Commits · bb714f72c79e6f45fbb3b7dbfc93907bb04436ae · Chen Yisong / angle

20 May, 2015 8 commits

Release egl::Surface on Texture image changes. · bb714f72

authored May 20, 2015

We would get into a broken state if the user would bind a Texture
to an egl::Surface, then change the Texture. This is valid in egl,
and should release the Surface when it happens.

BUG=485543

Change-Id: Idfaa305ac704f2bc579e79be816e88a23e69351b
Reviewed-on: https://chromium-review.googlesource.com/271986Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

bb714f72

Fix most occurences of -Winconsistent-missing-override · a588ef5b

authored May 19, 2015

Other occurences are in code generated by macro in googlemock.

BUG=angleproject:892

Change-Id: I8b2c6ad4b793be2d9a440687b1b010c9feb8bdae
Reviewed-on: https://chromium-review.googlesource.com/272224Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

a588ef5b

FunctionsGLX: Remove all but one of the ANGLE dependencies · 4f691690

authored May 19, 2015

A follow-up CL will make it so the dependencies on
TokenizeExtensionString doesn't pull the GL headers, this will then
allow use to remove the hack for the inclusion of glx.h

BUG=angleproject:892

Change-Id: I51c5466617057e2c580bd5e8056307515bcecc4b
Reviewed-on: https://chromium-review.googlesource.com/272206Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

4f691690

Completely isolate GLX in FunctionsGLX · 3cd0ca3c

authored May 15, 2015

This makes it so it is the only file including glx.h.

BUG=angleproject:892

Change-Id: I5c45b659ec102b7a2c92e010a6843fa21b7c664b
Reviewed-on: https://chromium-review.googlesource.com/271418Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

3cd0ca3c

Ensure that conditional blocks and loop bodies are sequence nodes · 7d7f8c44

authored May 19, 2015

AST transformations such as unfolding logical operations can only
create nodes inside sequence (block) nodes. This patch ensures that these
AST transformations work even inside conditional blocks and loop bodies
that were first parsed as single statements instead of blocks.

BUG=angleproject:971
TEST=WebGL conformance tests

Change-Id: If98cb7be653ec7b005537b89547f4f4cf1c07c72
Reviewed-on: https://chromium-review.googlesource.com/272141Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Tested-by: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

7d7f8c44

In D3D11, return 0 if not conformant, rather than GL_NONE · 08a5fedd

authored May 19, 2015

Change-Id: Ica6484d1b5960f162ad88648ff9b2b9b96710930
Reviewed-on: https://chromium-review.googlesource.com/272166Tested-by: Austin Kinross <aukinros@microsoft.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

08a5fedd

Updated generate_winrt_projects to produce Windows 10 projects · aa5cbdc8

authored May 18, 2015

BUG=angleproject:1004

Change-Id: I6f50226b21ff92ac85dbe5ed2e39dcc4f7e39306
Reviewed-on: https://chromium-review.googlesource.com/272160Tested-by: Cooper Partin <coopp@microsoft.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

aa5cbdc8

Refactoring: make tracking parent block position in AST traversal reusable · 56eea884

authored May 18, 2015

Add a helper function to make it easier for traverser classes to insert
statements, and use it in UnfoldShortCircuitToIf.

BUG=angleproject:971
TEST=angle_end2end_tests, WebGL conformance tests

Change-Id: I141bdd8abf4b01988581e6cb27c2320bf38370ac
Reviewed-on: https://chromium-review.googlesource.com/272140Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Tested-by: Olli Etuaho <oetuaho@nvidia.com>

56eea884

19 May, 2015 11 commits

Add ANGLE support for ESSL3 variant of mix() · d94f6647

authored May 18, 2015

This change adds support for ESSL3 variant of built-in function mix()
that takes last argument as a boolean vector. lerp() - HLSL equivalent
of mix() doesn't look to be supporting a variant that takes last
argument as a boolean vector so emulated it for HLSL.

