Fix for the bounding box tests with GL 4.5

Affects:
dEQP-GL*functional.primitive_bounding_box.*

Components: AOSP
VK-GL-CTS issue: 3104

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Change-Id: I38db9fc893a685893dbd6c2de8ab15ab3e0db11e

Fix stencil data unpacking in dEQP-VK.fragment_operations test

The test copies only one aspect (depth or stencil, not both) of the image
to the buffer, so host unpacking code should not use combinded DS format
 - instead S8_UIN format should be used in stencil tests.

The problem occurs in stencil tests, on implementations not supporing
S8_UINT format, where D+S format is used.

Components: Vulkan

VK-GL-CTS Issue: 3160

Affects: dEQP-VK.fragment_operations.*stencil

Change-Id: I27b69c02ebf08fa19f512cdd786ebe596d91cfeb

Merge "Merge vk-gl-cts/vulkan-cts-1.2.7 into vk-gl-cts/master"

Merge vk-gl-cts/vulkan-cts-1.2.7 into vk-gl-cts/master

Change-Id: I53860788b7f8844254a39df3ed14f621f34e90f2

Fix rectangular line tests

When we're testing against rectangular lines, we need to generate the
rectangular line geometry as well, otherwise we'll test against
parallelogram lines instead.

Components: Vulkan, Framework

Public issue: 274

Affected Tests: dEQP-VK.rasterization.primitives.*.rectangular_line_*

Change-Id: I514646e94d2faf1e996950fd6b3a738797f84ff0

Merge vk-gl-cts/opengl-es-cts-3.2.7 into vk-gl-cts/master

Change-Id: I719241f2424bd8e9f3c0d4b6898706289181b552

Merge vk-gl-cts/vulkan-cts-1.2.6 into vk-gl-cts/vulkan-cts-1.2.7

Change-Id: I015e15ff85941f0a8756bd4b13e93fcd9161d62e

Merge vk-gl-cts/opengl-es-cts-3.2.6 into vk-gl-cts/opengl-es-cts-3.2.7

Change-Id: Ic894ed63f8c389f9f5485e3f573a66fd9901f5aa

Fix vector access in EGL partial update tests

The vector of damage regions was being accessed even when the vector
was empty, which is invalid.

This changes the behaviour of the test, which used to invalidate the
whole framebuffer when it found no damage regions, but not invalidates
none of it. This seems like the correct behaviour.

Components: EGL
Affects: dEQP-EGL.functional.partial_update.*

Change-Id: I8c5500a85d000d16b240359de7eadc67144fb7d3

Fix dEQP-VK.glsl.opaque_type_indexing.ssbo*_compute validation error

Tests were not checking maxPerStageDescriptorStorageBuffers limits
in the compute shaders, which use more (6) than the minimum limit
set by the spec for that (4).

Components: Vulkan
VK-GL-CTS issue: 3137

Affected tests:

 dEQP-VK.glsl.opaque_type_indexing.ssbo*_compute

Signed-off-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>
Change-Id: I016fb5024d637e36869ae7e3f78467c91d0e8817

Fix dEQP-VK.image.extended_usage_bit.texture_* validation errors

Shader requires fragmentStoresAndAtomics but there wasn't any support
check for this physical device feature.

Components: Vulkan

VK-GL-CTS Issue: 3144

Affects:
dEQP-VK.image.extended_usage_bit.texture_read.*
dEQP-VK.image.extended_usage_bit.texture_write.*

Change-Id: I87f69256501519fe524159f02eb14ee6ac0e0092

Check depth bias clamp feature when needed

The inverted depth range tests were using the depth bias clamp feature
with a nonzero depth bias clamp value without checking if the feature
was available in the device.

Affected tests:
dEQP-VK.draw.inverted_depth_ranges.*

Components: Vulkan
VK-GL-CTS issue: 3128

Change-Id: I471235a0da38eb4c244e87e0d5851b8cb810f990

Add barrier between the atomic-compare-exchanges

Each of the second atomic-compare-exchange instructions relies on fact
that all the invocations have executed the first atomic-compare-exchange
instruction.  So insert a workgroup barrier between the two.

Affects:
dEQP-VK.spirv_assembly.instruction.spirv1p4.entrypoint.comp_workgroup_entry_point

Components: Vulkan

VK-GL-CTS issue: 3120

Change-Id: I95f897ceebcea8fdd5d6bda885a701be20d07ab1

Check for integer texture support in negative teximage tests

Affects:
KHR-GLES2.texture_3d.filtering.combinations.negative

Components: OpenGL
VK-GL-CTS issue: 3134

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Change-Id: I352d408a486fa12b406fadd6ec6b8d24e1cd211d

Merge vk-gl-cts/vulkan-cts-1.2.5 into vk-gl-cts/vulkan-cts-1.2.6

Change-Id: I16fde981fb8a0d98d4ed0522cb69a429a68545d0

Merge vk-gl-cts/vulkan-cts-1.2.4 into vk-gl-cts/vulkan-cts-1.2.5

Change-Id: I133af15a49b220d94c26f9db70e7830fbe167b72

Allow the use of the standard types

Removes the script that enforces the use of
the de-wrappers instead of the standard types.

VK-GL-CTS issue: 2963

Components: Framework

Change-Id: I5e2ff38a2b2cbc63a3c756a1e0260833ecd9381e
(cherry picked from commit ce1ae17c4785a8a57214a16fac46e28832956fa9)

Update vkDeviceFeatureTest.inl

Due to merge order issue, vkDeviceFeatureTest.inl was not updated
properly.
This CL updates the file.

