Commits · 1da7ac0ff43540aca0effdec3d24550997ec2b2d · Chen Yisong / angle

14 Oct, 2020 11 commits

GLX: Workaround to disable GLX_OML_sync_control support. · 1da7ac0f

authored Oct 13, 2020

Some users on Linux/Wayland are reporting issues rendering Chrome
unusable when ANGLE exposes GLX_OML_sync_control. This CL adds a new
feature to ANGLE to disable this extension, but for now we leave the
feature off by default, so users can test it via command line flags.

Bug: chromium:1137851
Change-Id: I804ea935fe441c4794b93ee62ca335e78b9e453a
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2468216Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jonah Ryan-Davis <jonahr@google.com>

1da7ac0f

Reimplement WindowSurfaceVkAndroid::getCurrentWindowSize() · 82817c9a

authored Oct 12, 2020

Use vkGetPhysicalDeviceSurfaceCapabilitiesKHR() instead of
ANativeWindow_getWidth() & ANativeWindow_getHeight(). The former
gives the proper size of the window, where the latter gives the size
of the device's display (which is the default size of a window).

Bug: b/168327817
Change-Id: I524116b7b07bf2ec7637fe2874806c5eb8318171
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2464920Reviewed-by: Tim Van Patten <timvp@google.com>
Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com>
Commit-Queue: Ian Elliott <ianelliott@google.com>

82817c9a

Tests: Use script type in isolate map. · 4c9f892f

authored Oct 10, 2020

This makes trigger.py use the same code paths as the bots
when they trigger standalone ANGLE tests.

Fixes launching tests via trigger.py on Linux because of
using the 'windowed_test_launcher' method that used xvfb.

Bug: angleproject:5114
Change-Id: I073631270b5b69aac8f3f0f6378bd71f951ce8e6
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2464425
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

4c9f892f

Roll Chromium from a6b7134c40b1 to 20d60a8527d7 (447 revisions) · c7dae0ea

authored Oct 14, 2020

https://chromium.googlesource.com/chromium/src.git/+log/a6b7134c40b1..20d60a8527d7

If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/chromium-angle-autoroll
Please CC ynovikov@google.com on the revert to ensure that a human
is aware of the problem.

To report a problem with the AutoRoller itself, please file a bug:
https://bugs.chromium.org/p/skia/issues/entry?template=Autoroller+Bug

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+doc/master/autoroll/README.md

Changed dependencies
* build: https://chromium.googlesource.com/chromium/src/build.git/+log/28f65f7116..4af5c075ce
* testing: https://chromium.googlesource.com/chromium/src/testing/+log/225e4dc4d6..a8c1307999
No update to Clang.

Bug: angleproject:5124
Tbr: ynovikov@google.com
Change-Id: I82d32eb5b80daa8d26491b4399289e00283c3f65
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2470543Reviewed-by: angle-autoroll <angle-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: angle-autoroll <angle-autoroll@skia-public.iam.gserviceaccount.com>

c7dae0ea

Vulkan: Fix vector size issue with clearWithCommand · 39b5e771

authored Oct 09, 2020

gl::DrawBuffersVector was used to hold the clear values, but that didn't
have enough space for depth/stencil clear values if MAX draw buffers
where used and cleared.

The added test in this change exposes the vkCmdClearAttachment Qualcomm
bug (previously presumed to affect color clears only) with depth/stencil
buffers, so the workaround is expanded to avoid vkCmdClearAttachment
entirely.

Bug: b/159808300
Change-Id: I27c58d9b534bce0bdd27cc53fc64e139f1363c1a
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2455166
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com>

39b5e771

Vulkan: Regression test for content defined of AHB images · bd0d3f9c

authored Oct 13, 2020

Bug: b/167275320
Bug: angleproject:4836
Change-Id: Ic5da615931f13e1ce34f486f163d4f28c8af9287
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2468458
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>

bd0d3f9c

Vulkan: Don't break the render pass on scissored clears · df8f71d1

authored Oct 09, 2020

clearImmediatelyWithRenderPassOp is removed and the draw path is used
for the scissor.  That path was added to avoid creating a large number
of graphics pipelines due to the scissor state.  This is now done by
using dynamic state for scissor in the draw path for clear.

