assignment of
SHADER_STAGE_VERTEX for the PERVERTEX_DECLARATION_MISMATCH_TC_TE_VS
iteration. The result is that used_shader_stages gets or’d with an
unitialised value. This was causing a valgrind error and depending on
what the original value happened to be it can make the test
erroneously fail.
Modules: OpenGL
VK-GL-CTS issue: 853
Affects:
KHR-GL*.CommonBugs.CommonBug_PerVertexValidation
Change-Id: I1094de5c2114c035e339cdb524808f45cbe55782
Daniel Koch [Fri, 10 Nov 2017 19:42:51 +0000 (14:42 -0500)]
Documentation updates for OpenGL 4.6
Change-Id: I2d0a29ee9de24ae5aa616e0a5450c923ba3ba26d
Component: OpenGL
Affects: None
Alexander Galazin [Fri, 10 Nov 2017 15:13:33 +0000 (16:13 +0100)]
Remove mod from GL mustpass
Components: OpenGL
VK-GL-CTS public issue: 51
Affects:
KHR-GL*.gpu_shader_fp64.builtin.mod_*
Change-Id: Idac62351b9ca926cdcb411990c33e7332e8a9db7
Alexander Galazin [Fri, 10 Nov 2017 12:11:07 +0000 (13:11 +0100)]
Update KC CTS reference
Components: OpenGL
Change-Id: I241503a713fc4a15cdcf359c40e2b119b5b5fd4f
Juan A. Suarez Romero [Mon, 30 Oct 2017 12:50:23 +0000 (12:50 +0000)]
Use proper name for array-of-array varyings.
varying_structure_location tests have the piece of shader code:
```
struct Data {
dvec2 single;
dvec2 array[1];
};
layout (location = 0) flat in Data tes_gs_output[][1];
```
When test tries to verify "single" variable, it tries to access through
"tes_gs_output[0].single", which is not correct. Rather it must use
"tes_gs_output[0][0].single"
Affects:
* KHR-GL46.enhanced_layouts.varying_structure_locations
Components: OpenGL
VK-GL-CTS issue: 796
Change-Id: Ib1c2ac6877f46f0be8530156a184b8fa41349c40
Juan A. Suarez Romero [Mon, 6 Nov 2017 18:50:28 +0000 (19:50 +0100)]
GL_ARB_compute_shader required for OpenGL < 4.3
GL_ARB_compute_shader is part of OpenGL 4.3 and above.
For older versions, it is required to enable it explicitly.
Fixes:
KHR-GL31.CommonBugs.CommonBug_ParenthesisInLayoutQualifierIntegerValue
KHR-GL32.CommonBugs.CommonBug_ParenthesisInLayoutQualifierIntegerValue
KHR-GL33.CommonBugs.CommonBug_ParenthesisInLayoutQualifierIntegerValue
KHR-GL40.CommonBugs.CommonBug_ParenthesisInLayoutQualifierIntegerValue
KHR-GL41.CommonBugs.CommonBug_ParenthesisInLayoutQualifierIntegerValue
KHR-GL42.CommonBugs.CommonBug_ParenthesisInLayoutQualifierIntegerValue
Components: OpenGL
VK-GL-CTS issue: 818
Change-Id: I8e7ab78e29861547f0379feebce0375e1dabf47c
Juan A. Suarez Romero [Tue, 7 Nov 2017 12:24:25 +0000 (12:24 +0000)]
Load functions from ARB_uniform_buffer_object
While this extension is part of OpenGL 3.1 and above, it is not part of
OpenGL 3.0.
As we need to run the tests for OpenGL 3.0 too, and we need to use some
functions exposed by this extension, let's load them from the extension
itself, instead of relying in the core functions.
Affects:
* KHR-GL30.transform_feedback.draw_xfb_instanced_test
* KHR-GL30.transform_feedback.draw_xfb_stream_instanced_test
Components: OpenGL
VK-GL-CTS issue: 819
Change-Id: I2cd1652b4c2d3b55e009c44b85b451d207387b96
James Helferty [Thu, 19 Oct 2017 18:24:35 +0000 (14:24 -0400)]
Relax XFB begin check to match recent spec edit
Update tests to reflect a recent change to the GL 4.6 specification.
