Mika Isojrvi [Thu, 14 Sep 2017 02:18:56 +0000 (02:18 +0000)]
Fix atomic ssbo xor test am:
f0fa05e898 am:
14cd264501 am:
6e80057a20 am:
cae1fe52ac am:
754e8f0192 am:
d422ad8111 am:
a8804ca914
am:
802854925b
Change-Id: I1be013ce9a238f81e973ecff1eac9ad45dfaaa44
Mika Isojrvi [Thu, 14 Sep 2017 02:18:38 +0000 (02:18 +0000)]
Fix atomic ssbo xor test am:
f0fa05e898 am:
14cd264501 am:
d4c01761d3 am:
d20475d813 am:
66bd71c817 am:
c06d76b076
am:
5e5d57c9d9
Change-Id: Ia837c1a5524aa4935fefcf015f3f832aba1e739e
Mika Isojrvi [Thu, 14 Sep 2017 02:18:18 +0000 (02:18 +0000)]
Mask compared bits in atomic operation tests. am:
b2ed9fbe1a am:
4d1d6446da am:
b78f9b0e57 am:
c2180f74a8 am:
76b48a0db2 am:
ff2205c90c
am:
bde0d133a4
Change-Id: Id371d2842d97e3f5d3d52d2b107e8e3d74b80719
Mika Isojrvi [Thu, 14 Sep 2017 02:18:01 +0000 (02:18 +0000)]
Mask compared bits in atomic operation tests. am:
b2ed9fbe1a am:
4d1d6446da am:
6c0e0ebde7 am:
371f033807 am:
e64be738f6 am:
a242f668a7 am:
7ed7199caf
am:
a921fc5252
Change-Id: I0825b4f59d2fbc8485ff24f6eec748757e7f8f09
Mika Isojrvi [Thu, 14 Sep 2017 02:17:29 +0000 (02:17 +0000)]
Mask compared bits in atomic operation tests. am:
b2ed9fbe1a am:
e37b18673d am:
97ce4c6f8c am:
4c56c58687 am:
0ddc9faea1 am:
d2d7d49296 am:
301312d147 am:
4dd9680c5d am:
bf288e2971 am:
1678011b33
am:
14f0575d74
Change-Id: I0e0ae598f390c38b1d57cb83d161ab6921020168
Courtney Goeltzenleuchter [Tue, 12 Sep 2017 23:11:21 +0000 (17:11 -0600)]
Catch BAD_MATCH from createWindowSurface
On some platforms, it's possible to fail createWindowSurface
for some attribute combinations. Catch that case and return
NotSupported rather than a test failure.
Test: adb -d shell am start \
-n com.drawelements.deqp/android.app.NativeActivity \
-e cmdLine '"deqp --deqp-case=dEQP-EGL.functional.wide_color.* \
--deqp-log-filename=/sdcard/dEQP-Log.qpa"'
Bug:
65376495
Change-Id: I6546cbdfc011c7d26cfa3cdad7cc6f0df1b67122
Courtney Goeltzenleuchter [Tue, 12 Sep 2017 23:10:51 +0000 (17:10 -0600)]
Require EGL_KHR_gl_colorspace for wide_color tests
These tests utilize enum EGL_GL_COLORSPACE_KHR which is introduced by
EGL_KHR_gl_colorspace extension, check first for support and skip the
test if extension is not supported.
v2: loosen the check, only check when enum in use (Courtney)
v3: use EGL_NONE instead of DE_NULL when indicating that test
should use the default colorspace. That is do not include
the EGL_GL_COLORSPACE_KHR attribute in the appropriate
create function.
Test: adb -d shell am start \
-n com.drawelements.deqp/android.app.NativeActivity \
-e cmdLine '"deqp --deqp-case=dEQP-EGL.functional.wide_color.* \
--deqp-log-filename=/sdcard/dEQP-Log.qpa"'
Bug:
65291400
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
(cherry picked from commit
fbf132dcf9467388408627bee73d90f7e8a1d401)
Change-Id: I6bd6423e6eb5fe4bb4df0222ab624c41446b0b60
Courtney Goeltzenleuchter [Mon, 11 Sep 2017 19:06:48 +0000 (13:06 -0600)]
Fix handling of EGL_NONE as color space
Color space of EGL_NONE was getting passed down and
driver threw EGL_BAD_ATTRIBUTE in response.
Issue was caused when changing the colorspace from nullptr
to EGL_NONE, missed a couple of places that needed to
test against EGL_NONE instead of nullptr.
Bug:
65376495
Test: adb -d shell am start \
-n com.drawelements.deqp/android.app.NativeActivity \
-e cmdLine '"deqp --deqp-case=dEQP-EGL.functional.wide_color.* \
--deqp-log-filename=/sdcard/dEQP-Log.qpa"'
Change-Id: If8812725be9d3858bf610c647ba53ff38944dd50
(cherry picked from commit
0d4e616c35fd3cbad82eb3c1f0a604dbbc6c8662)
Treehugger Robot [Mon, 11 Sep 2017 19:30:30 +0000 (19:30 +0000)]
Merge changes Id437e220,I9d6b3fba
* changes:
Allow ES 3.2 contexts to run sRGB write control unsupported test.