Running the following dEQP tests without and with dynamic state for
scissor:

    dEQP-GLES3.functional.fragment_ops.depth_stencil.stencil_ops.*

the number of graphics pipelines is reduced from 95392 to 16.

Bug: angleproject:4617
Bug: angleproject:4836
Change-Id: Ib373d8cd23ca2b67e6fd26aa2a1103f281f7e473
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2463985
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

df8f71d1

Vulkan: Regression test for content defined of external images · 62bba4ae

authored Oct 13, 2020

Bug: b/167275320
Bug: angleproject:4836
Change-Id: Iedbe1dfc8562c4c28477bf6a5a298620fcd48a67
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2468457
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

62bba4ae

Test Runner: Accept GTEST shard environment vars. · 9580f3b4

authored Oct 10, 2020

This makes interoperation with Chromium tests easier. For instance when
triggering tests with scripts when using sharding. And also using the
run/ wrapper scripts instead of run_gtest_angle_test.

Bug: angleproject:5158
Change-Id: Ia5b1f94c02a081caf576615175791fff4e0d3fb3
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2464030Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

9580f3b4

Vulkan: Remove Arm from perFrameWindowSizeQuery · 5dd82cac

authored Oct 13, 2020

The perFrameWindowSizeQuery feature was enabled for Arm, based on a
faulty dEQP test, that says that it is resizing a window on Android,
but doesn't actually.  The Arm team has verified that this feature is
no longer needed for Arm/Mali, and so it is being removed.

Bug: b/170733470
Bug: angleproject:5150
Change-Id: I8b9d7cb45983b6bbf8d52d37bf5b237f2c3e4fe4
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2468176Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com>
Reviewed-by: Ian Elliott <ianelliott@google.com>
Commit-Queue: Ian Elliott <ianelliott@google.com>

5dd82cac

trigger.py: Remove GTest shard env when running one shard. · 0059e75c

authored Oct 12, 2020

Bug: angleproject:5158
Change-Id: I3c29adf36f6f31f23bf662c2d7d6c4d05d8e9028
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2466781
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

0059e75c

13 Oct, 2020 11 commits

Suppress multisample "max samples" tests on Win/NV/GL. · 707af6e5

authored Oct 12, 2020

These produce low memory warnings on Quadro when run multi-process.

Bug: angleproject:5162
Change-Id: Ibef3a8ea9119d923a538181387c30562d503bcb6
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2466782
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

707af6e5

Test Runner: List tests explicitly without GTest. · ad80b184

authored Oct 13, 2020

This dramatically speeds up Android shards that use batches of
hudreds of tests. Because ANGLE already has the test list, it no
longer needs to call GTest with an enourmous gtest_filter.

Bug: angleproject:5164
Change-Id: I28fee2f36c50006f2a35a4dcd90f44f8ebe4f78c
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2468464
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

ad80b184

Vulkan: Fix command reordering on release to external · 426fa735

authored Oct 13, 2020

When releasing buffer/image to external, the necessary barrier was
recorded on the "outside render pass command buffer".  However, if the
resource was used in the current render pass, that render pass should
have closed before this, otherwise the barrier was reordered before it
which is incorrect.

Bug: chromium:1136367
Bug: chromium:1135792
Bug: angleproject:5002
Change-Id: I237d5e6bb46189a4ce61c2f4602e356955cc87a9
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2468456
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>

426fa735

Add PowerVR Vendor ID & Vulkan GL Line Emulation Exclusion · 298c2768

authored Oct 12, 2020

Adds the PowerVR VendorID into the accepted vendor list, and prevent
basic GL line emulation from being ran on PowerVR hardware as it's
not required on those platforms, yet causes CTS failures when enabled.

Fixes a variety of line failure tests across OGLES3/3.1 test suites
seen on PowerVR hardware.

Bug: angleproject:5145
Change-Id: I7383f1c09d026b8d33e6fbda66418c5f8b7f90bd
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2462094Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>

298c2768

Update IsARM64 · d5296d99

authored Oct 13, 2020

Updates IsARM64 to return true if __aarch64__ is defined, as the current
check only using _M_ARM64 only evaluates to true on Windows platforms.

