We can keep a shared reference to the Pipeline and Descriptor Set
layouts in the Program. This ensures they are not in use when they are
deleted. Note that they are allowed to be deleted as long as no command
buffers are currently recording with them. If the Program is deleted
then there should be no further commands using these layouts.

Bug: angleproject:2462
Change-Id: I75161b3ce1ee8eae33dd6becee79b4262b844cdd
Reviewed-on: https://chromium-review.googlesource.com/1089807
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>

This will be necessary to support more than one PipelineLayout per
instance of ANGLE. Sampler array handling requires different layouts
for different sampler uses.

Bug: angleproject:2462
Change-Id: I1d8b4919eed1a589002ad1898b05186f420061c7
Reviewed-on: https://chromium-review.googlesource.com/1089806
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Luc Ferron <lucferron@chromium.org>

Separated the Vulkan Validation Layers, Loader, Tools and Headers into
separate dirs to map to the new separate repo structure.

Bug: angleproject:2558
Change-Id: I422038f7dd7efe8c5b7a49dc5074de5caf40edfa
Reviewed-on: https://chromium-review.googlesource.com/1071880
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

Bug: angleproject:2658
Change-Id: I32ff9aa27b064d9977eea0b83b18c52c4e42e38d
Reviewed-on: https://chromium-review.googlesource.com/1096054
Commit-Queue: Luc Ferron <lucferron@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

All children of binary and ternary nodes are guaranteed to be
non-null. We don't need to check for their existence when traversing
the tree.

BUG=angleproject:2662
TEST=angle_unittests

Change-Id: I5575058e7213d0c4b4554ca616b4298e535842d6
Reviewed-on: https://chromium-review.googlesource.com/1098670Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>

It's not used anywhere and removing it will make changing traversal
code a bit simpler.

BUG=angleproject:2662
TEST=angle_unittests

Change-Id: I4a430a09ceb538c8b0e5d1bb0a95f3fd7657c276
Reviewed-on: https://chromium-review.googlesource.com/1098671Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>

Bug: angleproject:2637
Change-Id: I2cec129ed1aa0e55e0dbf1d353caa65a51c528a3
Reviewed-on: https://chromium-review.googlesource.com/1096017
Commit-Queue: Luc Ferron <lucferron@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

PointWithoutAttributesCompliance,
PointCoordRegressionTest,
PointSpriteAlternatingDrawTypes
were failing on Nexus 5X Vulkan, because pointSizeRange is [1,1] there,
and the tests assume they can use point sizes larger than 1.

Fix the tests to check whether the point size they want to use is available.

Note: PointWithoutAttributesCompliance and PointSpriteAlternatingDrawTypes
can actually work with size 1, though they need to calculate center
pixel more accurately. However, looks like there is a bug in Nexus 5X
driver, and no pixels are rendered. The tests do pass on Pixel 2.

Bug: angleproject:2447
Change-Id: I511574b1b11f2516c10209b85f86b3e7052e6686
Reviewed-on: https://chromium-review.googlesource.com/1026979
Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

These pack very well in some structs. Support is required so the
templates work as expected.

Also packs the Shader map into 8 bits. This includes a workaround for
an unusual warning generated by MSVC.

Bug: angleproject:2462
Change-Id: I804d1b9370b3951343718385210dec7d37700d73
Reviewed-on: https://chromium-review.googlesource.com/1091135
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>

Makes some methods const or static. Fixes parameter names. explicit
single argument constructors. Other small style fixes.

Bug: angleproject:2462
Change-Id: Ie2bf96603e9eda81d6f2aa788867978de10e9992
Reviewed-on: https://chromium-review.googlesource.com/1089805
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Luc Ferron <lucferron@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>

The gl_PointCoord.y coordinate needs to be flipped
around the X axis in Vulkan to have correct results.

