This patch intends to implement glSampleMaski on D3D11 backends.

In D3D11 the sample mask word can be set by OMSetRenderState. This
function can only accept an UINT as its parameter SampleMask, so
the value MAX_SAMPLE_MASK_WORDS in D3D11 backends is set to 1.

BUG=angleproject:1592
TEST=dEQP-GLES31.functional.state_query.*.sample_mask*
TEST=dEQP-GLES31.functional.texture.multisample.*.sample_mask*

Change-Id: I27f509cf81624d95948b06739bb9fe21288d2d31
Reviewed-on: https://chromium-review.googlesource.com/691259
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

PMurHash comes from the smhasher repository at chromium/src/third_party/smhasher

Bug: 697758
Change-Id: Id2859edf37ae66bf27509d53db7f22db8831fe44
Reviewed-on: https://chromium-review.googlesource.com/687970
Commit-Queue: Kai Ninomiya <kainino@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

This is a bit of a hack, similar to how D3D11 works. We need to write
output locations in the GLSL shader before we send them to glslang,
so we wait until the link call, then string-replace some hard-coded
identifeir code to the attribute location determined by ANGLE.

This CL also fills in some of the vertex format conversion tables in
formatutilsvk.cpp.

BUG=angleproject:2167

Change-Id: I2424d0d990bdbcd831a4dd130e61e87d8f8f479f
Reviewed-on: https://chromium-review.googlesource.com/677555
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>

This should give us the ability to send off more than one frame to
the presentation engine at once. Instead of using a single pair of
Semaphores to lock each surface, we make a Semaphore pair per
Swapchain image.

BUG=angleproject:1898

Change-Id: I9e833ed9969a79617d0a8968b0d5a25c27139e87
Reviewed-on: https://chromium-review.googlesource.com/672149
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>

This patch intends to implement glSampleMaski on OpenGL
backends.
Refers to: https://chromium-review.googlesource.com/c/487603

BUG=angleproject:1592
TEST=dEQP-GLES31.functional.state_query.*.sample_mask*
TEST=dEQP-GLES31.functional.texture.multisample.*.sample_mask*

Change-Id: If5ddd6ab57259593919a482be80fbdbe29f6f54d
Reviewed-on: https://chromium-review.googlesource.com/573727
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

BUG=angleproject:2137

Change-Id: I385eab17b60ae434b06c8235d7ee7849c145e765
Reviewed-on: https://chromium-review.googlesource.com/688681
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>

BUG=angleproject:2137

Change-Id: If1b186de7aed3464427864b3a68dffd74cb497ed
Reviewed-on: https://chromium-review.googlesource.com/687871
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>

BUG=angleproject:2137

Change-Id: I0d434f6b92164fecb10ee61c8cb662ccab8470f1
Reviewed-on: https://chromium-review.googlesource.com/685937
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>

Multisampled texture with unfilterable internalformat and default filter state shouldn't
be justified as incomplete texture in angle. However, the default filter state is set to
LINEAR and internalformat like R8I or R8UI is not filterable, which lead to texture
incomplete.

In es 3.1 spec 8.16, 'The effective internal format specified for the texture arrays is a
sized internal color format that is not texture - filterable(see table 8.13), and either
the magnification filter is not NEAREST or the minification filter is neither NEAREST nor
NEAREST_MIPMAP_NEAREST'.

As to Table 20.11, The default value of minFilter is NEAREST_MIPMAP_LINEAR, magFilter is LINEAR.

Because in 11.1.3.3,'Multisample textures are not filtered when samples are fetched, and filter
state is ignored.' So we ignore filtering for multismaple texture completeness.

BUG=angleproject:1590
TEST=dEQP-GLES31.functional.shaders.builtin_functions.texture_size.samples_*_texture_2d
TEST=dEQP-GLES31.functional.shaders.builtin_functions.texture_size.samples_*_texture_int_2d
TEST=dEQP-GLES31.functional.shaders.builtin_functions.texture_size.samples_*_texture_uint_2d
TEST=dEQP-GLES31.functional.texture.multisample.samples_*.use_texture_color_2d
TEST=dEQP-GLES31.functional.texture.multisample.samples_*.use_texture_int_2d
TEST=dEQP-GLES31.functional.texture.multisample.samples_*.use_texture_uint_2d
TEST=dEQP-GLES31.functional.texture.multisample.samples_*.use_texture_depth_2d

Change-Id: I0533bd028e0bbb66d9f731a2395f6c8ae7db8938
Reviewed-on: https://chromium-review.googlesource.com/564843
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

BUG=765953

Change-Id: I4dc79921766975cd75c489887b7e57ec4666fbbb
Reviewed-on: https://chromium-review.googlesource.com/685897Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

Applies to UNREACHABLE and UNIMPLEMENTED.
This helps track down the offending line more easily.