Require GL_EXT_sRGB_write_control for GL_FRAMEBUFFER_SRGB enable/disable
Treehugger Robot [Mon, 11 Sep 2017 18:58:52 +0000 (18:58 +0000)]
Merge "Use 64-bit math in work group size assertion."
Treehugger Robot [Mon, 11 Sep 2017 18:57:22 +0000 (18:57 +0000)]
Merge "Delete incorrect invalid map tests."
Kenneth Graunke [Sat, 9 Sep 2017 06:30:15 +0000 (23:30 -0700)]
Allow ES 3.2 contexts to run sRGB write control unsupported test.
GL_EXT_sRGB_write_control functionality is not part of ES 3.2.
So there's no need to restrict this to older contexts.
Affects:
dEQP-GLES31.functional.fbo.srgb_write_control.framebuffer_srgb_unsupported_enum
Components: AOSP
Change-Id: Id437e220958c66acf65fd5c40004e1fde13fdf64
Kenneth Graunke [Sat, 9 Sep 2017 06:26:15 +0000 (23:26 -0700)]
Require GL_EXT_sRGB_write_control for GL_FRAMEBUFFER_SRGB enable/disable
This test required GL_EXT_sRGB_write_control or ES 3.2. But glEnable
and glDisable of GL_FRAMEBUFFER_SRGB requires the extension - ES 3.2
doesn't provide that functionality. ES 3.2 implementations without the
extension would fail. Now they report NotSupported.
Affects:
dEQP-GLES31.functional.fbo.srgb_write_control.framebuffer_srgb_enabled
Components: AOSP
Change-Id: I9d6b3fba892288354620ecbbc51ef7a351507e43
Kenneth Graunke [Fri, 8 Sep 2017 22:37:38 +0000 (15:37 -0700)]
Delete incorrect invalid map tests.
The citation given in these tests comes from the OpenGL man pages, which
are not authoritative. The ES 3.2 specification says:
"6.3.2 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:
- dEQP-GLES31.functional.debug.negative_coverage.*.vertex_array.*_invalid_map
Components: AOSP
Change-Id: I26eb6ac0dfffbfc50612bed44d1beb6eb3d366c9
Treehugger Robot [Fri, 8 Sep 2017 21:25:48 +0000 (21:25 +0000)]
Merge "Relax EGL functional test thresholds"
Kenneth Graunke [Thu, 13 Jul 2017 01:44:11 +0000 (18:44 -0700)]
Use 64-bit math in work group size assertion.
If the maximum workgroup size in each dimension is large enough, this
can overflow, causing the comparison to fail. For example, on Intel
Haswell, we get:
assert(2048 * 2048 * 2048 > 2048) -> assert(0 > 2048) -> fail
Casting one value to deInt64 causes the multiplications to occur as
64-bit values, so we don't overflow.
Affects:
dEQP-GLES31.functional.debug.negative_coverage.*.compute.invalid_maximum_work_group_sizes
Components: AOSP
Change-Id: Ie03627a378f4fdc46ce241bbdd3693ae829c217d
james.glanville [Tue, 29 Aug 2017 10:40:53 +0000 (11:40 +0100)]
Relax EGL functional test thresholds
The current threshold of 0.02 was selected as suitable for 8-bit color depths.
However, for smaller bit depths such as RGB565 this threshold is too strict
and it has been relaxed to allow a larger error for low bit depth formats.
Affects:
dEQP-EGL.functional.render.*
Change-Id: I6d6e061be02bc8e63cb5d9662feb804bc839d567
Courtney Goeltzenleuchter [Tue, 5 Sep 2017 23:48:27 +0000 (17:48 -0600)]
Check for EGL_KHR_gl_colorspace before using it
Bug:
65376495
Test: adb -d shell am start \
-n com.drawelements.deqp/android.app.NativeActivity \
-e cmdLine '"deqp --deqp-case=dEQP-EGL.functional.wide_color.* \
--deqp-log-filename=/sdcard/dEQP-Log.qpa"'
Change-Id: Ief654c998e93d72899225120b0ca1ab2fa3ca624
Treehugger Robot [Tue, 5 Sep 2017 21:35:29 +0000 (21:35 +0000)]
Merge "Add more allowed 'dot' alternatives"
Mika Isojärvi [Thu, 31 Aug 2017 19:13:05 +0000 (12:13 -0700)]
Fix unused uniform bug in EGL robustness tests
Bug:
64845766
Bug:
65175554
Test: Ran tests on couple of Google devices
Change-Id: Iee2918c1e45ad305dd9ca30f1b81998bdac06127
Treehugger Robot [Tue, 5 Sep 2017 17:18:37 +0000 (17:18 +0000)]
Merge "Fix missing SSBO binding in EGL robustness tests"
Pyry Haulos [Tue, 5 Sep 2017 15:22:27 +0000 (15:22 +0000)]
Fix issues in pipeline.timestamp.transfer_tests am:
0f672f2a20
am:
9e85a126d2
Change-Id: I86a852a2216971095d4aa256a97f0b8e68e73646
Pyry Haulos [Tue, 5 Sep 2017 15:14:23 +0000 (15:14 +0000)]
Fix issues in pipeline.timestamp.transfer_tests
am:
0f672f2a20
Change-Id: Ib7e13db1c7801095b925890f226a38969a1b418e
Alexander Galazin [Tue, 22 Aug 2017 13:59:49 +0000 (15:59 +0200)]
Add more allowed 'dot' alternatives
Expanded dot may be re-associated in multiple ways.