Affects:

dEQP-VK.api.device_init.create_device_unsupported_features

Components: Framework, Vulkan

VK-GL-CTS issue: 3166

Change-Id: I242e387b707094b124b1f368b0e11d8a17f92dfc

Notice of withdrawal of Vulkan CTS 1.2.4.x

Components: Vulkan
Change-Id: I82e747ce2898064804c88bfa98eab3c8c0f48953

Fix GL_MAX_FRAGMENT_INTERPOLATION_OFFSET limit check

The spec is fixed to align with Vulkan, requiring the minimum value for
GL_MAX_FRAGMENT_INTERPOLATION_OFFSET to be 0.5-ULP, where

    ULP = 1/pow(2,GL_FRAGMENT_INTERPOLATION_OFFSET_BITS)

Affects:
dEQP-GL45.functional.state_query.multisample_interpolation.max_fragment_interpolation_offset.get_float
dEQP-GLES31.functional.state_query.multisample_interpolation.max_fragment_interpolation_offset.get_float
KHR-GL44.shader_multisample_interpolation.api.api
KHR-GL45.shader_multisample_interpolation.api.api
KHR-GL46.shader_multisample_interpolation.api.api
KHR-GLES31.core.shader_multisample_interpolation.api.api
KHR-GL45.limits.max_fragment_interpolation_offset
KHR-GL46.limits.max_fragment_interpolation_offset

Components: OpenGL
Khronos OpenGL issue: #149

Change-Id: I153a05ef6db8f50a6aa337188e5aeb8d504fbbdd

Test non-Sample Shading FragCoord.xy

In these test we are testing that FragCoord reflect the location of the
center of the fragment when Sample Shading is disabled. There is also a
set of tests which are using Centroid interpolation decoration on
FragCoord.

Components: Vulkan

VK-GL-CTS Issue: 3040

New tests: dEQP-VK.glsl.builtin_var.fragcoord_msaa_input.*no_sample*

Change-Id: Ic84bd57d8118e001bcd9e52ecb2e4193ff332ba6

Test vkCreateDevice with unsupported features

vkCreateDevice should return VK_ERROR_FEATURE_NOT_PRESENT if any
requested feature is not supported.

The old test covered only Vulkan 1.0 core features.
This CL extends the test to cover also Vulkan 1.1 and 1.2 core features
and features provided by extensions.

Affects:

dEQP-VK.api.device_init.create_device_unsupported_features

Components: Framework, Vulkan

VK-GL-CTS issue: 3031

Change-Id: I319ac28bad3d4524ef4a1477c4680e4e8f91f4da

Test anisotropic filtering of single-level images

Anisotropic filtering tests did not take into account
test cases that use images with only a single level.
This change adds a couple of new tests that test this functionality.

Affects:

dEQP-VK.texture.filtering_anisotropy.*

New tests:

dEQP-VK.texture.filtering_anisotropy.single_level.*

Components: Vulkan

VK-GL-CTS issue: 3039

Change-Id: I9656cfba1e94a112ac84ca57f00c697d47637ff1

Merge vk-gl-cts/vulkan-cts-1.2.7 into vk-gl-cts/master

Change-Id: Icc4ff9e543c2dbb7582df992e59ce3ae644b67fe

Limit storage buffer sizes in reconvergence tests

Return a not supported error if the storage buffer size is greater
than maxStorageBufferRange gpu limits.

Components: Vulkan

VK-GL-CTS Issue: 3114

Affected tests: dEQP-VK.reconvergence.*

Change-Id: Iecfbe0138b7d70420a353c2176110b843a49b091

Fix wrong VkAccessFlag when copying color image to buffer

Components: Vulkan
VK-GL-CTS issue: 3083

Affected tests:

   dEQP-VK.fragment_operations.early_fragment.*

Change-Id: Idb809fdd9efad586166b3d94c2a305e8c234274f
Signed-off-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>

Merge vk-gl-cts/vulkan-cts-1.2.6 into vk-gl-cts/vulkan-cts-1.2.7

Change-Id: Ibbd5912f0843ba22ee8402ee4a96aaf6ed53b303

Merge vk-gl-cts/vulkan-cts-1.2.5 into vk-gl-cts/vulkan-cts-1.2.6

Change-Id: Iac3e8ce0a318caed752aeb933994227278d9582f

Merge vk-gl-cts/vulkan-cts-1.2.4 into vk-gl-cts/vulkan-cts-1.2.5

Change-Id: Ifc19ab1fcdcfa681d0fef9b3c5e3448d12813fdf

Merge vk-gl-cts/vulkan-cts-1.2.3 into vk-gl-cts/vulkan-cts-1.2.4

Change-Id: Iab8e5c27a9f07ca0459c97bc8d72cd93b8ce675a

Merge vk-gl-cts/vulkan-cts-1.2.2 into vk-gl-cts/vulkan-cts-1.2.3

Change-Id: Icdbb121ec7a828457ff9df34efe018fbf845e3f9

Accept NaN as valid pass value in shader precision atan2 tests

Spec mentions that implementation may assume arguments are not
Inf if the entrypoint is not declared with SignedZeroInfNanPreserve
and can perform optimizations.

Components: Vulkan

VK-GL-CTS Issue: 3073

Affected tests:
dEQP-VK.spirv_assembly.instruction.compute.float16.*.atan2
dEQP-VK.spirv_assembly.instruction.graphics.float16.*.atan2*
dEQP-VK.glsl.builtin.precision.atan2.*

Change-Id: I6e1fbc14a5654422cf88e1453fae0774bfce0bb6

Add missing checking of integer dot product features

The requested features for dot products were being set up by the tests
but they were never actually validated. (In practice this probably won't
be a problem because they are compulsory with the extension and also in
the new core version, but it makes the code quite confusing).