Bug: angleproject:2457

Change-Id: I0d87ad28366623c2be0867c610cc35678a4af43f
Reviewed-on: https://chromium-review.googlesource.com/1090824Reviewed-by: Luc Ferron <lucferron@chromium.org>
Reviewed-by: Olli Etuaho <oetuaho@nvidia.com>
Commit-Queue: Luc Ferron <lucferron@chromium.org>

This patch intends to implement FramebufferTextureEXT on OpenGL
back-ends.
1. Support layered framebuffer attachments.
2. Add new framebuffer completeness rules on layered framebuffer
   attachments.
3. Support FRAMEBUFFER_ATTACHMENT_LAYERED_EXT as a valid <pname>
   parameter of GetFramebufferAttachmentParameteriv.

Note that for an entire level of a cube map:
1. It has no TextureTarget because TEXTURE_CUBE is not a valid
   target for TexImage*D.
2. It corresponds to 6 ImageDescs (that represents its faces) in
   class Texture, so when the cube map is cube complete, we return
   the ImageDesc of its first face, meanwhile we do not allow
   querying ImageDesc if it is not cube complete.

BUG=angleproject:1941
TEST=angle_end2end_tests
     dEQP-GLES31.functional.geometry_shading.query.framebuffer_attachment_layers
     dEQP-GLES31.functional.geometry_shading.query.framebuffer_incomplete_layer_targets
     dEQP-GLES31.functional.geometry_shading.layered.*
     dEQP-GLES31.functional.geometry_shading.instanced.invocation_per_layer_*
     dEQP-GLES31.functional.geometry_shading.instanced.multiple_layers_per_invocation_*

Change-Id: I44393b513ec8f1a682fd1c47d3eaa6f3b3fae877
Reviewed-on: https://chromium-review.googlesource.com/1075811
Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>

Change-Id: I83b39e1f4c429326a98694198411cf64a7361edb
Reviewed-on: https://chromium-review.googlesource.com/1096074Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

Bug: angleproject:2615
Change-Id: I918cc18984b2e5b22b5e13398355a2fd60e4eb00
Reviewed-on: https://chromium-review.googlesource.com/1093564
Commit-Queue: Luc Ferron <lucferron@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

- Update test expectations

BUG=angleproject:2306

Change-Id: I4a0376db1d095d7b14e00a5779631dcf2a6c427c
Reviewed-on: https://chromium-review.googlesource.com/1093795
Commit-Queue: Lingfeng Yang <lfy@google.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

This is a follow-up of this change:
https://chromium-review.googlesource.com/c/angle/angle/+/1092101

Just to fix a warning that prevents building in Visual Studio
successfully.

Bug: angleproject:2306
Change-Id: Ib87d5cb349019fb6b2d7f75b46552146d41787de
Reviewed-on: https://chromium-review.googlesource.com/1095480Reviewed-by: Lingfeng Yang <lfy@google.com>
Commit-Queue: Luc Ferron <lucferron@chromium.org>

ANGLE used to describe the abitily to attach textures and renderbuffers
of a specific format to a framebuffer using a single notion of "renderable".
However, for some formats, only one can be supported, but not the other.

Split TextureCaps::renderable into textureAttachment and renderbuffer.
Also, split InternalFormat::renderSupport into
textureAttachmentSupport and renderbufferSupport.

The only functional change is in a few places which now explicitly check
for texture or renderbuffer attachement support.
Information in format support tables was duplicated for the two capabilities,
so behavior should remain the same. It should be corrected in future CLs.

Note: additional information in those tables may need to be added
in order to properly support GenerateMipmap and TexStorage2DMultisample,
this is beyond the scope of this CL.

Bug: angleproject:2567
Change-Id: I18bce4100525be35709d8bbf4de08ec812aab502
Reviewed-on: https://chromium-review.googlesource.com/1086491
Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

We we're unable to cleanup the unused uniforms if we did not keep
a list of them while parsing them in ProgramLinkResource. Now
that we keep a history of them, we're able to clean them up and
fix a few dEQP tests.

