Mika Isojärvi [Tue, 7 Jun 2016 19:39:37 +0000 (12:39 -0700)]
Improve memory tracking in vulkan memory mapping tests
- Round up allocation size to multiple of page size for
memory usage tracking.
- Update reference memory usage tracking to follow new reference memory
implementation.
Bug: b/
69677943
Change-Id: I219c2b3abe0ee0731f5a7e2bb44492e5b85b95c2
Mika Isojärvi [Tue, 7 Jun 2016 18:21:46 +0000 (11:21 -0700)]
Optimize random memory mapping tests reference memory
Bug: b/
69677943
Change-Id: I7af6fe67a25462dcb0acc66c238a8708157060df
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
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
Mika Isojrvi [Wed, 19 Jul 2017 17:57:28 +0000 (17:57 +0000)]
Fix atomic ssbo xor test
am:
f0fa05e898
Change-Id: Ibc16a0ff1a6b1dbe764d4beafa5d9b63d9c090e9
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
Mika Isojrvi [Wed, 12 Jul 2017 22:51:27 +0000 (22:51 +0000)]
Mask compared bits in atomic operation tests.
am:
b2ed9fbe1a
Change-Id: I2f7b4cfbf2d926966af7c9956db670080771c164
Graeme Leese [Thu, 23 Feb 2017 16:56:10 +0000 (16:56 +0000)]
Fix extension support checks in negative api tests
Add checks for tessellation and geometry shader support for negative
shader storage and primitive bounding box tests.
Move all uses of cubemap array textures inside the extension checks in
the texture tests.
Bug:
29835218
Change-Id: I6051fff231285ed7d9c71015c2bb11a9fc7b909b
Signed-off-by: Pierre Couillaud <pierre@broadcom.com>
Mika Isojärvi [Tue, 5 Apr 2016 17:42:09 +0000 (10:42 -0700)]
Mask compared bits in atomic operation tests.
Bug:
27975797
Change-Id: Id24a784b5b8a665f509218358d58d7eca9a7bb93
Treehugger Robot [Wed, 3 May 2017 16:04:13 +0000 (16:04 +0000)]
Merge "Fix color change verification in dithering tests" into nougat-cts-dev
Treehugger Robot [Wed, 3 May 2017 16:01:05 +0000 (16:01 +0000)]
Merge "DEPTH_STENCIL_OES as tex format requires OES_depth_texture" into nougat-cts-dev
Treehugger Robot [Wed, 3 May 2017 15:59:50 +0000 (15:59 +0000)]
Merge "Fix error mask generation in checkLineContinuity" into nougat-cts-dev
Alexander Galazin [Thu, 20 Apr 2017 12:08:44 +0000 (14:08 +0200)]
Fix error mask generation in checkLineContinuity
Even if some pixels are missing the function could return
all-zero errorMask if the messageLimitcounter was exhausted.
Components: AOSP
Affects: dEQP-GLES31.functional.primitive_bounding_box.lines.*
Google bug:
35987764
Pyry Haulos [Tue, 18 Apr 2017 23:13:59 +0000 (16:13 -0700)]
Fix color change verification in dithering tests
Tests were supposed to allow color change once per row/column depending
on gradient direction but instead allowed color change in constant
direction only once over whole image.
Affects:
dEQP-GLES2.functional.dither.disabled.gradient_*
dEQP-GLES3.functional.dither.disabled.gradient_*
Bug:
37477346
Change-Id: I37ef06b4ff527f8c316eae56c8a99c88cdaaf4b4
Pyry Haulos [Mon, 17 Apr 2017 21:34:27 +0000 (14:34 -0700)]
Try to determine renderable format in lifetime tests
Affects:
dEQP-GLES2.functional.lifetime.attach.deleted_input.renderbuffer_framebuffer
dEQP-GLES2.functional.lifetime.attach.deleted_input.texture_framebuffer
dEQP-GLES2.functional.lifetime.attach.deleted_output.renderbuffer_framebuffer
dEQP-GLES2.functional.lifetime.attach.deleted_output.texture_framebuffer
Bug:
36724173
Change-Id: I46196317c144f294adce65ff4690571d28e1caf3
Pyry Haulos [Mon, 17 Apr 2017 19:34:26 +0000 (12:34 -0700)]
DEPTH_STENCIL_OES as tex format requires OES_depth_texture
Affects:
dEQP-GLES2.functional.fbo.completeness.*
dEQP-GLES3.functional.fbo.completeness.*
Bug:
36727598
Merged-In: Ic79408c21b80ea458baecf79c042e2694e72e0b2
Change-Id: Ib302b21b2b28d4f13aa6df8bdcac878f22ae51be
Treehugger Robot [Tue, 11 Apr 2017 16:56:42 +0000 (16:56 +0000)]
Merge "Try harder to defeat GLSL compiler dead-code optimizations" into nougat-cts-dev
Ian Romanick [Thu, 10 Nov 2016 01:29:01 +0000 (17:29 -0800)]
Try harder to defeat GLSL compiler dead-code optimizations
A number of CTS tests generate shaders like like:
#version 310 es
buffer TargetInterface
{
highp float target;
};
highp vec4 readInputs()
{
highp vec4 retValue = vec4(0.0);
retValue += vec4(float(target));
return retValue;
}
void writeOutputs(in highp vec4 dummyValue)
{
target = float(dummyValue.y);
}
void main()
{
writeOutputs(readInputs());
}
After various common optimizations this becomes:
buffer TargetInterface
{
highp float target;
};
void main()
{
target = target;
}
In the absence of memoryBarrier() or qualifiers on the buffer, there is
no guarantee about the order of writes to the buffer. Since this write
is not guaranteed to be visible either on the GPU or the CPU, we
eliminate it. Since there is no access to target in the shader, we
report GL_REFERENCED_BY_FRAGMENT_SHADER = GL_FALSE. The tests expect
GL_TRUE.
