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>

GetInteger64v wasn't introduced until GL 3.2, but some of the enums
for caps that are 64-bit in 3.2 (eg, max uniform block size) were
first introduced as 32-bit values in extensions. This comes up when
trying to use RenderDoc's OpenGL simulator, since it exposes some
relevant extensions and only uses core version 3.1.

BUG=None

Change-Id: Ie4be71b5c8656aae0fe08c270a53f5ef86c99710
Reviewed-on: https://chromium-review.googlesource.com/599030Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

BUG=angleproject:2041

Change-Id: Id5a85c40358b018c17984ac26ee7f97f10584d4b
Reviewed-on: https://chromium-review.googlesource.com/669642Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Olli Etuaho <oetuaho@nvidia.com>
Commit-Queue: Kai Ninomiya <kainino@chromium.org>

These tests crash the OSX shader compiler. Skip them instead of
expecting a failure.

BUG=angleproject:2137

Change-Id: I4476ba93390ce8dfc1e609e4297d6ce1236d2686
Reviewed-on: https://chromium-review.googlesource.com/671285
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>

Apple Clang 8.1 doesn't allow implicit conversion from gl::Error to
egl::Error with the egl::Error::Error(gl::Error&&) constructor. Same
thing for gl and egl reversed. This commits add conversion constructors
taking errors by value. There should be not performance impact for
non-error code paths.

BUG=

Change-Id: I91acf094af923080780b91850146d71016ec5ebc
Reviewed-on: https://chromium-review.googlesource.com/671284Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>

The program pipeline object management entries are:
GenProgramPipelines
DeleteProgramPipelines
BindProgramPipeline
IsProgramPipeline

BUG:angleproject:2123

Change-Id: I114d054b90caf2ee3f9befef7439552a1c309bc4
Reviewed-on: https://chromium-review.googlesource.com/629978
Commit-Queue: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

When switching VAOs, if we switch to a VAO which has disabled
attributes, we could occasionally in some edge cases not have a buffer
initialized to render with. Fix this by re-checking the current
value (disabled) attributes every VAO switch.

Probably a regression caused by d28758de:
"D3D11: Re-enable updateVertexBuffer dirty bits."

BUG=angleproject:2156
BUG=angleproject:1156

Change-Id: Ic1795f914b9b4fa846241b0b4f9a8fe9c1183320
Reviewed-on: https://chromium-review.googlesource.com/669963Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This is better in many ways:
1. It doesn't confuse clang format
2. \n doesn't need to be included after preprocessor directives like
   the version directive.
3. It's using built-in functionality instead of something custom.

Raw string literals should be the preferred way to include shader
source in C++ files going forward.

BUG=angleproject:2157
TEST=angle_end2end_tests

Change-Id: I8b236a6e2d5c25d920297e5bc5b5b143eddeba1f
Reviewed-on: https://chromium-review.googlesource.com/671046Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>

This is required to pass some dEQP GLES 3.1 tests for arrays of
arrays, and WebGL conformance tests were also recently fixed to
require this behavior. The intent of the GLSL ES spec was not to
restrict usage of length().

In practice GL drivers don't implement array length() on expressions
with side effects correctly in all cases. HLSL doesn't have an array
length operator either. Because of this we always remove array length
ops from the AST before output.

BUG=angleproject:2142
TEST=angle_unittests, angle_end2end_tests, WebGL conformance tests

Change-Id: I863a92e83ac5315b013af9a5626348482bad72b3
Reviewed-on: https://chromium-review.googlesource.com/643190
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

The patch cleans up MultiviewDrawTest.cpp by removing some calls to
glUseProgram as that would be done by drawQuad anyway.

BUG=angleproject:2062
TEST=angle_end2end_tests

Change-Id: If9aff686b0ea25e63467852845c46582fdf741c5
Reviewed-on: https://chromium-review.googlesource.com/657678Reviewed-by: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>

We should upstream this fix to dEQP once we have the chance.

BUG=chromium:759402

Change-Id: I64e5df9bc6552e6fabe2b4b60c877fa30fd4c1f2
Reviewed-on: https://chromium-review.googlesource.com/670101
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

BUG=angleproject:2137
TBR=geofflang@chromium.org

Change-Id: I16e0ca6422deed0c225f3eca11189ec7c62fb5c2
Reviewed-on: https://chromium-review.googlesource.com/669361Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>

The patch updates the cubemap completeness check to use the base level
instead of always level 0. Without this change the cubemap faces at
level 0 would have to be created in order to specify another base level
through TexParameteri.