Components: Vulkan
Affects: dEQP-VK.spirv_assembly.*

Change-Id: I62558d27f9bb360e3912b90964739bb1356df4b8

Don't silently drop features that were requested

The helper functions would just silently not request features that the
actual test code requested if it didn't think they were useful. This has
the potential for weird bugs, so stop doing it. Instead assert that
tests are not asking for features that don't seem useful. This will
still prevent tests being quietly skipped for features they won't use,
but will show up immediately if a test really does need these features
for something unexpected.

Some of the existing tests were requesting features that they weren't
going to use (mostly StoresAndAtomics for different stages), and so
these are fixed not to request them.

Components: Vulkan
Affects: dEQP-VK.spirv_assembly.*

Change-Id: I7acfc86c13067ee231cf71389d002886272081dd

Update checking of VulkanFeatures structs

The spirv_assembly tests defined various helpers for checking whether
certain features are supported but they worked inconsistently and were
repeated in various places. This change pulls the checking into a single
function which can be used to check instead of having to call a while
set of functions to achieve the same end.

The higher-level checking functions have always masked out a subset of
the features that are requested, and this is not changed, so requests
for certain things that look correct will silently not check all the
features.

The extension feature enums that were defined just for the checking
system are replaced with the standard structs. They're easier to use
because they don't need to be defined afresh when extensions are added
here, and because the names are automatically consistent with the names
that are known from the API. In at least 1 case the existing names were
pretty hard to decode correctly.

Components: Vulkan
Affects: dEQP-VK.spirv_assembly.*

Change-Id: Icd2fc3f7d37e0ec3cfed9a80f3152617c8f599a3

Merge vk-gl-cts/opengl-es-cts-3.2.7 into vk-gl-cts/master

Change-Id: I4affbd86cc7b03860722848f5234401ed68bb9d0

Tests for OpenGL ES 3.1 and 3.2 compressed formats

Adds texture specification and simple draw tests for all
OpenGL ES 3.1 and 3.2 compressed texture formats.

Adds tests for error values from compressed texture
related api calls.

New tests:

KHR-GLES31.core.compressed_format.*
KHR-GLES32.core.compressed_format.*

Components: OpenGL

VK-GL-CTS issue: 2174

Change-Id: I5a05962e79fca31ae526ec4e06f22c43d0a0f5b4

dEQP-VK.subgroups.*ray_tracing* missing extension check

Components: Vulkan
VK-GL-CTS issue: 3003

Affected:

dEQP-VK.subgroups.*ray_tracing*

Change-Id: I25719d00c6d0180c03726d73497727754f9539d2
Signed-off-by: Samuel Iglesias Gonsálvez <siglesias@igalia.com>

Re-enable protected memory tests

AHB protected memory tests were previously removed due to
validation errors. This change adds them back while also
fixing the errors.

VK-GL-CTS Issue: 3030

Affects:

dEQP-VK.api.external.memory.android_hardware_buffer.image_formats.*

Components: Vulkan
Change-Id: I0ba4809eec82b069d9f36b624c65c7fe7f8df7a3

Depth/Stencil descriptor tests

New set of tests that try to use depth/stencil images as descriptors
and/or as read-only attachments with using layouts that were previously
not covered by other CTS tests.

New tests:
dEQP-VK.image.depth_stencil_descriptor.*

Components: Vulkan
VK-GL-CTS issue: 2925

Change-Id: I00e022afb0b7bc9561e938e3c35a026d8ade2ca9

Test multiview with different depth ranges

This change extends existing tests to also cover creation of
viewports with different depth ranges, and maked sure that
depth clipping/clamping works correctly across the different
viewports.

Issue raported by layers in existing tests were also fixed.

Components: Vulkan

VK-GL-CTS issue: 2500

Affects:
dEQP-VK.multiview.*

New tests:
dEQP-VK.multiview.depth_different_ranges.3_6_12_9_6_12_9_3_6_12_9_3
dEQP-VK.multiview.renderpass2.depth_different_ranges.3_6_12_9_6_12_9_3_6_12_9_3

Change-Id: I0085eeda3aac2e36ab7013bf5a83009ce9ea94b9

Merge vk-gl-cts/vulkan-cts-1.2.7 into vk-gl-cts/master

Change-Id: Ib651fa8492c869d02b1185442b1a7ed431a2b7c6

Merge vk-gl-cts/vulkan-cts-1.2.6 into vk-gl-cts/vulkan-cts-1.2.7

Change-Id: I0e58f89d8cfdfaaa400f700ea87882cb281f080f

Merge vk-gl-cts/vulkan-cts-1.2.5 into vk-gl-cts/vulkan-cts-1.2.6

Change-Id: I2d3ee8a323e62c13c9f3e5ff499ff3614c802b29

Merge vk-gl-cts/vulkan-cts-1.2.4 into vk-gl-cts/vulkan-cts-1.2.5

Change-Id: I521f0b1f6f4c3a86e8612239ce2134be76e5d16a

Merge vk-gl-cts/vulkan-cts-1.2.3 into vk-gl-cts/vulkan-cts-1.2.4

Change-Id: Ic1ab6043f9d70b8491192137c6d8c2fc12f2112d

Merge vk-gl-cts/vulkan-cts-1.2.7 into vk-gl-cts/master

Change-Id: I09891e0da9f54df761a1d22474d04ac0ae5fe498

Merge vk-gl-cts/vulkan-cts-1.2.6 into vk-gl-cts/vulkan-cts-1.2.7

Change-Id: I7bd0d85348a194cda0ec5922fa93a2cb5dd02e19

Test blitting from ASTC formats

This change extends existing tests to also test blitting of all
ASTC compressed formats onto all compatible uncompressed formats.

VK-GL-CTS issue: 2366

Components: Vulkan

Affects:
dEQP-VK.api.copy_and_blit.core.blit_image.all_formats.color.2d.astc*

Change-Id: I62a9fcf056c9468f02ee0999734f6d1b5a438beb

Fix workgroup memory tests for int16s

Most of these tests were fixed to request the correct features in
634882efd17, but the int16 and uin16 variants still didn't request the
shaderInt16 feature.