The vectored versions of this test swizzle the value read from the
buffer before writing it back. These writes are not eliminated.
Adding a uniform instead of a literal constant also prevents the reads
and writes of the SSBO from being eliminated.
v2: Ignore the uniform named "zero" in
ResourceListTestCase::verifyResourceList. The alternative was to add
zero to the resource list, but that required making small changes
(mostly removing const) from over a dozen places in the code. This
slightly hacky, but localized, change seemed better.
v3: Various coding standards fixes suggested by Alexander Galazin and
Pyry. Add getDummyZeroUniformName to query name of the zero uniform and
a lot more documentation. Both suggested by Pyry.
The following tests are affected:
- dEQP-GLES31.functional.program_interface_query.buffer_variable.random.6
- dEQP-GLES31.functional.program_interface_query.buffer_variable.referenced_by.compute.unnamed_block.float
- dEQP-GLES31.functional.program_interface_query.buffer_variable.referenced_by.separable_fragment.unnamed_block.float
- dEQP-GLES31.functional.program_interface_query.buffer_variable.referenced_by.vertex_fragment_only_fragment.unnamed_block.float
- dEQP-GLES31.functional.program_interface_query.buffer_variable.referenced_by.vertex_fragment.unnamed_block.float
- dEQP-GLES31.functional.program_interface_query.buffer_variable.referenced_by.vertex_geo_fragment_only_fragment.unnamed_block.float
- dEQP-GLES31.functional.program_interface_query.buffer_variable.referenced_by.vertex_geo_fragment.unnamed_block.float
- dEQP-GLES31.functional.program_interface_query.buffer_variable.referenced_by.vertex_tess_fragment_only_fragment.unnamed_block.float
- dEQP-GLES31.functional.program_interface_query.buffer_variable.referenced_by.vertex_tess_fragment.unnamed_block.float
- dEQP-GLES31.functional.program_interface_query.buffer_variable.referenced_by.vertex_tess_geo_fragment_only_fragment.unnamed_block.float
- dEQP-GLES31.functional.program_interface_query.buffer_variable.referenced_by.vertex_tess_geo_fragment.unnamed_block.float
Change-Id: I867ad32476269ac1272c09672be0a6d6fe37e31e
Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Bugzilla: https://cvs.khronos.org/bugzilla/show_bug.cgi?id=16114
Google bug:
37228062
Pyry Haulos [Wed, 15 Mar 2017 19:23:06 +0000 (12:23 -0700)]
Fix warning introduced in
2c9e3ec9
Change-Id: I42bef21b2cc7274f107e5455dea1e382d8a419f4
(cherry picked from commit
d8b452a7533c195b1328918bb696a819542ff044)
Treehugger Robot [Wed, 15 Mar 2017 18:30:21 +0000 (18:30 +0000)]
Merge "Further relax line verification in primitive bbox tests" into nougat-cts-dev
Alexander Galazin [Mon, 6 Mar 2017 18:44:38 +0000 (19:44 +0100)]
Further relax line verification in primitive bbox tests
The original relaxation patch
ba45591407d0207e8ba0634ea5636fd721ed781f
allowed lines to be 1 pixel thinner or thicker compared
to the expected width. Unfortunately it missed the case
of a line with the width of 1 pixel that has extra verification
in the framework. This change covers this gap.
Google bug:
35987764
Change-Id: I193d76ae13738702c9d7361d1957f15ea7600c4f
Alexander Galazin [Thu, 23 Feb 2017 12:01:18 +0000 (13:01 +0100)]
Use glReadnPixels only if KHR_robustness is supported
glReadnPixels is introduced in KHR_robustness/GLES 3.2.
Use of this function in GLES 3.1 tests will result in segfault
on implementations that don't support KHR_robustness.
Components: AOSP, Framework
Affects: dEQP-GLES31.functional.debug.negative_coverage.callbacks.buffer.read_pixels_format_mismatch
VK-GL-CTS issue: 166
Google bug:
36122027
Change-Id: Idbe8734f7d82882ea99c75a8b74d693ac6dd831f
(cherry picked from commit
219019495770c221e92f43df3e76a954c613210d)
Mika Isojärvi [Wed, 16 Nov 2016 23:10:29 +0000 (15:10 -0800)]
Limit changes by xor to upper 8 bits in mixed atomic tests
Bug:
31270281
Test: Ran tests on multiple devices
Change-Id: I64393591171b304cd071f7a37b91efd2c7a853a6
Tina Zhou [Mon, 23 Jan 2017 19:09:19 +0000 (19:09 +0000)]
Merge "x11: Call XInitThreads()" into nougat-cts-dev
Tina Zhou [Mon, 23 Jan 2017 19:06:21 +0000 (19:06 +0000)]
Merge "x11: Fix deadlock" into nougat-cts-dev
Chad Versace [Fri, 20 Jan 2017 22:59:10 +0000 (14:59 -0800)]
x11: Fix deadlock
When waiting for a window's MapNotify event, drain only that window's
event queue. Do NOT drain the global event queue, which includes all
windows' events.