BUG=angleproject:1006
TESTS=dEQP tests
Fixes "no matching overloaded function found " errors in below tests:
dEQP-GLES3.functional.shaders.constant_expressions.builtin_functions.common.mix_*_b*
(Note: These tests still fail because of constant expression issues)

Change-Id: I79b353933cb450516b8678b1fdaeabe60417e9a7
Reviewed-on: https://chromium-review.googlesource.com/271751Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Olli Etuaho <oetuaho@nvidia.com>

d94f6647

Combine D3D11 FL10_0+ and FL9_3 format support tables · 3f65f2e9

authored May 14, 2015

BUG=angleproject:1002

Change-Id: I4305db01c0ca40fd0ccf0dd441744b6184d5e945
Reviewed-on: https://chromium-review.googlesource.com/270770Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Austin Kinross <aukinros@microsoft.com>

3f65f2e9

Combine D3D11 FL10_0+ and FL9_3 LoadFunctionMaps · 0ab152ab

authored May 14, 2015

BUG=angleproject:1002

A 'LoadFunction' is used to load data of a given <GL format, GL type> pair into
a specific DXGI_FORMAT. Before this change, BuildD3D11LoadFunctionMap mapped
<GL format, GL type> pairs to LoadFunctions. This meant each
<GL format, GL type> pair could only loaded using one LoadFunction, and
therefore could only be loaded into one specific DXGI_FORMAT.

We now want to map <GL format, GL type, DXGI_FORMAT> triples to LoadFunctions.
This allows us to use different DXGI_FORMATS for each <GL format, GL type> pair,
depending on the capabilities of the underlying D3D11 device.

Change-Id: I3b5e24c4bf98f18360ddd45b2022a90b5000f866
Reviewed-on: https://chromium-review.googlesource.com/270561Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Austin Kinross <aukinros@microsoft.com>

0ab152ab

DisplayGLX: rework the GLXFBConfig selection · 9ce754eb

authored May 15, 2015

This allows us to return multiple EGL configs, in particular we can have
configs with no depth-stencil buffer. Also includes a description of the
investigation I did for this patch.

BUG=angleproject:892

Change-Id: I921e254e91e818396dd4c9bd0ee2a43b00a043b9
Reviewed-on: https://chromium-review.googlesource.com/271426Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

9ce754eb

Roll GYP for Windows 10 project generation. · 327c18cd

authored May 15, 2015

BUG=angleproject:1004

Change-Id: Ide2f71f1aee70d0d962ff275ac9bd0f852b218ac
Reviewed-on: https://chromium-review.googlesource.com/271482Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Geoff Lang <geofflang@chromium.org>

327c18cd

eglGetConfigAttrib: handle EGL_NATIVE_VISUAL_ID · f0fce530

authored May 15, 2015

BUG=angleproject:892

Change-Id: Idf5496de7006f8fd0658e9d40a4f16dba00d90ea
Reviewed-on: https://chromium-review.googlesource.com/271473Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

f0fce530

Fix more warnings in samples and tests · 91c2fadf

authored May 15, 2015

These warnings are currently disabled because these files are not with
angle_code set to 1 in the gyp files to avoid gtest's code warnings.
We hope to either fix warnings in gtest or change our build system to
not use warnings for gtest.

BUG=angleproject:1003

Change-Id: I281781b77f92ef8c394dd53fbbd50d525e4da5c9
Reviewed-on: https://chromium-review.googlesource.com/271412Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

91c2fadf

Enable more warnings on GCC and clang · bc99bb6b

authored May 14, 2015

BUG=angleproject:892

Change-Id: I74ca341f29b245f698d1e1ad43149a91db46817f
Reviewed-on: https://chromium-review.googlesource.com/271411Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

bc99bb6b

Roll dEQP. · 0326a3ac

authored May 15, 2015

Includes fixes for some tests, and a new null platform class which
makes generating case lists easier.