BUG=angleproject:2167

Change-Id: I97eb80677a6dd432a04f5e7516c3db3de8e7128a
Reviewed-on: https://chromium-review.googlesource.com/679999
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

Some code refactoring in this change:
1) It moves some public methods to static.
These methods are only used in validationES3.cpp. They are not necessay to be
exposed to public.

2) The error messages to check context version are inconsistent. It makes them
to be consistent.

BUG:angleproject:2005

Change-Id: I5af4c3300634ccc44aac386c90dcb0522acbff83
Reviewed-on: https://chromium-review.googlesource.com/661324
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

This method was removed in a prior patch.

BUG=angleproject:1898

Change-Id: I2ab3c36bbef4ce6c978cd4c09cc191c4e7ff7991
Reviewed-on: https://chromium-review.googlesource.com/677547Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This saves some time spent in the driver, by making multiple draw
calls happen inside a single RenderPass.

This also makes the ReadPixels impl method non-const. I think in
the future we should avoid making const Impl methods unless they're
totally trivial.

BUG=angleproject:1898

Change-Id: I39172270a2f7dc5c1c2e3d4cc50af3bac8a29fa1
Reviewed-on: https://chromium-review.googlesource.com/672148Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>

Although this adds a boolean check (and state flag maintenance) cost
to each draw call, it makes ANGLE's internal life a lot simpler
because it doesn't have to process a framebuffer change until the
draw call. It turns out there are a few dependent checks of the
Framebuffer that aren't always easy to do. In one test, Context
destruction was triggering RenderTarget invalidation, after the
Context had already freed the Framebuffer manager.

This also fixes the problem in feature level 9_3 with framebuffer
invalidation affecting the internal dirty bit set. (Note that it
still dirties the Framebuffer for the next frame).

BUG=chromium:767279
BUG=angleproject:2151

Change-Id: I74d61bddf9926004a04f712a9f9eb1205d5df0e1
Reviewed-on: https://chromium-review.googlesource.com/676657Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This flag will only render the first frame of each perf test,
regardless of their preferences for how many seconds to run.
It will be useful for speeding up the run time of the perf tests
on testing infrastructure that only cares about correctness.

BUG=chromium:725308
BUG=chromium:765321

Change-Id: I926f488c42f27ef23ef06a0159902613cff04080
Reviewed-on: https://chromium-review.googlesource.com/677306Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

Introduce a dirty bit for current value attribs, and try to speed up
the check for dirty vertex attribs.

This series of small optimizations gives about a 15% improvement on
the draw call benchmark for the D3D11 backend with the null driver.

BUG=angleproject:1155

Change-Id: Idf560efa3af62776e8bdbdf693f8b06db8792c21
Reviewed-on: https://chromium-review.googlesource.com/666048Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This cleans up the point sprites dirtying logic into applyIndexBuffer.

This series of small optimizations gives about a 15% improvement on
the draw call benchmark for the D3D11 backend with the null driver.

BUG=angleproject:1155

Change-Id: I210dd408d11ef5a0b58b0ad32d1255c787a55fba
Reviewed-on: https://chromium-review.googlesource.com/666047Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

Only iterate the enabled draw buffers, and the depth/stencil buffer.

This series of small optimizations gives about a 15% improvement on
the draw call benchmark for the D3D11 backend with the null driver.

BUG=angleproject:1155

Change-Id: I5b29362f93c016c146d2a6527b378853bc070239
Reviewed-on: https://chromium-review.googlesource.com/666046Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This moves the skipDraw logic into applyPrimitiveType, since it's
more efficient to only check the primitive type once. Also merges
the draw*Impl and genericDraw* methods, since the generic* methods
weren't really doing anything anymore, and all the state logic lives
in StateManager::updateState.

This series of small optimizations gives about a 15% improvement on
the draw call benchmark for the D3D11 backend with the null driver.

BUG=angleproject:1155

Change-Id: I299213da6d1bbcb08691d5b50162e6cae16cb4bb
Reviewed-on: https://chromium-review.googlesource.com/666044Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

The workaround for calling clearRenderTargetView on small targets
isn't needed on the newly released Intel D3D driver 15.46-4771
because the fix of this bug has been merged into this driver.

BUG=chromium:655534

Change-Id: I90125971fd4b782b7363b8ca3104bd500e21a034
Reviewed-on: https://chromium-review.googlesource.com/668223
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

"Add test for GL_KHR_robust_buffer_access_behavior"
introduced a failure in Windows 10 NVIDIA OpenGL configs.