Bug: chromium:1132295
Change-Id: I4853012e077217162ee3439c17fe35dcc6157e71
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2468787
Commit-Queue: Brian Sheedy <bsheedy@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

d5296d99

Broaden Clear test failure to all OSX. · 9e6bfaff

authored Oct 13, 2020

ClearTest.ChangeFramebufferAttachmentFromRGBAtoRGB/ES2_OpenGL

Fails on AMD and all vendors.

Bug: angleproject:5165
Change-Id: Id98c14a695cd9b3dd0c5f57d77932908833f9406
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2468357Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Ian Elliott <ianelliott@google.com>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

9e6bfaff

Capture/Replay: Embed ANGLE version in replay. · 580f40a6

authored Oct 04, 2020

This is embedded as a #define. It can be used for book-keeping and can
also be used to control integration when there's incompatibility.

Bug: angleproject:5135
Change-Id: Ie0bb8ff9509c5f973860cb9e6ae48714a27c7398
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2449163
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com>
Reviewed-by: Cody Northrop <cnorthrop@google.com>

580f40a6

Roll glslang from 2067d1a93e6e to f4f1d8a352ca (2 revisions) · bc2dd313

authored Oct 13, 2020

https://chromium.googlesource.com/external/github.com/KhronosGroup/glslang/+log/2067d1a93e6e..f4f1d8a352ca

If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/glslang-angle-autoroll
Please CC ynovikov@chromium.org on the revert to ensure that a human
is aware of the problem.

To report a problem with the AutoRoller itself, please file a bug:
https://bugs.chromium.org/p/skia/issues/entry?template=Autoroller+Bug

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+doc/master/autoroll/README.md

Bug: None
Tbr: ynovikov@chromium.org
Change-Id: I35866acde56f6fa340cd5bf68612952c4ef2c678
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2466846Reviewed-by: angle-autoroll <angle-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: angle-autoroll <angle-autoroll@skia-public.iam.gserviceaccount.com>

bc2dd313

Roll Chromium from efbcef254191 to a6b7134c40b1 (408 revisions) · 06b956fe

authored Oct 13, 2020

https://chromium.googlesource.com/chromium/src.git/+log/efbcef254191..a6b7134c40b1

If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/chromium-angle-autoroll
Please CC ynovikov@chromium.org on the revert to ensure that a human
is aware of the problem.

To report a problem with the AutoRoller itself, please file a bug:
https://bugs.chromium.org/p/skia/issues/entry?template=Autoroller+Bug

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+doc/master/autoroll/README.md

Changed dependencies
* build: https://chromium.googlesource.com/chromium/src/build.git/+log/8dae5d8e04..28f65f7116
* testing: https://chromium.googlesource.com/chromium/src/testing/+log/b4cc6eb191..225e4dc4d6
* tools/clang: https://chromium.googlesource.com/chromium/src/tools/clang.git/+log/7e5979b1dd..d4827bfe10
* tools/mb: https://chromium.googlesource.com/chromium/src/tools/mb/+log/d3154dd7d9..1b83be28f4
No update to Clang.

Bug: angleproject:5161
Tbr: ynovikov@chromium.org
Change-Id: Ie9a048daf8b35cd6af7a33d0c462c9235618619e
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2467656Reviewed-by: angle-autoroll <angle-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: angle-autoroll <angle-autoroll@skia-public.iam.gserviceaccount.com>

06b956fe

Vulkan: Use StoreOp_None_QCOM for read only depth stencil buffer · df6b7298

authored Oct 12, 2020

For read only depth stencil buffers, there is no need to store depth or
stencil value. But we can not use DontCare for storeOp because vulkan
core spec says DontCare indicates data is undefined after this.
VK_QCOM_render_pass_store_ops extension introduces a new store op that
will leave data defined but skip the store. This CL utilize this if the
extension is available.

Bug: angleproject:5055
Change-Id: I104f3d01eb342a2d0cc900f342430e901bde1bff
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2462604
Commit-Queue: Charlie Lao <cclao@google.com>
Reviewed-by: Tim Van Patten <timvp@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

df6b7298

Vulkan: Expand PackedAttachmentOpsDesc from uint16_t to two uint16_ts · 4c6903a4

authored Oct 12, 2020

This is preparation for crrev.com/c/2462604. We run out of bits in the
struct PackedAttachmentOpsDesc. This expands it from 16 bits to 32 bits
(two uint16_t), gives us a lot more room for current and future needs.