Fixes deadlock in test 'dEQP-EGL.functional.multithread.window_context'
on Mesa master@
f57bdd48 with Intel Skylake.
Change-Id: I8c7f2db6c145f59493cb7572b20c4799e9a2cd0a
Depends-On: Ib3b535cec8295e062994fd759ae083c78641cf13
Chad Versace [Tue, 27 Dec 2016 21:47:33 +0000 (13:47 -0800)]
x11: Call XInitThreads()
Because the X11 manual says we must.
From man:XInitThreads(3):
The XInitThreads function initializes Xlib support for concurrent
threads. This function must be the first Xlib function
a multi-threaded program calls, and it must complete before any other
Xlib call is made.
Fixes crash in multithreaded test:
Test case 'dEQP-EGL.functional.multithread.window_context'..
[xcb] Unknown sequence number while processing queue
[xcb] Most likely this is a multi-threaded client and XInitThreads has not been called
[xcb] Aborting, sorry about that.
deqp-egl: xcb_io.c:259: poll_for_event: Assertion `!xcb_xlib_threads_sequence_lost' failed.
Change-Id: Ib3b535cec8295e062994fd759ae083c78641cf13
Kalle Raita [Thu, 1 Sep 2016 20:21:50 +0000 (13:21 -0700)]
CP: Remove two read-backs from copy image tests
Cherry pick commit
b2fb8ad5ceefe26964101bd403e7786ae83274a9
from master.
Bug:
33965234
Remove two read-backs from copy image tests
Test: Ran copy_image on N6 & N9 with L MR1 image to check for
false-negatives. PixelC and Pixel XL for passing with current OS image.
Change-Id: I9f7028eb550012f8d9abe2f7afee2d1402f5abf3
Kalle Raita [Wed, 7 Dec 2016 17:41:41 +0000 (09:41 -0800)]
CP: Non-compressed copy image target size (64,64,8
Cherry-pick commit
8076c97ac0fc706288ef266b0cbe748012c4e0c4 from master:
Non-compressed copy image tests target size reduced to (64,64,8). The
compressed cases retain old size of (128,128,16) to ensure coverage of
all block size combinations.
Bug:
33965234
Test: Run tests on N6 and N9 on L MR 1 with known failures
Change-Id: If22e8722c1de58aa239cf6fe29e79d3bdd13c589
Tina Zhou [Fri, 20 Jan 2017 18:28:49 +0000 (18:28 +0000)]
Merge "Remove tests affected by NV/OES_viewport_array extension" into nougat-cts-dev
Tina Zhou [Fri, 20 Jan 2017 18:28:42 +0000 (18:28 +0000)]
Merge "Generate flat triangles in polygon offset tests." into nougat-cts-dev
Tina Zhou [Fri, 20 Jan 2017 18:28:34 +0000 (18:28 +0000)]
Merge "Relax line width verification in primitive bbox tests" into nougat-cts-dev
Tina Zhou [Fri, 20 Jan 2017 18:28:26 +0000 (18:28 +0000)]
Merge "Fix eglBindAPI negative test" into nougat-cts-dev
Tina Zhou [Fri, 20 Jan 2017 18:28:18 +0000 (18:28 +0000)]
Merge "Drop incorrect framebuffer texture subtests." into nougat-cts-dev
Tina Zhou [Fri, 20 Jan 2017 18:27:50 +0000 (18:27 +0000)]
Merge "Optimize swapchain OOM tests" into nougat-cts-dev
Tina Zhou [Fri, 20 Jan 2017 18:27:40 +0000 (18:27 +0000)]
Merge "Fix internal format/type for ES3 3D + depth/stencil negative API tests." into nougat-cts-dev
Tina Zhou [Fri, 20 Jan 2017 18:27:14 +0000 (18:27 +0000)]
Merge "Add support for EGL_EXT_pixel_format_float" into nougat-cts-dev
Pyry Haulos [Wed, 18 Jan 2017 20:28:32 +0000 (12:28 -0800)]
Remove tests affected by NV/OES_viewport_array extension
Tests were not properly taking into account viewport clamping behavior
introduced by GL_NV/OES_viewport_array extensions.
Bug:
34349411
Change-Id: I3eae36359e6408191e0fbefb20cededec2a5bd39
Merged-In: I3d3c2e32d7dd96d1db960aa53dc18cc690187ec0
Alexander Galazin [Wed, 11 Jan 2017 07:50:13 +0000 (08:50 +0100)]
Generate flat triangles in polygon offset tests.
The tests setup triangles with a depth slope and then call
glPolygonOffset(factor, units) with variable units and
always 0 slope scaled factor.
According to the GLES 3.2 spec 13.7.2 Depth Offset:
"void PolygonOffset( float factor, float units );
factor scales the maximum depth slope of the polygon, and units scales an
implementation-dependent constant that relates to the usable resolution of the
depth buffer."
If the polygon has a depth slope,
the depth slope factor has to be not equal to zero.
The rationale of the slope scaled factor is to ensure that the offset
(equation 13.11) is greater than max(dz/dx, dz/dy) for all resolutions,
which may not be the case for the constant offset since it doesn't take
the screen resolution into account.
Changed tests:
dEQP-GLES2.functional.polygon_offset.*_displacement_with_units
dEQP-GLES3.functional.polygon_offset.*_displacement_with_units
Bug:
34386628
Change-Id: Ie8009ba1766eb50c3c5a750122aa8bb231144210
Kenneth Graunke [Fri, 6 Jan 2017 08:25:28 +0000 (00:25 -0800)]
Drop incorrect framebuffer texture subtests.