BUG=angleproject:998

Change-Id: I1641c579bb63713fccde05e48f0cea31324f7d1b
Reviewed-on: https://chromium-review.googlesource.com/271202Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

0326a3ac

Revert "Clear image descs on Impl errors." · fb074dd8

authored May 19, 2015

Didn't fix the crash, and the crash experiment is complete.

BUG=485543

This reverts commit 96748507.

Change-Id: If7d97901b87506bf306ab5558575d0413c76cb68
Reviewed-on: https://chromium-review.googlesource.com/272150Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

fb074dd8

Set mUsesDiscardRewriting more conservatively · fd3b9be6

authored May 18, 2015

This fixes a regression from recent refactoring of unfolding short-circuiting
operators. mUsesDiscardRewriting used to be set based on detecting patterns
that the AST traverser now overwrites before the code reaches OutputHLSL. With
this patch mUsesDiscardRewriting is set more conservatively, which ensures that
using discard works correctly also in case the input shader contains the same
kind of pattern as the short-circuiting operator unfolding creates.

TEST=WebGL conformance tests on D3D9
BUG=chromium:488552

Change-Id: Iea0103822a268079593b0913f0b6fc54c745c6e3
Reviewed-on: https://chromium-review.googlesource.com/271761Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Olli Etuaho <oetuaho@nvidia.com>

fd3b9be6

18 May, 2015 1 commit

Add constant folding support for min,max and clamp · 274f0709

authored May 05, 2015

This change adds necessary mechanism to support constant folding of
built-ins that take more than one parameter and also adds constant
folding support for min, max and clamp built-ins.

BUG=angleproject:913
TESTS=dEQP tests
(126 tests started passing with this change)
dEQP-GLES3.functional.shaders.constant_expressions.builtin_functions.common.min_*
dEQP-GLES3.functional.shaders.constant_expressions.builtin_functions.common.max_*
dEQP-GLES3.functional.shaders.constant_expressions.builtin_functions.common.clamp_*

Change-Id: Iccc9bf503a536f2e3c144627e64572f2f95db9db
Reviewed-on: https://chromium-review.googlesource.com/271251Reviewed-by: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Olli Etuaho <oetuaho@nvidia.com>

274f0709

15 May, 2015 11 commits

WindowSurfaceGLX: set the border pixel attribute · 8befcff5

authored May 15, 2015

This is so we can use a different visual depth than our parent window so
that we can create a subwindow whichever the parent's visual.

BUG=angleproject:892

Change-Id: I9bab676761b1f80fad1599adb8e97a3ec710f67f
Reviewed-on: https://chromium-review.googlesource.com/271425Reviewed-by: Kenneth Russell <kbr@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

8befcff5

Enable postSubBuffer on D3D11 with DXGI 1.2. · d176dbd7

authored May 07, 2015

BUG=147291

Change-Id: Iafddb83b4949168f1c59272a99ae0244d86ce3f9
Reviewed-on: https://chromium-review.googlesource.com/270572Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Sunny Sachanandani <sunnyps@chromium.org>

d176dbd7

Revert "Call XSync in the GLX EGL impl if we have a new X server connection" · 9c0a4218

authored May 15, 2015

broke linux builders

This reverts commit cedb8ca2.

Change-Id: I57143ec207b454bce7498a4d33f6d54464b55ab8
Reviewed-on: https://chromium-review.googlesource.com/271424Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Tested-by: Zhenyao Mo <zmo@chromium.org>

9c0a4218

Call XSync in the GLX EGL impl if we have a new X server connection · cedb8ca2

authored May 14, 2015

This prevents races between the application's connection and ANGLE's one
that could lead to X errors crashing the program.