Bug: angleproject:5055
Change-Id: I492cbd18c3b71b76b5b48652094409e76c5721b1
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2465615Reviewed-by: Tim Van Patten <timvp@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Charlie Lao <cclao@google.com>

4c6903a4

12 Oct, 2020 8 commits

Vulkan: TexSubImage3D using PBO should use correct layerCount · 4c6b4794

authored Oct 12, 2020

The layerCount for 3D texture and 2DArray should be 1 if it does not
have layers.

Bug: b/170657065
Change-Id: I974ddda7c0eaa92da8033f15c682820e3d4eb32a
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2466616Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Tim Van Patten <timvp@google.com>
Commit-Queue: Charlie Lao <cclao@google.com>

4c6b4794

Test Runner: Pass batch ID to child processes. · 336202d7

authored Oct 09, 2020

This will let both the ANGLETest and dEQP test runners handle child
processes correctly. The existing ANGLETest runner will reuse displays,
and the dEQP test runner will write to a batch-unique filename.

This solves the problem of the multi-process --bot-mode writing to the
same file from multiple children simultaneously.

Long-term it will be good to include the dEQP QPA files into the test
artifacts.

Also disables flushing after every log write when running as a child
process. This hugely improves performance on machines with no SSD.

Test: https://chromium-swarm.appspot.com/task?id=4f2b87c4f8234910
Bug: angleproject:5157
Change-Id: I226d24adf55e0f8b98e5a8e7947862e6906b4c40
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2464424Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com>
Reviewed-by: Ian Elliott <ianelliott@google.com>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

336202d7

FrameCapture: Move compressed texture cache to shared location · 3e8305f5

authored Oct 06, 2020

In support of tracking compressed texture data across multiple
threads and contexts, move our cache to FrameCaptureShared.
Make this accessible only via egl::ShareGroup, which will
ensure unique texture IDs.

Test: Asphalt 8 MEC
Bug: b/150484427
Bug: angleproject:4048
Change-Id: I4e75b323d7aedc0333ceffba305439e54c4d6ac1
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2455266
Commit-Queue: Cody Northrop <cnorthrop@google.com>
Reviewed-by: Courtney Goeltzenleuchter <courtneygo@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

3e8305f5

Reland "Remove redundant BlendStateArray tracking" · 0f55e3e5

authored Sep 11, 2020

Rebase the original CL and update capture/replay state
serialization

Original CL reviewed-on:
https://chromium-review.googlesource.com/c/angle/angle/+/2169093

Bug: angleproject:4394
Bug: chromium:1085996
Bug: chromium:1086582
Bug: chromium:1086585
Bug: chromium:1086586
Change-Id: If0fba8b6e185540aed57d22eaf0ff79ec142209e
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2404442
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

0f55e3e5

Roll SwiftShader from 51b03d58fb17 to e72c6099f946 (5 revisions) · 30d067cf

authored Oct 12, 2020

https://swiftshader.googlesource.com/SwiftShader.git/+log/51b03d58fb17..e72c6099f946

2020-10-11 natsu@google.com Revert "Update virlg_hw path to reflect Minigbm"
2020-10-10 swiftshader.regress@gmail.com Regres: Update test lists @ 139f5c35
2020-10-10 natsu@google.com Update virlg_hw path to reflect Minigbm
2020-10-09 capn@google.com Fix depth bias calculations
2020-10-09 capn@google.com Move interpolant clamping out of interpolate()

If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/swiftshader-angle-autoroll
Please CC ynovikov@chromium.org on the revert to ensure that a human
is aware of the problem.