Add all possible alternatives to prevent precision failures.
Components: AOSP
VK-GL-CTS issue: 634
Google issue:
28767510
Affects:
dEQP-GLES3.functional.shaders.builtin_functions.precision.dot.*
dEQP-GLES3.functional.shaders.builtin_functions.precision.reflect.*
dEQP-GLES31.functional.shaders.builtin_functions.precision.dot.*
dEQP-GLES31.functional.shaders.builtin_functions.precision.reflect.*
Change-Id: I2a91add1b92363acab2172179320cd880da779ec
Mika Isojärvi [Thu, 31 Aug 2017 18:54:48 +0000 (11:54 -0700)]
Fix missing SSBO binding in EGL robustness tests
Bug:
64845581
Bug:
65175642
Test: Ran the tests on couple of Google devices
Change-Id: I3a98fe12802e6951483499e800cd3fe7412e37f0
Pyry Haulos [Thu, 9 Mar 2017 17:46:29 +0000 (09:46 -0800)]
Fix issues in pipeline.timestamp.transfer_tests
This change fixes two issues:
* Buffer overrun in vkCmdCopyBufferToImage and vkCmdCopyImageToBuffer
tests.
* Validation error (srcStageMask must not be 0).
VK-GL-CTS issue: 185
Components: Vulkan
Change-Id: I9679566f08c4a914eb6066bab248abd6d51e0f7e
Tapani Pälli [Mon, 28 Aug 2017 06:20:53 +0000 (09:20 +0300)]
Require EGL_KHR_gl_colorspace for wide_color tests.
These tests utilize enum EGL_GL_COLORSPACE_KHR which is introduced by
EGL_KHR_gl_colorspace extension, check first for support and skip the
test if extension is not supported.
v2: loosen the check, only check when enum in use (Courtney)
Change-Id: I891a6fdcce49ac6dcd42577076d4cf64a143dd38
Signed-off-by: Tapani Pälli <tapani.palli@intel.com>
Treehugger Robot [Thu, 24 Aug 2017 22:57:17 +0000 (22:57 +0000)]
Merge "Check only what's needed and use threshold"
Mika Isojärvi [Thu, 24 Aug 2017 20:13:19 +0000 (13:13 -0700)]
Merge Vulkan CTS 1.0.2.5 into aosp/master
Bug:
64932237
Change-Id: Ic2758a85f3a8f8cf00985b03a89830ba509f824d
Courtney Goeltzenleuchter [Fri, 28 Jul 2017 19:51:31 +0000 (13:51 -0600)]
Check only what's needed and use threshold
Check for only the functionality needed for a specific test.
E.g. if testing floating point color, don't need to check
for support of 10:10:10:2 pixel format.
Some devices are not able to hit the exact values calculated.
Not unexpected, update tests to use thresholds.
Note: waiting on results of b/
64487716. Seeing failure on
Pixel devices. Need to make sure it's not a test issue.
Test: adb -s
0424415804161004f80d -d shell am start \
-n com.drawelements.deqp/android.app.NativeActivity \
-e cmdLine '"deqp --deqp-case=dEQP-EGL.functional.wide_color.* \
--deqp-log-filename=/sdcard/dEQP-Log.qpa"'
Change-Id: I2adb31eaa3b189179aeceb9e8f6b1ac8587b6027
(cherry picked from commit
041a1b5e82db7a6945aaefd791639de1c8c9e84d)
Treehugger Robot [Tue, 22 Aug 2017 18:13:36 +0000 (18:13 +0000)]
Merge "Avoid thread state races in EGL multithread tests"
Treehugger Robot [Tue, 22 Aug 2017 18:10:07 +0000 (18:10 +0000)]
Merge "Run Draw*BaseVertex if EXT_draw_elements_base_vertex is present"
Alexander Galazin [Wed, 16 Aug 2017 09:36:01 +0000 (11:36 +0200)]
Run Draw*BaseVertex if EXT_draw_elements_base_vertex is present
Draw*BaseVertex methods are enabled if
the EXT_draw_elements_base_vertex extension is present
Follow-up on
1b1ea05b
Components: AOSP
Affects: dEQP-GLES31.functional.draw_base_vertex.*
Change-Id: If8a4860a956c771ec5ea3341e2227f437b12a286
Mika Isojrvi [Thu, 17 Aug 2017 00:08:48 +0000 (00:08 +0000)]
Fix texture mip level size in framebuffer fetch tests
am:
be0d165a5d
Change-Id: I7fb9c075ee79e411ea295999e9edd7f38a5e5bc9
Mika Isojärvi [Wed, 16 Aug 2017 20:40:22 +0000 (13:40 -0700)]
Fix texture mip level size in framebuffer fetch tests
Bug:
64280744
Test: Ran tests with multiple Google devices
Change-Id: I369a3c9b16270655332d5a608207f0da2633e0fe
james.glanville [Wed, 16 Aug 2017 19:04:43 +0000 (19:04 +0000)]