BUG=angleproject:2153
TEST=angle_end2end_tests

Change-Id: Iee1fdc6adf0e69d797821a2ce2f2b2b85dfcdfc1
Reviewed-on: https://chromium-review.googlesource.com/668439
Commit-Queue: Martin Radev <mradev@nvidia.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

This forces all return values to be checked for gl::Error.
Requires quite a bit of minor refactoring. I also added a macro to
swallow an error without returning from a function.

We could look at storing the errors in the Context at some point,
since almost always when we're generating errors that we need to
discard we have access to the Context as a parameter.

BUG=angleproject:2150

Change-Id: I457e48a30c002eda0993acbcd3180ba87bf169fb
Reviewed-on: https://chromium-review.googlesource.com/665173
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

The object owned by the BindingPointer does not necessarily derive
from RefCountObject; it could also just derive from RefCountObjectNoID
(e.g. Compiler).

Found with Clang's CFI bad cast checker.

BUG=chromium:507755

Change-Id: I7e431746b2783e2fc0f2d347a4a27bd60da18473
Reviewed-on: https://chromium-review.googlesource.com/667218Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This will prioritize re-allocating smaller handles. It should not
affect performance in most cases, but will prefer allocating
handles that are in the "fast" portion of the resource map.

BUG=angleproject:1458

Change-Id: Ib2853be936f09fc1e6b5bfb870c360ce8424ab5f
Reviewed-on: https://chromium-review.googlesource.com/665993
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

This was already implemented, it simply hooks it up to the
rest of the code. Could improve performance on some badly
behaved benchmarks.

BUG=angleproject:1390

Change-Id: I539df611d51ca085712fa8022bf8a7c1990afc65
Reviewed-on: https://chromium-review.googlesource.com/663896
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>

Also use a single UpdateSubresource call to update the buffer.
Should improve performance on some benchmarks.

BUG=angleproject:1390

Change-Id: I70d54d86d3d3beb0e2caee86338ee03081070ac8
Reviewed-on: https://chromium-review.googlesource.com/663895Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

Now that we have support for the Debug runtime on the Chromium bots,
we can enable the feature in angle_end2end_tests that checks for any
runtime messages after test execution. They should now show up in the
Chromium CQ.

BUG=angleproject:1878

Change-Id: Ie7502b031a49bcb6a68cf7e3f5e40760fa076ec1
Reviewed-on: https://chromium-review.googlesource.com/667724Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

BUG+angleproject:2137

Change-Id: Ib33210cc89c846db665887f0e59e4b1c91fe29b7
Reviewed-on: https://chromium-review.googlesource.com/665363Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Corentin Wallez <cwallez@chromium.org>

This probably was left out accidentally. Also this fixes the return
type of the prototype in the extension header (should be void, not
boolean).

BUG=angleproject:1523

Change-Id: I7bf0b36b05a4cba4cb6fb2411fc53103dda54bfe
Reviewed-on: https://chromium-review.googlesource.com/657898Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

The ANGLE_multiview specification is modified so that Draw* generates
an error if the number of views in the active program does not match
with the number of views in the active draw framebuffer object.
The tests and validation are modified accordingly.

The patch also sets a contact person, updates the contributor list and
sets the correct enum values in the ANGLE_multiview specification.

BUG=angleproject:2062
TEST=angle_end2end_tests

Change-Id: I15fee4c5e729605bb1d6292f7ad1155637578dea
Reviewed-on: https://chromium-review.googlesource.com/663160
Commit-Queue: Martin Radev <mradev@nvidia.com>
Reviewed-by: Corentin Wallez <cwallez@chromium.org>

BUG=angleproject:2148

Change-Id: If3fa4a42a7343ed133f85be1a4d9d0fa48b427cd
Reviewed-on: https://chromium-review.googlesource.com/665158Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

BUG=angleproject:2075

Change-Id: Ic5e061d4908c072f92f80f70fa6da27b1346cd83
Reviewed-on: https://chromium-review.googlesource.com/663944Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Geoff Lang <geofflang@chromium.org>

Also remove StateManager11::invalidateEverything. With all state
application consolidated in the StateManager, it should be aware of
all state changes, and shouldn't need outside notifications.