"ES31: Enable shader storage buffer support for OpenGL backend"
introduced a failure on Linux Intel HD 630 configs.

BUG=angleproject:1463
BUG=angleproject:1951

Change-Id: I791bc729893b7f31093ab1678dc8d01edfbd0a5a
Reviewed-on: https://chromium-review.googlesource.com/676473Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This allows us to skip calling the dynamic shader generation and
program shader application when there haven't been any state
changes. It builds on the previous work that immediately update
state caches in the VertexArray11 and Framebuffer11. It should
improve performance in draw-call limited applications by a small
margin.

For reference, here are the conditions under which the shaders
are refreshed:

1. Directly changing the program executable
2. The vertex attribute layout
3. The fragment shader's rendertargets
4. Enabling/disabling rasterizer discard
5. Enabling/disabling transform feedback
6. An internal shader was used
7. Drawing with/without point sprites

Improves the score of the draw call stress test for the D3D11
back-end (with null driver) by about 40% on my test machine.

The 9_3 back-end seems to have an issue where the getSRV call
to a texture storage can change the "use level zero workaround"
status of the storage, which in turn will invalidate the state.
Since this is localized to 9_3 only, put in a hack to disable
an assert check for now.

BUG=angleproject:2151

Change-Id: Idbd0a31376691b33972e735d5833a9b02a8a4aa9
Reviewed-on: https://chromium-review.googlesource.com/666278Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

The used flag was redundant with the index (which used MAXUINT). The
name was redundant with the stored uniform. Removing these gives a
very minor performance speed up when iterating and retrieving
uniform locations.

BUG=angleproject:1390

Change-Id: Ieeccdff7c131e1359e754e246d3648b73aad5baf
Reviewed-on: https://chromium-review.googlesource.com/659224
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

For some reason we were using the second GLint value passed into the
method, when the spec says we should have a single clear value.

Noticed when adding tests for lazy robust resource init (tests will
be added in a subsequent patch).

BUG=angleproject:2107

Change-Id: Iadfc5072796255072bfb71ff1918253045e29576
Reviewed-on: https://chromium-review.googlesource.com/675049
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>

This allows us to call StateManager11::invalidateRenderTarget from the
Framebuffer11::signal function, which will then trigger state refresh
on the next draw call.

This requires passing Context through a few more Renderbuffer methods,
and reorganizing the RenderTarget signalling so that it doesn't signal
dirty in the destructor. Instead they are signaled as they are
destroyed in the containing classes.

BUG=angleproject:2151

Change-Id: I4cf575e4a01b48275ff78d75bc55b2d1fced591d
Reviewed-on: https://chromium-review.googlesource.com/673139
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>

This cleans up some of the iteration logic.

BUG=angleproject:2151

Change-Id: I8d80a8d732ee808babbb4859290b648b4fa67b4d
Reviewed-on: https://chromium-review.googlesource.com/673138
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

This ensures they are auto-freed when the Storage is destroyed.
Also remove the 'delete this' design pattern.

BUG=angleproject:2151

Change-Id: I784b94c7125a1bbc15f5d6ae90e55317e199faa6
Reviewed-on: https://chromium-review.googlesource.com/673137
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>

This will allow us to pull out the Renderer from the Context in more
places in Buffer11, for state update. Impacts a few method calls in
a few places.

BUG=angleproject:2151

Change-Id: I1360caea65a94d3de4cd9f52d1b74b10439b02b3
Reviewed-on: https://chromium-review.googlesource.com/673136Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This patch enables 'location' layout qualifier for vertex outputs and
fragment shader inputs when the shader version is 3.1 in ANGLE GLSL
compiler and adds the check on location conflicts for these varyings.

According to GLSL ES 3.1 SPEC (Chapter 4.4.1 and Chapter 4.4.2),
'location' layout qualifier is allowed on both inputs and outputs of
vertex and fragment shaders.

'location' layout qualifier on shader interfaces is only valid on shaders
whose version is 3.1 and above. According to GLSL ES 3.0 SPEC, vertex shader
cannot have output layout qualifiers (Chapter 4.3.8.2) and fragment shader
cannot have input layout qualifiers (Chapter 4.3.8.1).

The 'location' qualifier on varyings is used in the shader interface
matching defined in OpenGL ES 3.1. (OpenGL ES 3.1 SPEC Chapter 7.4.1). This
new link rule will be added to Program.cpp in another patch.