Merge "Use safe glFramebufferTexture* function"
am:
ec5ccbe144
Change-Id: I156ecc2a6ed17e3e23ef17ebbe026be27c1670c3
Michael Chock [Wed, 16 Aug 2017 19:04:22 +0000 (19:04 +0000)]
Merge "Do not use degenerate frames in EGL tests"
am:
f5971fe17f
Change-Id: I9807e27047a0d0280e06bc826cab3e6cb88b5ab6
Treehugger Robot [Wed, 16 Aug 2017 18:59:45 +0000 (18:59 +0000)]
Merge "Use safe glFramebufferTexture* function"
Treehugger Robot [Wed, 16 Aug 2017 18:59:42 +0000 (18:59 +0000)]
Merge "Do not use degenerate frames in EGL tests"
Alexander Galazin [Wed, 16 Aug 2017 18:47:50 +0000 (18:47 +0000)]
Merge "Report tests using Draw*BaseVertex as NotSupported"
am:
f96636fdfa
Change-Id: Iccb20250f9b98f983fc831911a1f9b4d4a98fbfb
Treehugger Robot [Wed, 16 Aug 2017 18:36:54 +0000 (18:36 +0000)]
Merge "Report tests using Draw*BaseVertex as NotSupported"
Alexander Galazin [Wed, 16 Aug 2017 18:20:18 +0000 (18:20 +0000)]
Merge "Fix error double accounting in fuzzyCompare()"
am:
0cf17c4bf8
Change-Id: Ife3e4f1bd13c6f4cb939950635880ce32d2b1df3
Treehugger Robot [Wed, 16 Aug 2017 18:06:52 +0000 (18:06 +0000)]
Merge "Fix error double accounting in fuzzyCompare()"
Michael Chock [Mon, 14 Aug 2017 15:36:11 +0000 (08:36 -0700)]
Avoid thread state races in EGL multithread tests
Previously, threads objects inside dEQP-EGL.functional.multithread.*
tests would conflate thread execution status with test results. The
former would only be set inside the thread, but the latter could be set
by other threads (e.g., a test thread might set its status to RUNNING
while another test might attempt to set the status of NOT_SUPPORTED).
These race conditions could lead to incorrect results.
Fix this by moving error and not-supported states out of the thread
execution state, and instead storing them as single-purpose flags inside
the test case object, avoiding the possibility of conflicting writes.
Change-Id: I2b117aa98c1c2b69b0d134610d6fd37656083e54
Michael Chock [Mon, 14 Aug 2017 18:55:14 +0000 (11:55 -0700)]
Do not use degenerate frames in EGL tests
The dEQP-EGL.functional.buffer_age.* and
dEQP-EGL.functional.partial_update.* tests use randomly-generated frames
for rendering. The test allows the generation of frames with a width or
height of 0, but the tcu::PixelBufferAccess interface does not permit
0-sized subregions (see assertions inside getSubregion overloads in
tcuTextureUtil.cpp).
Skip the rendering of degenerate frames in the reference renderer, as
they are unsupported but would not affect the rendering if they were.
Change-Id: I3977c3c293715db633dde74132e56b5d9303ab28
james.glanville [Wed, 16 Aug 2017 10:50:19 +0000 (11:50 +0100)]
Use safe glFramebufferTexture* function
glFramebufferTexture() is only supported when GL_EXT_geometry_shader is
present. Use glFramebufferTextureLayer() instead so that testing is
possible on unextended ES3.1 contexts.
Components: AOSP
Affects:
dEQP-GLES31.functional.shaders.framebuffer_fetch.basic.framebuffer_texture_level
Alexander Galazin [Wed, 16 Aug 2017 09:36:01 +0000 (11:36 +0200)]
Report tests using Draw*BaseVertex as NotSupported
Draw*BaseVertex methods were added in GLES 3.2.
Report tests using these methods as NotSupported if
GLES 3.1 context is returned.
Components: AOSP
Affects: dEQP-GLES31.functional.draw_base_vertex.*
Change-Id: I87c918d5f66a0a5b547f8e625c662203a7f85048
Alexander Galazin [Wed, 16 Aug 2017 00:32:11 +0000 (00:32 +0000)]
Check for tessellation/geometry extensions being exposed
am:
573154bb7b
Change-Id: I86d05362cf94c693d4000969d2c177c692f07686
Alexander Galazin [Fri, 11 Aug 2017 09:03:50 +0000 (11:03 +0200)]
Check for tessellation/geometry extensions being exposed
Some sample_variables tests use tessellation/geometry shaders.
Need to check that the relevant extensions are exposed.
