- 31 Jul, 2018 6 commits
-
-
Jamie Madill authored
This moves out some shared logic into a more accessible place. We don't need to validate buffer overflows when using robust resource access but we do need to validate we are using client side data correctly. Bug: angleproject:1391 Change-Id: I7a3dca8409c5a1faf1ff7bc732d5ed1bd62eb3b1 Reviewed-on: https://chromium-review.googlesource.com/1148817 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by:
Geoff Lang <geofflang@chromium.org>
-
Courtney Goeltzenleuchter authored
Android build system wants to see pre-built components like this with naming that indicates what it is. Bug: angleproject:2702 Change-Id: I84acee78a6ceef5b04c87436c68d41510f6e66c7 Reviewed-on: https://chromium-review.googlesource.com/1155004 Commit-Queue: Courtney Goeltzenleuchter <courtneygo@google.com> Reviewed-by:
Jamie Madill <jmadill@chromium.org>
-
Jamie Madill authored
This cache is updated in the following locations: 1. GLES1: clientActiveTexture. 2. GLES1: disableClientState/enableClientState. 3. Context: linkProgram/useProgram/programBinary. 4. Context: bindVertexArray. 5. Vertex Array: most state changes. Improves performance by about 6% in the GL no-op test. Also includes fixes for keeping the client memory attribs mask in sync. The cache also includes a boolean if there are any enabled client attributes. Bug: angleproject:1391 Change-Id: I93b6a2c8492355958fd5483f14b70535729091d6 Reviewed-on: https://chromium-review.googlesource.com/1147437 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by:
-
Ian Elliott authored
Bug: angleproject:2749 Change-Id: Ie8fdb67918ea1c1054527fe1e0c1be4054fe9cb3 Reviewed-on: https://chromium-review.googlesource.com/1154189Reviewed-by:
Ian Elliott <ianelliott@google.com> Reviewed-by:
-
Olli Etuaho authored
It's safe to read from a multiview framebuffer if it is layered and has just one view. The native OVR_multiview spec supports this as well. BUG=angleproject:2062 TEST=angle_end2end_tests Change-Id: I04e1364390574075f7e06e39a64e3bf05a539a05 Reviewed-on: https://chromium-review.googlesource.com/1156509 Commit-Queue: Olli Etuaho <oetuaho@nvidia.com> Reviewed-by:
-
Jamie Madill authored
This will enable more caching for ValidateDrawAttribs. It requires some minor refactoring of the GLES 1 code. Bug: angleproject:1391 Change-Id: I52b73c9384d14cdb90ba6337bfc1ab345866fff0 Reviewed-on: https://chromium-review.googlesource.com/1147436 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by:
Frank Henigman <fjhenigman@chromium.org> Reviewed-by:
-
- 30 Jul, 2018 2 commits
-
-
Frank Henigman authored
Override fixed to float, generate code, enable tests. BUG=angleproject:2405 Change-Id: Ic3e9a31eaf22372023b94081b0f4a83770dcabbe Reviewed-on: https://chromium-review.googlesource.com/1144455 Commit-Queue: Frank Henigman <fjhenigman@chromium.org> Reviewed-by:
-
Frank Henigman authored
When we get a buffer with vertex data in an unsupported format or alignment, correct that as we copy it to a DynamicBuffer, then use the copy. Enable tests. BUG=angleproject:2405 Change-Id: I2132abea4d936f6b53d9209be7f99a0e2d8de219 Reviewed-on: https://chromium-review.googlesource.com/1141277 Commit-Queue: Frank Henigman <fjhenigman@chromium.org> Reviewed-by:
-
- 28 Jul, 2018 1 commit
-
-
Zhenyao Mo authored
BUG=angleproject:2507 TBR=geofflang@chromium.org Change-Id: I6788303bdca8739c42b1f43e45b5b2c16cea4ed3 Reviewed-on: https://chromium-review.googlesource.com/1154127Reviewed-by:
Zhenyao Mo <zmo@chromium.org> Commit-Queue: Zhenyao Mo <zmo@chromium.org>
-
- 27 Jul, 2018 10 commits
-
-
Jamie Madill authored
This CL improves performance on the draw call microbenchmark by 10% when no-oping driver calls. Bug: angleproject:2738 Change-Id: I4f5c11db90d9056ce4557b2a4432bc55b42b5bba Reviewed-on: https://chromium-review.googlesource.com/1150093 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by:
Yuly Novikov <ynovikov@chromium.org>
-
Geoff Lang authored
ChromeOS devices share some of the same driver bugs as Qualcomm. NVIDIA devices require the image to be preserved. BUG=angleproject:2507 Change-Id: I2750f78c219490fcd7e6abd224adf8d8e96b7848 Reviewed-on: https://chromium-review.googlesource.com/1153606Reviewed-by:
-
Geoff Lang authored