For the OpenGL ES 3.1 extension GL_OES_geometry_shader, according to
GL_OES_shader_io_blocks SPEC (Chapter 4.4.1 and Chapter 4.4.2), 'location'
layout qualifier is both valid on geometry shader inputs and outputs. This
feature will be implemented together with other rules on geometry shader
inputs and outputs.

BUG=angleproject:2144
TEST=angle_unittests

Change-Id: I62d85f7144c177448321c2db36ed7aaeaa1fb205
Reviewed-on: https://chromium-review.googlesource.com/645366
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

BUG=angleproject:1951
TEST=angle_end2end_tests:ShaderStorageBuffer

Change-Id: I1afc3cd005ad2e595c6ce937fc53e17423f8ec8b
Reviewed-on: https://chromium-review.googlesource.com/618132
Commit-Queue: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

This test is ported from webgl test element-index-uint.html.

BUG=angleproject:1393, angleproject:1463

Change-Id: I165e3dd2913968c8cc035c570a7bcaf91aed095a
Reviewed-on: https://chromium-review.googlesource.com/651239
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

The patch adds new built-ins and extends the semantic parser to add
support for textureGather and textureGatherOffset.

BUG=angleproject:1442
TEST=angle_unittests
TEST=angle_deqp_gles31_tests.exe
     --deqp-case=dEQP-GLES31.functional.texture.gather*
     --deqp-egl-display-type=angle-gl

Change-Id: Iaf98c3420fbd61193072fdec8f5a61ac4c574101
Reviewed-on: https://chromium-review.googlesource.com/660124
Commit-Queue: Martin Radev <mradev@nvidia.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

When using FRAGMENT_SHADER_DERIVATIVE_HINT_OES as a hint with
OES_standard_derivatives disabled, we should return INVALID_ENUM, not
INVALID_OPERATION.

Bug: angleproject:2158

Change-Id: I5759f1e8bb19d2caed278506054aebc5d82d431c
Reviewed-on: https://chromium-review.googlesource.com/673374Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This will ensure that new methods (placed at the end) don't collide
with the context pointer when using a mismatched platform.

BUG=angleproject:1892

Change-Id: Ice6a3ccaf3cc2a1e36a04bc217386cf096612176
Reviewed-on: https://chromium-review.googlesource.com/672147Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

We could dereference a null pointer in a couple edge cases.

BUG=angleproject:2151

Change-Id: I0d40b930c5008b80928bfe8c93588f80127fd166
Reviewed-on: https://chromium-review.googlesource.com/665995
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

Instead of using a compile-time define, use an OS call to override
the environment variable the loader uses to look for layers. This
should allow us to have a run-time override mechanism, so we can
more easily use ANGLE with RenderDoc and other tools that hook
into the layers for debugging and profiling purposes.

This should also allow the developer to install and use their own
layers with ANGLE if desired.

This patch removes the angle_loader.h generation since it is no
longer necessary.

It also fixes an unrelated loader warning that occured when releasing
the current pipeline object.

BUG=angleproject:1898

Change-Id: Ic4a5120a6b73745397451ef9e3897e157da1feda
Reviewed-on: https://chromium-review.googlesource.com/671490Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This warning was popping up with a new version of the layers when
running the hello_triangle sample. Fix it by assigning the current
Serial ID to the Context for the current pipeline. A more robust
fix in the future would probably be to assign Serial IDs to the
Pipelines themselves.

BUG=angleproject:1898

Change-Id: Ib5e8ea7c4c26907d1987529d8127249f9b18e17f
Reviewed-on: https://chromium-review.googlesource.com/672146Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

We had a performance regression in the Textures benchmark. What the
test was doing was iterating over all possible texture state,
ensuring the active texture was dirty every frame. This is an attempt
to improve on the speed by not doing as much resetting work in
State::syncProgramTextures. It introduces an active textures mask to
speed iteration over the active texture set.

Also makes a refactoring change to Context to make it easier to limit
caps to an implementation maxium. The number of active textures is
limited to 64 so they easily fit in the bitset mask, with a limit of
32 per shader stage. No mask is currenly kept for compute shaders.

With the fix the performance should be about the same as before (which
is good, as the test always sets the textures dirty).

Test: TexturesBenchmark.Run/gl_8_textures_5_rebind_3_state_8_mips

BUG=chromium:765363
BUG=angleproject:1387

Change-Id: I8bcf95be3671195373573f89f406edaba40aa1be
Reviewed-on: https://chromium-review.googlesource.com/670279Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

To prevent collisions when linking with other copies of MurmurHash3.

BUG=697758

Change-Id: Id8a5c709ba972812ffa3ca143e7553cbf05fc57a
Reviewed-on: https://chromium-review.googlesource.com/671194
Commit-Queue: Kai Ninomiya <kainino@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>