Bug: angleproject:2582
Bug: angleproject:2585
Bug: angleproject:2587
Bug: angleproject:2589
Bug: angleproject:2590
Bug: angleproject:2593

Change-Id: Ic1f9151e356a3d05e83f1031cc7b187b370284e5
Reviewed-on: https://chromium-review.googlesource.com/1085644
Commit-Queue: Luc Ferron <lucferron@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

BUG=angleproject:2306

Change-Id: Ibb168d5c9f7aa96a48c96ffbe96ecead2276975e
Reviewed-on: https://chromium-review.googlesource.com/1092101
Commit-Queue: Lingfeng Yang <lfy@google.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Make it easy to test with different vertex data formats.
Add test cases for all formats.

BUG=angleproject:2405

Change-Id: I147643a0912e97588588e67946ac326ed449cf07
Reviewed-on: https://chromium-review.googlesource.com/1053564
Commit-Queue: Frank Henigman <fjhenigman@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

After [1], a manual dependency on exe_and_shlib_deps is no longer necessary
since it's automatically added.  This CL removes all remaining manual references
to exe_and_shlib_deps.

[1] https://chromium.googlesource.com/chromium/src.git/+/d7ed1f0a9c28c932fddc834ca5de44f28266c7f5

BUG=chromium:845700

Change-Id: I66bcead6586050bf952c4aa1f4d3f726c88d2277
Reviewed-on: https://chromium-review.googlesource.com/1091231Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

To support both virtualized and unvirtualized contexts, the RendererGL object
represents the native context.  Update ContexGL and the various DisplayGLs to
hold RendererGL  with a shared_ptr so that deletion of the renderer happens at
the correct time.

Update RendererGL to take ownership of FunctionsGL.

BUG=angleproject:2464

Change-Id: Id040a8053973d73936c0a7ff0ab5edb1a3f16dc6
Reviewed-on: https://chromium-review.googlesource.com/1085851Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

We now always use custom parsing code for parsing floats in GLSL
shaders. Previously this code was only used in corner cases that
stringstream parsing did not handle according to the GLSL spec.

This is slightly faster in compiler perftests, and results in a
smaller binary as well.

Some new test cases are added to make sure that the custom float
parsing behaves correctly.

BUG=chromium:849245
TEST=angle_unittests, angle_perftests

Change-Id: I2a88ec6a8b427016e34519d72bc98216947a4c64
Reviewed-on: https://chromium-review.googlesource.com/1092697Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>

This test is ported from webgl conformance test
WebglConformance_conformance2_glsl3_short_circuiting_in_loop_condition.
Since HLSL and ESSL have different rules to process short-circuit evaluation,
this patch will use the final flattened shader so that the translated HLSL
is almost same with the current ES shader.

BUG=843369
TEST=angle_end2end_tests

Change-Id: Ibc742f8ac4592a83c4c7dad37a64a5563eb58bd5
Reviewed-on: https://chromium-review.googlesource.com/1090328
Commit-Queue: Jiajia Qin <jiajia.qin@intel.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

This patch is the first one of using ShaderMap in both DynamicHLSL and
StateManager11 to make the code cleaner and more straightforward.

BUG=angleproject:2169

Change-Id: I2b206d0250f5ced071cc1c3632367b16e5e02dfb
Reviewed-on: https://chromium-review.googlesource.com/1089473Reviewed-by: Jiajia Qin <jiajia.qin@intel.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jiawei Shao <jiawei.shao@intel.com>

Change-Id: Id6ea245849696d4c6d7eabc6860c0ac424dd8013
Reviewed-on: https://chromium-review.googlesource.com/1091309Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

- Revise entry point definitions to use packed enums

BUG=angleproject:2306

Change-Id: I06ad95f475d1dbaf07ec24ff2544503c4a44e826
Reviewed-on: https://chromium-review.googlesource.com/1090996Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Lingfeng Yang <lfy@google.com>