Components: AOSP
Affects:
dEQP-GLES31.functional.debug.negative_coverage.*.oes_sample_variables.access_built_in_types_inside_other_shaders
Change-Id: Iff74939d474082938955e5b7ebfc41ea3b4a6ec5
Chih-Hung Hsieh [Thu, 10 Aug 2017 16:22:23 +0000 (16:22 +0000)]
Suppress non-critical warnings in external/deqp.
am:
b7ae0d8484
Change-Id: I203833d39406f0530796d10e6d7a4d8f9beb2dbb
Alexander Galazin [Tue, 8 Aug 2017 11:15:40 +0000 (13:15 +0200)]
Fix error double accounting in fuzzyCompare()
fuzzyCompare() has a loop to go through all the pixels and generate error sum.
It skips some pixels using random number generator between 0-8.
Random number generator sometime generate 0 and it ends up
calculating error sum twice for the same pixel.
If this pixel has error, then this add up error sum twice.
Components: Framework
Change-Id: I08e06e29faedb99cfd8fa479e97a437c06008848
Chih-Hung Hsieh [Wed, 9 Aug 2017 23:23:07 +0000 (16:23 -0700)]
Suppress non-critical warnings in external/deqp.
* We should enable those warnings after fixing the upstream source code.
Bug:
64487164
Test: build with WITH_TIDY=1
Change-Id: Ie4961d23b8f57e8d3a3ada0e64df82ed18dd366f
Mika Isojarvi [Fri, 4 Aug 2017 23:10:09 +0000 (23:10 +0000)]
Merge "Release EGL resources in CTS worker threads"
am:
fde96926aa
Change-Id: I8502a522e44c310dfb96710c7b301ebe798668ec
Mika Isojarvi [Fri, 4 Aug 2017 22:55:11 +0000 (22:55 +0000)]
Merge "Release EGL resources in CTS worker threads"
Graeme Leese [Fri, 4 Aug 2017 22:21:06 +0000 (22:21 +0000)]
Merge "Fix extension support checks in negative api tests" into nougat-cts-dev am:
4f1051ad32 am:
5bd0fb31a1
am:
ea24f81118
Change-Id: I02de92c1f3a63550bc24575d73e67dce8c6d6923
Graeme Leese [Fri, 4 Aug 2017 22:09:03 +0000 (22:09 +0000)]
Merge "Fix extension support checks in negative api tests" into nougat-cts-dev am:
4f1051ad32
am:
5bd0fb31a1
Change-Id: I2ab773e341fe8a0819963ca9351894ee740591cf
Graeme Leese [Fri, 4 Aug 2017 22:03:02 +0000 (22:03 +0000)]
Merge "Fix extension support checks in negative api tests" into nougat-cts-dev
am:
4f1051ad32
Change-Id: I4f54481292ffaaf5e1cf56a05ef24f7dfa4983f3
Treehugger Robot [Fri, 4 Aug 2017 21:56:36 +0000 (21:56 +0000)]
Merge "Fix extension support checks in negative api tests" into nougat-cts-dev
Michael Chock [Fri, 4 Aug 2017 19:38:53 +0000 (19:38 +0000)]
Check for EGLimage support in EGLimage test
am:
1886241d23
Change-Id: Ia4d21d273030b87dd9aa6b5e3eae5323bfe41dab
Michael Chock [Wed, 26 Jul 2017 17:07:16 +0000 (10:07 -0700)]
Release EGL resources in CTS worker threads
For worker threads in multithreaded EGL tests, call eglReleaseThread
before thread termination.
Affects:
dEQP-EGL.functional.color_clears.multi_thread.*
dEQP-EGL.functional.multithread.*
dEQP-EGL.functional.render.multi_thread.*
dEQP-EGL.functional.thread_cleanup.*
dEQP-EGL.performance.gles2_shared_render.*
Change-Id: I1e47fae1b9a3d1f148a80215c1c2baa0b684c57e
Michael Chock [Tue, 25 Jul 2017 20:16:38 +0000 (13:16 -0700)]
Check for EGLimage support in EGLimage test
Verify support for EGL 1.5+, EGL_KHR_image, or EGL_KHR_image_base before
attempting to call eglCreateImage[KHR] in test
dEQP-EGL.functional.image.api.invalid_create_image.
Change-Id: Icde0f5c568a00da7b7aaa07bc2fb58b7d62df165
Daniel Koch [Wed, 26 Jul 2017 17:38:21 +0000 (13:38 -0400)]
Allow VK_KHR_relaxed_block_layout extension
VK-GL-CTS Issue: 596
Affects:
dEQP-VK.api.info.device.extensions
Change-Id: I25395cbc76e2a4a176c74e9d46f16f3e105c4669
Component: Vulkan
Arkadiusz Sarwa [Wed, 15 Mar 2017 15:02:07 +0000 (16:02 +0100)]
Add VK_KHR_relaxed_block_layout
New tests:
dEQP-VK.ssbo.layout.multi_basic_types.*.relaxed_block*
dEQP-VK.ssbo.layout.random.relaxed.*
VK-GL-CTS issue: 117
Components: Vulkan
Change-Id: Ieda2aa291dcbff250b81e0da06c39781800f4af7
(cherry picked from commit
2a11db7c2e23afd4bef2f9a0eab84356d912ff3c)
Arkadiusz Sarwa [Tue, 9 May 2017 14:45:32 +0000 (16:45 +0200)]
Add GlslBuildOptions::FLAG_ALLOW_RELAXED_OFFSETS
Enabling the flag allows GLSL to use layout(offset = N) qualifiers that
follow VK_KHR_relaxed_block_layout rules instead of GLSL std140/std430
rules.
