platform/upstream/VK-GL-CTS.git
6 years agoeglGetFrameTimestamps: Don't require reads done to be monotonic.
Brian Anderson [Mon, 12 Mar 2018 17:54:36 +0000 (17:54 +0000)]
eglGetFrameTimestamps: Don't require reads done to be monotonic.
am: e8634c9254

Change-Id: I68606ee22f516db51b60db205cd7ebb28cc563f6

6 years agoeglGetFrameTimestamps: Don't require reads done to be monotonic.
Brian Anderson [Fri, 9 Mar 2018 18:25:08 +0000 (10:25 -0800)]
eglGetFrameTimestamps: Don't require reads done to be monotonic.

Reads done is usually monotonic, but only under the assumption
that every buffer of the surface is being consumed by the same
unit. It is possible for a surface to switch between being an
overlay and being composited by the GPU. It's also possible a
driver will maintain a shadow copy of a frame that will further
affect when reads from a buffer might finish.

Remove the overly strict check.

Bug: 7033485174184639
Test: --deqp-case=dEQP-EGL*get_frame_timestamps*
Change-Id: I8e242a9fd649720b4d8059797c71163329b7b92c

6 years agoDO NOT MERGE: Fix dEQP-VK.api.copy_and_blit.image_to_image_depth value range. am...
Bas Nieuwenhuizen [Fri, 23 Feb 2018 22:09:07 +0000 (22:09 +0000)]
DO NOT MERGE: Fix dEQP-VK.api.copy_and_blit.image_to_image_depth value range. am: f03effb5a0  -s ours
am: 7c8b312fbe  -s ours

Change-Id: I7302bfbf8d9c38c8a01e604253a9cca1bba8b76c

6 years agoDO NOT MERGE: Fix dEQP-VK.api.copy_and_blit.image_to_image_depth value range.
Bas Nieuwenhuizen [Fri, 23 Feb 2018 22:03:07 +0000 (22:03 +0000)]
DO NOT MERGE: Fix dEQP-VK.api.copy_and_blit.image_to_image_depth value range.
am: f03effb5a0  -s ours

Change-Id: I26792ef93ead582aabb8bd5beb386588846c7489

6 years agoDO NOT MERGE: Fix dEQP-VK.api.copy_and_blit.image_to_image_depth value range.
Bas Nieuwenhuizen [Thu, 15 Feb 2018 10:33:22 +0000 (11:33 +0100)]
DO NOT MERGE: Fix dEQP-VK.api.copy_and_blit.image_to_image_depth value range.

Per spec:

"Because depth or stencil aspect buffer to image copies may require format conversions on some
implementations, they are not supported on queues that do not support graphics. When copying to
a depth aspect, the data in buffer memory must be in the the range [0,1] or undefined results occur."

Initializing it to integer values falls afoul of this rule
quickly. As the image is 256 by 256, x/255.0 should always fall in the range [0,1].

Not entirely sure this would initialize combined depth stencil correctly,
but this version of CTS does not have a test using those formats with
this code.

This has been fixed upstream for a while with

"9909b9f3 Fix combined depth/stencil clear colors."

but this is not a straight backport as these tests saw some significant rework between
this CTS version and the fix.

Bug: b/73449389

Change-Id: Iaf0ce14ed4338d437f1b20033be5d0f8c0cbd699

6 years agoMerge "Merge remote-tracking branch 'origin/vulkan-cts-1.0.2' into HEAD" into oreo...
Chris Forbes [Tue, 13 Feb 2018 21:20:03 +0000 (21:20 +0000)]
Merge "Merge remote-tracking branch 'origin/vulkan-cts-1.0.2' into HEAD" into oreo-cts-dev
am: 6d6ebc1b4b

Change-Id: I71f98787bf29843efbe66343187c3a67c7ca39f9

6 years agoMerge "Merge remote-tracking branch 'origin/vulkan-cts-1.0.2' into HEAD" into oreo...
Treehugger Robot [Tue, 13 Feb 2018 21:08:38 +0000 (21:08 +0000)]
Merge "Merge remote-tracking branch 'origin/vulkan-cts-1.0.2' into HEAD" into oreo-cts-dev

6 years agoDO NOT MERGE: Sparse resources fixes am: ae793c389b -s ours
Kantoch [Fri, 9 Feb 2018 19:00:10 +0000 (19:00 +0000)]
DO NOT MERGE: Sparse resources fixes am: ae793c389b  -s ours
am: ce25f76fbb  -s ours

Change-Id: I8fa668d939b6c45df4590f823b029536895b4574

6 years agoDO NOT MERGE: Sparse resources fixes
Kantoch [Fri, 9 Feb 2018 18:54:40 +0000 (18:54 +0000)]
DO NOT MERGE: Sparse resources fixes
am: ae793c389b  -s ours

Change-Id: Ib87ed9843313e5ba76ccae32161ed5719615fa55

6 years agoDO NOT MERGE: Sparse resources fixes
Kantoch [Tue, 19 Jul 2016 11:37:33 +0000 (13:37 +0200)]
DO NOT MERGE: Sparse resources fixes

1. Do not allow to request more queues than available
2. Allow to reuse queue if it supports more than one type of operations

(cherry picked from VK-GL-CTS commit 5f2cf2f7139c2f603a02fe3c284921dd9152dee3)

Bug: b/72734354

Change-Id: Ia063df59c2220a082ea5a73a0fc968cfb97fda2c

6 years agoHandle unsupported no_config contexts
Mark Adams [Wed, 7 Feb 2018 17:31:36 +0000 (17:31 +0000)]
Handle unsupported no_config contexts
am: b60ae978ad

Change-Id: Idf51783fac99d14fa5ed28dbffd0db36b80e998b

6 years agoMerge remote-tracking branch 'origin/vulkan-cts-1.0.2' into HEAD
Chris Forbes [Tue, 6 Feb 2018 01:12:53 +0000 (17:12 -0800)]
Merge remote-tracking branch 'origin/vulkan-cts-1.0.2' into HEAD

Bug: b/72972234

Change-Id: I537a4f3207c1f88d68a51a23f21a5fed0975b037

6 years agoPass in proper pWaitDstStageMask in vktWsiSharedPresentableImageTests
Vamsidhar Gaddam [Wed, 31 Jan 2018 10:23:48 +0000 (11:23 +0100)]
Pass in proper pWaitDstStageMask in vktWsiSharedPresentableImageTests

One of the submitInfos for a queue submit had
a semaphore passed in without passing in any stage mask. This is a
spec violation.

