When a EGL pbuffer is created based on a D3D SRGB texture using
EGL_ANGLE_d3d_texture_client_buffer, SRGB conversions are performed as
if it was an SRGB surface. The value of EGL_GL_COLORSPACE now
reflects that correctly. If the pbuffer is bound to a texture and used
as a framebuffer attachment, querying GL_FRAMEBUFFER_COLOR_ATTACHMENT
also reflects that correctly.

The behavior is the same on both the D3D backend, where there is no
native interop involved, and on the GL backend using
WGL_NV_DX_interop(2).

There are a few limitations on the GL backend that relies on native
interop:

1. SRGB conversion for textures created this way can't be disabled
using the GL_FRAMEBUFFER_SRGB_EXT toggle that's exposed in ANGLE
through EXT_sRGB_write_control. This is now documented in the
EGL_ANGLE_d3d_texture_client_buffer spec. On the D3D backend this is
not a problem since EXT_sRGB_write_control is not supported either
way.

2. Creating a pbuffer out of a D3D11 texture with the format
DXGI_FORMAT_B8G8R8A8_UNORM_SRGB does not work, even though it was
listed as one of the supported formats in the
EGL_ANGLE_d3d_texture_client_buffer spec. It's now mentioned that
support for this format is optional.

BUG=angleproject:2300
TEST=angle_white_box_tests

Change-Id: I70ee0646680805e4469291a5b2ce59e92fda009e
Reviewed-on: https://chromium-review.googlesource.com/866743Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>

BUG=angleproject:2267
TEST=angle_unittests

Change-Id: I25bd8baded9c13e75555578e4b61b99a56e0c702
Reviewed-on: https://chromium-review.googlesource.com/850974Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>

Add #!/usr/bin/python2 and the executable permission bit to all scripts
where missing.

BUG=angleproject:2209

Change-Id: Ib33017c17e579c371b89bbfbdb7136b870027dc5
Reviewed-on: https://chromium-review.googlesource.com/862987Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Frank Henigman <fjhenigman@chromium.org>

Note: by the time the old code used to call `getBaseLevelInternalFormat`, the base level image has already been updated with `redefine()` call above, thus the check `internalformat != storageFormat` wasn't correct.

Change-Id: I3da6df54490d6e72e5094388ed7e39a4c7d920d2
Reviewed-on: https://chromium-review.googlesource.com/862256Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Dzmitry Malyshau <dmalyshau@mozilla.com>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

Indicate scripts/bootstrap.py should be run by python2.

TBR=jmadill@chromium.org
BUG=angleproject:2209

Change-Id: I2ccfee1407cd4ce457aa28a374ac202b8281d40b
Reviewed-on: https://chromium-review.googlesource.com/862618Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Commit-Queue: Frank Henigman <fjhenigman@chromium.org>

This makes sure that the Device to DeviceImpl ratio is always 1:1 and
avoids any potential double-deletion or unexpected deletion of
DeviceImpl objects.

BUG=742034

Change-Id: I778068ccd09b7478d3683123456062b94be242a1
Reviewed-on: https://chromium-review.googlesource.com/854627
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

A virtual function to get the constant value of an AST node is added
to TIntermTyped. This way a constant value can be retrieved
conveniently from multiple different types of nodes. TIntermSymbol
nodes pointing to a const variable can return the value associated
with the variable, constructor nodes can build a constant value from
their arguments, and indexing nodes can index into a constant array.

This enables constant folding operations on constant arrays, while
making sure that large amounts of data are not duplicated in the
output shader. When folding an operation makes sense, the values of
the arguments can be retrieved by using the new
TIntermTyped::getConstantValue(). When folding an operation would
result in duplicating data, the AST can just be left to be written out
as is.

For example, if the code contains a constant array of arrays, indexing
into individual elements of the inner arrays can be folded, but
indexing the top level array is left in place and not replaced with
duplicated array literals.

Constant folding is supported for indexing and comparisons of arrays.

In case constant arrays are only referenced through foldable
operations, the variable declarations will be pruned from the AST by
the RemoveUnreferencedVariables step.

BUG=angleproject:2298
TEST=angle_unittests

Change-Id: I5b3be237b7e9fdba56aa9bf0a41b691f4d8f01eb
Reviewed-on: https://chromium-review.googlesource.com/850973Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>

Only exact version of Windows SDK works.
Clarify how to do a release build.

BUG=angleproject:2297

Change-Id: I2566b7e807a84c4edc3b457299036e67a64b5a36
Reviewed-on: https://chromium-review.googlesource.com/852553
Commit-Queue: Frank Henigman <fjhenigman@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

Add an initialize function to match other impl objects and simplify the
creation of DeviceImpls.

