src


Log

Author Commit Date CI Message
Geoff Lang aef1663f 2015-06-05T16:09:55 Add missing GL_STENCIL_INDEX8 format to the formatutilsgl table. Fixes conformance/renderbuffers/framebuffer-object-attachment.html BUG=angleproject:884 Change-Id: I421b563fa8549eb2edc42a9ac1c798fc457256ee Reviewed-on: https://chromium-review.googlesource.com/275690 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Brandon Jones <bajones@chromium.org> Tested-by: Geoff Lang <geofflang@chromium.org>
Corentin Wallez 91f828e3 2015-07-13T11:08:19 LineLoopTest: Fix the test not rendering anything. BUG=angleproject:1063 Change-Id: Id0772b5c8479150a1e736aa7cd35ec6444753278 Reviewed-on: https://chromium-review.googlesource.com/285126 Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org>
Austin Kinross 61d46d87 2015-07-09T13:47:09 Fix memory layout of imageformats.h's 16-bit structs Change-Id: I87c363f68394d868d2a72e84bfdfd64db0c6e1d0 Reviewed-on: https://chromium-review.googlesource.com/284652 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Austin Kinross <aukinros@microsoft.com>
Jamie Madill 2ee580f1 2015-07-10T11:14:58 D3D11: Futher optimize input layout cache. *re-land with fix for matrix attributes* Using the new vertex format type enum, we can shrink the size of the input layout tables and reduce draw call overhead further. BUG=angleproject:959 Change-Id: I6d8ad78a003c41f40e7e1caa5972838f8ff4fce8 Reviewed-on: https://chromium-review.googlesource.com/284811 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Kenneth Russell <kbr@chromium.org> Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Cooper Partin 75da1974 2015-06-16T15:03:14 Changed d3d11 desktop renderer to use d3d9 debug annotation interfaces. The D3D11 renderer must choose the D3D9 debug annotator because the D3D11 interface method ID3DUserDefinedAnnotation::GetStatus on desktop builds doesn't work with the Graphics Diagnostics tools in Visual Studio 2013. The D3D9 annotator works properly for both D3D11 and D3D9. Incorrect status reporting can cause ANGLE to log unnecessary debug events. Change-Id: I9a31c63cbc506904eb39577826fc4df8d503f03a Reviewed-on: https://chromium-review.googlesource.com/278162 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Cooper Partin <coopp@microsoft.com> Tested-by: Cooper Partin <coopp@microsoft.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Minmin Gong 03786d77 2015-05-27T11:22:30 After a create-on-bind, the subsequent glGen* should not waste handle space. Consider this calling sequence: glBindTexture(GL_TEXTURE_2D, 3); // create-on-bind glGenTextures(1, tex); Before this fix, mUnallocatedList in HandleAllocator is in reverse sorted order. It's split into (4, MAX_UINT) and (1, 3). The glGen gets tex = 4, and handles 1 and 2 will not be allocated until it reaches MAX_UINT. With this fix, the elements in mUnallocatedList is in sorted order (1, 3) and (4, MAX_UINT). So tex is 1, and we can have all rest handles. Change-Id: If408ea92f7c243791119fe386e3e0ea27954e55c Reviewed-on: https://chromium-review.googlesource.com/273886 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Minmin Gong <mgong@microsoft.com>
Jamie Madill a8bb9cdc 2015-07-13T09:59:53 Lazily load the HLSL compiler DLL. If the user is loading resources straight from the disk with the binary shader cache, they shouldn't need to load the compiler DLL at all. Note: might be a good idea to track more statistics here. BUG=angleproject:1014 Change-Id: Ie75dcce3eb249347c545f1bbca7d0977df7addee Reviewed-on: https://chromium-review.googlesource.com/282554 Reviewed-by: Kenneth Russell <kbr@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Olli Etuaho e0f623a4 2015-07-10T11:58:30 Check that block instance name isn't reserved BUG=angleproject:1060 TEST=dEQP-GLES3.functional.shaders.uniform_block.invalid.* Change-Id: I6de3aa0f714d585e7821798744530f3e51ddee7f Reviewed-on: https://chromium-review.googlesource.com/284259 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Zhenyao Mo <zmo@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com>
Corentin Wallez 7ef89571 2015-07-10T11:08:03 EGLContextCompatibility: suppress test on Linux non-standalone The tests turn out to fail before we have a context current on AMD so we cannot use isAMD and have to resort to disabling for all Linuxes on the FYI bots. Using MAYBE like Chrome doesn't seem to work with TEST_P. BUG=angleproject:892 Change-Id: Ibabd2c359ce55c7677af0f8bd5723f845935f2be Reviewed-on: https://chromium-review.googlesource.com/284872 Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org> Reviewed-on: https://chromium-review.googlesource.com/284749
Geoff Lang c3b57529 2015-07-10T20:22:42 Revert "EGLContextCompatibility: suppress test on Linux non-standalone" Causing gyp generation failures on FYI bots: https://build.chromium.org/p/chromium.gpu.fyi/builders/GPU%20Win%20Builder/builds/30490 This reverts commit af98f24aece19c3231d9e7300bbab5c323109b23. Change-Id: I0b3ea26ee02591588bda14676211658b42135cf2 Reviewed-on: https://chromium-review.googlesource.com/284910 Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Geoff Lang <geofflang@chromium.org>
Corentin Wallez af98f24a 2015-07-10T11:08:03 EGLContextCompatibility: suppress test on Linux non-standalone The tests turn out to fail before we have a context current on AMD so we cannot use isAMD and have to resort to disabling for all Linuxes on the FYI bots. Using MAYBE like Chrome doesn't seem to work with TEST_P. BUG=angleproject:892 Change-Id: Ie964168eb0b1d830dd1fd0988869adf39bc2309a Reviewed-on: https://chromium-review.googlesource.com/284872 Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org>