Previously, if a shader layout qualifier specified that xfb_stride was
non-zero for a given xfb_buffer, then a buffer object was required to
be bound at that transform feedback buffer binding point. After the
spec change, it is now legal to not have a buffer bound, provided no
outputs are associated with the buffer.
Components: OpenGL
VK-GL-CTS issue: 785
Affects:
* KHR-GL44/45/46.enhanced_layouts.xfb_stride_of_empty_list
* KHR-GL44/45/46.enhanced_layouts.xfb_stride_of_empty_list_and_api
Change-Id: I38622bad1af3bc8a2a4c683eadc184e6231f7c85
Adam Czupryna [Fri, 7 Jul 2017 09:42:37 +0000 (11:42 +0200)]
Add CTS_ARB_gl_spirv test implementation
Add CTS_ARB_gl_spirv test implementation that contains:
SpirvModulesPositiveTest,
SpirvModulesShaderBinaryMultipleShaderObjectsTest,
SpirvModulesStateQueriesTest,
SpirvModulesErrorVerificationTest,
SpirvGlslToSpirVEnableTest,
SpirvGlslToSpirVBuiltInFunctionsTest,
SpirvGlslToSpirVSpecializationConstantsTest,
SpirvValidationBuiltInVariableDecorationsTest
Affects:
KHR-GL45.gl_spirv.*
Components: Framework, OpenGL
VK-GL-CTS issue: 554
Change-Id: I0d54c70df619dbbd80b28350d2ac5e31aca08927
(cherry picked from commit
a24448cdd72ffdbd8f7f571886625b8a53100979)
Iago Toral Quiroga [Mon, 2 Oct 2017 11:10:20 +0000 (13:10 +0200)]
Fix attribute mapping in enhanced layouts tests for input arrays
Some tests in KHR-GL45.enhanced_layouts.varying_* (specifically
KHR-GL45.enhanced_layouts.varying_array_{locations,components})
use input arrays on the vertex shader, however, the code
that uploads vertex data for these (and other test cases) doesn't
account for arrays at all.
Also, the tests in this category that call
TextureTestBase::prepareAttributes to do the VAO setup seem to
provide inconsistent values for variable.m_descriptor.m_offset,
which also leads to incorrect vertex attribute mappings.
Specifically, in the 'component' variants of these tests
(such as varying_array_components), two different input variables
have m_offset = 0, which has the consequence that we end up
computing a vertex buffer with a size that's only big enough to
fit one of the inputs and we overwrite the data for the first
input with the data for the second input when we setup the VAO,
however, the 'location' variants (such as varying_array_locations),
see different offsets for different input variables, which leads to
a different VAO setup.
Components: OpenGL
VK-GL-CTS issue: 739
Affects:
KHR-GL45.enhanced_layouts.varying_*
KHR-GL46.enhanced_layouts.varying_*
Change-Id: Icec238a970e2cd67d6109a25fba980f756abb71b
Juan A. Suarez Romero [Wed, 18 Oct 2017 09:37:27 +0000 (09:37 +0000)]
Several fixes for texture_size_promotion
Specifically, this changeset fixes:
- Check for unsupported framebuffer configurations
- Setup data for depth/stencil textures
- Explicitly set minification/magnification filters
- Deal with the corner cases in precision when using fixed-point
textures.
Affects:
KHR-GL45.texture_size_promotion.functional
KHR-GL46.texture_size_promotion.functional
Components: OpenGL
VK-GL-CTS issue: 783
Change-Id: I3f191084b2a604f1a743759905d641d2b13db2ad
Iago Toral Quiroga [Tue, 17 Oct 2017 07:58:57 +0000 (09:58 +0200)]
Respect GL_MAX_TRANSFORM_FEEDBACK_INTERLEAVED_COMPONENTS
KHR-GL45.enhanced_layouts.xfb_override_qualifiers_with_api
generates shaders that capture XFB for varyings of all types.