BUG=742034

Change-Id: I569c8252d5d23c8af98835f6c08e7a3b640fc3f3
Reviewed-on: https://chromium-review.googlesource.com/854626Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

It can be inferred from the presence of an owning display.

BUG=742034

Change-Id: I6a33378f4256733da6d9658d47763153e431ad70
Reviewed-on: https://chromium-review.googlesource.com/854625Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

BUG=angleproject:1920
TEST=angle_end2end_tests

Change-Id: I90bcd3bf5a078623b3a739615fbc7b157b6fb94c
Reviewed-on: https://chromium-review.googlesource.com/808144Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>

The PipelineDesc class is a 512-byte packed description of the entire
Vulkan pipeline state. It uses the alignas keyword and some static
asserts to verify that the structures are packed. This ensures that
when ANGLE uses MurmurHash to hash the entire struct, and memcmp to
check for identity, that there are no garbage padding bits.

This CL does not implement the Pipeline cache, but it will help, since
now we have a packed type that can be used as the key to a hash map.

Bug: angleproject:2163
Change-Id: I16efa927f08d30d89a9c4c8943edd211c6878ac8
Reviewed-on: https://chromium-review.googlesource.com/829893
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>

Adreno's driver in Android Oreo follows the spec to the letter,
and applies "#pragma STDGL invariant(all)" only to the outputs of
a fragment shader, causing InvariantAllBoth test to fail.

Since the wording of the spec doesn't match the intended usage of this
pragma, ANGLE should flatten it to workaround spec conformant drivers.

BUG=angleproject:1293

Change-Id: I9f8ad44adc3b958204b88805dba8c5b7964b3d9f
Reviewed-on: https://chromium-review.googlesource.com/860920Reviewed-by: Kenneth Russell <kbr@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Yuly Novikov <ynovikov@chromium.org>

This will be useful for the Vulkan pipeline state cache. This cache
wants to pack the pipeline description very carefully into a specific
number of bits, using the "alignas" keyword. Using this keyword
requires an aligned allocator, which we can lift from Chromium's
base/memory submodule.

Also includes a unittest copied from Chrome.

Bug: angleproject:2163
Change-Id: I25976be4610636db3f43552bba23d823f5a49a24
Reviewed-on: https://chromium-review.googlesource.com/837944Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

Bug: angleproject:1595

Change-Id: I031beab818daca8a20726d7bff564649feefc1f3
Reviewed-on: https://chromium-review.googlesource.com/844214
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

Treat aggregate constructors as having no side effects, which enables
pruning declarators that have a constructor initializer.

Some logic in RemoveUnreferencedVariables is fixed to make this work
correctly for structs. The bugs were previously not exposed since
constructors were treated as having side effects, but now that those
can be pruned the logic needs to be correct.

BUG=angleproject:2298
TEST=angle_unittests

Change-Id: I6fbe61a9e82065196baa29c200bf556fc21d8962
Reviewed-on: https://chromium-review.googlesource.com/856499Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>

BUG=angleproject:2290
TEST=dEQP-GLES31.functional.texture.multisample.samples_*.sample_position
TEST=TextureMultisampleTestES31.CheckSamplePositions*

Change-Id: If8b74c16d5c104215456e35b8922279be972cee3
Reviewed-on: https://chromium-review.googlesource.com/844062
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Instead of referring to the gl_in symbol in the symbol table, create
a gl_in symbol with the right array size once the array size is known.
This makes the type of TIntermSymbol nodes pointing to gl_in
consistent with the variable type.

BUG=angleproject:2267
TEST=angle_unittests

Change-Id: I31673d33526a91f8d069ee8d7d2f181a49665fd0
Reviewed-on: https://chromium-review.googlesource.com/857004Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>

NVIDIA OpenGL drivers at least up to version 388.59 don't clamp
gl_FragDepth when it is written to a floating point depth buffer.
This bug is now worked around by clamping gl_FragDepth in the shader
if it is statically used.

BUG=angleproject:2299
TEST=angle_end2end_tests on NVIDIA

Change-Id: I61589b2b0dd2813c4901a157c8d37e470063773c
Reviewed-on: https://chromium-review.googlesource.com/840842
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>

Create new TVariables to change struct declaration qualifiers in
PruneNoOps so that the node doesn't end up with an inconsistent
qualifier from the variable.

BUG=angleproject:2267
TEST=angle_unittests

Change-Id: I7f6cd557eb3e46f4ead4ffbe37d8960ff7776d10
Reviewed-on: https://chromium-review.googlesource.com/850673Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>

Was flaky.