Components: Vulkan
Change-Id: I4f168b04ad0916b740891b603ddde2de0f4bcf50
(cherry picked from commit
4b82e12d3a7bf0947887e0b9d586d0ab63bd4888)
Tony Zlatinski [Wed, 19 Jul 2017 21:21:53 +0000 (16:21 -0500)]
Select correct memory heap index for memory import
When importing memory, the application (CTS) should follow
the spec rules about memory imports. I.e. The application must use the
same memory heap index for the import that has been used for the
exported memory. In addition, the application can't
import a buffer or image to a heap that does not support CPU mapping and
expect the memory to be CPU mappable/accessible.
This change modifies the export memory routines, now, returning the
index of the heap memory used for the export. Then the import routines are
modified to use those indexes for the memory import operations.
Components: Vulkan
Affects:
dEQP-VK.api.external.memory.*
dEQP-VK.synchronization.cross_instance.*
VK-GL-CTS issue: 585
Change-Id: I48ab85b7fd2b67655163f055fb03e8f2a3a42fd6
Alexander Galazin [Fri, 21 Jul 2017 17:51:48 +0000 (19:51 +0200)]
Merge vk-gl-cts/vulkan-cts-1.0.1 into vk-gl-cts/vulkan-cts-1.0.2
Change-Id: Ifac39caba5598cbb75d5fb25d271ad0b709140a6
Alexander Galazin [Fri, 21 Jul 2017 17:44:17 +0000 (19:44 +0200)]
Merge vk-gl-cts/vulkan-cts-1.0.0 into vk-gl-cts/vulkan-cts-1.0.1
Change-Id: If1266ef43a81d4d015c3f656315ea629acbbd55e
Alexander Galazin [Fri, 21 Jul 2017 16:55:45 +0000 (18:55 +0200)]
Merge vk-gl-cts/vulkan-cts-1.0.1 into vk-gl-cts/vulkan-cts-1.0.2
Change-Id: I65f618365df474653f68d311647f601f91a38e4b
Igor Ostrowski [Tue, 4 Jul 2017 16:01:23 +0000 (18:01 +0200)]
Fix layout transition in blitting tests
Affects:
dEQP-VK.api.copy_and_blit.*.blit_image.*
Components: Vulkan
VK-GL-CTS issue: 571
Change-Id: I780374a8fbff336a7789bd6a745e635ac1202b62
Igor Ostrowski [Wed, 12 Jul 2017 13:15:14 +0000 (15:15 +0200)]
Add metadata bindings in sparse tests
All metadata must be bound prior to device use of the sparse image
Affects:
dEQP-VK.sparse_resources.image_sparse_residency.*
dEQP-VK.sparse_resources.mipmap_sparse_residency.*
dEQP-VK.sparse_resources.shader_intrinsics.*
dEQP-VK.glsl.texture_functions.*sparse*
dEQP-VK.glsl.texture_gather.*sparse*
Components: Vulkan
VK-GL-CTS issue: 572
Change-Id: I4bbf2b74ed7fc0cea000084769cbed0dbe4a802e
Kalle Raita [Thu, 20 Jul 2017 19:42:12 +0000 (19:42 +0000)]
Merge "Add triangle edge guardband for Tex2D lookup diff"
am:
da231f19ab
Change-Id: I5162ef3b9e6d763e55b7e6a9049ae80523cafdb3
Kalle Raita [Thu, 20 Jul 2017 19:30:15 +0000 (19:30 +0000)]
Merge "Add triangle edge guardband for Tex2D lookup diff"
Mika Isojrvi [Wed, 19 Jul 2017 18:19:03 +0000 (18:19 +0000)]
Fix atomic ssbo xor test am:
f0fa05e898 am:
14cd264501 am:
6e80057a20 am:
cae1fe52ac am:
754e8f0192 am:
d422ad8111
am:
a8804ca914
Change-Id: I6518e6abcc83b89b14d187e2d16fec2754337298
Mika Isojrvi [Wed, 19 Jul 2017 18:16:00 +0000 (18:16 +0000)]
Fix atomic ssbo xor test am:
f0fa05e898 am:
14cd264501 am:
6e80057a20 am:
cae1fe52ac am:
754e8f0192
am:
d422ad8111
Change-Id: I6046d017e6528f4e80282ba82c58344c9d51085b
Mika Isojrvi [Wed, 19 Jul 2017 18:13:49 +0000 (18:13 +0000)]
Fix atomic ssbo xor test am:
f0fa05e898 am:
14cd264501 am:
d4c01761d3
am:
674ba44413
Change-Id: I182f8ee0b409eaa2000f3d18aa9c7d10b076b578
Mika Isojrvi [Wed, 19 Jul 2017 18:13:34 +0000 (18:13 +0000)]
Fix atomic ssbo xor test am:
f0fa05e898 am:
14cd264501 am:
d4c01761d3 am:
d20475d813 am:
66bd71c817
am:
c06d76b076