BUG=angleproject:892

Change-Id: Ic09462e76724d1a32c35b32708de97258d99241c
Reviewed-on: https://chromium-review.googlesource.com/271178Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

cedb8ca2

Clear image descs on Impl errors. · 96748507

authored May 11, 2015

This is a speculative fix for a possible crash, when we could get
out-of-sync between the Impl and the GL state.

BUG=485543

Change-Id: Id26475f1aae7da05b5ffd802c4a7600e21b40315
Reviewed-on: https://chromium-review.googlesource.com/270194Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

96748507

Add -lrt and -ldl to link flags on Linux builds · 4b6945a4

authored May 14, 2015

BUG=angleproject:892

Change-Id: Ic06b493cbfdc061565d94d98aabfc3b1f2d882c6
Reviewed-on: https://chromium-review.googlesource.com/271275Reviewed-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

4b6945a4

translator: Reject shaders that use both FragColor+FragData. · 14e95b38

authored May 07, 2015

*re-land with fix for unused var in release*

We checked this at link time in the D3D back-end, but this
restriction applies to all shaders.

TEST=dEQP-GLES2.functional.shaders.fragdata.*
BUG=angleproject:995
BUG=478572

Change-Id: I63258f4de47e658812822f31601cc235f48c0826
Reviewed-on: https://chromium-review.googlesource.com/271470Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

14e95b38

Unfold sequence operator when operations inside need unfolding · 0ea959f4

authored May 13, 2015

Unfolding the sequence operator ensures correct evaluation order when parts of
the expression need to get hoisted out of the expression to the surrounding
scope.

BUG=angleproject:1001
TEST=WebGL conformance tests

Change-Id: Ifd3093c9fb63d3e9842ebb4b9531b2f46875ad8a
Reviewed-on: https://chromium-review.googlesource.com/270660Tested-by: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

0ea959f4

Defer executing if statements in the global scope · d81ed841

authored May 12, 2015

Unfolding of short-circuiting operators (ternary and logical operators) may
create if statements in the global scope, which is not valid HLSL. Use existing
deferred global initialization function to defer execution of if statements in
the global scope.

TEST=WebGL conformance tests
BUG=angleproject:819

Change-Id: I2b0afcc6824dab6bb87eb6abed609e75b1384dab
Reviewed-on: https://chromium-review.googlesource.com/270461Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Olli Etuaho <oetuaho@nvidia.com>

d81ed841

Make UnfoldShortCircuit to change AST instead of writing output · a6f22096

authored May 08, 2015

This is needed to make way for further AST transformations to handle array
expressions that need to work correctly together with unfolding short-
circuiting operators. This also improves the maintainability of HLSL output
by isolating the unfolding into a separate compilation step.

The new version of UnfoldShortCircuit traverser will traverse the tree until
an expression that needs to be unfolded is encountered. It then unfolds it and
gets reset. The traverser will be run repeatedly until no more operations to
unfold are found. This helps with keeping the traverser's design relatively
simple.

All declarations are separated to single declarations before short-circuit
unfolding is run. Previously OutputHLSL already output every declaration
separately.

BUG=angleproject:960
TEST=WebGL conformance tests, angle_unittests, angle_end2end_tests

Change-Id: Id769be396adbd4c0223e418980dc464dd855f019
Reviewed-on: https://chromium-review.googlesource.com/270460Tested-by: Olli Etuaho <oetuaho@nvidia.com>
Reviewed-by: Jamie Madill <jmadill@chromium.org>

a6f22096

Revert "translator: Reject shaders that use both FragColor+FragData." · 531e3d22

authored May 15, 2015

The change introduced a warning in Windows release build.

This reverts commit b8e3a568.