Components: Vulkan
Affects: dEQP-VK.spirv_assembly.instruction.compute.workgroup_memory.int16
         dEQP-VK.spirv_assembly.instruction.compute.workgroup_memory.uint16

Change-Id: If848202fc96bd191d67be3d11cb09b7890f936d7

Check support for EGL 1.5 or higher when using GLES3

Layered AHB tests are using GLES3 tests, but the
check for a required  EGL version supporting that
was missing. Now these tests throw not supported when
the EGL version is below 1.5.

VK-GL-CTS Issue: 3076

Affects:

dEQP-EGL.functional.image.*android_native_array*

Components: EGL
Change-Id: I4e79e8447ec6a3a243b30ab1f64dbb00a4fa4e96

Accept NaN as valid in atan2 precision tests

The Vulkan CTS has been updated
(https://gerrit.khronos.org/c/vk-gl-cts/+/7857) so that it is not
required to generate correct results for atan(inf, inf), and this makes
the corresponding change in GL. The GLSL spec is not as clear on what is
required here, but does say:
    Built-in functions not listed above and not defined as equations of
    the above have undefined precision. These include, for example, the
    trigonometric functions

In general, I think that it is expected that the default state of Vulkan
is a match for GL.

Components: OpenGL
VK-GL-CTS issue: 3073
Affects: dEQP-GLES3.functional.shaders.builtin_functions.precision.atan2.highp_*

Change-Id: Ifae97a2e86962f5aff68d5eed09ba38069445468

Fix AHB external memory test mip levels

AHB external memory tests were testing mipmaps using a
queried maximum number of supported mipmap levels, but
at the same time using an image size of 64x64.
The number of mip levels need to match the number for a
full mip chain, in this case 7. Now a correct number of
mip levels is used, which also gets rid of a validation
error.

VK-GL-CTS Issue: 3112

Affects:

dEQP-VK.api.external.memory.android_hardware_buffer.image_formats.*

Components: Vulkan
Change-Id: I1c6d09f93c06ac1c52cea283fd86359ccde01e4f

Update Vulkan headers

Components: Vulkan
Change-Id: I1fc2ec802183e2a1efb50915b66bae7422730f8d

Add tests for VK_KHR_shader_integer_dot_product

Components: Vulkan, Framework

VK-GL-CTS issue: 2729

New tests:
dEQP-VK.spirv_assembly.instruction.compute.opsdotkhr.*
dEQP-VK.spirv_assembly.instruction.compute.opudotkhr.*
dEQP-VK.spirv_assembly.instruction.compute.opsudotkhr.*
dEQP-VK.spirv_assembly.instruction.compute.opsdotaccsatkhr.*
dEQP-VK.spirv_assembly.instruction.compute.opudotaccsatkhr.*
dEQP-VK.spirv_assembly.instruction.compute.opsudotaccsatkhr.*

Affects:
dEQP-VK.api.info.get_physical_device_properties2.features
dEQP-VK.api.info.get_physical_device_properties2.properties

Signed-off-by: Stuart Brady <stuart.brady@arm.com>
Signed-off-by: Pablo Cascón <pablo.cascon@arm.com>
Signed-off-by: Neil Hickey <neil.hickey@arm.com>
Change-Id: Ia618d9eb33369e72eedc5845342e17e1ee12da3a
(cherry picked from commit 12d5b5fe34a4cde919e1d9d3949154bf33ef21f2)

Fulfill OpControlBarrier requirement

The valid usage requirement in vulkan specification 1.2.170
VUID-StandaloneSpirv-None-04641 requires: "If the Scope for
memory is Invocation, then memory semantics must be None".
New SPIRV-Tools enforces it.

Fix tests:
 * dEQP-VK.spirv_assembly.instruction.graphics.barrier.in_function
 * dEQP-VK.spirv_assembly.instruction.graphics.barrier.in_switch
 * dEQP-VK.spirv_assembly.instruction.graphics.barrier.in_if
 * dEQP-VK.spirv_assembly.instruction.graphics.barrier.after_divergent_if
 * dEQP-VK.spirv_assembly.instruction.graphics.barrier.in_loop

Update tests:
 * dEQP-VK.spirv_assembly.instruction.graphics.barrier.*

Components: Vulkan

Change-Id: Ie6997fdb36c0327d88a2b7b6b4826dfee690377d

Fix Clang -Wcomma warnings

Replace commas that were most likely intended to be semicolons
instead.

Components: Vulkan, OpenGL ES

VK-GL-CTS issue: 3097

Change-Id: I106292084a12db1517da33c26a88df5232ea33fd

Fix qpReleaseInfo generation with non-ASCII paths

On Windows CMake fails to generate qpReleaseInfo.inl if DE_GIT_DIR is a
path with non-ASCII characters. Specify ENCODING UTF8 for git output.

Components: Framework

VK-GL-CTS issue: 3096

Change-Id: I8368ba5e2670b723471e2d7a7beec1f11376c49f

Add tests for VK_EXT_load_store_op_none

VK-GL-CTS Issue: 2971

New tests:

dEQP-VK.renderpass*.load_store_op_none.*

Components: Vulkan
Change-Id: If3eb79e75b14f96037d14fcaa469230d3434a729
(cherry picked from commit b0025b814cad4fd699761578eaec575fe76f7f75)

Remove invalid SPIR-V decorations

Some SPIR-V asm tests used invalid Binding and DescriptorSet decorations.
This CL removes those decorations.