Affects:

dEQP-VK.wsi.android.shared_presentable_image.*

Components: Vulkan

VK-GL-CTS issue: 974

Change-Id: Iee2ea838e971a3610f44b8c1f765869a8a21fedb

6 years agoSpec relaxation for YCbCr Implicit chroma reconstruction
Panagiotis Apostolou [Wed, 17 Jan 2018 09:01:58 +0000 (10:01 +0100)]
Spec relaxation for YCbCr Implicit chroma reconstruction

Incorporates spec relaxation in the YCbCr tests which describes that when
doing implicit reconstruction with nearest chroma filtering, an
implementation may behave as if both xChromaOffset and yChromaOffset were
VK_CHROMA_LOCATION_MIDPOINT.

Tests affected:
dEQP-VK.ycbcr.conversion.*

Components: Vulkan

VK-GL-CTS issue: 979

Change-Id: I10e83d2f5845c3e74673aa9e42fc4f0950f9e6cd

6 years agoHandle unsupported no_config contexts
Mark Adams [Fri, 12 Jan 2018 21:21:00 +0000 (16:21 -0500)]
Handle unsupported no_config contexts

EGL_KHR_no_config_context doesn't mandate support for ES 1.1 or
for GL versions prior to 3.0. Update the test to handle
optional support, and also update the GL version to 3.0 in order
to test a case which is expected to work.

Bug: b/72041864
Components: AOSP

Test: dEQP-EGL.functional.create_context.no_config

Change-Id: Ibe7856000c67326fe23cd024435290fba536d127

6 years agoSuppress new warnings from gcc-6
Chris Forbes [Sat, 20 Jan 2018 00:43:46 +0000 (00:43 +0000)]
Suppress new warnings from gcc-6
am: 9a58d041fe

Change-Id: I176dd24b363db336f4f42f8d8c2c5aebb2632dcb

6 years agoSuppress new warnings from gcc-6
Chris Forbes [Fri, 19 Jan 2018 18:06:53 +0000 (10:06 -0800)]
Suppress new warnings from gcc-6

check_build_sanity currently fails with the system gcc on rodete. Let's
just suppress this warning for now. Intend to back out the suppression
and fix the warnings in next release.

Bug: b/72219443
Test: build
Change-Id: Ic842a3de3d247d53c49117562b51e8eece3752f9

6 years agoCheck GS support before testing adjacent primitives
Yanjun Zhang [Wed, 10 Jan 2018 00:12:47 +0000 (16:12 -0800)]
Check GS support before testing adjacent primitives

Affects: dEQP-VK.draw.basic_draw.draw.*

Components: Vulkan

VK-GL-CTS issue: 926

Change-Id: I02569b1944eed52f641966f9206d44983d1ff165

6 years agoFix buffer usage flags for index and indirect buffers
Chris Forbes [Wed, 10 Jan 2018 02:13:15 +0000 (18:13 -0800)]
Fix buffer usage flags for index and indirect buffers

Previously, createAndUploadBuffer would provide
VK_BUFFER_USAGE_VERTEX_BUFFER_BIT for all buffers. This is no good for
index buffers and indirect buffers.

VK-GL-CTS: 930

Change-Id: I774beef7c7226f4fbd5670591eb66475300ee208
Components: Vulkan
Affects: dEQP-VK.draw.instanced.*

6 years agoSet maxDescriptorSet* limit table based on shader stages
Yanjun Zhang [Thu, 11 Jan 2018 19:44:54 +0000 (11:44 -0800)]
Set maxDescriptorSet* limit table based on shader stages

Re-submit the patch based on WG review and feedback.

Affects: dEQP-VK.api.info.device.properties

Components: Vulkan

VK-GL-CTS issue: 923

Change-Id: Ie627b52cfae099d87f5b6183118d42bcccd313ca

6 years agoMerge "Update owners file" into oreo-mr1-cts-dev
Treehugger Robot [Wed, 17 Jan 2018 22:30:54 +0000 (22:30 +0000)]
Merge "Update owners file" into oreo-mr1-cts-dev

6 years agoUpdate owners file
Mika Isojärvi [Thu, 28 Sep 2017 17:25:24 +0000 (10:25 -0700)]
Update owners file

Bug: 72113298
Test: build/make/tools/checkowners.py -c -v OWNERS
Change-Id: I117459d91f39055aaa5f0bc6abfbda32f31cdab2
(cherry picked from commit 412471299c0fac5cb5af37fcdeadbf39a4a36744)

6 years agoModify compute shader grid setup.
Anders Frostad Pedersen [Wed, 17 Jan 2018 11:05:45 +0000 (12:05 +0100)]
Modify compute shader grid setup.

Modify how the grid is set up for compute shader draw indirect test,
in order to avoid verfication failures due to difference in
floating-point computations on the CPU vs the GPU.

The existing code assumes infinite FP precision.

This change will guarantee the next value of pos(X) to be equal to
the first value of pos(X+1), avoiding gaps/discontinuities.