This patch removes two broken subcases of:
dEQP-GLES31.functional.debug.negative_coverage.log.buffer.framebuffer_texture_layer
These were fixed in dEQP master by commit
f28187809176f5eeb1ee25ccfa70d3,
but we cannot backport that to release branches because it changes the
existing expectations, which would break previously passing
implementations. See buganizer
30287182 and
28090252.
The first subtest incorrectly checked layer (array slice) when it should
have checked miplevel. The subtest description says:
"GL_INVALID_VALUE is generated if texture is a 2D multisample array
texture and layer not 0."
which is wrong - array textures can certainly have multiple slices.
The ES 3.1 specification says that:
"An INVALID_VALUE error is generated if texture is not zero and level is
not a supported texture level for texture, as described above."
and miplevel 0 is the only valid miplevel for a multisample texture.
The second subtest is no longer valid according to the resolution of
https://cvs.khronos.org/bugzilla/show_bug.cgi?id=15968.
Bug:
34281627
Merged-In: I500074c1777d88291c0b395498bb0f7d5613d27d
Pyry Haulos [Tue, 15 Nov 2016 18:28:48 +0000 (10:28 -0800)]
Relax line width verification in primitive bbox tests
Originally contributed by ARM to Khronos OpenGL ES CTS. Cherry-picked
with permission.
Google bug:
22713865
Khronos OpenGL issue: #19
Change-Id: I95359847611144e179819aaab6ac5dfe4bf4e4da
Pyry Haulos [Wed, 4 Jan 2017 22:30:07 +0000 (14:30 -0800)]
Optimize swapchain OOM tests
Swapchain OOM tests were hitting timeouts on some platforms. This commit
optimizes the tests in two ways:
* VkInstance, VkDevice, and VkSurface are now re-used across all
sub-cases.
* Loop is re-written to avoid re-creating swapchains for parameters
0..N-1 when testing parameter combination N.
Affects: dEQP-VK.wsi.*.swapchain.simulate_oom.*
Bug:
33555898
(cherry picked from commit
2beac9057d9113ac306632d819ded852691a6842)
Change-Id: I19e9c7c8012a82cd12322a733afab961f4c7a1ae
Kenneth Graunke [Mon, 12 Sep 2016 19:37:08 +0000 (12:37 -0700)]
Fix internal format/type for ES3 3D + depth/stencil negative API tests.
This is a port of commit
ae7f8e0a07730e693b24d3dc7a23d2372319145e from
the ES 3.1 tests to the ES 3.0 tests.
According to the ES 3.2 specification:
"Textures with a base internal format of DEPTH_COMPONENT, DEPTH_STENCIL
or STENCIL_INDEX are supported by texture image specification commands
only if target is TEXTURE_2D, TEXTURE_2D_MULTISAMPLE, TEXTURE_2D_ARRAY,
TEXTURE_2D_MULTISAMPLE_ARRAY, TEXTURE_CUBE_MAP or TEXTURE_CUBE_MAP_ARRAY.
Using these formats in conjunction with any other target will result in
an INVALID_OPERATION error."
This subtest tried to check the above error condition, but it specified
GL_DEPTH_STENCIL / GL_DEPTH_COMPONENT as format, rather than internalFormat.
Since the above text calls out "base internal format", we should specify
it as internalFormat.
We also change GL_DEPTH_STENCIL to use GL_UNSIGNED_INT_24_8 rather than
GL_UNSIGNED_BYTE, as that combination was illegal for a different reason
than the one the test intended to check.
Affects dEQP-GLES3.functional.negative_api.texture.teximage3d.
Bug:
34103293
Change-Id: Ie01e2d130bb1cadc821153487e3e41593e3ca15e
Nicolas Boichat [Sun, 18 Dec 2016 16:13:28 +0000 (00:13 +0800)]
Fix eglBindAPI negative test
It is ok for an implementation to not fail eglBindAPI if the current
display does not support the specified client API. However, in that
case, getConfig must not return any configuration for that specific
API.
Bug:
33278378
Affects: dEQP-EGL.functional.negative_api.bind_api
Test: Deploy to elm Chromebook, run with Cherry, test passes.
Change-Id: Idb86971de7a30e1aa78bb0951f971ed2829e2794
Mark Adams [Wed, 14 Dec 2016 22:02:52 +0000 (17:02 -0500)]
Add support for EGL_EXT_pixel_format_float
This updates the egl.xml version and identifies/skips the float configs
Bug:
32969314
Change-Id: Iacfdf6b1c32c86ea0e6f80ca7ebaaeef583d1135
Merged-In: Iadeb0d4a76a06b2be716f3e76a1cc9fb69aff16a
Pyry Haulos [Wed, 4 Jan 2017 16:20:43 +0000 (08:20 -0800)]
DO NOT MERGE: Remove changed texel buffer access tests from mustpass
NYC CTS bug-fix changes the SPIR-V for these tests causing drivers that
passed old CTS fail the new version.
Bug:
33041922
Change-Id: I7e23a700693d5e6ee7c4264e05d449f49f820516
Pyry Haulos [Wed, 4 Jan 2017 00:45:19 +0000 (16:45 -0800)]
DO NOT MERGE: Remove a few changed Vulkan tests from mustpass
Original NYC CTS was decorating struct member locations differently and
this was later fixed causing drivers that pass original NYC CTS fail.