Affects:

dEQP-VK.spirv_assembly.instruction.compute.physical_pointers.complex_types_compute.*
dEQP-VK.spirv_assembly.instruction.compute.pointer_parameter.buffer_memory
dEQP-VK.spirv_assembly.instruction.graphics.pointer_parameter.buffer_memory*
dEQP-VK.spirv_assembly.instruction.graphics.16bit_storage.input_output_float_16_to_16x2.scalar
dEQP-VK.spirv_assembly.instruction.graphics.16bit_storage.input_output_float_16_to_16x2.vec2
dEQP-VK.spirv_assembly.instruction.graphics.16bit_storage.input_output_int_16_to_16x2.scalar_int
dEQP-VK.spirv_assembly.instruction.graphics.16bit_storage.input_output_int_16_to_16x2.scalar_uint
dEQP-VK.spirv_assembly.instruction.graphics.16bit_storage.input_output_int_16_to_16x2.ivec2
dEQP-VK.spirv_assembly.instruction.graphics.16bit_storage.input_output_int_16_to_16x2.uvec2

Components: Vulkan

VK-GL-CTS issue: 3063

Change-Id: Icfee33f9ce041bf60a19fba55aef9fdab0747148

Merge branch 'vulkan-cts-1.2.7'

Change-Id: I206c4059e348c7163a597e47e76a2146e23d7fbc

Merge branch 'vulkan-cts-1.2.6' into vulkan-cts-1.2.7

Change-Id: Iec55ddd12f12935ab70d00af5841a8b22f9a3c89

Add tests for VK_KHR_shader_integer_dot_product

Components: Vulkan, Framework

VK-GL-CTS issue: 2729

New tests:
dEQP-VK.spirv_assembly.instruction.compute.opsdotkhr.*
dEQP-VK.spirv_assembly.instruction.compute.opudotkhr.*
dEQP-VK.spirv_assembly.instruction.compute.opsudotkhr.*
dEQP-VK.spirv_assembly.instruction.compute.opsdotaccsatkhr.*
dEQP-VK.spirv_assembly.instruction.compute.opudotaccsatkhr.*
dEQP-VK.spirv_assembly.instruction.compute.opsudotaccsatkhr.*

Affects:
dEQP-VK.api.info.get_physical_device_properties2.features
dEQP-VK.api.info.get_physical_device_properties2.properties

Signed-off-by: Stuart Brady <stuart.brady@arm.com>
Signed-off-by: Pablo Cascón <pablo.cascon@arm.com>
Signed-off-by: Neil Hickey <neil.hickey@arm.com>
Change-Id: Ia618d9eb33369e72eedc5845342e17e1ee12da3a

Allow GL_RGB9_E5 as a valid fbo format

If GL_QCOM_render_shared_exponent is available then treat
GL_RGB9_E5 as a valid internal format for fbo. Also, allow
copy of RGB9_E5 pixel data.

Components: OpenGL
VK-GL-CTS Issue: 2898

Affects:
KHR-GLES3.packed_pixels.rectangle.rgb9_e5
KHR-GLES3.packed_pixels.pbo_rectangle.rgb9_e5
KHR-GLES3.packed_pixels.varied_rectangle.rgb9_e5

Change-Id: Ie57d74b9bbd11cf7ccca175813077c22405a0926
(cherry picked from commit db8e2899b1f0ab4677aa24af6fad4119f5fe1fa2)

Merge branch 'vulkan-cts-1.2.7'

Change-Id: I0034a268b5c4dd313953bf4cd8b6511fd3b5635e

Remove unreachable breaks

Fix warnings for Clang builds with -Wunreachable-code-break.

Components: Framework, OpenGL, OpenGL ES, Vulkan

VK-GL-CTS issue: 3074

Change-Id: Ib658e631bad3d06d787c6151f744e8db3e850fb4

Add GL_RGB9_E5 as renderable ext format in fbo tests

-If GL_QCOM_render_shared_exponent is supported then treat
GL_RGB9_E5 as a renderable format for fbo completeness tests

Components: OpenGL
VK-GL-CTS Issue: 3066

Affects:
dEQP-GLES3.functional.fbo.completeness.renderable.renderbuffer.color0.rgb9_e5
dEQP-GLES3.functional.fbo.completeness.renderable.texture.color0.rgb9_e5

Change-Id: I74566a5b56c7528c89ac97423c5b6f9207397c70
(cherry picked from commit d18b37f7319839e2be970a96fecdeb2cdb1db7e4)

Add a batch of GraphicsFuzz coverage tests

This commit adds a batch of GraphicsFuzz coverage tests.

The tests have been separated by a request.

Components: Vulkan

New Tests:

dEQP-VK.graphicsfuzz.cov-dfdx-dfdy-after-nested-loops
dEQP-VK.graphicsfuzz.cov-nested-loops-float-array-select-by-fragcoord
dEQP-VK.graphicsfuzz.cov-nested-loops-set-struct-data-verify-in-function

Change-Id: I33b01c2f14318fb65b1f895eea89433828c82d51

Merge branch 'vulkan-cts-1.2.6' into vulkan-cts-1.2.7

Change-Id: I5d1960f36c445e344a36442d03b7a71a50732038

Fix GLES3 Clang UBSan errors

Fix or avoid runtime errors detected by Clang's undefined behavior
sanitizer in the GLES3 module.