Affects:
dEQP-GLES31.functional.draw_indirect.compute_interop.large.drawarrays_combined_grid_1000x1000_drawcount_1
dEQP-GLES31.functional.draw_indirect.compute_interop.large.drawarrays_combined_grid_1000x1000_drawcount_5000
dEQP-GLES31.functional.draw_indirect.compute_interop.large.drawarrays_combined_grid_1000x1000_drawcount_8
dEQP-GLES31.functional.draw_indirect.compute_interop.large.drawarrays_separate_grid_1000x1000_drawcount_1
dEQP-GLES31.functional.draw_indirect.compute_interop.large.drawarrays_separate_grid_1000x1000_drawcount_5000
dEQP-GLES31.functional.draw_indirect.compute_interop.large.drawarrays_separate_grid_1000x1000_drawcount_8

Bug: 71515808
Change-Id: I21fe8ec9ea48b15f6171265df802bc47be775f56

6 years agoLimit maximum allocation size in memory mapping tests am: adff582b50 am: 71ef361f3e
Pyry Haulos [Sat, 13 Jan 2018 00:04:17 +0000 (00:04 +0000)]
Limit maximum allocation size in memory mapping tests am: adff582b50 am: 71ef361f3e
am: fe861b61a5

Change-Id: Id7e2ade9ceaad34bee00b2b36bfde50ca2492dd5

6 years agoImprove memory tracking in vulkan memory mapping tests am: 3a8ec35d52 am: b1f38f97cf...
Mika Isojrvi [Sat, 13 Jan 2018 00:04:00 +0000 (00:04 +0000)]
Improve memory tracking in vulkan memory mapping tests am: 3a8ec35d52 am: b1f38f97cf  -s ours
am: 346da766ee

Change-Id: Iece2a8d3c40b646e982008d82f40b16dfdbf94ca

6 years agoOptimize random memory mapping tests reference memory am: 7e69d67a2c am: b2dcf86f17...
Mika Isojrvi [Sat, 13 Jan 2018 00:03:38 +0000 (00:03 +0000)]
Optimize random memory mapping tests reference memory am: 7e69d67a2c am: b2dcf86f17  -s ours
am: 1883d6ef48

Change-Id: I9f6246177cc3651b84370fedd7f48fd9dddf1f94

6 years agoLimit maximum allocation size in memory mapping tests am: adff582b50
Pyry Haulos [Fri, 12 Jan 2018 23:57:24 +0000 (23:57 +0000)]
Limit maximum allocation size in memory mapping tests am: adff582b50
am: 71ef361f3e

Change-Id: I138e6025b0cfb77ffd8197b796e32b06011cbc8a

6 years agoImprove memory tracking in vulkan memory mapping tests am: 3a8ec35d52
Mika Isojrvi [Fri, 12 Jan 2018 23:57:09 +0000 (23:57 +0000)]
Improve memory tracking in vulkan memory mapping tests am: 3a8ec35d52
am: b1f38f97cf  -s ours

Change-Id: I8fc045601456dd0b8d0026c4e70db43cf4998886

6 years agoOptimize random memory mapping tests reference memory am: 7e69d67a2c
Mika Isojrvi [Fri, 12 Jan 2018 23:56:55 +0000 (23:56 +0000)]
Optimize random memory mapping tests reference memory am: 7e69d67a2c
am: b2dcf86f17  -s ours

Change-Id: Iaa15b3997481353a88aef19cdde4348ae2cc3f75

6 years agoLimit maximum allocation size in memory mapping tests
Pyry Haulos [Fri, 12 Jan 2018 22:57:15 +0000 (22:57 +0000)]
Limit maximum allocation size in memory mapping tests
am: adff582b50

Change-Id: Iff465c2be4d7c27702036f8f060a458c55830732

6 years agoImprove memory tracking in vulkan memory mapping tests
Mika Isojrvi [Fri, 12 Jan 2018 22:57:12 +0000 (22:57 +0000)]
Improve memory tracking in vulkan memory mapping tests
am: 3a8ec35d52

Change-Id: I54613d0c408d80a65cc175e86495d0ffd4c650b5

6 years agoOptimize random memory mapping tests reference memory
Mika Isojrvi [Fri, 12 Jan 2018 22:57:10 +0000 (22:57 +0000)]
Optimize random memory mapping tests reference memory
am: 7e69d67a2c

Change-Id: I2c91c3720102d1c4c949bf19a2b6af51c59581f1

6 years agoMultiple draw buffers in blend_equation_advanced
Mark Adams [Fri, 12 Jan 2018 21:44:17 +0000 (21:44 +0000)]
Multiple draw buffers in blend_equation_advanced
am: 123de4f307

Change-Id: I1ce06acaa4f5f725097bcc248b7dbdf048bac61e

6 years agoMultiple draw buffers in blend_equation_advanced
Mark Adams [Fri, 20 Oct 2017 13:49:08 +0000 (09:49 -0400)]
Multiple draw buffers in blend_equation_advanced

GL_NVX_blend_equation_advanced_multi_draw_buffers adds support for
multiple draw buffers with advanced blend equations. This change
modifies the below test to check for this extension and to expect
GL_NO_ERROR if it is present.

Components: AOSP

VK-GL-CTS issue: 770
Bug: b/71901344

Affects:
dEQP-GLES31.functional.debug.negative_coverage.get_error.advanced_blend.attachment_advanced_equation
(cherry picked from commit 2cb35a235c1a7bd3c63b296c37e186eec34e08e2)

Change-Id: I865fa30e527b5899197bade465602bafc92a54a1

6 years agoCheck sampleRateShading for sample_qualifier_distinct_values.* tests
Yanjun Zhang [Wed, 10 Jan 2018 00:26:50 +0000 (16:26 -0800)]
Check sampleRateShading for sample_qualifier_distinct_values.* tests

Affects: dEQP-VK.pipeline.*.sample_qualifier_distinct_values.*

Components: Vulkan

VK-GL-CTS issue: 925

Change-Id: Ife1166e8ed88699c67ee6acde79857f34c013398

6 years agoUse unique indices for each binding in out_of_pool_memory test
Chris Forbes [Tue, 9 Jan 2018 23:26:34 +0000 (15:26 -0800)]
Use unique indices for each binding in out_of_pool_memory test

Previously, all the layout bindings would use binding=0, which is
invalid. Give them their own binding indices.