Dmitry Skiba 01971113 2015-07-10T14:54:00 Cache TTypes. *re-land with build fix for Win/Release* This change saves us ~70KiB per compiler with just ~3KiB of cache. BUG=492725 Change-Id: I4382c55b2480f70b00c5d117fcb7e0c51d0dfbb4 Reviewed-on: https://chromium-review.googlesource.com/284735 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Dmitry Skiba <dskiba@google.com> Reviewed-by: Alexis Hétu <sugoi@chromium.org>
Jamie Madill 53b76107 2015-07-10T18:52:58 Revert "Cache TTypes." Unused variable warning in Release: warning C4189: 'MaxEnumValue' : local variable is initialized but not referenced I'll handle fixing this. BUG=492725 This reverts commit b25d14e4b39ee273fad3bb585c3a3a929b3f794c. Change-Id: I502fc5288d5d3c48ecd43f84acdf66b7e300ad22 Reviewed-on: https://chromium-review.googlesource.com/284863 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Corentin Wallez c70e1208 2015-07-10T10:53:01 egl::MakeCurrent: report platform's makeCurrent errors This caused the context to be current from ANGLE's point of view while it wasn't from the driver's point of view. BUG=angleproject:892 Change-Id: I0719b4702721916e9eb3a9f4cafc7a86dbc7a625 Reviewed-on: https://chromium-review.googlesource.com/284871 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org>
Nico Weber 44897140 2015-07-10T09:50:00 clang/win: Fix -Wunused-private-field warnings in debug builds. No intended behavior change. BUG=505317 Change-Id: I5dfa1d67715b18133f3373ca00a6d1d96b678043 Reviewed-on: https://chromium-review.googlesource.com/284850 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Nico Weber <thakis@chromium.org>
Alexis Hetu 07e57df7 2015-06-16T16:55:52 Added non square matrix construction Added new construction operations for non square matrices, along with the required changes to the related translation functions. Change-Id: I04ae7d4b2d1bb363b35d088cea45c0e7c4bc8a13 Reviewed-on: https://chromium-review.googlesource.com/277729 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Geoff Lang <geofflang@chromium.org>
Dmitry Skiba b25d14e4 2015-06-23T17:43:14 Cache TTypes. This change saves us ~70KiB per compiler with just ~3KiB of cache. BUG=492725 Change-Id: I7b8053c032c9c68f646162baf61abf6db3afe52e Reviewed-on: https://chromium-review.googlesource.com/284003 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Jamie Madill 8e34494f 2015-07-09T14:22:07 Share data between VertexArray and Impl. Using the same design as for the Framebuffer::Data helper, we can use a struct to share between the object and the Impl. This also gives the Impl access to the maxEnabledAttrib, and saves some duplicated storage. BUG=angleproject:1040 TEST=WebGL CTS, end2end_tests, unittests Change-Id: I55c91e8a5f3dcae302cab441182320aafd5375ef Reviewed-on: https://chromium-review.googlesource.com/283930 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Olli Etuaho 0d208976 2015-07-10T15:45:09 Revert "D3D11: Futher optimize input layout cache." This seems to have broken some dEQP tests in debug build, possibly uncovering a previously undetected bug. TEST=dEQP-GLES3.functional.shaders.linkage.varying.basic_types.* This reverts commit d2328a5246d8e0ab8a3dde1b30b80ce5e161e6f3. Change-Id: I9f082d7a6fca38b4c076fecc342ac40d5416ebef Reviewed-on: https://chromium-review.googlesource.com/284780 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com>
Olli Etuaho fb6ab2ce 2015-07-09T20:55:28 Matrix layout qualifiers are valid for non-matrices Uniform block members may be structs containing matrices, and there's no explicit mention in the spec that the matrix qualifiers would only be valid for matrix fields. Only generate a warning when an extraneous matrix layout qualifier is used. TEST=dEQP-GLES3.functional.shaders.*layout* (all pass with this patch) BUG=angleproject:1058 Change-Id: Ifb2f2392d84100d997cc2d68fd55e9ad17a17cea Reviewed-on: https://chromium-review.googlesource.com/284112 Reviewed-by: Zhenyao Mo <zmo@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com>
Nico Weber cb8ff7fb 2015-07-09T21:52:54 clang/win: Fix a -Wreorder warning ..\..\third_party\angle\src\libANGLE\renderer\d3d\d3d11\SwapChain11.cpp(30,7) : error: field 'mDepthStencilRenderTarget' will be initialized after field 'mPassThroughResourcesInit' [-Werror,-Wreorder] mDepthStencilRenderTarget(this, renderer, true), ^ While here, also reorder the fields that are initialized in the constructor to match the order the fields are declared in, to make it easier to see if any are missing. No intended behavior change. BUG=505304 Change-Id: If48195c84161868d33d4361088d3ad22b7a0fc31 Reviewed-on: https://chromium-review.googlesource.com/284682 Reviewed-by: Nico Weber <thakis@chromium.org> Tested-by: Nico Weber <thakis@chromium.org>
Jamie Madill 06c5a646 2015-07-09T14:48:12 Suppress dEQP failure, only showing on the bot. Unsure why this is failing: functional.shaders.struct.local.dynamic_loop_struct_array_fragment BUG=angleproject:1059 Change-Id: I72e05269f831769a763c44354b95a7224edb72fa Reviewed-on: https://chromium-review.googlesource.com/284636 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Jamie Madill d2328a52 2015-07-09T11:30:30 D3D11: Futher optimize input layout cache. Using the new vertex format type enum, we can shrink the size of the input layout tables and reduce draw call overhead further. BUG=angleproject:959 Change-Id: Iea595fa64edbbd91f669138dfdeb9d2543b83929 Reviewed-on: https://chromium-review.googlesource.com/277291 Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Kenneth Russell <kbr@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Jamie Madill 0342253e 2015-07-09T11:30:29 D3D11: Refactor vertex format tables. Use switch statements based on the format type enum, instead of a map from the old VertexFormat structure to the info payload. This saves on lookup and initialization time. BUG=angleproject:959 Change-Id: I6d1bf282c88f67272ca067cb64bc71e77824a12e Reviewed-on: https://chromium-review.googlesource.com/277290 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Olli Etuaho b5841ef2 2015-07-08T17:25:22 Disallow texture lookup functions in global initializers Do this by simply disallowing all function calls in global initializers - all built-in math ops are handled as other ops, not function calls. This change was tested extensively with popular WebGL content, with no regressions found. TEST=angle_unittests BUG=angleproject:988 Change-Id: Id1107fa294ae4012d5dd3949539d0b7b4cd21943 Reviewed-on: https://chromium-review.googlesource.com/283703 Reviewed-by: Zhenyao Mo <zmo@chromium.org> Tested-by: Olli Etuaho <oetuaho@nvidia.com>