Bug:
34060750
Bug:
33041922
Change-Id: I3fe33960a74fb8f568aa7e1327bc09de8bc3df41
Tina Zhou [Wed, 21 Dec 2016 22:32:08 +0000 (22:32 +0000)]
Merge "Blacklist match_different_[member_]struct_names SSO validation tests." into nougat-cts-dev
Tina Zhou [Wed, 21 Dec 2016 22:31:04 +0000 (22:31 +0000)]
Merge "Change GL_MAX_TESS_CONTROL_TOTAL_OUTPUT_COMPONENTS minmax to 2048" into nougat-cts-dev
Tina Zhou [Wed, 21 Dec 2016 22:30:03 +0000 (22:30 +0000)]
Merge changes I7c2ffd66,Ib53b245d,I4862cca3 into nougat-cts-dev
* changes:
Update glslang to d02dc5d
Cherry-pick SPIR-V ClipDistance validation fixes
Cherry-pick dEQP-VK.pipeline.cache fixes to nougat-cts-dev
Nicolas Boichat [Tue, 20 Dec 2016 18:55:24 +0000 (18:55 +0000)]
GetDevices: Also allow `\.` to be part of the host name
am:
dbf9715509
Change-Id: I659ce579867c7ca0c9b47e3b53a734265e0fbb26
Kenneth Graunke [Thu, 8 Dec 2016 22:09:31 +0000 (14:09 -0800)]
Blacklist match_different_[member_]struct_names SSO validation tests.
This patch blacklists the following tests:
- dEQP-GLES31.functional.separate_shader.validation.varying.match_different_struct_names
- dEQP-GLES31.functional.separate_shader.validation.io_blocks.match_different_member_struct_names
The resolution of Khronos bug #15866 is that the spec should be changed
to require that "structures have the same name" in order for the types
to match during SSO validation. These structures differ in name,
so they should not match, and we should get an SSO validation failure.
See https://cvs.khronos.org/bugzilla/show_bug.cgi?id=15866 and Khronos
legacy CTS commit
1607a3ce7caa5736471fb460d08130e62d689351.
Bug:
33457655
Nicolas Boichat [Wed, 30 Nov 2016 04:14:52 +0000 (12:14 +0800)]
GetDevices: Also allow `\.` to be part of the host name
Useful when connecting over TCP to a device with a given IP.
Change-Id: Ibe3b75e5ba3ab0ff90219bb31d45e21d8f5fd61b
Pyry Haulos [Tue, 29 Nov 2016 23:24:17 +0000 (15:24 -0800)]
Update glslang to d02dc5d
Bug:
33041922
(cherry picked from commit
4e1f56d2457e0c8736e3e92c48bd8df0c309d600)
Change-Id: I7c2ffd66f55cafa402ffc7a0222973e282f4825f
Pyry Haulos [Wed, 23 Nov 2016 00:13:32 +0000 (16:13 -0800)]
Cherry-pick SPIR-V ClipDistance validation fixes
These fixes are needed in order for SPIR-V validation to pass when
shaders are compiled with a newer glslang version.
Bug:
33041922
Explicitly declare gl_PerVertex block
This fixes SPIR-V validation error in shaders generated from GLSL (not ES)
due to missing ClipDistance capability.
(cherry picked from commit
0a6fe1448b303a6a4e4ab48712eaf1c0b16d75b1)
Explicitly declare gl_PerVertex in push constant tests
(cherry picked from commit
cba5d02d184b771bc4a21b8aecbbde3377d86167)
Change tes/geom shaders in timestamp tests use 310 es
Vertex and fragment shaders were already using that GLSL version.
(cherry picked from commit
d47309690aa9b2e685b267ea6395e88b32c855a8)
Change ubo and ssbo tests to use 310 es shaders
(cherry picked from commit
4d36051923691c5f92cf3615e82d30d54ed3b09b)
Explicitly declare gl_PerVertex in dynamic viewport state tests
(cherry picked from commit
816c46d518d08a4ff321ca93780f2786fa86ba60)
Explicitly declare gl_PerVertex in occlusion query tests
(cherry picked from commit
a1257f69aa6008d4e5a22148a43d59692eb24822)
Change-Id: Ib53b245d25c6cbc345a88e8b247f01ac7cf8a541
Pyry Haulos [Mon, 21 Nov 2016 23:54:55 +0000 (15:54 -0800)]
Cherry-pick dEQP-VK.pipeline.cache fixes to nougat-cts-dev
Bug:
33041922
Validation fixes in dEQP-VK.pipeline
- Corrected out of range minDepthBounds value
- Added missing image layout transitions
- timestamp, cache tests: pass data through geometry shader
- cache test: use correct initialLayout in attachment when starting
a second render pass
(cherry picked from commit
1d3585c5ec71cf90d6519913db6a01ef2c0f840c)
Fix use of out-of-scope struct in pipeline cache tests
Fixes #363
(cherry picked from commit
1ec71c633e5755a1932c1da7692aef69b66999db)
Pipeline cache test: fix SPIRV validation error
(cherry picked from commit
89cce75bae9beb123311281386fb59b8da59052b)
Change-Id: I4862cca352955c33d04540f2f881760f4e4ffd65
Pyry Haulos [Tue, 15 Nov 2016 18:34:57 +0000 (10:34 -0800)]
Change GL_MAX_TESS_CONTROL_TOTAL_OUTPUT_COMPONENTS minmax to 2048
Google bug:
33105204
Khronos bug: 12823
(cherry picked from commit
fe6bb67987a4d7fc59739c878e658f2a6d071cde)
Change-Id: I56fcc4ce527c31e48004c43e43e2ec994675e1b4
Tina Zhou [Thu, 17 Nov 2016 17:10:28 +0000 (17:10 +0000)]
Merge "Fix sample_mask_in.bit_count_per_two_samples tests for 2x MSAA." into marshmallow-cts-dev
am:
a5af1b076f
Change-Id: I01c65d693d7ad5befe9244cb0d24814750a80bd2
Kenneth Graunke [Thu, 17 Nov 2016 17:10:28 +0000 (17:10 +0000)]
Fix sample_mask_in.bit_count_per_two_samples tests for 2x MSAA.