BUG=angleproject:2301

Change-Id: Id71cf6c43e133810c64535bbde494d12546b0adc
Reviewed-on: https://chromium-review.googlesource.com/854978Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Yuly Novikov <ynovikov@chromium.org>

The SPIRV-Tools 'external revision generate' script is coded such that
it assumes a .git folder is present. This causes a problem for the
tarball build. We can fix this by using the same tooling we use for
ANGLE's commit id script to check if the .git directory is present
before running the generator. If it is missing, we insert a hard-coded
dummy header. Also use the 'DEPS' file as a placeholder for .git/HEAD
to check if the current revision needs updating.

Bug: chromium:799620
Bug: angleproject:2237
Change-Id: Icea8e9c66f1600df7dca2aaa45fe449f687f5b55
Reviewed-on: https://chromium-review.googlesource.com/854255Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This is now updated in Chrome, so we can remove the last uses.

Bug: angleproject:2159
Change-Id: Ia31e42b8a685756ee70450be19a52248e3efa92f
Reviewed-on: https://chromium-review.googlesource.com/853119Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

Added missing quotes.

BUG=angleproject:2297

Change-Id: Ieb6ca4ae37786f48ad3c6f5dccefef25afce23a2
Reviewed-on: https://chromium-review.googlesource.com/852636Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Commit-Queue: Frank Henigman <fjhenigman@chromium.org>

This was causing a heavy performance regression in the Vulkan perf
test. Local testing shows it recovers most of the performance gap
with this fix.

Bug: angleproject:2264
Bug: chromium:799372
Change-Id: Ia2bbf526b305e1b2e901d904574dbe5618e6e5ef
Reviewed-on: https://chromium-review.googlesource.com/852365Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

The clamping doesn't happen on NVIDIA so this is a test that will
confirm a follow-up compiler workaround patch works properly.

BUG=angleproject:2299

Change-Id: I006b522d014e29d2a5e634dfb9cf827ee3bc0536
Reviewed-on: https://chromium-review.googlesource.com/852252
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Olli Etuaho <oetuaho@nvidia.com>

for compute shaders on D3D

BUG=angleproject:1987
TEST=angle_end2end_tests.ComputeShaderTest.*

Change-Id: I075296ac3b6796a334929699c16f2399d7915e51
Reviewed-on: https://chromium-review.googlesource.com/844063
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

In case folding a ternary op or a comma op would change the qualifier
of the expression, the folding is deferred to a separate traversal
step.

After this there are no more cases where the type of a TIntermSymbol
node needs to differ from the type of the variable it is referring to.
There are still some cases where some parts of TIntermSymbol type are
changed while keeping the TVariable type the same though, like when
assigning array size to gl_PerVertex nodes or sanitizing qualifiers of
struct declarations.

BUG=angleproject:2267
TEST=angle_unittests, angle_end2end_tests

Change-Id: I1501c8d361f5f765f43ca810d1b7248d9e2c5986
Reviewed-on: https://chromium-review.googlesource.com/850672
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>

This patch labels geometry_shading.vertex_transform_feedback.* as
"SKIP" instead of "FAIL" to fix some failures in dEQP tests.

OpenGL ES 3.1 extension GL_EXT_geometry_shader requires additional
validations on transform feedback, which hasn't been implemented
in ANGLE, so geometry_shading.vertex_transform_feedback.* fail just
after beginTransformfeedback(). As endTransformfeedback() isn't
called in dEQP function resetStateES():
1. An error occurs in this function when trying to clear transform
   feedback.
2. Some of the tests afterwards also fail because it isn't allowed
   to change active program by useProgram() when transform feedback
   is on.

So it is better to skip these tests before ANGLE supports these new
features.

BUG=angleproject:1941

Change-Id: I8834f71e6a655f18118ee4ecf26f4118da6de2f1
Reviewed-on: https://chromium-review.googlesource.com/843955
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jiajia Qin <jiajia.qin@intel.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>

This makes the device type selection part of the ANGLE platform
extension. We currently support NULL driver selection on all
available back-ends (although on the NULL back-end, it already
has no device type). Optionally we could expose certain features
of this as separate extensions.

This currently also supports the old hidden enum, until we can
update Chrome and consolidate to the new exposed official enum.

Bug: angleproject:2159

Change-Id: I85d0811098e644e8192c207673af9e18ed7c1da2
Reviewed-on: https://chromium-review.googlesource.com/846021
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

From third_party/vulkan-validation-layers/src/layers/CMakeLists.txt
Otherwise, there is a conflict between global functions with same name
from different layers and the loader.
Conflict is resolved by -Bsymbolic.