Corentin Wallez e9cc78a1 2015-07-06T07:17:00 Make Buffer11::mBuffersStorages an array instead of a map BUG=angleproject:516 Change-Id: I7f969435a89c0063b504628d2ec3321cc02df7ce Reviewed-on: https://chromium-review.googlesource.com/283243 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org>
Corentin Wallez bc3b5e63 2015-07-07T10:08:05 D3D11: map index buffers only when needed. Before this patch, index buffers where always mapped in case we needed the index data for the indexed point sprites workaround. This patch makes it so the index buffer is only mapped in this case, when we need to stream index data or when we need to fill the static copies of the index buffers. This make the memory usage of http://alteredqualia.com/xg/examples/mammoth.html go down from 41MB to 28MB. BUG=angleproject:516 Change-Id: I937506d06fd6f074ef2120469dbd235e20245fca Reviewed-on: https://chromium-review.googlesource.com/283626 Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org>
Dmitry Skiba 58832204 2015-07-06T16:11:13 Calculate TFunction::mangledName lazily. This saves us ~15KiB per compiler instance on 32-bit Android. BUG=492725 Change-Id: I9db2998482941b5ab2eb5cb4925cbd1fb58ab4be Reviewed-on: https://chromium-review.googlesource.com/283584 Reviewed-by: Zhenyao Mo <zmo@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Arun Patole 0c5409f8 2015-07-08T15:17:53 Support constant folding of fragment processing built-ins This change adds constant folding support for following derivative functions: - dFdx, dFdy and fwidth. As per spec, derivatives of constant arguments should be 0, so this change just sets result components to zero when the above operations are performed on constant argument. BUG=angleproject:913 TEST=dEQP Tests dEQP-GLES3.functional.shaders.constant_expressions.builtin_functions. fragment_processing.* (All 12 tests started passing with this change.) Change-Id: I67b7c5abd970b740e15e3bd1ee56721e81bbecc4 Reviewed-on: https://chromium-review.googlesource.com/283759 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Arun Patole 551279e5 2015-07-07T18:18:23 Support constant folding of more common built-ins This change adds const folding support for below mentioned common built-ins: - isnan, ininf, floatBitsToInt, floatBitsToUint, intBitsToFloat and uintBitsToFloat BUG=angleproject:913 TEST=angle_unittests(new: MathUtilTest.isNan/inInf), dEQP Tests dEQP-GLES3.functional.shaders.constant_expressions.builtin_functions.common.* (20 more tests started passing with this change) Change-Id: Ifdedb251cd8b183b4999314c0f5de857bc20702f Reviewed-on: https://chromium-review.googlesource.com/283767 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Corentin Wallez bcab6426 2015-07-06T07:35:14 Fix typo in Buffer11::getContantBufferRangeStorage BUG=angleproject:516 Change-Id: I907ec41ce5691300d17ca5b08a62b0d4579b01e5 Reviewed-on: https://chromium-review.googlesource.com/283244 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org>
Corentin Wallez 2e62a9a8 2015-07-03T14:13:50 Simplify the flow of IndexDataManager::prepareIndexData This will simplify the work on follow up CLs a lot BUG=angleproject:516 Change-Id: Ie8374ec28fce2e01e76c711a4d280e1ed004a21f Reviewed-on: https://chromium-review.googlesource.com/283380 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Corentin Wallez 0b5a4a49 2015-07-08T11:59:46 EGLContextCompatibilityTest: check for suppression only when a context is bound. Otherwise glGetString(GL_RENDERER) returns "ANGLE" and doesn't provide information on the vendor. BUG=angleproject:892 Change-Id: I01899ca78656aaa8db860829c78bf3c61c355eae Reviewed-on: https://chromium-review.googlesource.com/284051 Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org>
Jamie Madill d3dfda2b 2015-07-06T08:28:49 Refactor how we store vertex formats. Instead of storing a vertex format as a struct with the full info, instead use an enum, and look up the info when we need it. This saves a lot of constructor initialization time, operator comparison time, and storage. It also will allow us to look up D3D format info more quickly. BUG=angleproject:959 Change-Id: I202fd1ea96981073bc1b5b232b1ec3efa91485cb Reviewed-on: https://chromium-review.googlesource.com/277289 Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Jamie Madill 2810a20c 2015-07-07T14:23:58 Fix the InputLayoutCache test on Windows 8. On some configs, this would use an RGBA4 16-bit backbuffer, which is not enough resolution for the test. Fix this by requiring 32-bit backbuffers. We might also want to figure out if we want to default to 16-bit on Windows 8 desktop, which seems less than ideal. BUG=angleproject:959 Change-Id: I53b3a9ef6cbecf76c7c764623e5eb1f454a6b72d Reviewed-on: https://chromium-review.googlesource.com/283773 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Austin Kinross <austin.kinross@gmail.com> Reviewed-by: Cooper Partin <coopp@microsoft.com> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Daniel Cheng fda3ec74 2015-07-08T08:22:04 Fix another clang warning with -Wmissing-braces in ANGLE. Clang warns if there are missing braces around a subobject initializer. The most common idiom that triggers this is: STRUCT s = {0}; which can be more simply written as: STRUCT s = {}; BUG=505297 Change-Id: Ie9f9037a1ba4b6303daa30424cdc24c1ecf18896 Reviewed-on: https://chromium-review.googlesource.com/284160 Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Geoff Lang <geofflang@chromium.org>