am:
ea5589c748
Change-Id: I7358276f100608371da5b44ebc8e102a54706197
Tina Zhou [Thu, 17 Nov 2016 17:10:12 +0000 (17:10 +0000)]
Merge "Fix internal format/type for 3D + depth/stencil negative API tests." into marshmallow-cts-dev
am:
d962b86909
Change-Id: I6a82d71f2fed9f848a953ae21c30079bacb07052
Kenneth Graunke [Thu, 17 Nov 2016 17:10:05 +0000 (17:10 +0000)]
Fix internal format/type for 3D + depth/stencil negative API tests.
am:
d2f3b468db
Change-Id: Id43b369d67e04cf3a6d6b5b4e3303f02c7a35b87
Tina Zhou [Thu, 17 Nov 2016 17:03:36 +0000 (17:03 +0000)]
Merge "Fix sample_mask_in.bit_count_per_two_samples tests for 2x MSAA." into marshmallow-cts-dev
Tina Zhou [Thu, 17 Nov 2016 17:02:44 +0000 (17:02 +0000)]
Merge "Fix internal format/type for 3D + depth/stencil negative API tests." into marshmallow-cts-dev
Tina Zhou [Thu, 17 Nov 2016 00:43:40 +0000 (00:43 +0000)]
Merge "Add the support to device connection via TCP/IP" into marshmallow-cts-dev
am:
6cea9dbc62
Change-Id: I3525d88dbc3e04b702cef54d1759f406222047b2
Chun-Ta Lin [Thu, 17 Nov 2016 00:43:39 +0000 (00:43 +0000)]
Add the support to device connection via TCP/IP
am:
4ccc9fd56c
Change-Id: I16ab5038ff64448458aaafc7bda43128e0941e4e
Tina Zhou [Thu, 17 Nov 2016 00:38:13 +0000 (00:38 +0000)]
Merge "Add the support to device connection via TCP/IP" into marshmallow-cts-dev
Aaron Holden [Tue, 15 Nov 2016 01:47:00 +0000 (17:47 -0800)]
Add runtime to CtsDeqpTestCases
Test: cts-tradefed run cts -m CtsDeqpTestCases
Bug:
32843094
Change-Id: I6ea71f0d6658b1f06354374ece54743609472345
Tina Zhou [Wed, 19 Oct 2016 15:59:19 +0000 (15:59 +0000)]
Merge "Check for EXT_color_buffer_float in read_pixels_fbo_format_mismatch" into nougat-cts-dev
Tina Zhou [Wed, 19 Oct 2016 15:58:07 +0000 (15:58 +0000)]
Merge "CP: Use empty region when buffer age is not supported" into nougat-cts-dev
Tina Zhou [Wed, 19 Oct 2016 15:56:53 +0000 (15:56 +0000)]
Merge "Fix deStrnlen fall-back implementation" into nougat-cts-dev
Tina Zhou [Wed, 19 Oct 2016 15:56:21 +0000 (15:56 +0000)]
Merge "Remove a few preprocessor cases testing undefined behavior" into nougat-cts-dev
Tina Zhou [Wed, 19 Oct 2016 15:52:15 +0000 (15:52 +0000)]
Merge "Check for shader type support in negative precise tests" into nougat-cts-dev
Tina Zhou [Wed, 19 Oct 2016 15:50:37 +0000 (15:50 +0000)]
Merge "Check for shader type support in negative shader directive tests" into nougat-cts-dev
Kenneth Graunke [Sun, 4 Sep 2016 07:34:47 +0000 (00:34 -0700)]
Fix sample_mask_in.bit_count_per_two_samples tests for 2x MSAA.
The dEQP-GLES31.functional.shaders.sample_variables.sample_mask_in.
bit_count_per_two_samples.multisample_{texture,renderbuffer}_{1,2}
tests report:
Verifying gl_SampleMaskIn.
Fragment shader may be invoked [ceil(numSamples/2), numSamples] times.
=> gl_SampleMaskIn should have the number of bits set in range
[1, numSamples - ceil(numSamples/2) + 1]:
which suggests that maxBitCount should be ceil(2/2) + 1 = 1 + 1 = 2.
However, it then says:
Setting minBitCount = 1, maxBitCount = 1.
Having two bits should be acceptable for 2x MSAA.
This patch drops the special case for 1x/2x MSAA, making them work
like all the other MSAA levels.
Change-Id: Ie615c82150588dab53226dc46670182a29d85ef5
(cherry picked from commit
724519b8fda9c2a92d646c3dfd1478e0daa282e9)
Kenneth Graunke [Mon, 12 Sep 2016 19:37:08 +0000 (12:37 -0700)]
Fix internal format/type for 3D + depth/stencil negative API tests.