VK-GL-CTS: 929

Change-Id: I28e62db6c257ed0843dd9b77b2d8cb01d02e3eba
Components: Vulkan
Affects: dEQP-VK.api.descriptor_pool.out_of_pool_memory

6 years agoMerge "Reduce GPU workload of pipeline_barrier tests" into oreo-cts-dev
Jeff Leger [Fri, 12 Jan 2018 01:31:43 +0000 (01:31 +0000)]
Merge "Reduce GPU workload of pipeline_barrier tests" into oreo-cts-dev
am: 4c26442e3d

Change-Id: I32a3a35837be16886acebd28e30ccdadffa73811

6 years agoMerge "Reduce GPU workload of pipeline_barrier tests" into oreo-cts-dev
Treehugger Robot [Fri, 12 Jan 2018 01:10:31 +0000 (01:10 +0000)]
Merge "Reduce GPU workload of pipeline_barrier tests" into oreo-cts-dev

6 years agoglshared: Add watchdog touches to avoid timeouts on slow hardware
Chris Forbes [Thu, 11 Jan 2018 22:22:49 +0000 (22:22 +0000)]
glshared: Add watchdog touches to avoid timeouts on slow hardware
am: 0d9250e9eb

Change-Id: If675c974a2389787e53d2d618509ca81a48990d9

6 years agoRound up vertex buffer size to stride multiple to fix OOB attribute reads
Noah Fredriks [Fri, 3 Nov 2017 18:20:06 +0000 (14:20 -0400)]
Round up vertex buffer size to stride multiple to fix OOB attribute reads

Partial stride multiple vertex buffer elements are now considered OOB
to accomodate implementations that fetch stride-sized elements

Affects: dEQP-VK.texture.*

Components: Vulkan

VK-GL-CTS issue: 812

Change-Id: I72a2143146b4ea0d893d70d41b7ca2c84c34542a
(cherry picked from commit 50c0544ec8c2e99158bf8a375eacfc9baf364a91)

6 years agoglshared: Add watchdog touches to avoid timeouts on slow hardware
Chris Forbes [Thu, 21 Dec 2017 18:22:58 +0000 (10:22 -0800)]
glshared: Add watchdog touches to avoid timeouts on slow hardware

Based on diffs posted to deqp-external-requests

Bug: b/71753331
Bug: b/67854695

Change-Id: Id25e334814caf65f72b76e66d3b07fccba5ad355
Component: AOSP

6 years agoRun 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
Bug: 70716357
Test: run on google devices

Affects: dEQP-GLES31.functional.draw_base_vertex.*

Change-Id: If8a4860a956c771ec5ea3341e2227f437b12a286
(cherry picked from commit 079ca561114713d31e05c4124b8daeef764cf9b8)
Merged-In: 079ca561114713d31e05c4124b8daeef764cf9b8

6 years agoReport tests using Draw*BaseVertex as NotSupported
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.*

Bug: 70716326
Test: run on google devices

Change-Id: I87c918d5f66a0a5b547f8e625c662203a7f85048
(cherry picked from commit 1b1ea05b4e8d6e9dbb07187b0af6bc12c315fb4e)
Merged-In: 1b1ea05b4e8d6e9dbb07187b0af6bc12c315fb4e

6 years agoRemove dEQP-GLES31.*vertex_array.*_invalid_map tests
David Sodman [Thu, 18 May 2017 16:30:17 +0000 (09:30 -0700)]
Remove dEQP-GLES31.*vertex_array.*_invalid_map tests

Tests made assumptions about GL error returned from glDrawElements*
functions when passed a mapped buffer object, but, that error is not
supported by the specification.

Affects: dEQP-GLES31.*vertex_array.*
Components: AOSP
Bug: 37942569
Bug: 70716290
Test: run on google devices

Change-Id: Ic066dfbe717ff1e7741bec8ee870b09b99362dc7
(cherry picked from commit 428a2256159883c651a409c4e0cbe031953f4c4e)
Merged-In: 428a2256159883c651a409c4e0cbe031953f4c4e

6 years agoMerge "Avoid watchdog timeout in Vulkan builtin precision tests" into oreo-cts-dev
Chris Forbes [Sat, 16 Dec 2017 00:41:49 +0000 (00:41 +0000)]
Merge "Avoid watchdog timeout in Vulkan builtin precision tests" into oreo-cts-dev
am: aa801278e8

Change-Id: I5818c937676ad5c4c79814be1c39c8c264949977

6 years agoMerge "Avoid watchdog timeout in Vulkan builtin precision tests" into oreo-cts-dev
Treehugger Robot [Sat, 16 Dec 2017 00:34:23 +0000 (00:34 +0000)]
Merge "Avoid watchdog timeout in Vulkan builtin precision tests" into oreo-cts-dev

6 years agoFix Invalid VkImageSubresourceRange
Noah Fredriks [Thu, 14 Dec 2017 19:48:48 +0000 (14:48 -0500)]
Fix Invalid VkImageSubresourceRange

Pipeline image barrier was using the arraySize of the wrong image.

Affects: dEQP-VK.api.copy_and_blit.core.resolve_image.*

Components: Vulkan

VK-GL-CTS Issue: 910

Change-Id: I021bdd0ba5eff9e894f21186b3d85bb2f2c99f16

6 years agoModify spirv_assembly tests adding dependency on 16bit_storage extension
Jeff Leger [Sat, 2 Dec 2017 02:16:28 +0000 (21:16 -0500)]
Modify spirv_assembly tests adding dependency on 16bit_storage extension

Several tests were using 16bit_storage, but were not checking the
extension was supported nor declaring the required opCapability.

Affects:
dEQP-VK.spirv_assembly.instruction.compute.sconvert.*
dEQP-VK.spirv_assembly.instruction.compute.uconvert.*

Components: Vulkan

VK-GL-CTS issue: 875

Change-Id: I07fda13c69cdef09967d4aab5067edff07c47fa1

6 years agoAvoid creating vkInstances for not supported cases
Noah Fredriks [Wed, 29 Nov 2017 23:27:52 +0000 (18:27 -0500)]
Avoid creating vkInstances for not supported cases

Defer vkInstance creation for cross instance tests until not supported
checks complete to shorten the runtime of these 10k total test cases.

Affects: dEQP-VK.synchronization.cross_instance.*

Components: Vulkan

VK-GL-CTS Issue: 885

Change-Id: I481de90d8c0363c23a287aaf457ac101c327c558

6 years agoFix mediump zero division in matrix inverse tests
Igor Ostrowski [Wed, 15 Nov 2017 17:33:51 +0000 (18:33 +0100)]
Fix mediump zero division in matrix inverse tests

To avoid zero division all input matrix determinants should be same signed.