Arun Patole ab2b9a23 2015-07-06T18:27:56 Refactor unary operation constant folding code This change splits TIntermConstantUnion::foldUnary into two functions: - foldUnaryWithDifferentReturnType: - Handles constant folding of operations where the return type has a different number of components compared to the operand type. - foldUnaryWithSameReturnType: - Handles constant folding of unary operations where the return type is the same as operand type. BUG=angleproject:913 TEST=angle_unittests, dEQP Tests (dEQP-GLES3.functional.shaders.constant_expressions.builtin_functions.*) Change-Id: I675891138d4e17fd2390c03e9f710e0be0b7c7b6 Reviewed-on: https://chromium-review.googlesource.com/283464 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Arun Patole cdfa8f50 2015-06-30T17:48:25 Constant fold float pack/unpack functions This change adds constant folding support for following floating point pack and unpack functions: - packSnorm2x16, unpackSnorm2x16, packUnorm2x16, unpackUnorm2x16, packHalf2x16, and unpackHalf2x16. BUG=angleproject:913 TEST=angle_unittests(new: MathUtilTest.packAndUnpack*), dEQP Tests dEQP-GLES3.functional.shaders.constant_expressions.builtin_functions.float_pack_unpack.* (all 10 tests started passing with this change) Change-Id: I2b69faebee4127e5e13900b3a9485b7145950277 Reviewed-on: https://chromium-review.googlesource.com/282683 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Olli Etuaho 846fe05f 2015-07-07T17:41:21 Disable using globals as l-values in global initializers It should not be possible to use globals as l-values in global initializers. This change was tested extensively with popular WebGL content, with no regressions found. TEST=angle_unittests BUG=angleproject:988 Change-Id: I21ab731eb1d92aeae25795856ccae280792ad1f0 Reviewed-on: https://chromium-review.googlesource.com/283910 Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Kenneth Russell 6affeef6 2015-07-07T15:16:49 Suppress angle_end2end_tests failures on Linux/NVIDIA and Linux/AMD. Verified the NVIDIA suppression locally; assuming the AMD suppression will take effect on the bot. Thanks to geofflang@ and cwallez@ for the suggestion on the approach to take. BUG=angleproject:892 Change-Id: Ida4b6c1395185b0dfb0ea890395e35cb6db7ccd5 Reviewed-on: https://chromium-review.googlesource.com/283974 Tested-by: Kenneth Russell <kbr@chromium.org> Reviewed-by: Brandon Jones <bajones@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Jamie Madill 2f232370 2015-07-07T21:57:55 Revert "Cache TTypes." Lots of compile errors. BUG=492725 This reverts commit 5377720aae042c5bfae0e8a37032c90dc3ab78cf. Change-Id: I64889b99b1f1f48d39b87ebb668f6a32a3abac45 Reviewed-on: https://chromium-review.googlesource.com/283945 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Corentin Wallez e9336c0d 2015-07-07T16:46:26 EGLSurfaceTest: run d3d specific tests only on d3d. This is to help with the FYI bot and pending a refactor to use our test instantiation. BUG=angleproject:892 Change-Id: I56e772d7988a2172ad2384bdf408a8eaeab7129a Reviewed-on: https://chromium-review.googlesource.com/283971 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Kenneth Russell <kbr@chromium.org> Reviewed-by: Cooper Partin <coopp@microsoft.com> Tested-by: Corentin Wallez <cwallez@chromium.org>
Dmitry Skiba 5377720a 2015-06-23T17:43:14 Cache TTypes. This change saves us ~70KiB per compiler with just ~3KiB of cache. BUG=492725 Change-Id: I2ee748f7271b1afe0834083cbff6f0b84104aa53 Reviewed-on: https://chromium-review.googlesource.com/281411 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Dmitry Skiba <dskiba@google.com>
Austin Kinross da056468 2015-06-16T17:00:11 For WinRT, add EGLRenderSurfaceScaleProperty to specify swapchain scaling factor Change-Id: Iede6682306082346cf2a3a55fe58f732b7a453df Reviewed-on: https://chromium-review.googlesource.com/277931 Tested-by: Austin Kinross <aukinros@microsoft.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Olli Etuaho ce39f6ff 2015-07-06T15:25:19 Disallow user-defined function calls in global variable init Generate an error message when an user-defined function call is found in a global variable initializer. Even before this patch, the call graph already marked functions that were only called from the global scope as unused. This change was tested extensively with popular WebGL content, with no regressions found. TEST=angle_unittests BUG=angleproject:988 Change-Id: Iec1b16d2af386f1e5c383f86926d80cef553b694 Reviewed-on: https://chromium-review.googlesource.com/283291 Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Jamie Madill 8fc55c8e 2015-06-30T11:04:18 Lazily load SwapChain11 pass-through resources. This allows Chromium to avoid initializing a whole bunch of resources and a shader when it collects GPU information. This should speed up startup. BUG=angleproject:1014 Change-Id: Id1e777c6ea86d66ad812540f9d8494744cda64de Reviewed-on: https://chromium-review.googlesource.com/282553 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Brandon Jones <bajones@chromium.org> Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Jamie Madill e62cedd0 2015-06-30T11:03:54 Load Clear11 resources lazily. Again this saves on startup time and memory use. BUG=angleproject:1014 Change-Id: I3890dc2ea50ff4e4f1ff9d7cb223e1ab57c330ce Reviewed-on: https://chromium-review.googlesource.com/282552 Reviewed-by: Brandon Jones <bajones@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Jamie Madill 86a07aec 2015-06-30T11:03:18 Add LazyInputLayout to defer loading input layouts. Lazily loading input layouts can also save memory and startup time. There are several cases where we don't need these resources until later, or sometimes, at all. BUG=angleproject:1014 Change-Id: I4e0d45353b5d3969bd1ed86ad26f47d34e542ed5 Reviewed-on: https://chromium-review.googlesource.com/282551 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Brandon Jones <bajones@chromium.org>