BUG=angleproject:2296

Change-Id: I40859326626c3865ba346c6e817b7c512735054e
Reviewed-on: https://chromium-review.googlesource.com/850727
Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

Improves ValidateVertexShaderAttributeTypeMatch by storing vertex
attributes types into masks for quick comparisons when needed. This shows
2% improvement to glDrawElements for the aquarium workload.

BUG=angleproject:2202

Change-Id: I87fa3d30c3d8cdba6dfd936cd1a41fd27b1c6b77
Reviewed-on: https://chromium-review.googlesource.com/814795
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Returns an OOM error to avoid
query the context for status should Map fail.

BUG=angleproject:2284

Change-Id: I4d35ef05c76fcc69bdab19f88c39a415d13ee116
Reviewed-on: https://chromium-review.googlesource.com/837519
Commit-Queue: Yuly Novikov <ynovikov@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

This is partially complete - it is missing the auto-gen for custom
ANGLE and Chromium extensions that aren't present in gl.xml. We
can upstream some of these extensions, but will also need to make
custom handling for some extensions that are too volatile or under-
specced to upstream to Khronos.

This also tweaks some of the Context method names to match what the
auto-generator expects. It also updates some TexStorage entry points.

Also includes a specialized error return value for glTestFenceNV.

Also removes the TexImage3DOES entry point which was neither used
nor accessible to the application.

Bug: angleproject:2263
Change-Id: I3667c22b1b15f84ab4c7dfecb5745aaf73d11e76
Reviewed-on: https://chromium-review.googlesource.com/846420
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>

The previous code was hard to read since the referenced interface
blocks stored a different type of node depending on if the interface
block was instanced or not.

BUG=angleproject:2267
TEST=angle_unittests

Change-Id: Ie8fdb61a17280ca0875159702f819b884d08706b
Reviewed-on: https://chromium-review.googlesource.com/839443
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

This patch intends to fix a bug in logging link errors.

OpenGL ES requires glGetProgramInfoLog return a string
that contains information about last link or validation
attempt on a program object (GLES 2.0 Chapter 6.1.8,
GLES 3.0 Chapter 6.1.12). So all the link error logs
should be cleared when a new link process is begun.

This patch also removes several redundant allocations
of mLazyStream in ProgramD3D::compileProgramExecutables.
Calling "<<" on InfoLog objects will initialize its
member mLazyStream from heap, so we will skip using "<<"
if there is actually nothing to output.

BUG=angleproject:2295
TEST=angle_end2end_tests
Change-Id: Ib81fffd3d05919a8ebccd9145ff780548ca86a70
Reviewed-on: https://chromium-review.googlesource.com/848324
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

This removes unnecessary indirection. It's easier to just create
TVariables in createSamplerSymbols, and to track referenced variables
using TVariable pointers instead of TIntermSymbol pointers.

BUG=angleproject:2267
TEST=angle_unittests, angle_end2end_tests

Change-Id: Id1e75e04da084eb9026f581f22070b27a45615ba
Reviewed-on: https://chromium-review.googlesource.com/839442Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>

These were broken by command re-ordering. The first problem was that
the Framebuffer RenderNode was not correctly being flagged as dirty.
This is fixed by keeping a serial instead of a bool to track the
RenderNode's cleanliness. The second issue was that an image layout
transition was being set with incorrect bits. Fix this by using the
correct access mask.

This fixes angle_perftests's Vulkan render test.

Bug: angleproject:2264
Bug: chromium:798866
Change-Id: I268b0f7aeb2c5e22892f6ef59dec62391b30bfb4
Reviewed-on: https://chromium-review.googlesource.com/848539Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This allows enabling and disabling parts of the build via gn args,
similar to what gyp allowed.

BUG=angleproject:1569

Change-Id: Ia95c134f98f8941408df96b1d4e2563e713a9ef2
Reviewed-on: https://chromium-review.googlesource.com/838260
Commit-Queue: Frank Henigman <fjhenigman@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

The default heap ordering is to return the greatest element in the
heap. Since the handle allocator expects a minimal return value on
a new allocation, this caused a bug. The bug is triggered by reserving
handles, allocating new handles, then freeing the handles and
allocating again with the same allocator. Fix the bug by using
std::greater instead of std::less, which will make the heap
return the smallest value instead of largest.

Also adds some logging debugging code for the handle allocators.

Bug: angleproject:1458
Change-Id: Ibef5dcbed0a664ccad0e0335f081e2355162584b
Reviewed-on: https://chromium-review.googlesource.com/848644
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>