Change-Id: I911e3e8460be2205119bb53ebbe8f56677e9e4c6
Reviewed-on: https://chromium-review.googlesource.com/1090760Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Luc Ferron <lucferron@chromium.org>

BUG=angleproject:2629
BUG=angleproject:2349
BUG=angleproject:2327
BUG=angleproject:2328
BUG=angleproject:2630
BUG=angleproject:2631
BUG=angleproject:2349

Change-Id: Ie0752dd07ec8ee151e56c1d4362c47249b7393fc
Reviewed-on: https://chromium-review.googlesource.com/1088876Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

This bug is caught because we haven't implemented
caps.shaderStorageBufferOffsetAlignment on D3D backend. So the default
value is 0. This change will set its value for D3D11 backend.

BUG=angleproject:2625
TEST=angle_deqp_gles31_tests

Change-Id: I03c045c08903e4da41659133e7dbd9c4133186cc
Reviewed-on: https://chromium-review.googlesource.com/1088192Reviewed-by: Yunchao He <yunchao.he@intel.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

BUG=angleproject:2378
TEST=dEQP-GLES31.functional.debug.negative_coverage.get_error.texture.copytexsubimage2d_texture_internalformat

Change-Id: I4ad0ec6636c0312a3ecfeb47df3297e4bf399c35
Reviewed-on: https://chromium-review.googlesource.com/938328Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Yunchao He <yunchao.he@intel.com>

R=cwallez@chroimium.org, fjhenigman@chromium.org, geofflang@chromium.org, jmadill@chromium.org, ynovikov@chromium.org

Change-Id: Iae4cfeb4db83c777d3c9a0a4c115c583113c5b70
Reviewed-on: https://chromium-review.googlesource.com/1087567
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

Quite a substantial update.

Includes ANGLE-related fixes contributed by Google and Intel.

Also includes some test suppressions for ES 3.1 D3D11.

Bug: angleproject:2617
Bug: angleproject:2619
Bug: angleproject:2625
Change-Id: Iea90036996b1a49278443e5d88dbe7ace32de6d4
Reviewed-on: https://chromium-review.googlesource.com/1087179
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

- Update test expectations

BUG=angleproject:2306

Change-Id: Ic5aa5f052bcbe9c5adaf0eb0c6c06df66fd1720c
Reviewed-on: https://chromium-review.googlesource.com/1082978
Commit-Queue: Lingfeng Yang <lfy@google.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

- Add unit tests for the API
- Update test expectations

+ Pass through point size from the vertex array to the shader,
required for new tests to pass.

BUG=angleproject:2306

Change-Id: Ib19436c1f4cb12873adea94f734c821363f9e27d
Reviewed-on: https://chromium-review.googlesource.com/1079993
Commit-Queue: Lingfeng Yang <lfy@google.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Also enable the dEQP test that was blocked on that UNIMPLEMENTED call.

Bug: angleproject:2600
Change-Id: Iae1280d6e4f52c5b8e66c9f7fbaa9b2730c4eb0e
Reviewed-on: https://chromium-review.googlesource.com/1085693Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Luc Ferron <lucferron@chromium.org>