Jamie Madill 531d5f45 2015-07-06T08:26:45 D3D11: Refactor InputLayout cache. *re-land with fix for cache overflowing* Using a much more compact input layout structure allows us to save quite a bit of time comparing input layouts, or computing hashes. A subsequent patch shrinks the size of the structure further. BUG=angleproject:959 Change-Id: If240bb7c84d78fc8c9fb6f9049bf71d8a81c97c6 Reviewed-on: https://chromium-review.googlesource.com/283227 Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Corentin Wallez 01e4bfe7 2015-07-02T15:07:13 Update dEQP test expectations with Linux failures BUG=angleproject:1051 Change-Id: I5f397e230f0e56b1e6832c3dae5ae5c0c9a8b7c4 Reviewed-on: https://chromium-review.googlesource.com/283226 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Jamie Madill 16f99b74 2015-07-02T14:09:06 Fix warnings with code analysis tools. Variable shadowing and a few other non-bug warnings. BUG=None Change-Id: Ibf62d6e7fef91fc213a53a6902f494147ed30ebc Reviewed-on: https://chromium-review.googlesource.com/283223 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Arun Patole 56eee79f 2015-07-01T15:23:54 Move Float16ToFloat32 implementation to common Float16ToFloat32 is now required by constant folding of float pack/unpack built-ins as well, moving it to commom so that it is available in libangle_common instead of libANGLE. BUG=angleproject:913 TEST=angle_unittests Change-Id: I1e23bb98c1b725c79f933a66ddb9986133fec19c Reviewed-on: https://chromium-review.googlesource.com/282961 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com>
Ben Wells e6e8c7c5 2015-07-01T16:48:31 Fix 'unused private field' warnings for clang. BUG=505317 Change-Id: I731b6e5ae5d5eba50f556314eeeebc6a3d29fc7a Reviewed-on: https://chromium-review.googlesource.com/282932 Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Geoff Lang <geofflang@chromium.org>
Cooper Partin 1bd7dd45 2015-06-11T08:58:53 Added 16-bit formats to GLConfigs to reduce offscreen texture memory usage. BUG=angleproject:1042 Change-Id: I7024bd47601a21b08cafbf6460d512151b53d035 Reviewed-on: https://chromium-review.googlesource.com/277002 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Corentin Wallez 54c34e06 2015-07-02T15:06:55 Fix compilation on Linux BUG=angleproject:892 Change-Id: I135e356e1ec78c95e69a260bd0576e5285e759b5 Reviewed-on: https://chromium-review.googlesource.com/283225 Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org>
Cooper Partin 558f2b5a 2015-06-02T09:34:11 Added emulated indexed pointsprite rendering support. This emulation is for renderers that do not support Geometry Shaders. BUG=angleproject:949 Change-Id: I7acf003e83ea6661f10a703486e6d07eb28786f8 Reviewed-on: https://chromium-review.googlesource.com/274851 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Cooper Partin <coopp@microsoft.com>
Corentin Wallez ccda698b 2015-07-02T08:26:22 Remove redundant chdir in angle_deqp_test_main Fixes the tests not running directly from the VS UI. BUG=angleproject:1051 Change-Id: Ia299d8e7ad676330902725c114e8f5766db49d82 Reviewed-on: https://chromium-review.googlesource.com/283158 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org>
Corentin Wallez c491e0fb 2015-07-02T10:41:11 Fix command line being ignored in non-gtest dEQP tests BUG=angleproject:1051 Change-Id: If10abaea31ad29af9f8d438403f67b9dfbfd7e45 Reviewed-on: https://chromium-review.googlesource.com/283167 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org>
Corentin Wallez f5cb6aa5 2015-06-30T11:31:54 Port gpu_test_expectations and angle_deqp_googletest to Linux BUG=angleproject:1051 Change-Id: I46c3a5651d36750558735eb626ed5c9424b250b4 Reviewed-on: https://chromium-review.googlesource.com/282596 Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org>
Jamie Madill 2c70e64b 2015-07-02T15:09:58 Revert "D3D11: Refactor InputLayout cache." Seems to be failing maps_pixel_test on the bots. BUG=angleproject:959 This reverts commit 53fd1552ba7dbf91a726fce139d748860bd8a5f5. Change-Id: I1f2593caca58f701d284993d2ee313846806491b Reviewed-on: https://chromium-review.googlesource.com/283157 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Jamie Madill 53fd1552 2015-07-02T07:26:19 D3D11: Refactor InputLayout cache. Using a much more compact input layout structure allows us to save quite a bit of time comparing input layouts, or computing hashes. A subsequent patch shrinks the size of the structure further. BUG=angleproject:959 Change-Id: I7a0f10428cc7e96967e60cc085ff2e8779718978 Reviewed-on: https://chromium-review.googlesource.com/277288 Reviewed-by: Kenneth Russell <kbr@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Jamie Madill 03260faa 2015-06-22T13:57:22 D3D11: Optimize checks for active UBOs. We can reduce the amount of iterations we use by using std::vectors for storing UBO indexes, and resizing them to the max active index. This allows us to skip inactive ranges. Reduces draw call overhead. BUG=angleproject:959 Change-Id: I2577e5dbdce0e5ab102ca916eb7f1f35cf03f9ad Reviewed-on: https://chromium-review.googlesource.com/277287 Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Kenneth Russell <kbr@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Daniel Cheng f33ab835 2015-06-30T19:08:16 Fix some clang warnings with -Wmissing-braces in ANGLE. Clang warns if there are missing braces around a subobject initializer. The most common idiom that triggers this is: STRUCT s = {0}; which can be more simply written as: STRUCT s = {}; BUG=505297 Change-Id: Ib4513ee195d43e7c4a4a2f46c328866f0023a2a5 Reviewed-on: https://chromium-review.googlesource.com/282921 Reviewed-by: Kenneth Russell <kbr@chromium.org> Tested-by: Kenneth Russell <kbr@chromium.org>