Affects:
dEQP-VK.glsl.matrix.inverse.*

Components: Vulkan
VK-GL-CTS issue: 881

Change-Id: Ib702f81c4eb7f25ba9081db6ead67da02f484ca9

6 years agoDon't assume rendering complete before buffer latch
Mark Adams [Wed, 6 Dec 2017 18:28:29 +0000 (18:28 +0000)]
Don't assume rendering complete before buffer latch
am: 9e0214cf61

Change-Id: Iee3ccf2afef375ac45e60133537a141d685d607f

6 years agoLimit maximum allocation size in memory mapping tests
Pyry Haulos [Fri, 19 Aug 2016 12:20:38 +0000 (13:20 +0100)]
Limit maximum allocation size in memory mapping tests

Fixes also ReferenceMemory::m_flushed memory usage counting when
estimating reference memory usage (size should be divided by atomSize).

Fixes #468
Bug: b/69677943

Change-Id: Iad754e211e8070b09254d2451e03418fc2be541b

6 years agoImprove memory tracking in vulkan memory mapping tests
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

6 years agoOptimize random memory mapping tests reference memory
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

6 years agoDon't assume rendering complete before buffer latch
Mark Adams [Mon, 4 Dec 2017 22:26:54 +0000 (17:26 -0500)]
Don't assume rendering complete before buffer latch

It's valid for the buffer to latch before rendering is completed, and
this is causing intermittent failures.

Bug: b/69981429

Test: --deqp-case=dEQP-EGL*get_frame_timestamps*

Change-Id: Ia9f1d0e4ffa7f3e16a46fc4566157a651d1b0601

6 years agoAvoid watchdog timeout in Vulkan builtin precision tests
Chris Forbes [Mon, 16 Oct 2017 19:42:35 +0000 (12:42 -0700)]
Avoid watchdog timeout in Vulkan builtin precision tests

Calculating the reference intervals can take a substantial amount of
time. On some very slow targets, this can cause us to trip the watchdog
in the verification loop.

Same approach and limit as in modules/glshared/glsBuiltinPrecisionTests.cpp

Affected tests:
dEQP-VK.glsl.builtin.precision.*

Component: Vulkan

Bug: 67854695
Test: run on google devices
Change-Id: I492b656e7e2a9aa724bb3785bdf1416a4c22ec04

6 years agoReduce GPU workload of pipeline_barrier tests
Jeff Leger [Wed, 22 Nov 2017 17:12:58 +0000 (12:12 -0500)]
Reduce GPU workload of pipeline_barrier tests

Tests with large buffer sizes will now clamp the
GPU workload to avoid triggering execution timeout
on implementations.

Affects: dEQP-VK-memory.pipeline_barrier.*

Components: Vulkan

VK_GL_CTS issue: 813

Change-Id: I479f6ad04ef2ec6f9d87b65a39a2bc13bcbcc6a2

6 years agoReduce GPU workload of pipeline_barrier tests
Jeff Leger [Wed, 22 Nov 2017 17:12:58 +0000 (12:12 -0500)]
Reduce GPU workload of pipeline_barrier tests

Tests with large buffer sizes will now clamp the
GPU workload to avoid triggering execution timeout
on implementations.

b/68360750

Affects: dEQP-VK-memory.pipeline_barrier.*

Components: Vulkan

VK_GL_CTS issue: 813

Change-Id: If8772973fd49cac0b8cc5fa66c33af4de38f2f68

6 years agoAdded constraints for random access flags
Jari Komppa [Tue, 14 Nov 2017 12:21:06 +0000 (14:21 +0200)]
Added constraints for random access flags

The pipeline barrier tests randomize access flags and stages, sometimes
resulting in access flags being enabled for stages that do not support
said accesses. This fix removes the illegal access flags while leaving
the random flag generation intact.

Affects:

dEQP-VK.memory.pipeline_barrier.*

Components: Vulkan

VK-GL-CTS issue: 814

Change-Id: I135ea8a9e41b5d8428ef77fe6ecbdf2380d80c7d

6 years agoDo not use depthClampEnable unconditionally.
Jorg Wagner [Mon, 6 Nov 2017 13:56:49 +0000 (14:56 +0100)]
Do not use depthClampEnable unconditionally.

Depth clamping is not needed by this test, so
do not rely on the optional feature when creating
the pipeline object.

Affects:
dEQP-VK.api.smoke.*

Components: Vulkan

VK-GL-CTS issue: 817

Change-Id: I6794f7f5d17d7610ff43dcb4080a6a30e12ed95b

6 years agoLower correlation threshold in flush-finish tests again am: 6455e6f987 am: 2e18b48b04...
Gurchetan Singh [Tue, 7 Nov 2017 20:45:44 +0000 (20:45 +0000)]
Lower correlation threshold in flush-finish tests again am: 6455e6f987 am: 2e18b48b04 am: b31b60abe6 am: da172c6afd am: ed080cc354
am: 4d5a25b142

Change-Id: Iece967058049e50e63f7cc4235bf6468eddffc97

6 years agoLower correlation threshold in flush-finish tests again am: 6455e6f987 am: 2e18b48b04...
Gurchetan Singh [Tue, 7 Nov 2017 20:42:12 +0000 (20:42 +0000)]
Lower correlation threshold in flush-finish tests again am: 6455e6f987 am: 2e18b48b04 am: b31b60abe6 am: da172c6afd
am: ed080cc354

Change-Id: I3c63098007f8634c51ecd9fbf8844b5c5f2f2c26

6 years agoLower correlation threshold in flush-finish tests again am: 6455e6f987 am: 2e18b48b04...
Gurchetan Singh [Tue, 7 Nov 2017 20:38:43 +0000 (20:38 +0000)]
Lower correlation threshold in flush-finish tests again am: 6455e6f987 am: 2e18b48b04 am: b31b60abe6
am: da172c6afd

Change-Id: Iecdc967997849e2377fb480454f89a75e072c298

6 years agoLower correlation threshold in flush-finish tests again am: 6455e6f987 am: 2e18b48b04
Gurchetan Singh [Tue, 7 Nov 2017 20:36:05 +0000 (20:36 +0000)]
Lower correlation threshold in flush-finish tests again am: 6455e6f987 am: 2e18b48b04
am: b31b60abe6