According to the ES 3.2 specification:
"Textures with a base internal format of DEPTH_COMPONENT, DEPTH_STENCIL
or STENCIL_INDEX are supported by texture image specification commands
only if target is TEXTURE_2D, TEXTURE_2D_MULTISAMPLE, TEXTURE_2D_ARRAY,
TEXTURE_2D_MULTISAMPLE_ARRAY, TEXTURE_CUBE_MAP or TEXTURE_CUBE_MAP_ARRAY.
Using these formats in conjunction with any other target will result in
an INVALID_OPERATION error."
This subtest tried to check the above error condition, but it specified
GL_DEPTH_STENCIL / GL_DEPTH_COMPONENT as format, rather than internalFormat.
Since the above text calls out "base internal format", we should specify
it as internalFormat.
We also change GL_DEPTH_STENCIL to use GL_UNSIGNED_INT_24_8 rather than
GL_UNSIGNED_BYTE, as that combination was illegal for a different reason
than the one the test intended to check.
Change-Id: I44bf607533b59f93fa42415bd6cad8783352582d
(cherry picked from commit
ae7f8e0a07730e693b24d3dc7a23d2372319145e)
Pyry Haulos [Thu, 7 Jul 2016 21:58:52 +0000 (14:58 -0700)]
Fix deStrnlen fall-back implementation
Fixes #423
Bug:
31952754
(cherry picked from commit
038db60bcbb1cdaf0157e5176af6422bda13ff28)
Change-Id: I7f22c2f0d325ecd2a57462f324214b7009c18782
Pyry Haulos [Tue, 11 Oct 2016 17:24:27 +0000 (10:24 -0700)]
Cherry-pick image_format_properties changes from upstream
Bug:
31732496
Contains 4 commits:
1: Fix dEQP-VK.api.info.image_format_properties - TRANSIENT usage flag used with DMA flags
(cherry picked from commit
83f90be9118a121b47f5e16f04029e205c68f994)
2: Fix VK_IMAGE_USAGE_TRANSIENT_ATTACHMENT_BIT handling in query tests
83f90be9 added a check that was intended to make tests treat combination
of TRANSIENT_ATTACHMENT_BIT with non-RT usages invalid. However, that
change was buggy and lead the tests to treat _all_ usages of
TRANSIENT_ATTACHMENT_BIT invalid and thus skip testing them.
(cherry picked from commit
570326e5c8d85fe16fa6ce36de73c8f9bcba4d3e)
3: Do not require TRANSIENT usage support without other RT usages
Affects dEQP-VK.api.info.image_format_properties.*
See CTS issue #507 and spec issue #540
(cherry picked from commit
377ad0ad5962351d14badb1857b55dc8e8e49218)
4: Fix required sample counts per spec issue 478
Affects dEQP-VK.api.info.image_format_properties.*
See CTS issue #507 and spec issue #478
(cherry picked from commit
02c61e30f50cb04505f421b1ebe07c4a42864acf)
Change-Id: I4c714f887c1554994f90ec361ed44876465cd5a3
Pyry Haulos [Thu, 29 Sep 2016 17:11:52 +0000 (10:11 -0700)]
Remove a few preprocessor cases testing undefined behavior
Bug:
27126954
Change-Id: I1ee7ca0c0102b6273b1bc165b3d1ad24cee449e2
Daniel Xie [Wed, 14 Sep 2016 18:15:10 +0000 (18:15 +0000)]
Merge "CP: Split load/store image tests into sub-groups" into nougat-cts-dev
Daniel Xie [Tue, 13 Sep 2016 16:43:43 +0000 (16:43 +0000)]
Merge "CP: Require 'master' config to be supported in Android CTS" into nougat-cts-dev
Daniel Xie [Tue, 13 Sep 2016 16:43:39 +0000 (16:43 +0000)]
Merge "CP: Handle EGL_EXT_yuv_surface in eglChooseConfig() tests" into nougat-cts-dev
Daniel Xie [Tue, 13 Sep 2016 16:43:35 +0000 (16:43 +0000)]
Merge "Remove a few EGL tests using RGBA4 format from mustpass" into nougat-cts-dev
Daniel Xie [Tue, 13 Sep 2016 16:43:30 +0000 (16:43 +0000)]
Merge "CP: Reduce max iter count in alloc_callback_fail.device" into nougat-cts-dev
Daniel Xie [Tue, 13 Sep 2016 16:43:20 +0000 (16:43 +0000)]
Merge changes I212a8bc2,I476b6586 into nougat-cts-dev
* changes:
DO NOT MERGE Allow old behavior in image format property tests
CP: Fix required sample count tests
Daniel Xie [Tue, 13 Sep 2016 16:43:15 +0000 (16:43 +0000)]
Merge "CP: Fix ext check in EXT_draw_buffers_indexed negative tests" into nougat-cts-dev
Daniel Xie [Tue, 13 Sep 2016 16:43:11 +0000 (16:43 +0000)]
Merge "CP: Fix geometry shader support check" into nougat-cts-dev
Daniel Xie [Tue, 13 Sep 2016 16:43:06 +0000 (16:43 +0000)]
Merge "CP: Check and enable VK_KHR_sampler_mirror_clamp_to_edge" into nougat-cts-dev
Daniel Xie [Tue, 13 Sep 2016 16:42:54 +0000 (16:42 +0000)]
Merge changes I85492417,I93389a2c into nougat-cts-dev
* changes:
CP: Limit VkInstance count in object_management.multithreaded*
CP: Do not test VkDevice, VkInstance creation OOM paths in WSI tests
Pyry Haulos [Thu, 8 Sep 2016 22:27:06 +0000 (15:27 -0700)]
Check for shader type support in negative shader directive tests
Bug:
31312897
Change-Id: I40c2b47b0a6e6d0478cb6b63051ff474aae02f1f
Pyry Haulos [Thu, 8 Sep 2016 22:18:16 +0000 (15:18 -0700)]
Check for shader type support in negative precise tests
Bug:
31312494
Change-Id: I8bc1536e238efcef23fea0164731b9851ddf35b9
Kalle Raita [Sat, 13 Aug 2016 00:09:41 +0000 (17:09 -0700)]
CP: Split load/store image tests into sub-groups
Split load/store negative tests into sub-groups per texture type to
reduce the time required per single test case.