To report a problem with the AutoRoller itself, please file a bug:
https://bugs.chromium.org/p/skia/issues/entry?template=Autoroller+Bug

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+doc/master/autoroll/README.md

Bug: None
Tbr: ynovikov@chromium.org
Change-Id: I5ea04e45273dfd26395635c66a10095f0cb410dc
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2465864Reviewed-by: angle-autoroll <angle-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: angle-autoroll <angle-autoroll@skia-public.iam.gserviceaccount.com>

30d067cf

Roll Vulkan-ValidationLayers from 0046c31f2f96 to 9d4b40ae645e (10 revisions) · 46d4522b

authored Oct 12, 2020

https://chromium.googlesource.com/external/github.com/KhronosGroup/Vulkan-ValidationLayers.git/+log/0046c31f2f96..9d4b40ae645e

2020-10-10 locke@lunarg.com tests: UnnormalizedCoordinates Sampler
2020-10-10 locke@lunarg.com layers: Reduce loop
2020-10-10 locke@lunarg.com layers: VUID-02704
2020-10-10 locke@lunarg.com layers: VUID-02703
2020-10-10 locke@lunarg.com layers: VUID-vkCmdDraw-None-02702
2020-10-10 locke@lunarg.com layers: VUID-vkCmdDraw-None-02699
2020-10-09 tony@lunarg.com layers: Use debug printf create infos
2020-10-09 rgarcia@igalia.com layers: Fix depth/stencil resolve false positives
2020-10-09 mark@lunarg.com dispatch: Add secondary commandbuffer tracking
2020-10-09 mark@lunarg.com dispatch: Move CB APIs to manually generated

If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/vulkan-validation-layers-angle-autoroll
Please CC ynovikov@chromium.org on the revert to ensure that a human
is aware of the problem.

To report a problem with the AutoRoller itself, please file a bug:
https://bugs.chromium.org/p/skia/issues/entry?template=Autoroller+Bug

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+doc/master/autoroll/README.md

Bug: None
Tbr: ynovikov@chromium.org
Change-Id: Iffa368652d3e1a4a85c0bfd148267cd54620384b
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2465863Reviewed-by: angle-autoroll <angle-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: angle-autoroll <angle-autoroll@skia-public.iam.gserviceaccount.com>

46d4522b

Roll Chromium from 77bfcdcde874 to efbcef254191 (569 revisions) · dec35c56

authored Oct 12, 2020

https://chromium.googlesource.com/chromium/src.git/+log/77bfcdcde874..efbcef254191

If this roll has caused a breakage, revert this CL and stop the roller
using the controls here:
https://autoroll.skia.org/r/chromium-angle-autoroll
Please CC ynovikov@chromium.org on the revert to ensure that a human
is aware of the problem.

To report a problem with the AutoRoller itself, please file a bug:
https://bugs.chromium.org/p/skia/issues/entry?template=Autoroller+Bug

Documentation for the AutoRoller is here:
https://skia.googlesource.com/buildbot/+doc/master/autoroll/README.md

Changed dependencies
* build: https://chromium.googlesource.com/chromium/src/build.git/+log/f12b0ac8bf..8dae5d8e04
* testing: https://chromium.googlesource.com/chromium/src/testing/+log/26fd14f32d..b4cc6eb191
* third_party/abseil-cpp: https://chromium.googlesource.com/chromium/src/third_party/abseil-cpp/+log/c89bad5e6d..66b870c3b1
* tools/mb: https://chromium.googlesource.com/chromium/src/tools/mb/+log/ed409570c0..d3154dd7d9
* tools/skia_goldctl/linux: ehm_yIO4CdwsvaJZJPQ2xBXAFv3KXEgrDpTY3AaK7EYC..sEWxSos77EKQLWziVoCV6Sa52qijVB0fz9uN1EGoUd0C
* tools/skia_goldctl/mac: wIHfWG-DQOw0ISbpkWZrm2f2GoKa3H1Z4I2w8MxhlVIC..rp50J-ChpY5Dyvjwl71mK28FOzgevSspS8YE1qxgkrAC
* tools/skia_goldctl/win: ixP2POw1dVLvzG-YPiGAJdFrTLwpr--nwyQr_S5FoRkC..6eX8c167sU_avc61blSZpNkSEa5HwWNYfWCk-88QTNsC
No update to Clang.