BUG=angleproject:2052

Change-Id: I6748b3944b25f958fa093052bcb7336e66255dd4
Reviewed-on: https://chromium-review.googlesource.com/659400
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

BUG=angleproject:1474

Change-Id: Ie80385a9c89453694f1411e49dba7b298c3a1a6c
Reviewed-on: https://chromium-review.googlesource.com/664478Reviewed-by: Jamie Madill <jmadill@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

Was used in a couple other places.

BUG=angleproject:2052

Change-Id: Ib335271a42c9569bbb452b6de8b683023cfc5900
Reviewed-on: https://chromium-review.googlesource.com/659399
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

BUG=angleproject:2052

Change-Id: Ib6f55be3d71d083a87e845447f174a55413c8a2f
Reviewed-on: https://chromium-review.googlesource.com/659398Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This adds an internal and external version of the setShaderResource
method. The external version sets the state as dirty.

BUG=angleproject:2052

Change-Id: I6d2d47490c0af89ff5592d4e9c53eb69f8a3264d
Reviewed-on: https://chromium-review.googlesource.com/659397Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

BUG=angleproject:2052

Change-Id: I1cbf2adc09c82b9de68785265def9361eff418b3
Reviewed-on: https://chromium-review.googlesource.com/659236Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

BUG=angleproject:2052

Change-Id: I11094744bab4570712632a424f29dff8b19a8dea
Reviewed-on: https://chromium-review.googlesource.com/659235Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This will make it easier to implement fast state switching on Context
change.

BUG=angleproject:2052

Change-Id: I045cc2164200a93215629a2746068e686d7c99ff
Reviewed-on: https://chromium-review.googlesource.com/659234Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This will make it easier to do state update on context switch.

BUG=angleproject:2052

Change-Id: Ia73cfd07ced40a9e22d6b34a5619250ede9e8844
Reviewed-on: https://chromium-review.googlesource.com/659233Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

There were a few calls that were using the context directly.
Consolidating them in the StateManager11 class should make efficient
state update after a Context switch possible. Also remove the Context
parameter from the setRenderTargets methods, since these are only
used internally and should be dirtying other states themselves.

BUG=angleproject:2052

Change-Id: I878c99fbde7a467f30cc89f2ee5aca476b8cf506
Reviewed-on: https://chromium-review.googlesource.com/659232Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

Only reset when the resource is valid. This will prevent a memory
allocation that could happen when triggering suprious reset() calls.

BUG=angleproject:1155

Change-Id: I3b9bc1f9e0542c14ed5fd4a03f82ad23c94b734b
Reviewed-on: https://chromium-review.googlesource.com/659231
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

Move this down into the D3D11 renderer. Achieve this by passing a
mutable pointer to the memory program cache to the ContextImpl.

This will allow the D3D11 back-end to more easily sync state then
apply state changes. It also cleans up the gl-side Context a bit.

BUG=angleproject:1155

Change-Id: Ia2c63c05cf414e0d0b22b69c3ed7128f0e405933
Reviewed-on: https://chromium-review.googlesource.com/659230
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

This was a missing part of the validation. Makes us pass the WebGL
test conformance2/rendering/blitframebuffer-outside-readbuffer.

Was necessary for lazy robust resource init.

BUG=angleproject:2107
BUG=chromium:644740

Change-Id: I54c50012fc09ec80a65a2e75f5bde05101c8a1a7
Reviewed-on: https://chromium-review.googlesource.com/663212Reviewed-by: Frank Henigman <fjhenigman@chromium.org>
Reviewed-by: Yuly Novikov <ynovikov@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This also changes the dirty TF object to use a Serial, which is more
secure for very edge-care reallocation issues. It also moves the
StateManager11::updateState call to be the very first thing that
happens in a draw call. This prepares the back-end for making the
state sync actually happen in Context11::syncState, instead of
inside the draw call.

Also moves a bit more TF management code out of RendererD3D and
Renderer9.

BUG=angleproject:2052

Change-Id: I93d033a07be2049023111975a31637c53893e8c8
Reviewed-on: https://chromium-review.googlesource.com/659229Reviewed-by: Geoff Lang <geofflang@chromium.org>
Commit-Queue: Jamie Madill <jmadill@chromium.org>

This eliminates some of the redundant work we do in uniform state
updates. Driver uniforms and constant buffers are no longer synched
with every draw call, but only when the StateManager11 thinks they
might be dirty. Should improve overall draw call throughput.

BUG=angleproject:1155
BUG=angleproject:1390

Change-Id: I351c23fb18dcb3ef2f79e3b4f411e1e09c6300dc
Reviewed-on: https://chromium-review.googlesource.com/659228
Commit-Queue: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>