Change-Id: I77bbc35876043c9a164aa2665965f5189ee90052
Reviewed-on: https://chromium-review.googlesource.com/271430Reviewed-by: Olli Etuaho <oetuaho@nvidia.com>
Tested-by: Olli Etuaho <oetuaho@nvidia.com>

531e3d22

14 May, 2015 9 commits

Fix "unused private variable" warning in ParseContext · e8675195

authored May 14, 2015

BUG=angleproject:892

Change-Id: I3d0c23f1c2d43477c8b4ff28a6db470b3491de0b
Reviewed-on: https://chromium-review.googlesource.com/271276Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

e8675195

Make X11Window::setVisible block until the window is mapped. · 3730f644

authored May 14, 2015

That way the code calling setVisible knows it can safely create a
framebuffer for this window without having its content be undefined.

BUG=angleproject:830

Change-Id: Ia694f50ac20e1d510d0c0f899226ed241d9a5a3a
Reviewed-on: https://chromium-review.googlesource.com/271156Reviewed-by: Kenneth Russell <kbr@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

3730f644

FunctionsGLX: add support for glXWaitX/GL · d64a38b3

authored May 14, 2015

The functions are not used in the code but are useful during development
to try to see where race condition with the X server are.

BUG=angleproject:892

Change-Id: Ie1267e46ac251bd9ac62e2a146db68ad84644fb8
Reviewed-on: https://chromium-review.googlesource.com/271177Reviewed-by: Geoff Lang <geofflang@chromium.org>
Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

d64a38b3

FunctionsGLX: make most functions take the display implicitly · 582173d9

authored May 13, 2015

In later CLs FunctionsGLX will need to store the X display so we take
advantage of it to reduce the verbosity of the other GLX files slightly.

BUG=angleproject:892

Change-Id: I42ea00d0a37055e5e0752a860978b8ef5afb7a0b
Reviewed-on: https://chromium-review.googlesource.com/271163Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

582173d9

translator: Reject shaders that use both FragColor+FragData. · b8e3a568

authored May 07, 2015

We checked this at link time in the D3D back-end, but this
restriction applies to all shaders.

TEST=dEQP-GLES2.functional.shaders.fragdata.*
BUG=angleproject:995
BUG=478572

Change-Id: I99111cc6aa05b9352693f9c3b5bc70d56c9842d4
Reviewed-on: https://chromium-review.googlesource.com/269846Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Tested-by: Jamie Madill <jmadill@chromium.org>

b8e3a568

Make TParseContext a class, with private data. · 6e06b1f4

authored May 14, 2015

*re-land with build fix*

BUG=angleproject:995

Change-Id: I67d3ded8f6c705b54fb372857e07ce1a86b58475
Reviewed-on: https://chromium-review.googlesource.com/271162Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Zhenyao Mo <zmo@chromium.org>

6e06b1f4

Use a sparse histogram for D3D11CreateDevice errors. · 558d5eb4

authored May 14, 2015

This avoid us needing to capture an enum value for every possible
error code. Leave the existing enum values in for now.

BUG=477701

Change-Id: Iccaa4d9b732fe4d56432e17578421cba11bcf9ce
Reviewed-on: https://chromium-review.googlesource.com/271159Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Brandon Jones <bajones@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

558d5eb4

Add collection of D3D11 and DXGI stats. · 72537ef1

authored May 14, 2015

Will require a chromium-side patch to enable boolean histograms.

BUG=angleproject:966
BUG=436191

Change-Id: I3026c73894fbbff521503b377955816a8e831df8
Reviewed-on: https://chromium-review.googlesource.com/271174Tested-by: Jamie Madill <jmadill@chromium.org>
Reviewed-by: Brandon Jones <bajones@chromium.org>
Reviewed-by: Geoff Lang <geofflang@chromium.org>

72537ef1

Fix typo in a comment in PbufferTest · 21ce9b02

authored May 14, 2015

BUG=angleproject:892

Change-Id: I8383b0b061f613ee3592c3d7cd41135164223604
Reviewed-on: https://chromium-review.googlesource.com/271158Reviewed-by: Jamie Madill <jmadill@chromium.org>
Tested-by: Corentin Wallez <cwallez@chromium.org>

21ce9b02