Bug: None
Tbr: ynovikov@chromium.org
Change-Id: I9bd5f5fbdd62000f87fbb0cde9b54f7a8ff8440b
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2465384Reviewed-by: angle-autoroll <angle-autoroll@skia-public.iam.gserviceaccount.com>
Commit-Queue: angle-autoroll <angle-autoroll@skia-public.iam.gserviceaccount.com>

dec35c56

Reland: "4 Vulkan content defined CLs." · 68bd685a

authored Oct 10, 2020

Reland "Vulkan: Avoid content restore by detecting no-op stencil"
This relands commit 243d0f89.

Reland "Vulkan: Restore at the end of RP if write-after-invalidate"
This relands commit e5d52ac3.

Reland "Vulkan: Invalidate/restore depth/stencil separately."
This relands commit 61fa0878.

Reland "Vulkan: Move content-defined tracking to ImageHelper"
This relands commit 2392e6b3.

Reason for revert: Caused crashes in Fuchsia x64 and on ARM.
Reland fixes content defined for external images.

Original CL message:

Content-defined tracking was done in render targets prior to this
change.  This had multiple drawbacks:

- When a framebuffer attachment is changed (including the first time
  it's set), it's unknown whether the contents of the attachment is
  defined.
- Invalidate takes effect at the end of render pass, at which point the
  render target objects may be gone.  Attachment ImageHelpers are
  however correctly tracked.

This change moves content-defined tracking to the ImageHelper itself,
and tracks it per subresource.  ImageHelper::onWrite() now receives the
subresource that is being written, and marks it as having defined
content.

A future optimization can make use of this change to
ImageHelper::onWrite to track "dirty" subresources.  This can lead to
the removal of unnecessary barriers when same-kind writes are done on
different subresources of the image.  See http://anglebug.com/3347#c15

Bug: b/167275320
Bug: angleproject:4836
Bug: angleproject:5159
Change-Id: If5c1ae7152657fd7c94db7d55bea4fb9ddf835ba
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2464825Reviewed-by: Shahbaz Youssefi <syoussefi@chromium.org>
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>

68bd685a

10 Oct, 2020 10 commits

Sync extended dirty bit on texture copies. · 3345e5fd

authored Oct 10, 2020

This works around an issue in the GL back-end where the
extended dirty bit wasn't synced in internal blit draw calls.

This bug is preventing progress with the standalone test runner.
This temporarily solution can be put in place until we have
a more complete solution for internal draw state sync.

Bug: angleproject:5160
Change-Id: Id8ae5efa66b249b008a1535b60493c880f5a06a8
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2464824
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

3345e5fd

Revert 4 Vulkan content defined CLs. · d6b1c17b

authored Oct 10, 2020

Revert "Vulkan: Avoid content restore by detecting no-op stencil"
This reverts commit 243d0f89.

Revert "Vulkan: Restore at the end of RP if write-after-invalidate"
This reverts commit e5d52ac3.

Revert "Vulkan: Invalidate/restore depth/stencil separately."
This reverts commit 61fa0878.

Revert "Vulkan: Move content-defined tracking to ImageHelper"
This reverts commit 2392e6b3.

Causes crashes in Fuchsia x64 and on ARM.

Original CL message:

Content-defined tracking was done in render targets prior to this
change.  This had multiple drawbacks:

- When a framebuffer attachment is changed (including the first time
  it's set), it's unknown whether the contents of the attachment is
  defined.
- Invalidate takes effect at the end of render pass, at which point the
  render target objects may be gone.  Attachment ImageHelpers are
  however correctly tracked.

This change moves content-defined tracking to the ImageHelper itself,
and tracks it per subresource.  ImageHelper::onWrite() now receives the
subresource that is being written, and marks it as having defined
content.

A future optimization can make use of this change to
ImageHelper::onWrite to track "dirty" subresources.  This can lead to
the removal of unnecessary barriers when same-kind writes are done on
different subresources of the image.  See http://anglebug.com/3347#c15

Bug: b/167275320
Bug: angleproject:4836
Bug: angleproject:5159
Change-Id: I93d9dfe973caa7ce70aefa46b5b7d04a8637efb3
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2464822Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

d6b1c17b

Suppress more tests on Intel/GL/Mac. · f3f80036

authored Oct 10, 2020

These tests crash/fail when reusing EGL Displays.