Change-Id: I8c4a4123d2258dd3bd577ce309204141a3a20d7e
Mika Isojrvi [Wed, 19 Jul 2017 18:13:30 +0000 (18:13 +0000)]
Fix atomic ssbo xor test am:
f0fa05e898 am:
14cd264501 am:
6e80057a20 am:
cae1fe52ac
am:
754e8f0192
Change-Id: I8dfffa1cad4108bf3fd6f9cc8136cf5215d61ad6
Mika Isojrvi [Wed, 19 Jul 2017 18:10:30 +0000 (18:10 +0000)]
Fix atomic ssbo xor test am:
f0fa05e898 am:
14cd264501 am:
6e80057a20
am:
cae1fe52ac
Change-Id: Iad44b21021d6f24b0d6c126cf81cf24632a0aaab
Mika Isojrvi [Wed, 19 Jul 2017 18:10:29 +0000 (18:10 +0000)]
Fix atomic ssbo xor test am:
f0fa05e898 am:
14cd264501 am:
d4c01761d3 am:
d20475d813
am:
66bd71c817
Change-Id: If7fb06d8679a3de6238cad0b726b5fe96bddd036
Mika Isojrvi [Wed, 19 Jul 2017 18:07:01 +0000 (18:07 +0000)]
Fix atomic ssbo xor test am:
f0fa05e898 am:
14cd264501
am:
6e80057a20
Change-Id: I85568945b0bb02c5b82dbba6ba449cd851358694
Mika Isojrvi [Wed, 19 Jul 2017 18:07:00 +0000 (18:07 +0000)]
Fix atomic ssbo xor test am:
f0fa05e898 am:
14cd264501 am:
d4c01761d3
am:
d20475d813
Change-Id: Ib23d8a8b67847c4444f7c1e2e05af16c023f9cb4
Mika Isojrvi [Wed, 19 Jul 2017 18:04:31 +0000 (18:04 +0000)]
Fix atomic ssbo xor test am:
f0fa05e898 am:
14cd264501
am:
d4c01761d3
Change-Id: I3fe580bed5dcdb5f41742c0e6e0ee1db05eae572
Mika Isojrvi [Wed, 19 Jul 2017 18:04:31 +0000 (18:04 +0000)]
Fix atomic ssbo xor test am:
f0fa05e898
am:
14cd264501
Change-Id: Ide39036c1f956b64464d21eced9a16c0a2022f2b
Mika Isojrvi [Wed, 19 Jul 2017 18:03:03 +0000 (18:03 +0000)]
Fix atomic ssbo xor test am:
f0fa05e898 am:
14cd264501
am:
d4c01761d3
Change-Id: I32a66d488ccd7f1544f23b299eaf4d0642cebc4a
Mika Isojrvi [Wed, 19 Jul 2017 18:00:27 +0000 (18:00 +0000)]
Fix atomic ssbo xor test am:
f0fa05e898
am:
14cd264501
Change-Id: Ib554d7464534203cbcde53f0ebd6931082900a82
Mika Isojrvi [Wed, 19 Jul 2017 17:57:28 +0000 (17:57 +0000)]
Fix atomic ssbo xor test
am:
f0fa05e898
Change-Id: Ibc16a0ff1a6b1dbe764d4beafa5d9b63d9c090e9
Kalle Raita [Tue, 18 Jul 2017 21:52:15 +0000 (14:52 -0700)]
Add triangle edge guardband for Tex2D lookup diff
Previously 3D and cube textures allowed pixels along the seam of
triangles to use data from either triangle. This change copies that
allowance to the 2D texture computeTextureLookupDiff. The projection
values chosen for the mipmap.2d.projected group make the data
discontinous over the edge, making it sensitive to tie-breaker
resolution. This change resolves the issue.
Affects: framework
Test: Executed mipmap.2d.projected group with failing result images
Bug:
63335787
Change-Id: Ibaed590fb5e1bc9eb837c9027ddadcea2081f629
Courtney Goeltzenleuchter [Tue, 18 Jul 2017 20:25:01 +0000 (20:25 +0000)]
EGL: Add wide-color tests
am:
6a26fb1fd0
Change-Id: Ib496be50b20bce758c441f08a34b0e10998b2e98
Mika Isojärvi [Thu, 14 Apr 2016 17:28:16 +0000 (10:28 -0700)]
Fix atomic ssbo xor test
Allow any values in high order bits in xor tests when using uint types.
Bug:
28144118
Change-Id: I0670aa12642882a329cb923487198fdf045b912b
Courtney Goeltzenleuchter [Mon, 19 Jun 2017 16:35:01 +0000 (10:35 -0600)]
EGL: Add wide-color tests
Verify necessary extensions and functionality
is present to support wide-color. Specifically, pixel formats
FP16 and 10:10:10:2 and Display-P3 color space.
Test: adb -d shell \
am start -n com.drawelements.deqp/android.app.NativeActivity \
-e cmdLine '"deqp --deqp-case=dEQP-EGL.functional.wide_color.* \
--deqp-log-filename=/sdcard/dEQP-Log.qpa"'
TODO:
- Do we need test for EGL_EXT_gl_colorspace_scrgb_linear?
That should be no-op for driver.