Affected tests:
dEQP-GLES3.functional.shaders.operator.unary_operator.minus.*
dEQP-GLES3.functional.fragment_out.basic.*
dEQP-GLES3.functional.texture.filtering.2d.formats.*

Components: OpenGL ES, Framework

VK-GL-CTS issue: 3018

Change-Id: Ic10ae433deb2f835ceb2fe3ab9664b05f72cd3d0

Merge branch 'vulkan-cts-1.2.5' into vulkan-cts-1.2.6

Change-Id: Ia705e141206a3751d4abc2d6eccd958d7e3f2f0e

Merge branch 'vulkan-cts-1.2.4' into vulkan-cts-1.2.5

Change-Id: I952af5b744206c52b4161d8c1102b6a8c2a53f58

Merge branch 'vulkan-cts-1.2.3' into vulkan-cts-1.2.4

Change-Id: Ica9ea69514cd36415ad89f4b34e8384b0523181c

Merge remote-tracking branch 'origin/vulkan-cts-1.2.2' into vulkan-cts-1.2.3

Change-Id: I3c5b289149d9f2601b5d5ab8986d41b3f33f8bab

Add a batch of GraphicsFuzz coverage tests

This commit adds a batch of GraphicsFuzz coverage tests.

Components: Vulkan

New Tests:

dEQP-VK.graphicsfuzz.cov-condition-bitfield-extract-integer
dEQP-VK.graphicsfuzz.cov-condition-matrix-determinant-uniform
dEQP-VK.graphicsfuzz.cov-descending-loop-index-temporary-array
dEQP-VK.graphicsfuzz.cov-do-while-switch-case-bitcount-findmsb
dEQP-VK.graphicsfuzz.cov-dummy-function-loop-array-element-increment-never-read
dEQP-VK.graphicsfuzz.cov-fragcood-multiple-conditions-function-loop-global-counter
dEQP-VK.graphicsfuzz.cov-fragcoord-multiple-conditions-function-global-loop-counter-simplified
dEQP-VK.graphicsfuzz.cov-function-nested-do-whiles-looped-once
dEQP-VK.graphicsfuzz.cov-int-initialize-from-multiple-large-arrays
dEQP-VK.graphicsfuzz.cov-loop-break-after-first-iteration-set-array-element
dEQP-VK.graphicsfuzz.cov-loop-condition-bitfield-extract-set-array-elements
dEQP-VK.graphicsfuzz.cov-loop-condition-logical-or-never-iterated
dEQP-VK.graphicsfuzz.cov-loop-fragcoord-identical-condition
dEQP-VK.graphicsfuzz.cov-loop-set-vector-components-pow-two
dEQP-VK.graphicsfuzz.cov-multiple-fragcoord-conditions-false-never-return-sample-texture
dEQP-VK.graphicsfuzz.cov-multiple-fragcoord-conditions-never-return-color-uninitialized
dEQP-VK.graphicsfuzz.cov-nested-loop-continue-inner-copy-array-element
dEQP-VK.graphicsfuzz.cov-nested-loop-not-greater-than-increment-array-element
dEQP-VK.graphicsfuzz.cov-nested-loops-float-bits-to-int-increment-array
dEQP-VK.graphicsfuzz.cov-nested-loops-global-counter-increment-single-element
dEQP-VK.graphicsfuzz.cov-nested-loops-sample-opposite-corners
dEQP-VK.graphicsfuzz.cov-two-nested-loops-switch-case-matrix-array-increment

Change-Id: Ia02cb23ec103be95184ad65da475d8b95a8d8504

Add test for finding unknown memory property flag sets

This test retrieves all the found propertyFlags from available memory
types, and then it makes a comparison between the found memory
propertyFlags and the acceptable propertyFlag sets. If an unknown
combination of memory propertyFlags is found, the test will fail.

Components: Vulkan

VK-GL-CTS Issue: 2704

New tests: dEQP-VK.memory.requirements.memory_property_flags.check_all

Change-Id: I9c456076bb7364572183ca4976f56775ffd5e286

Fix loop over cube map faces in GL4

Found when building with Clang with -Wunreachable-code-loop-increment
enabled. TextureViewUtilities::initTextureStorage is supposed to call
glTexImage2D for all cube map faces but stops after the first one
(GL_TEXTURE_CUBE_MAP_POSITIVE_X).

Affects:
KHR-GL4*.texture_view.gettexparameter
KHR-GL4*.texture_view.errors

Components: OpenGL

VK-GL-CTS issue: 3085

Change-Id: I3d85351ab50ad0bd8b779537b1a8b3bef162f80f

Fix fragment shading rate validation layer warning

We passing a number of layer > 1 for an image view type of
VK_IMAGE_VIEW_TYPE_2D.

Also if we're not using ATTACHMENT_MODE_2DARRAY, then there is only
one layer for the fragment rate shading attachment, so we should check
against layer 0.

Components: Vulkan

Affects:
dEQP-VK.fragment_shading_rate.srlayered.*

VK-GL-CTS issue: 3079

Change-Id: I351f880a9bdd934f773e517713839c67ad453766

Add run mode that verifies amber requirements

This change adds a new run mode that verifies that .amber file
requirements match the ones defined in CTS.

Running CTS with --deqp-runmode=amber-verify goes through all
amber test cases and outputs any mismatches in the log file.
If any mismatches are found, the run ends with internal_error.

This change also contains fixes for all of the mismatches
that exist in CTS today. A couple required update in Amber,
and thus this change also updates it.

Affects:
dEQP-VK.compute.workgroup_memory_explicit_layout.copy_memory.variable_pointers
dEQP-VK.spirv_assembly.instruction.spirv1p4.*
dEQP-VK.spirv_assembly.instruction.terminate_invocation.*

Components: Vulkan, Framework
VK-GL-CTS issue: 2967

Change-Id: Ic30b5d555d9cdf293a2ece8a690a4a2a943b1b80

Remove extra semicolons

Fix warnings for Clang builds with -Wextra-semi, -Wextra-semi-stmt and
-Wc++98-compat-extra-semi.