Change-Id: I0c83a1932bb47c1b295cf2f0f19447b7c7950557

6 years agoLower correlation threshold in flush-finish tests again am: 6455e6f987 am: 84373d7748...
Gurchetan Singh [Tue, 7 Nov 2017 20:34:09 +0000 (20:34 +0000)]
Lower correlation threshold in flush-finish tests again am: 6455e6f987 am: 84373d7748 am: 7776db1369
am: 83cec0e56e

Change-Id: Ie4217eb2aa68077a3fc53820c8de2931eac705c8

6 years agoLower correlation threshold in flush-finish tests again am: 6455e6f987
Gurchetan Singh [Tue, 7 Nov 2017 20:34:04 +0000 (20:34 +0000)]
Lower correlation threshold in flush-finish tests again am: 6455e6f987
am: 2e18b48b04

Change-Id: Id0dd6ef88d6a9f1a53a0b72e7fc6912ade51878d

6 years agoLower correlation threshold in flush-finish tests again am: 6455e6f987 am: 84373d7748
Gurchetan Singh [Tue, 7 Nov 2017 20:31:07 +0000 (20:31 +0000)]
Lower correlation threshold in flush-finish tests again am: 6455e6f987 am: 84373d7748
am: 7776db1369

Change-Id: I8ab737ed4a43886a1ecc7656ef84b265c1b2f7bc

6 years agoLower correlation threshold in flush-finish tests again
Gurchetan Singh [Tue, 7 Nov 2017 20:29:56 +0000 (20:29 +0000)]
Lower correlation threshold in flush-finish tests again
am: 6455e6f987

Change-Id: I06395cdc2b3b19e7426f35e80981d0547309ede4

6 years agoLower correlation threshold in flush-finish tests again am: 6455e6f987
Gurchetan Singh [Tue, 7 Nov 2017 20:29:09 +0000 (20:29 +0000)]
Lower correlation threshold in flush-finish tests again am: 6455e6f987
am: 84373d7748

Change-Id: I0bda19c76be8c905ce711c089e5576fd7bfee09e

6 years agoLower correlation threshold in flush-finish tests again
Gurchetan Singh [Tue, 7 Nov 2017 20:24:36 +0000 (20:24 +0000)]
Lower correlation threshold in flush-finish tests again
am: 6455e6f987

Change-Id: I11a62d5d0af3612ed1cdcbedee1ba53d5b0e57ae

6 years agoMerge changes I5f7e56e3,I7f88e1da into oreo-cts-dev am: 67694b5635 am: aa425927d0
Brian Anderson [Tue, 7 Nov 2017 20:02:39 +0000 (20:02 +0000)]
Merge changes I5f7e56e3,I7f88e1da into oreo-cts-dev am: 67694b5635 am: aa425927d0
am: efb1ae2ab9

Change-Id: I5623468a2c31d7e265e89f5d860e61fe911e225b

6 years agoMerge changes I5f7e56e3,I7f88e1da into oreo-cts-dev am: 67694b5635
Brian Anderson [Tue, 7 Nov 2017 19:59:37 +0000 (19:59 +0000)]
Merge changes I5f7e56e3,I7f88e1da into oreo-cts-dev am: 67694b5635
am: aa425927d0

Change-Id: I4a353896ae1d58cfa206c37b4081010f7577899b

6 years agoMerge changes I5f7e56e3,I7f88e1da into oreo-cts-dev
Brian Anderson [Tue, 7 Nov 2017 19:57:39 +0000 (19:57 +0000)]
Merge changes I5f7e56e3,I7f88e1da into oreo-cts-dev
am: 67694b5635

Change-Id: I202b9be2b529cc3ba80a96f66a31b42252b47234

6 years agoMerge changes I5f7e56e3,I7f88e1da into oreo-cts-dev
Treehugger Robot [Tue, 7 Nov 2017 19:49:35 +0000 (19:49 +0000)]
Merge changes I5f7e56e3,I7f88e1da into oreo-cts-dev

* changes:
  eglGetFrameTimestamps: Consider timestamps of 0 as invalid.
  eglGetFrameTimestamps: Allow reads done to equal rendering complete.

6 years agoExclude incremental_present scale down from Android CTS am: da253a4fe0 am: 3d03b52e1f
Chris Forbes [Tue, 7 Nov 2017 03:07:01 +0000 (03:07 +0000)]
Exclude incremental_present scale down from Android CTS am: da253a4fe0 am: 3d03b52e1f
am: b2185cdcbb

Change-Id: I9047deda4c47d482d7e29c7c866a0cfa8321d9f1

6 years agoExclude shared_presentable_image scale down from Android CTS am: 4e9d89ac4a am: 8192b...
Chris Forbes [Tue, 7 Nov 2017 03:06:37 +0000 (03:06 +0000)]
Exclude shared_presentable_image scale down from Android CTS am: 4e9d89ac4a am: 8192b7059e
am: b3e8cd74fe  -s ours

Change-Id: I0429832d24cdd31462dd229b821b0a29e1c90dbe

6 years agoExclude incremental_present scale down from Android CTS am: da253a4fe0
Chris Forbes [Tue, 7 Nov 2017 02:57:37 +0000 (02:57 +0000)]
Exclude incremental_present scale down from Android CTS am: da253a4fe0
am: 3d03b52e1f

Change-Id: I55883fde5f84a8e332b4fea8ae5a8859b52ef9cb

6 years agoExclude incremental_present scale down from Android CTS
Chris Forbes [Tue, 7 Nov 2017 02:55:34 +0000 (02:55 +0000)]
Exclude incremental_present scale down from Android CTS
am: da253a4fe0

Change-Id: I9097d2fdf2783464974888010bb89b654b7e0e8b

6 years agoExclude shared_presentable_image scale down from Android CTS am: 4e9d89ac4a
Chris Forbes [Tue, 7 Nov 2017 02:40:06 +0000 (02:40 +0000)]
Exclude shared_presentable_image scale down from Android CTS am: 4e9d89ac4a
am: 8192b7059e

Change-Id: I04fc2b48d6830403add863a62e9cbfc65fd75899

6 years agoExclude shared_presentable_image scale down from Android CTS
Chris Forbes [Tue, 7 Nov 2017 02:37:04 +0000 (02:37 +0000)]
Exclude shared_presentable_image scale down from Android CTS
am: 4e9d89ac4a

Change-Id: I5a4c71c270694fa27cd8686b7baa7e31f3d39ca1

6 years agoeglGetFrameTimestamps: Consider timestamps of 0 as invalid.
Brian Anderson [Thu, 19 Oct 2017 21:59:28 +0000 (14:59 -0700)]
eglGetFrameTimestamps: Consider timestamps of 0 as invalid.