Change-Id: I537a04b1680e33e46215c4bb583f6987d20f0b4a
Mika Isojarvi [Mon, 17 Jul 2017 20:19:45 +0000 (20:19 +0000)]
Merge "Merge Vulkan CTS 1.0.2.4 into aosp/master"
am:
0f5658c72b
Change-Id: Id028fbdc5b596d55e044fb80ad34ecc95d0563a4
Mika Isojarvi [Mon, 17 Jul 2017 20:13:57 +0000 (20:13 +0000)]
Merge "Merge Vulkan CTS 1.0.2.4 into aosp/master"
JarredDavies [Fri, 14 Jul 2017 14:55:46 +0000 (15:55 +0100)]
Fix incorrect scale in pipeline_barrier tests
The ImageBlitFromImage verfication code incorrectly scales the refrence image it creates.
The test should add 0.5 to the x and y coords before scaling them.
Affects:
dEQP-VK.memory.pipeline_barrier.*
Components: Vulkan
VK-GL-CTS issue: 577
Change-Id: I28d0b4deec0f1837c21cdc2dc68a9169ca8ae96b
Slawomir Cygan [Thu, 6 Jul 2017 17:24:18 +0000 (19:24 +0200)]
copies_and_blitting: fix ambiguous signed normalized 1.0 value
Below tests copy floats to signed normalized 8-bit values.
In that process it happens that 1.0 float value is copied to 8_snorm
elements as: {0x00, 0x00, 0x80, 0x3f}
According to spec, on SNORM values:
"Note that while zero is exactly expressible in this representation,
one value (-128 in the example) is outside the representable range,
and must be clamped before use."
The problem occurs when 0x80 == 128 value is copied to buffer:
as it falls outside of <-1,1> range, it may be clamped to 0x81 (both values
represent -1.0 SNORM value).
The change here processes all result and expected data buffers to remove
all ambiguites by converting to float and back to signed normalized format.
Affects:
dEQP-VK.api.copy_and_blit.*.image_to_image.all_formats.color.r32_sfloat.a8b8g8r8_snorm_pack32.*
dEQP-VK.api.copy_and_blit.*.image_to_image.all_formats.color.r32_sfloat.r8g8b8a8_snorm.*
Components: Vulkan
VK-GL-CTS issue: 563
Change-Id: I175665f39811cf6a80862967ad0a9bb8e13527e9
Yanjun Zhang [Fri, 7 Jul 2017 20:45:59 +0000 (13:45 -0700)]
Relax threshold in cube map filtering tests
Use the same colorThreshold calculation as GLES CTS so Vulkan CTS
has the same behavior as GLES CTS.
Components: Vulkan
Affects:
dEQP-VK.pipeline.sampler.view_type.cube.format.r5g6b5_unorm_pack16.*
VK-GL-CTS issue: 561
Change-Id: I9bd632bb242f653b1f2d8bb50ee9867da926e95b
Slawomir Cygan [Fri, 7 Jul 2017 14:41:30 +0000 (16:41 +0200)]
cross_instance sync: run without VK_KHR_get_memory_requirements2
This extension is used in the test, to check if resource needs dedicated
allocations (to be used created by dependent VK_KHR_dedicated_allocation
extension).
If the extension is not supported assume there is no need for dedicated
allocations.
Affects: dEQP-VK.synchronization.cross_instance.*
VK-GL-CTS issue: 548
Components: Vulkan
Change-Id: Icb4f5a34b98da213d0cb665bcc4e6b048d7b8a3c
gkbldval [Fri, 14 Jul 2017 16:49:26 +0000 (18:49 +0200)]
Refactor synchronization test code to avoid compiler warning
vkt::synchronization::s_resources is unused in recently
added file, causing -Wunused warning (treated as error).
Affects: dEQP-VK.synchronization.*
Components: Vulkan
VK-GL-CTS issue: 578
Change-Id: Ifd8ce0812673e6b2da7c28bacb4384b3c80f65d1
Alexander Galazin [Wed, 5 Jul 2017 15:09:38 +0000 (17:09 +0200)]
Fix gcc 5.4 warnings
gcc 5.4 (default in Ubuntu 16.04) reports conversion to int
warnings/errors for bitwise or and assign operator.
GCC bug: 40752
Components: Framework, Vulkan
Change-Id: Ia857c44cf54d5571d857e1fada0e2aaa8f4b950d
(cherry picked from commit
0bd8828618d95d2a83bc7475179a9227bfd22f5f)
Alexander Galazin [Fri, 14 Jul 2017 07:48:05 +0000 (09:48 +0200)]
Add option to skip fetch external sources
Added option to check_build_sanity.py to skip running
fetch_sources.py. Allows to run multiple concurrent invocations
of check_build_sanity.py using the same source folder.
Components: Framework
Change-Id: I6289920860fb20287c405cefc0213c27f326857b
Mika Isojärvi [Fri, 14 Jul 2017 22:20:31 +0000 (15:20 -0700)]
Merge Vulkan CTS 1.0.2.4 into aosp/master
Bug:
62712168
Change-Id: Ib5dcdc7645193825171cb93c5a3b8a0820cbc5a1