Components: Framework, OpenGL, OpenGL ES, Vulkan

VK-GL-CTS issue: 3080

Change-Id: I1caa81b65415a1e94797c6df48fc8a4539e9635a

Fix warning in gcc-8

The code for setting up the output buffer was duplicated into each of
the cases for different input types, and the compiler doesn't understand
that DE_ASSERT marks an unreachable path. Changing to using standard
assert fixes the warning:

vktShaderFConvertTests.cpp:541:19: error: ‘outputBufferSizeInfo::memorySizeBytes’
may be used uninitialized in this function [-Werror=maybe-uninitialized]
   541 |  BufferSizeInfo   outputBufferSizeInfo;
       |                   ^~~~~~~~~~~~~~~~~~~~

Component: Vulkan
Affects: dEQP-VK.glsl.builtin.precision_fconvert.*

Change-Id: Icf6302166d8bce7c3ae1447973c895c34465a726

Enable Vulkan matrix determinant and inverse precision tests

A set of new matrix3x3/matrix4x4 determinant and inverse precision
tests for Amber.

Components: Vulkan

VK-GL-CTS Issue: 2878

New tests:
dEQP-VK.glsl.builtin.precision.determinant.highp.mat3
dEQP-VK.glsl.builtin.precision.inverse.highp.mat3
dEQP-VK.glsl.builtin.precision_fp16_storage16b.determinant.compute.mat3
dEQP-VK.glsl.builtin.precision_fp16_storage16b.inverse.compute.mat3
dEQP-VK.glsl.builtin.precision_fp16_storage32b.determinant.compute.mat3
dEQP-VK.glsl.builtin.precision_fp16_storage32b.inverse.compute.mat3
dEQP-VK.glsl.builtin.precision_double.determinant.compute.mat3
dEQP-VK.glsl.builtin.precision_double.inverse.compute.mat3
dEQP-VK.glsl.builtin.precision.determinant.highp.mat4
dEQP-VK.glsl.builtin.precision.inverse.highp.mat4
dEQP-VK.glsl.builtin.precision_fp16_storage16b.determinant.compute.mat4
dEQP-VK.glsl.builtin.precision_fp16_storage16b.inverse.compute.mat4
dEQP-VK.glsl.builtin.precision_fp16_storage32b.determinant.compute.mat4
dEQP-VK.glsl.builtin.precision_fp16_storage32b.inverse.compute.mat4
dEQP-VK.glsl.builtin.precision_double.determinant.compute.mat4
dEQP-VK.glsl.builtin.precision_double.inverse.compute.mat4

Change-Id: I5d748c1ecbaf4e8444a2d09eac8da60bfa54271c

Correct Vulkan multiview 'tesellation' test names

Tests are renamed from 'tesellation' to 'tessellation' but are otherwise
unchanged.

Affects:

dEQP-VK.multiview*tessellation*

Components: Vulkan
Change-Id: I1686bb6fb8a0e85e201853602f840e03e847169d

Do not overflow when copy data

As `vertexDataSize` is aligned with `nonCoherentAtomSize`, it could be
bigger than the actual size of m_data.

This is actually causing a heap-buffer overflow when executing in our
Rpi4 Vulkan driver.

Affected tests:
dEQP-VK.spirv_assembly.instruction.graphics.cross_stage.basic_type.flat

Components: Vulkan

VK-GL-CTS issue: 3062

Change-Id: I6b1f449e9f0a42740adb051ed037ff911dac763a

Fix protected memory ssbo pipeline barrier use

Incorrect stage flags are passed to vkCmdPipelineBarrier in the copy
buffer operation in the dEQP-VK.protected_memory.ssbo.ssbo_read.*

These tests were introduced in origin/vulkan-cts-1.1.0

Components: Vulkan

Affects:
dEQP-VK.protected_memory.ssbo.*

VK-GL-CTS Issue: 3041

Change-Id: Iccdb96fee30e32bf26dec1e26d65ea8d5600c384

Don't alias image views in separate_channels subpass dependency tests

This is not allowed by the Vulkan spec:

> Multiple attachments that alias the same memory must not be used in a
> single subpass.

Instead we should have the input attachment and color attachment point
to the same attachment, which is allowed.

Components: Vulkan

Affects:
dEQP-VK.renderpass.suballocation.subpass_dependencies.separate_channels.*
dEQP-VK.renderpass2.suballocation.subpass_dependencies.separate_channels.*

VK-GL-CTS Issue: 3052

Change-Id: Iddb1111be5147f9816104fa879e023a9a3445555

Issue in serial storage address computation

When serial storage and src/dst buildTypes are different, serial storage address computation is not correct. In the beginning of the test case, serial storage buildType is either set to host or device according to test case being run (cpu-built or gpu-built).
In the 2nd phase of this test case (deserialize all from the previous step to a new top-level AS) VkCopyMemoryToAccelerationStructureInfoKHR is filled with srcAddress according to testCase buildType, and dst acceleration structure is device.
In this stage, if testCase builtType is device, the serialStorage address computation is correct. However, if testCase buildType is host then the issue arises. The address is computed according to host, although device address is needed.

That is, for serial storage address computation: buildType of the acceleration structure (src or dst) should be used rather than serial storage buildType which is initially set according to test case being run.

Affects:
dEQP-VK.ray_tracing_pipeline.acceleration_structures.header_bottom_address.*

Components: Vulkan
VK-GL-CTS issue: 3028

Change-Id: If8c0387862ce6f1db00fc61016dd0e602ff72ba8

Add coverage for sampling a cubemap that has been rendered to

This test checks that a cubemap will be sampled correctly if its
contents have been modified by a draw call.