Fixes these dEQP tests:
dEQP-GLES2.functional.state_query.shader.uniform_value_array
dEQP-GLES2.functional.shaders.indexing.uniform_array.vec2_*
dEQP-GLES2.functional.shaders.indexing.uniform_array.vec3_*
dEQP-GLES2.functional.shaders.struct.uniform.nested_struct_array*
dEQP-GLES2.functional.shaders.struct.uniform.loop_nested_struct_array*
dEQP-GLES2.functional.shaders.struct.uniform.dynamic_loop_nested_struct_array*
dEQP-GLES2.functional.uniform_api.value.assigned.by_value.get_uniform.multiple_nested_structs_arrays.*
dEQP-GLES2.functional.uniform_api.value.assigned.basic_array_assign_partial.multiple_basic_array.*
dEQP-GLES2.functional.uniform_api.value.assigned.basic_array_assign_full.multiple_basic_array.*
dEQP-GLES2.functional.uniform_api.value.assigned.by_value.get_uniform.multiple_basic_array.vertex
dEQP-GLES2.functional.uniform_api.value.assigned.by_pointer.render.multiple_basic
dEQP-GLES2.functional.uniform_api.value.assigned.by_pointer.get_uniform.basic_array_first_elem_without_brackets
dEQP-GLES2.functional.uniform_api.value.assigned.by_pointer.get_uniform.multiple_basic.vertex
dEQP-GLES2.functional.uniform_api.value.assigned.by_pointer.get_uniform.multiple_basic.fragment
dEQP-GLES2.functional.uniform_api.value.assigned.by_pointer.get_uniform.multiple_basic.both
dEQP-GLES2.functional.uniform_api.value.assigned.by_pointer.get_uniform.multiple_basic_array.*
dEQP-GLES2.functional.uniform_api.value.assigned.by_pointer.get_uniform.multiple_nested_structs_arrays.*
dEQP-GLES2.functional.uniform_api.value.assigned.by_pointer.get_uniform.struct_in_array.bool_*
dEQP-GLES2.functional.uniform_api.value.assigned.by_pointer.get_uniform.basic_struct.bool_*
dEQP-GLES2.functional.uniform_api.info_query.unused_uniforms.*
dEQP-GLES2.functional.uniform_api.value.initial.get_uniform.multiple_basic.vertex
dEQP-GLES2.functional.uniform_api.value.initial.get_uniform.multiple_basic.fragment
dEQP-GLES2.functional.uniform_api.value.initial.get_uniform.multiple_basic.both
dEQP-GLES2.functional.uniform_api.value.initial.get_uniform.multiple_basic_array.*
dEQP-GLES2.functional.uniform_api.value.initial.get_uniform.multiple_nested_structs_arrays.*
dEQP-GLES2.functional.uniform_api.value.initial.render.basic_struct.*
dEQP-GLES2.functional.uniform_api.value.initial.render.struct_in_array.*
dEQP-GLES2.functional.uniform_api.value.initial.render.array_in_struct.*
dEQP-GLES2.functional.uniform_api.value.initial.render.multiple_basic_array.*
dEQP-GLES2.functional.uniform_api.value.initial.render.multiple_nested_structs_arrays.*
dEQP-GLES2.functional.uniform_api.value.initial.render.nested_structs_arrays.*
dEQP-GLES2.functional.uniform_api.value.assigned.unused_uniforms*

Bug: angleproject:2586

Change-Id: I51f619832424131b7c1eeff9589259b14cf4f358
Reviewed-on: https://chromium-review.googlesource.com/1085369Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Luc Ferron <lucferron@chromium.org>

Move the vertex buffer copying and conversion functions from the D3D back
end so the Vulkan back end can use them.  No functional change.

BUG=angleproject:2405

Change-Id: If3c04a6c0dc37c78c1779d8b5d3cece263e8a031
Reviewed-on: https://chromium-review.googlesource.com/1083817
Commit-Queue: Frank Henigman <fjhenigman@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

The constructor of gl::Context creates a bunch of zero and helper objects
but it is not current at this time.  If the backend if not using virtualized
contexts this causes the created objects to be created on the wrong native
context.

BUG=angleproject:2464

Change-Id: I9718df87d0afeb08729920363d362d5f891061ed
Reviewed-on: https://chromium-review.googlesource.com/1048114Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Luc Ferron <lucferron@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

If a uniform was already marked as dirty, and the call to
setMatrixUniform wasn't updating the data, the flag was flipped
to false, causing the descriptor set never to be initialized.

This enables these tests:
dEQP-GLES2.functional.shaders.conversions.scalar_to_matrix.*

Bug: angleproject:2583

Change-Id: I0bcb95cc66951e0ed347af5d874178c403a1c08a
Reviewed-on: https://chromium-review.googlesource.com/1085228Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Luc Ferron <lucferron@chromium.org>