For the cases of dmat4x3, dmat3x4 and dmat4, this requires
over 64 transform feedback components in total, which is larger
than the minimum value that implementations are required to
support for GL_MAX_TRANSFORM_FEEDBACK_INTERLEAVED_COMPONENTS,
which the OpenGL 4.6 spec sets in 64. For implementations that
support this minimum, the generated shaders are invalid and
expected to fail to link.
Check the required component size of the XFB for the generated
shaders and skip them if they exceed the implementation limit.
At worse, for implementations that only support the minimum
number of transform feedback components required by the spec
it would only skip the 3 largest types (out of 34 being tested).
Components: OpenGL
VK-GL-CTS issue: 776
Affects:
KHR-GL45.enhanced_layouts.xfb_override_qualifiers_with_api
Change-Id: I51c9da494f96a0746f74116f817d99afd070a9ee
Iago Toral Quiroga [Wed, 18 Oct 2017 10:08:54 +0000 (12:08 +0200)]
Component aliasing on vertex inputs is allowed
From the OpenGL 4.6 spec, section 11.1 Vertex shaders,
page 384:
"It is possible for an application to bind more than one
attribute name to the same location. This is referred
to as aliasing. This will only work if only one of the
aliased attributes is active in the executable program,
or if no path through the shader consumes more than one
attribute of a set of attributes aliased to the same
location. A link error can occur if the linker determines
that every path through the shader consumes multiple
aliased attributes, but implementations are not required
to generate an error in this case."
The varying_location_aliasing_with_mixed_types enhanced
layouts test generates negative tests to check that
component aliasing produces linker errors for all shader
stages, including vertex shaders when the aliasing occurs
on the vertex inputs. By the citation above, the spec
doesn't mandate implementations to produce an error in
this case, so these particular tests are invalid.
This change removes the test cases that expect an error
to be generated for shaders with component aliasing on
vertex shader inputs.
Components: OpenGL
VK-GL-CTS issue: 781
Affects:
KHR-GL45.enhanced_layouts.varying_location_aliasing_with_mixed_types
KHR-GL46.enhanced_layouts.varying_location_aliasing_with_mixed_types
Change-Id: Ie1cafc92c4b7023c3d911ec5fe883dad297029d5
Eleni Maria Stea [Tue, 13 Jun 2017 07:08:08 +0000 (09:08 +0200)]
Online compression for ETC2 compressed formats is not required
The DSA compressed texture tests currently require online compression for
ETC2 formats but drivers are not required to support this. Since the purpose
of the test is not to test the compression, when a compression format is not
supported we use already compressed data.
Components: OpenGL
VK-GL-CTS issue: 497
Affects:
KHR-GL45.direct_state_access.textures_compressed_subimage
Change-Id: Icff52745e9cf7a2acb39aea71d31f4ff827ea01c
Kenneth Graunke [Thu, 12 Oct 2017 04:39:05 +0000 (21:39 -0700)]
Delete incorrect invalid map transform feedback subtest.
The citation given in this test comes from the OpenGL man pages, which
are not authoritative. The ES 3.2 and GL 4.x specifications say:
"Effects of Mapping Buffers on Other GL Commands
Any GL command which attempts to read from, write to, or change the
state of a buffer object may generate an INVALID_OPERATION error if
all or part of the buffer object is mapped. However, only commands
which explicitly describe this error are required to do so. If an
error is not generated, using such commands to perform invalid
reads, writes, or state changes will have undefined results and may
result in GL interruption or termination."
Importantly, it says "_may_ generate an INVALID_OPERATION error". Some
drivers choose not to implement this to avoid the CPU overhead required
to enforce this optional restriction.
Affects:
- KHR-*.transform_feedback.api_errors_test
VK-GL-CTS issue: 765
Change-Id: I34ca78926fe0f311db096de98b71e0981ee91d53
Components: OpenGL
Jeannot Breton [Fri, 6 Oct 2017 13:56:17 +0000 (08:56 -0500)]
Skip non accelerated pixel format
They were previously skipped in older CTS version,
but when we reorganized the code to merge all CTS together the
related code got remove
We also fix the way we detect hardware acceleration
Affects:
pixel formats selected by cts-runner on Windows
Components: OpenGL
VK-GL-CTS issue: 692
Change-Id: I9968d85b555fe41aea8e64b885f182407e1ce95e