Dmitry Skiba 7f17a50c 2015-06-22T15:08:39 Store TFunction::returnType by a const pointer. On 32-bit Android this change saves ~30KiB per compiler instance. BUG=492725 Change-Id: I8bea48d57ee7eac0a0ee417035085c0d335aea09 Reviewed-on: https://chromium-review.googlesource.com/281047 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Dmitry Skiba <dskiba@google.com>
Dmitry Skiba efa3d8eb 2015-06-22T14:52:10 Change TSymbolTable::insertBuiltIn() TType* pointers to const. Const types make it possible to implement caching and other optimizations. BUG=492725 Change-Id: I64398bb9effcc909dd052a038acbb5ec0ca730e8 Reviewed-on: https://chromium-review.googlesource.com/281046 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Zhenyao Mo <zmo@chromium.org> Tested-by: Dmitry Skiba <dskiba@google.com>
Jamie Madill 517ccdf8 2015-06-30T11:02:55 Rename DeferredShader to LazyShader. Deferred rendering is a shading technique, and "lazy" is the more correct term. Also rename a few methods so we can use a shared interface for other types of lazy D3D11 resources. BUG=angleproject:1014 Change-Id: I2fd2c823291d414003690066eaec063f1cd0bc4c Reviewed-on: https://chromium-review.googlesource.com/282550 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Zhenyao Mo <zmo@chromium.org> Reviewed-by: Brandon Jones <bajones@google.com>
Corentin Wallez a55e4b7a 2015-06-09T13:30:25 Enable compilation of dEQP tests on Linux BUG=angleproject:1051 Change-Id: I158e59ca581b580825848e831d05064cdcc30d96 Reviewed-on: https://chromium-review.googlesource.com/276202 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org>
Minmin Gong d3e7aa33 2015-06-29T12:25:24 Allow calling HandleAllocator::reserve with MAX_UINT. BUG=angleproject:1052 Change-Id: I03883799ef334f39c9e855a0a4b7af1ab02d5da9 Reviewed-on: https://chromium-review.googlesource.com/282414 Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Minmin Gong <mgong@microsoft.com>
Jamie Madill 006909d6 2015-06-30T12:33:22 Update failing WebGL test expectations. Change-Id: I8c469dad3e6a084e5b37a76ea3a1f86b9a99d613 Reviewed-on: https://chromium-review.googlesource.com/282624 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Austin Kinross 137b1517 2015-06-17T16:14:53 Improve D3D GetProgramBinary chipset validation - Check chipset identifying info before trying to compile shaders - Check device feature level when loading a binary - Use chipset VendorID/DeviceID etc instead of LUID so that program binaries remain valid across system reboots Change-Id: I88ba4543bb990956d1d8fb324abf9784d72950cd Reviewed-on: https://chromium-review.googlesource.com/280428 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Austin Kinross 5a6dec50 2015-06-12T14:53:53 Ensure FencSyncTests::BasicOperations can't get stuck in an infinite loop BUG=angleproject:1053 Change-Id: If6661ce3d87cdf44b1d45911418ecf78b5e6175a Reviewed-on: https://chromium-review.googlesource.com/277370 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Jamie Madill 476682e6 2015-06-30T10:04:29 Use std::vector for TranslatedAttribs. This allows us to cache a std::vector between calls, and avoids us calling allocation/constructors for locals, which saves us some time. It also allows us to use the vector's size to limit the range of attribs we look at. BUG=angleproject:959 Change-Id: I799ed6c92fa8fca96e92e235b125a11d2d551aab Reviewed-on: https://chromium-review.googlesource.com/277286 Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Jamie Madill 46565a42 2015-06-22T13:57:21 Small optimization to VertexDataManager. Avoid calling a getter repeatedly in a loop, when we can instead store a local reference to an array. BUG=angleproject:959 Change-Id: I507ad1a6cf3bb6183dd3499df024dfec6950a6c8 Reviewed-on: https://chromium-review.googlesource.com/277285 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Kenneth Russell <kbr@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Sam McNally 5a0edc62 2015-06-30T12:36:07 Change intializer list order to match field order. BUG=505304 Change-Id: I38149e72732f0b4eb7d398f90b5d11660e2a7fbc Reviewed-on: https://chromium-review.googlesource.com/282465 Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Geoff Lang <geofflang@chromium.org>
Jamie Madill fdee4112 2015-06-26T14:27:31 Roll dEQP 554adf..92f7752. Includes updated Licences. https://android.googlesource.com/platform/external/deqp/+log/554adf..92f7752 BUG=angleproject:998 Change-Id: If2da8719addeef25ec46adc105cf63c8b861f55a Reviewed-on: https://chromium-review.googlesource.com/282271 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Brandon Jones <bajones@google.com> Reviewed-by: Corentin Wallez <cwallez@chromium.org>
Corentin Wallez ac2582d3 2015-06-26T13:18:18 WindowSurfaceGLX: swap before resizing the child window BUG=angleproject:892 Change-Id: I9b27777547624959fce6e045f9aac2652f139b9b Reviewed-on: https://chromium-review.googlesource.com/282058 Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org>