Bug:
30785598
Bug:
31343756
(cherry picked from commit
90763f1f7dfed4880fee7305b8f5789137951afc)
Change-Id: I4bd10004ee90f42ae616e654e425aa11df0cdd8d
Pyry Haulos [Thu, 1 Sep 2016 20:40:14 +0000 (13:40 -0700)]
Check for EXT_color_buffer_float in read_pixels_fbo_format_mismatch
Bug:
31244127
Change-Id: I0e25c5042aa832679f2da06b8401456058762750
Pyry Haulos [Wed, 31 Aug 2016 16:45:39 +0000 (09:45 -0700)]
DO NOT MERGE Allow old behavior in image format property tests
Vulkan API specification has changed in respect to what sample counts
vkGetPhysicalDeviceImageFormatProperties() should report. The change has
been made after initial Android Nougat drivers were frozen. For NYC CTS
we need to tolerate the old behavior and issue compatibility warning
instead.
Bug:
30739954
Change-Id: I212a8bc2b948760d3bf457e1c0e1199e0a8cc31e
Nanley Chery [Tue, 19 Apr 2016 19:13:58 +0000 (12:13 -0700)]
CP: Fix required sample count tests
According to,
Table 11.1. Image and image view parameter compatibility requirements
Only non-cube images with type VK_IMAGE_TYPE_2D can have sample counts
greater than one.
In addition, the following properties should hold for multisampled
images: the formats must support being used in a color or depth/stencil
attachment, or in a storage image when shaderStorageImageMultisample is
supported. Images not fitting the above descriptions may have a sample
count of one or zero (in error conditions). Align tests for sample
counts to this behavior.
Bug:
30739954
(cherry picked from commit
dc4227d5ec87080f5dad810b5e33ae3b3fa014a4)
Change-Id: I476b6586c351bddfb9f0b92cd7ec45fb9cc297c8
Unsuk Jung [Tue, 30 Aug 2016 20:44:03 +0000 (20:44 +0000)]
Merge "Don't require supported binary formats in negative tests." into marshmallow-cts-dev
am:
3aa5258338
Change-Id: I8cb9cdf8f87cacaa2cb4c586953f884a3cfee098
Haixia Shi [Tue, 30 Aug 2016 20:44:02 +0000 (20:44 +0000)]
Don't require supported binary formats in negative tests.
am:
7cd59a4409
Change-Id: If08ea28912fe10ddc570841c8862d0918b1c224d
Unsuk Jung [Tue, 30 Aug 2016 20:39:55 +0000 (20:39 +0000)]
Merge "Don't require supported binary formats in negative tests." into marshmallow-cts-dev
Kalle Raita [Thu, 25 Aug 2016 21:52:52 +0000 (14:52 -0700)]
CP: Use empty region when buffer age is not supported
In the dEQP-EGL.functional.partial_update.* tests the damage region
needs to set to empty instead of implicit full screen for frame that do
not draw when the EGL_KHR_buffer_age is not supported. Otherwise the
contents become invalid and cannot be verified.
Bug:
30735352
Test: Local run of dEQP-EGL.functional.partial_update.* on various Nexus devices
(cherry picked from commit
bc29554d7eed0bd996506fc1e074f264305c8143)
Change-Id: I17fc4fdca1da176e13e78b1ec2bebc8cb6d266b0
Pyry Haulos [Mon, 8 Aug 2016 18:08:47 +0000 (11:08 -0700)]
CP: Check and enable VK_KHR_sampler_mirror_clamp_to_edge
Several tests in dEQP-VK.pipeline group assume that
VK_SAMPLER_ADDRESS_MODE_MIRROR_CLAMP_TO_EDGE mode is supported. Spec
actually requires that VK_KHR_sampler_mirror_clamp_to_edge is enabled if
that mode is used.
This change enables VK_KHR_sampler_mirror_clamp_to_edge in default
device, if supported, and adds relevant checks in the test code.
In the future we'll want to clean this up by giving test cases more
control about the device configuration they need.
Fixes #339
Bug:
29996742
(cherry picked from commit
ed8561a07cdb1bb7730eccf9025b738838392464)
Change-Id: I12312f0e48d20724c082f31c4ab03792dbb44826
Pyry Haulos [Mon, 8 Aug 2016 19:34:35 +0000 (12:34 -0700)]
CP: Fix geometry shader support check
Bug:
30738286
(cherry picked from commit
83470b2dcbff396773f94d7a1448c12362ad3380)
Change-Id: Ife07ebd21f115114441ac8dc794a8d9e12c8acae