Bug: angleproject:4992
Change-Id: I72dd61543cb055130836428751167e7528cf55af
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2464722Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

f3f80036

Vulkan: Avoid content restore by detecting no-op stencil · 243d0f89

authored Oct 08, 2020

Previously, as long as stencil was enabled, it was considered that it is
also being modified. This caused stencil invalidate to be undone in a
number of situations, such as:

- glEnable(GL_STENCIL_TEST); // with func/ops default
- glDrawArrays();
- glInvalidateFramebuffer([GL_STENCIL_ATTACHMENT]);
- glClear(GL_DEPTH_BUFFER_BIT);
- Close render pass

In the above scenario, invalidation of stencil was undone at the end of
render pass.

In this change, the following cases are considered read-only stencil:

- Func = GL_NEVER, stencilFail = GL_KEEP
- Func = GL_ALWAYS, stencilPassDepth* = GL_KEEP
- stencilFail = GL_KEEP, stencilPassDepth* = GL_KEEP

Note that while the above scenario is fixed for no-op stencil, a similar
issue persists if stencil was not no-op.  The reason stencil invalidate
is undone in that case is due to the fact that it's assumed any command
after the invalidate call will be a draw call that outputs to stencil,
but that is not the case with the glClear call in this example.

Bug: angleproject:4836
Change-Id: Ie2ea2d52b7c8ee2394f5456773a7ef434e2b2b16
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2461465
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Ian Elliott <ianelliott@google.com>

243d0f89

Vulkan: Clear depth by shader if depthClamp not supported · 6c1c3bd9

authored Oct 09, 2020

This will avoid breaking the render pass when clearing depth through
clearWithDraw if the depthClamp Vulkan feature is not present.

Bug: angleproject:4836
Change-Id: I845fd5074dd95f6896da89f9e119ebc5000a5688
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2462719
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>

6c1c3bd9

Vulkan: Restore at the end of RP if write-after-invalidate · e5d52ac3

authored Oct 08, 2020

If a depth/stencil attachment is invalidated, but subsequently drawn to
in the same render pass, undo the invalidate when the render pass is
closed.  Adapted from
https://chromium-review.googlesource.com/c/angle/angle/+/2386478.

Bug: b/167275320
Bug: angleproject:4836
Change-Id: I17a35bfd692ddc403ceaa6ec44b5c4f16ff9eed6
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2461464
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Ian Elliott <ianelliott@google.com>

e5d52ac3

Vulkan: Use depthClamp to clear depth where available · f8070feb

authored Oct 09, 2020

This will avoid breaking render pass if clearing depth in clearWithDraw.

Bug: angleproject:4836
Change-Id: I50242d1115efc91059923143f6ae5fd25fb3d36f
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2462717
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

f8070feb

Vulkan: Boilerplace for vkCmdSetScissor · dd4b6445

authored Oct 09, 2020

Bug: angleproject:4836
Change-Id: I4fa5355fc3e7fcf3ecd091d299c5c0c8d3a74732
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2463984
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Charlie Lao <cclao@google.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

dd4b6445

Vulkan: Invalidate/restore depth/stencil separately. · 61fa0878

authored Oct 08, 2020

Depth/stencil content defined is already tracked separately in the
ImageHelper.  This change exposes this tracking from RenderTargetVk.

Bug: b/167275320
Bug: angleproject:4836
Change-Id: Ie6520e7a4ab557eb233c60c6ab0d4a8f8f098bf6
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2462039
Commit-Queue: Shahbaz Youssefi <syoussefi@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Ian Elliott <ianelliott@google.com>

61fa0878

Metal: Disable fence sync extension on Nvidia · 87f3f4b5

authored Oct 09, 2020

The following tests are flaky on Mac FYI Retina Release (NVIDIA), so the
Metal fence sync extension is being disabled for Nvidia.

Bug: chromium:1136673
Change-Id: I7e99f558950c6e867318e27f27bb526b4c209fcc
Reviewed-on: https://chromium-review.googlesource.com/c/angle/angle/+/2463516
Commit-Queue: Tim Van Patten <timvp@google.com>
Reviewed-by: Kai Ninomiya <kainino@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

87f3f4b5