Corentin Wallez 9de512e2 2015-06-26T13:16:50 WindowSurfaceGLX: Fix child window creation size This makes Chrome on ANGLE work on Linux without obvious bugs. BUG=angleproject:892 Change-Id: I595ac00a6646cb407941932b0d4f9f020d4cd059 Reviewed-on: https://chromium-review.googlesource.com/282057 Reviewed-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org>
Geoff Lang e6d4e12c 2015-06-29T13:33:55 Add remaining ES3 caps queries. BUG=angleproject:1049 Change-Id: Ib5076380c692f4fa94921692baa7da14440556d1 Reviewed-on: https://chromium-review.googlesource.com/282386 Tested-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Jamie Madill 27c08917 2015-06-22T13:57:20 Cache active attribute lists in VertexDataManager. This saves us re-creating vector memory every iteration, and allows us to iterate over much smaller lists of attributes. In the future it could allow us to update the cache more efficiently with state change updates. BUG=angleproject:959 Change-Id: Ie8ae7a31726468dc2184165380f1f3e5e0152936 Reviewed-on: https://chromium-review.googlesource.com/277284 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Kenneth Russell <kbr@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Jamie Madill 3d72cc79 2015-06-22T13:57:19 Pass current value type instead of the object. This allows us to simplify the logic in VertexDataManager enough that we can start to cache attribute information. BUG=angleproject:959 Change-Id: I7b53a137d73f40f86e3acb9caebb66f9cacf8b6f Reviewed-on: https://chromium-review.googlesource.com/277283 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Kenneth Russell <kbr@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Geoff Lang 2a46b4e8 2015-06-29T15:06:49 Remove the PackUnpackHalfOverflow test. Different drivers handle overflow differently. The overflow case is not defined in the spec. BUG=angleproject:1044 Change-Id: I9e3883af147c6c1a4fcdcddea3ff70c62f859c56 Reviewed-on: https://chromium-review.googlesource.com/282540 Reviewed-by: Geoff Lang <geofflang@chromium.org> Tested-by: Geoff Lang <geofflang@chromium.org>
Geoff Lang 8273e006 2015-06-15T13:40:19 Add new GLSL target versions. Test the emulated GLSL functions against multiple GL versions. BUG=angleproject:1044 Change-Id: I1e12523301042f0d541ab2f4e73f02319d1584ef Reviewed-on: https://chromium-review.googlesource.com/277702 Reviewed-by: Zhenyao Mo <zmo@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Tested-by: Geoff Lang <geofflang@chromium.org>
Geoff Lang 83ab236f 2015-06-23T13:18:03 Improve the test precision in PackUnpackTest. * Output to floating point framebuffers so that results are not truncated to 8 bit normalized values. * Tighten the epsilon from 1/255 to 0.0005. May need further tweaking as more GPUs are tested. * Explicitly check for -inf and inf values when they were previously rounded to 0 and 255. * Explicitly check for clamping of overflowed normalized floats by providing the expectations instead of computing them. BUG=angleproject:1044 Change-Id: I8410cf61fe79a9919fb0db7e415bf97b2d0abd9e Reviewed-on: https://chromium-review.googlesource.com/281191 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Geoff Lang <geofflang@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Geoff Lang b3a6a8f3 2015-06-23T16:10:14 Error when encountering non-preprocessor tokens before #extension in ESSL3. BUG=angleproject:1047 Change-Id: I4a548270f651e35b2c8d1ab5d0f46185230c5f74 Reviewed-on: https://chromium-review.googlesource.com/281216 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Geoff Lang <geofflang@chromium.org>
Jamie Madill 34d719b5 2015-06-26T15:28:06 Revert "Revert "dEQP: Compile dEQP targets in Chromium."" Should be fixed now with latest patches. BUG=angleproject:998 BUG=500736 This reverts commit d0071dd9c39d0ada72292d476d8e8e081731c953. Change-Id: I3a611bfa10429a1f97d18e9aa1f316eabb67d114 Reviewed-on: https://chromium-review.googlesource.com/282055 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Jamie Madill 4d06a24c 2015-06-26T11:12:24 dEQP: Extend GoogleTest support. *re-land after rebase on top of Chrome build fix* This patch will enable a Chromium-side CL to build dEQP/ANGLE with GoogleTest. BUG=angleproject:998 BUG=500736 Change-Id: Id02c66a260b0a7bab3de9cda624fe097322b2d3c Reviewed-on: https://chromium-review.googlesource.com/281824 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@google.com> Reviewed-by: Kenneth Russell <kbr@chromium.org>
Jamie Madill 86bc3a66 2015-06-26T11:12:23 dEQP: Move the DEPS dirs to mirror Chromium. This avoid us needing to replicate gyp variables in both Chromium's and ANGLE's common.gypi files. BUG=angleproject:998 Change-Id: Iba1938fd37eb4bee28ae7daad31bf85a2d3bd692 Reviewed-on: https://chromium-review.googlesource.com/282145 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@google.com> Reviewed-by: Kenneth Russell <kbr@chromium.org>
Arun Patole 7fa3355f 2015-06-10T15:15:18 Add constant folding support for matrix built-ins This change adds constant folding support for following matrix built-ins: - matrixCompMult, outerProduct, transpose, determinant and inverse. BUG=angleproject:913 TEST=angle_unittests(new: MatrixUtilsTest, ConstantFoldingTest.*Matrix*), dEQP Tests: dEQP-GLES3.functional.shaders.constant_expressions.builtin_functions.matrix.* (All 54 tests started passing with this change) Change-Id: I7b9bf04b9a2cbff72c48216cab04df58c5f008d6 Reviewed-on: https://chromium-review.googlesource.com/276574 Reviewed-by: Olli Etuaho <oetuaho@nvidia.com> Tested-by: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by: Jamie Madill <jmadill@chromium.org>