There's a bug in the kernel that doesn't merge fences properly,
resulting in "reads done" timestamps of 0 when the fences being
merged have already signaled.

This is a temporary relaxation of the test requirements until
the kernel fix is required.

Bug: 68811985
Test: --deqp-case=dEQP-EGL*get_frame_timestamps*
Change-Id: I5f7e56e3a0c9808237a1ebc3787cf788feef1a79

6 years agoeglGetFrameTimestamps: Allow reads done to equal rendering complete.
Brian Anderson [Tue, 17 Oct 2017 19:17:40 +0000 (12:17 -0700)]
eglGetFrameTimestamps: Allow reads done to equal rendering complete.

If there are no reads performed by the compositor, it is possible
for reads done timestamp to be the same as the rendering complete
timestamp.

Also add and use check_lt and check_le, to print out more detailed
debug info in the results.

Bug: 68811985
Test: --deqp-case=dEQP-EGL*get_frame_timestamps*
Change-Id: I7f88e1da98be97101e8b0d6cd130b75bda8d88ca

6 years agoKHR_swapchain: Trigger the fence off vkQueueSubmit
Chris Forbes [Mon, 23 Oct 2017 22:13:38 +0000 (15:13 -0700)]
KHR_swapchain: Trigger the fence off vkQueueSubmit

Another instance of 438ea0888e8ccae.

VK-GL-CTS issue: 756

Component: Vulkan

Affects:
dEQP-VK.wsi.*.swapchain.*

Change-Id: I87da6800e0c8402c2c50ca88eb3360b31c3a515d

6 years agoGOOGLE_display_timing: Trigger the fence off vkQueueSubmit
Chris Forbes [Mon, 23 Oct 2017 22:09:56 +0000 (15:09 -0700)]
GOOGLE_display_timing: Trigger the fence off vkQueueSubmit

Another instance of 438ea0888e8ccae.

VK-GL-CTS issue: 756

Component: Vulkan

Affects:
dEQP-VK.wsi.*.display_timing.*

Change-Id: If94644402579ba1f46347d155313f6ef52d125a0

6 years agoVK_KHR_variable_pointers: Don't use OpPtrAccessChain on a bare SSBO
Jason Ekstrand [Thu, 19 Oct 2017 16:07:52 +0000 (09:07 -0700)]
VK_KHR_variable_pointers: Don't use OpPtrAccessChain on a bare SSBO

You can't use OpPtrAccessChain on the direct result of an OpVariable
because it's not the element of any array.  Instead, we need to do our
OpPtrAccessChain on a pointer to an actual array element and that
pointer needs to be properly decorated with a stride.

Component: Vulkan

VK-GL-CTS issue: https://gitlab.khronos.org/Tracker/vk-gl-cts/issues/686

Affects:
dEQP-VK.spirv_assembly.instruction.graphics.variable_pointers.multi_buffer_*

Change-Id: I9c45fb1320099e404672ee169a6a8e5c41eb13f7

6 years agoVK_KHR_variable_pointers: Delete an unused and invalid variable
Jason Ekstrand [Thu, 19 Oct 2017 15:43:36 +0000 (08:43 -0700)]
VK_KHR_variable_pointers: Delete an unused and invalid variable

The %float_var variable was not being used in any way by the test.  It
was also invalid because it had the StorageBuffer storage class (so it's
in external memory) but also had a constant initializer.  The real
variable being tested in these tests is %f32_ptr_var which has the
Function storage class so a constant initializer is fine.

Component: Vulkan

VK-GL-CTS issue: https://gitlab.khronos.org/Tracker/vk-gl-cts/issues/686

Affects:
dEQP-VK.spirv_assembly.instruction.graphics.variable_pointers.nullptr_graphics.opvariable_initialized_null_*
dEQP-VK.spirv_assembly.instruction.graphics.variable_pointers.nullptr_graphics.opselect_null_or_valid_ptr_*

Change-Id: I9e2f118b1c774a84c0b6900ef0db081cb5256c2f

6 years agoAvoid watchdog timeout in Vulkan builtin precision tests
Chris Forbes [Mon, 16 Oct 2017 19:42:35 +0000 (12:42 -0700)]
Avoid watchdog timeout in Vulkan builtin precision tests

Calculating the reference intervals can take a substantial amount of
time. On some very slow targets, this can cause us to trip the watchdog
in the verification loop.

Same approach and limit as in modules/glshared/glsBuiltinPrecisionTests.cpp

Affected tests:
dEQP-VK.glsl.builtin.precision.*

Component: Vulkan

Bug: 67854695
Test: run on google devices
Change-Id: I492b656e7e2a9aa724bb3785bdf1416a4c22ec04

6 years agoeglGetFrameTimestamps: Consider timestamps of 0 as invalid.
Brian Anderson [Thu, 19 Oct 2017 21:59:28 +0000 (14:59 -0700)]
eglGetFrameTimestamps: Consider timestamps of 0 as invalid.

There's a bug in the kernel that doesn't merge fences properly,
resulting in "reads done" timestamps of 0 when the fences being
merged have already signaled.

This is a temporary relaxation of the test requirements until
the kernel fix is required.