VK-GL-CTS issue: 2506

New Tests:
dEQP-VK.image.sample_cubemap.write_face_0

Components: Vulkan

Change-Id: I1ee62842d6c952b5628822fe49ee6a91c26ead04

Merge vk-gl-cts/vulkan-cts-1.2.7 into vk-gl-cts/master

Change-Id: Ieb90dde42cb20ab43cebd0f2b5d3ad4091c3b57a

Merge vk-gl-cts/vulkan-cts-1.2.6 into vk-gl-cts/vulkan-cts-1.2.7

Change-Id: Ieaff21c7c6228a4507325afe13ec2a3807b345cb

Add missing vertexPipelineStoresAndAtomics checks in ray query tests

Affected tests:
dEQP-VK.ray_query.acceleration_structures.*.vertex_shader.*
dEQP-VK.ray_query.acceleration_structures.*.tess_control_shader.*
dEQP-VK.ray_query.acceleration_structures.*.tess_evaluation_shader.*
dEQP-VK.ray_query.acceleration_structures.*.geometry_shader.*
dEQP-VK.ray_query.ray_flags.vertex_shader.*
dEQP-VK.ray_query.ray_flags.tess_control_shader.*
dEQP-VK.ray_query.ray_flags.tess_evaluation_shader.*
dEQP-VK.ray_query.ray_flags.geometry_shader.*
dEQP-VK.ray_query.traversal_control.vertex_shader.*
dEQP-VK.ray_query.traversal_control.tess_control_shader.*
dEQP-VK.ray_query.traversal_control.tess_evaluation_shader.*
dEQP-VK.ray_query.traversal_control.geometry_shader.*

Components: Vulkan

VK-GL-CTS Issue: 3069

Change-Id: I49c37d292380fda47b113bf4fa14a6395f8e81a5

Check support in synchronization operation tests

Synchronization tests were creating pipelines before checking for support,
causing unnecessary use of CPU time when encountering an operation
later on that was not supported.

Affects:

dEQP-VK.synchronization.op.*

Components: Vulkan

VK-GL-CTS issue: 2731

Change-Id: Ia83afde4767ced1549a1b0573cdc9e8af8526870

Read stencil aspect from image in DEPTH_ATTACHMENT_STENCIL_READ_ONLY_OPTIMAL  layout

The test was using sampler to read depth aspect from image in
combined depth+stencil format and in
VK_IMAGE_LAYOUT_DEPTH_ATTACHMENT_STENCIL_READ_ONLY_OPTIMAL
layout - what is invalid.

Instead, the test now reads (and tests) stencil aspect of this
image. As the test code already supported testing stencil formats,
just a few conditions are changed/added to follow the stencil path
if VK_IMAGE_LAYOUT_DEPTH_ATTACHMENT_STENCIL_READ_ONLY_OPTIMAL
is used.

VK-GL-CTS Issue: 3048

Components: Vulkan

Affects: dEQP-VK.synchronization2.none_stage.*

Change-Id: I918a5cdad83bcd25b2f8b6cee642b56218c2b7fe

Optimize image clearing tests

This CL optimizes image clearing tests.
The tests did a lot redundant calculation for each pixel in the verify stage.
This CL moves those calculations to outside of the pixel verify function.

Affects:

dEQP-VK.api.image_clearing.*

Components: Vulkan

VK-GL-CTS issue: 2933

Change-Id: Ie3403c591c9f084115a13cc78e0a66c54c48817c

Ensure that destroying old swapchain has no effect

There should be no crash when recreating Vulkan swapchain with
old swapchain and later calling vkDestroySwapchainKHR on old swapchain.

New tests:

dEQP-VK.wsi.*.swapchain.destroy.old_swapchain

Components: Vulkan

VK-GL-CTS issue: 2935

Change-Id: I7f03e55767ececec8a6a70ef553f0676d3a3c616

Conservative rasterization with SampleMask

Additional test coverage for conservative rasterization Conservative
rasterization used in conjunction with various multisampling setup
including FullyCoveredEXT built-in, minSampleShading, postDepthCoverage
and SampleMask.

Components: Vulkan

VK-GL-CTS issue: 2868

Affects:
dEQP-VK.pipeline.multisample.conservative_with_full_coverage.*

Change-Id: I8c2858b81cfdf27e622acf74e273db9014499c9c

Add GL_RGB9_E5 as renderable ext format in fbo tests

-If GL_QCOM_render_shared_exponent is supported then treat
GL_RGB9_E5 as a renderable format for fbo completeness tests

Components: OpenGL
VK-GL-CTS Issue: 3066

Affects:
dEQP-GLES3.functional.fbo.completeness.renderable.renderbuffer.color0.rgb9_e5
dEQP-GLES3.functional.fbo.completeness.renderable.texture.color0.rgb9_e5

Change-Id: I74566a5b56c7528c89ac97423c5b6f9207397c70

Update Vulkan headers

Components: Vulkan

Change-Id: Iefdf1bc052c5f9c63bdf9062c01c96a4c11a92c6

Implement getProcAddress for surfaceless platform

Currently this virtual function is not overridden, causing a crash
whenever a test calls getProcAddress in a surfaceless context.

Affects:
KHR-GL3{0,1,2}.transform_feedback.draw_xfb_instanced_test
KHR-GL3{0,1,2}.transform_feedback.draw_xfb_stream_instanced_test

Components: Framework

VK-GL-CTS issue: 3021

Change-Id: Idb29b5ca9d70c5f3be6a32fdf2c827d5d7591050

Fix for the various multisample tests with GL 4.5

Affects:
dEQP-GLES31.functional.shaders.sample_variables.*
dEQP-GLES31.functional.sample_shading.*
dEQP-GLES31.functional.shaders.multisample_interpolation.*
dEQP-GL45.functional.shaders.sample_variables.*
dEQP-GL45.functional.sample_shading.*
dEQP-GL45.functional.shaders.multisample_interpolation.*

Components: AOSP
VK-GL-CTS issue: 3014

Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Change-Id: I6e4297a0558a31a1dc2258da7434d3fcc1900b73