Jamie Madill 9e582b93 2015-06-22T11:18:32 perf_tests: Normalize MS counts in EGL init test. The un-normalized values depend on the number of frames rendered, so if we optimize down one time, the other ms values all increase, because we are rendering more frames. This generates spurious perf alerts from the perf dashboard. BUG=None Change-Id: I1bc480141a8cad7eee96bcdf4d6747a3cbae30f2 Reviewed-on: https://chromium-review.googlesource.com/280980 Reviewed-by: Zhenyao Mo <zmo@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Jamie Madill 7d112bb9 2015-06-22T13:57:19 Refactor source pointer math out of VertexBuffer9/11. This math can live in a single place in the VertexDataManager. This cleans up the code and paves the way for future optimizations. BUG=angleproject:959 Change-Id: I7138c6e080d9c3d6507b55d981bfb62c2590a2a8 Reviewed-on: https://chromium-review.googlesource.com/277282 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Kenneth Russell <kbr@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Jamie Madill 9c38580b 2015-06-22T13:57:18 Refactor VertexDataManager to store translated attrib. Refactoring patch only. Reduces the number of parameters we need to pass around in internal functions. Clears the way for future optimization work. BUG=angleproject:959 Change-Id: Ic98ab5a07189eaa053dffce994546666bb07cede Reviewed-on: https://chromium-review.googlesource.com/277281 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Kenneth Russell <kbr@chromium.org> Reviewed-by: Geoff Lang <geofflang@chromium.org>
Kenneth Russell 3b345ebd 2015-06-24T17:24:20 Fix ANGLE build on Chrome OS ARM. These builders set use_x11=1 but don't have GLX headers. Avoid trying to compile ANGLE's GL, GLX and X11 sources on Chrome OS at all. BUG=503173,angleproject:892 Change-Id: I7f5862bb83b33f6ce302244fa5668295789a3a41 Reviewed-on: https://chromium-review.googlesource.com/281778 Reviewed-by: Zhenyao Mo <zmo@chromium.org> Tested-by: Kenneth Russell <kbr@chromium.org>
Kenneth Russell 0eb0a816 2015-06-19T20:18:16 Fix seg fault in IsPlatformAvailable. Some GLX implementations can't handle having libGL.so.1 dlclose'd out from under them while there's any X window still open against which a GLXWindow was ever created. Load the library once per process rather than once per FunctionsGLX instance. This showed up as a crash in X11Window::destroy, during XCloseDisplay. BUG=angleproject:892 Change-Id: I0da2674786a952865e5bea2513259c0a32d4453c Reviewed-on: https://chromium-review.googlesource.com/280902 Tested-by: Kenneth Russell <kbr@chromium.org> Reviewed-by: Corentin Wallez <cwallez@google.com>
Jamie Madill 57cdc69c 2015-06-23T21:03:06 Revert "dEQP: Extend GoogleTest support." Causing runhooks failures: gyp: Undefined variable deqp_path in C:\b\build\slave\GPU_Win_Builder\build\src\third_party\angle\src\tests\tests.gyp BUG=angleproject:998 BUG=500736 This reverts commit 8edd75d1bb6680e7c2018ece9f5c81e3b87bf938. Change-Id: I8cd83bb5871dea81abea61554b182825145298e6 Reviewed-on: https://chromium-review.googlesource.com/281208 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Jamie Madill d0071dd9 2015-06-23T21:01:01 Revert "dEQP: Compile dEQP targets in Chromium." Causing runhooks failures: gyp: Undefined variable deqp_path in C:\b\build\slave\GPU_Win_Builder\build\src\third_party\angle\src\tests\tests.gyp BUG=angleproject:998 BUG=500736 This reverts commit 3018c206abd1fbfb94ab049326ff89f1f8cf3be4. Change-Id: I8c0402fb7465101a9331cbb2020693469d644e7e Reviewed-on: https://chromium-review.googlesource.com/281194 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Corentin Wallez eed29ebb 2015-06-23T12:43:38 Simplify the assignment of use_x11. Hopefully fixes the flaky compile errors in ChromeOS BUG=angleproject:892 BUG=503173 Change-Id: Ifd79f1e382c12be3aecd9efb4860287df7d43060 Reviewed-on: https://chromium-review.googlesource.com/281180 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Corentin Wallez <cwallez@chromium.org>
Jamie Madill 3018c206 2015-06-22T15:35:54 dEQP: Compile dEQP targets in Chromium. This switches the primary targets for dEQP into the non-standalone namespace, so we can build dEQP ANGLE tests on Windows with Chrome. BUG=angleproject:998 BUG=500736 Change-Id: I2ae25f31ab123c49965412445f661b6b69d19109 Reviewed-on: https://chromium-review.googlesource.com/281020 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Tested-by: Jamie Madill <jmadill@chromium.org>
Jamie Madill 8edd75d1 2015-06-22T15:35:53 dEQP: Extend GoogleTest support. This patch will enable a Chromium-side CL to build dEQP/ANGLE with GoogleTest. BUG=angleproject:998 BUG=500736 Change-Id: Ie480e3319c1bd77dd1de2c964f78ef97297fcfe9 Reviewed-on: https://chromium-review.googlesource.com/280856 Tested-by: Jamie Madill <jmadill@chromium.org> Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Zhenyao Mo <zmo@chromium.org>
Geoff Lang e148f3d3 2015-06-17T11:41:28 Fix lots of compilation errors in generated GLSL functions. Reformatted to improve readability. BUG=angleproject:angleproject:947 Change-Id: I467f0396a0318b7a51d5cafce3f11e55d929c369 Reviewed-on: https://chromium-review.googlesource.com/278250 Reviewed-by: Corentin Wallez <cwallez@chromium.org> Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Geoff Lang <geofflang@chromium.org>
Geoff Lang 7b9b284b 2015-06-15T11:02:49 Refactor GLSL version calculations for emulated functions. Pass the target GLSL version to InitBuiltInFunctionEmulatorForGLSL so that it can be updated to generate emulated functions for multiple versions. BUG=angleproject:1044 Change-Id: I34c408df52fd0f7d6c2f66d0579ac854afd93b87 Reviewed-on: https://chromium-review.googlesource.com/277700 Reviewed-by: Jamie Madill <jmadill@chromium.org> Tested-by: Geoff Lang <geofflang@chromium.org>