Bug: 66916041
Test: Test: --deqp-case=dEQP-EGL*get_frame_timestamps*
Change-Id: I0b2a65303c425a74a48d6ae8c971c16652d727a3

6 years agoExclude incremental_present scale down from Android CTS
Chris Forbes [Fri, 29 Sep 2017 16:51:49 +0000 (09:51 -0700)]
Exclude incremental_present scale down from Android CTS

Bug: b/67022169
Bug: b/67295411
Test: check_build_sanity
Change-Id: I4be889f0bf64251069e829924faf4e8c113baa86

6 years agoExclude shared_presentable_image scale down from Android CTS
Chris Forbes [Fri, 29 Sep 2017 16:51:49 +0000 (09:51 -0700)]
Exclude shared_presentable_image scale down from Android CTS

Bug: b/66464066
Test: check_build_sanity
Change-Id: If815b53b8e998e720aca4fc1994e0151eef971a5

6 years agoeglGetFrameTimestamps: Allow reads done to equal rendering complete.
Brian Anderson [Tue, 17 Oct 2017 19:17:40 +0000 (12:17 -0700)]
eglGetFrameTimestamps: Allow reads done to equal rendering complete.

If there are no reads performed by the compositor, it is possible
for reads done timestamp to be the same as the rendering complete
timestamp.

Also add and use check_lt and check_le, to print out more detailed
debug info in the results.

Bug: 66916041
Test: --deqp-case=dEQP-EGL*get_frame_timestamps*
Change-Id: Ic77fb3871bfcd2ec2298b20db7247b6782088399

6 years agoLower correlation threshold in flush-finish tests again
Gurchetan Singh [Fri, 6 Oct 2017 16:06:21 +0000 (09:06 -0700)]
Lower correlation threshold in flush-finish tests again

This test expects predictability from glFlush/glFinish timing.
In particular,  the code from FlushFinishCase::analyzeResults()
checks that the result falls within specific timing thresholds.

Notice that in the GLES3 specification, there is no requirement for
timing, just a guarantee that "commands [...] complete in finite time".
This is a problem because it makes the test very flaky, for example
if the CPU/GPU clocks are not ramping up fast enough.

BUG=b:67331095
TEST=Ran dEQP-GLES3.functional.flush_finish.finish on Elm for over 3+ hours
     Observed 1.5% to 2% flakiness without this patch. Observed no flakiness
     with this patch.

(cherry picked from commit 59f4fcde5ac1402ab3b911cf6a9f66b886c10464)
Signed-off-by: Gurchetan Singh <gurchetansingh@google.com>
6 years agoVK_KHR_incremental_present: Trigger the fence off vkQueueSubmit
Jason Ekstrand [Fri, 6 Oct 2017 23:10:26 +0000 (16:10 -0700)]
VK_KHR_incremental_present: Trigger the fence off vkQueueSubmit

Previously, the tests were using AcquireNextImage to trigger the fence
and using it to guard the vkFreeCommandBuffers call.  However, the
acquire fence tells you when the PE is done with an image and when you
can start rendering to it.  This meant that the test was waiting for the
start of the previous frame's rendering not the end before deleting the
previous frame's command buffer.  This meant that the previous frame's
rendering could still be active on the GPU when the command buffer was
freed.  By triggering the fence off of the vkQueueSubmit, we wait for
the end of the previous frame's rendering before freeing its command
buffer.

VK-GL-CTS issue: 756

Component: Vulkan

Affects:
dEQP-VK.wsi.*.incremental_present.*

Change-Id: I87b25eb2a25e4ae48c8c527d5d63c45d89a31a30
(cherry picked from commit db916576a76a17c02c14390fc5d5b3db1488d1d2)

6 years agoDrawUtil: Fix depthBounds check
Igor Ostrowski [Mon, 9 Oct 2017 14:41:31 +0000 (16:41 +0200)]
DrawUtil: Fix depthBounds check

Affects:
dEQP-VK.glsl.builtin_var.fragdepth*

Components: Vulkan

VK-GL-CTS issue: 758

Change-Id: I4da79b236921e64f11566fe386b144497249123b

6 years agoExclude incremental_present scale down from Android CTS
Chris Forbes [Fri, 29 Sep 2017 16:51:49 +0000 (09:51 -0700)]
Exclude incremental_present scale down from Android CTS

Bug: b/65643778
Bug: b/67022169
Bug: b/67295411
Test: check_build_sanity
Change-Id: I4be889f0bf64251069e829924faf4e8c113baa86

6 years agoExclude shared_presentable_image scale down from Android CTS
Chris Forbes [Fri, 29 Sep 2017 16:51:49 +0000 (09:51 -0700)]
Exclude shared_presentable_image scale down from Android CTS

Bug: b/65643778
Bug: b/66464066
Test: check_build_sanity
Change-Id: If815b53b8e998e720aca4fc1994e0151eef971a5

6 years agoDisable some UBO tests which got broken by recent glslang change am: d81d8e2af7 am...
Chris Forbes [Tue, 3 Oct 2017 22:46:18 +0000 (22:46 +0000)]
Disable some UBO tests which got broken by recent glslang change am: d81d8e2af7 am: 188944cfd1
am: 0f95f68482

Change-Id: Ia49724d1cf3fade59f71e39d6f321398e042e5f6

6 years agoDisable some UBO tests which got broken by recent glslang change am: d81d8e2af7
Chris Forbes [Tue, 3 Oct 2017 22:41:02 +0000 (22:41 +0000)]
Disable some UBO tests which got broken by recent glslang change am: d81d8e2af7
am: 188944cfd1

Change-Id: Iab4679d945588de21bf329f114074d3bd0f92765

6 years agoDisable some UBO tests which got broken by recent glslang change
Chris Forbes [Tue, 3 Oct 2017 22:37:51 +0000 (22:37 +0000)]
Disable some UBO tests which got broken by recent glslang change
am: d81d8e2af7

Change-Id: Ic275c1ec2efeaf191a4537ca78a8b0109b8ac9fc

6 years agoDisable some UBO tests which got broken by recent glslang change
Chris Forbes [Tue, 3 Oct 2017 17:17:59 +0000 (10:17 -0700)]
Disable some UBO tests which got broken by recent glslang change

Bug: b/67013571
Change-Id: I54c3f9aba636381a57e398de603b09dad4d14188