The read framebuffer may not be complete and be incapable of syncing. Removed the generate syncDirtyObjects method so each caller must make sure they are only syncing objects that are known to be valid for the operation. BUG=angleproject:2737 Change-Id: Ia8edf3fca3a8369aa813be46ba99f6b50a36b2e6 Reviewed-on: https://chromium-review.googlesource.com/1151621 Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by:
Olli Etuaho <oetuaho@nvidia.com>
-
LRN authored
Setting EGL_WIDTH or EGL_HEIGHT results in mFixedWidth and mFixedHeight changing their values. Whenever the swap chain is tested for needing a resize, surfaces with mFixedSize are treated as if they should be resized to mFixedWidth and mFixedHeight from their current size. This way fixed-size surfaces can be resized - it's enough to do a couple of eglSurfaceAttrib() calls and then either let ANGLE check surface for needing a resize in due course, or force that by calling eglWaitNative(). BUG=angleproject:2725 Change-Id: I2932d8ecd5a2f10500e4aa8583a169aa09bf01bb Reviewed-on: https://chromium-review.googlesource.com/1139055 Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by:
-
Jamie Madill authored
This removes another use of the gl::ErrorOrResult pattern. Unblocks error refactoring. Bug: angleproject:2738 Change-Id: Ib611a3b68f8995469befd51797dfed34eaeee84e Reviewed-on: https://chromium-review.googlesource.com/1151450 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by:
-
Jamie Madill authored
This makes the Context more available for logging errors. Also includes more refactoring to VertexDataManager to ensure we can access the gl::Context. Bug: angleproject:2738 Change-Id: Iae3d22a1403078d236bfe63a3e2d203c13678dc4 Reviewed-on: https://chromium-review.googlesource.com/1151449Reviewed-by:
-
Jamie Madill authored
This gives us better control over the error returned by the init routine. Also changes the streaming buffer in VertexDataManager to not be a unique_ptr. Bug: angleproject:2738 Change-Id: I3193840dfb71c7574adbe65bc5f9227f4add1fd3 Reviewed-on: https://chromium-review.googlesource.com/1151448Reviewed-by:
-
Tom Anderson authored
Changelist: https://chromium.googlesource.com/external/github.com/KhronosGroup/Vulkan-Loader/+log/da58202e9569eff53560b4c3a007de2297be40d2..a8358804f3c5b4e284d5e9a88b232309bb3c333d BUG=chromium:427584 Change-Id: I758d2ba05aa731b07e5d336a6ebab45a4837e142 Reviewed-on: https://chromium-review.googlesource.com/1152356Reviewed-by:
Corentin Wallez <cwallez@chromium.org> Commit-Queue: Corentin Wallez <cwallez@chromium.org>
-
Olli Etuaho authored
Declaring HLSL outputs with SV_* semantics after the ones with TEXCOORD* semantics makes multiview shaders that contain a flat varying run correctly. This is a workaround for what seems to be a bug in the DX runtime, reproducible on at least NVIDIA and Intel. This also adds another, clearer test case for flat interpolation used together with multiview. BUG=angleproject:2062 TEST=angle_end2end_tests Change-Id: I127b85497aa4fee58b74893165a980cbdd4444d6 Reviewed-on: https://chromium-review.googlesource.com/1146809Reviewed-by:
-
Frank Henigman authored
After a release() or destroy() it is now possible continue using a DynamicBuffer because we reset the internal variables and the next allocate() restarts everything. Also removed the valid() method as it wasn't very useful and made reuse more complicated. No functional change. BUG=angleproject:2405 Change-Id: I5ca9b758fb300618f6578ebc406494ad281c1622 Reviewed-on: https://chromium-review.googlesource.com/1144229 Commit-Queue: Frank Henigman <fjhenigman@chromium.org> Reviewed-by:
-
- 26 Jul, 2018 2 commits
-
-
Geoff Lang authored
BUG=angleproject:2507 Change-Id: Ica33166e9e23e933977c3ab034d4f5a8cada9fb1 Reviewed-on: https://chromium-review.googlesource.com/1143454 Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by:
-
Jiawei Shao authored
This patch moves the implementation of translating compute shader builtin variables from link time to compile time. Unlike graphics shaders that require the information from other shader stages, we actually have enough information to translate compute shader builtin variables in compile time. Many redundant codes in DynamicHLSL have been removed after this refactor. BUG=angleproject:1442 Change-Id: I7458006785ff966a00a3825610adc5566652c75e Reviewed-on: https://chromium-review.googlesource.com/1149609Reviewed-by:
Jiajia Qin <jiajia.qin@intel.com> Reviewed-by:
-
- 25 Jul, 2018 14 commits
-
-
Omar El Sheikh authored
Currently disabled test on android Also made the default EGL device type be EGL_PLATFORM_ANGLE_DEVICE_TYPE_HARDWARE_ANGLE Updated destroy function of RendererVk to properly reset the mMemoryProperties member variable Bug: angleproject:2735 Change-Id: Ie1f8ab1ac93b18eda4c10c3b10ecd4756f1d8543 Reviewed-on: https://chromium-review.googlesource.com/1140508 Commit-Queue: Omar El Sheikh <theoking@google.com> Reviewed-by:
-
Jamie Madill authored
The Context needed for the error refactor and Vulkan implementation. Bug: angleproject:2738 Change-Id: Id816b61ebf342c953a320ab29b6386e356dc64e8 Reviewed-on: https://chromium-review.googlesource.com/1150092Reviewed-by:
-
Jamie Madill authored
This is needed for the error refactoring and also for the Vulkan implementation. Bug: angleproject:2738 Change-Id: I4e1bed7f67ef17feb5554b5838a2ed5feb22bba0 Reviewed-on: https://chromium-review.googlesource.com/1150091Reviewed-by:
-
Jamie Madill authored
Replace the struct type of binary and symbol parameter nodes with a lookup of the struct type in the tree. Bug: angleproject:2494 Change-Id: I1e892e0f1b1f97302d1cdce1035f51c01d0efd9e Reviewed-on: https://chromium-review.googlesource.com/1135695Reviewed-by:
-
Jamie Madill authored
We can re-use the same info for each level of the texture. Also use fixed sized arrays for the subresource data structure. Bug: chromium:867089 Change-Id: Ie43886f708d1141fb80d30a78cabdd37dfbf6f94 Reviewed-on: https://chromium-review.googlesource.com/1149082Reviewed-by:
-
Jamie Madill authored
Needed to get access to the scratch buffer. Bug: chromium:867089 Change-Id: I217bde149d24432e4baa276ead02ebc52a219958 Reviewed-on: https://chromium-review.googlesource.com/1149081Reviewed-by:
-
Jamie Madill authored
This moves a D3D specific type into a shared location. Bug: chromium:867089 Change-Id: Ieac9f1aa4a9de11fbc86550eee30fb242a2cb98e Reviewed-on: https://chromium-review.googlesource.com/1149080Reviewed-by:
-
Jamie Madill authored
Don't use ErrorOrResult. Also pass gl::Context around to more places. Bug: angleproject:2378 Change-Id: Id498471fc9ab4f4d6b2285e110381b10a1247689 Reviewed-on: https://chromium-review.googlesource.com/1148465Reviewed-by:
-
Jamie Madill authored
This will make it possible to record errors in the context immediately. Bug: angleproject:2378 Change-Id: Ib7ccdbcfa2bf51e3cc876496267c61cf5a26b7bd Reviewed-on: https://chromium-review.googlesource.com/1148464Reviewed-by:
-
Jamie Madill authored
Bug: angleproject:2738 Change-Id: I062b2fe3ff2b53bb361539c408b61626ac6c2f25 Reviewed-on: https://chromium-review.googlesource.com/1148463Reviewed-by:
-
Jamie Madill authored
Unblocks further return value refactoring for perf. Bug: angleproject:2738 Change-Id: Ie92aa656736f84b4e4ef6d4ab8e0377ad4afb2af Reviewed-on: https://chromium-review.googlesource.com/1142955Reviewed-by:
-
Jamie Madill authored
Unblocks further error refactoring. Also cleans up some checked math logic. Bug: angleproject:2738 Change-Id: I0c9fba9bb908dfc8424921d0db9871241b86e7de Reviewed-on: https://chromium-review.googlesource.com/1142954Reviewed-by:
-
Jamie Madill authored
This removes the use of the ErrorOrResult pattern in Renderer11. Bug: angleproject:2738 Change-Id: Icc046ddd86394df56ca1acd10b1804fd6afa8ad0 Reviewed-on: https://chromium-review.googlesource.com/1142953Reviewed-by:
-
Jamie Madill authored
This pattern isn't available when using angle::Result instead of gl::Error. Refactor the Buffer11 class to use gl::Error instead. This unblocks further return value refactoring for perf. Bug: angleproject:2738 Change-Id: I739044ae318c6b1fc40756a4f1fcadf609276c1a Reviewed-on: https://chromium-review.googlesource.com/1142952Reviewed-by:
-
- 24 Jul, 2018 5 commits
-
-
Omar El Sheikh authored
The mInFlightCommand vector could grow faster than the GPU can complete tasks. Check for this edge case and give the GPU a chance to catch up with work. Also allows us to enable performance tests for Draw Calls on vulkan since it was affected by this limit check before. Bug: angleproject:2742 Change-Id: I66a86ea6b5300fa3d74b07dc341aa597824b5f09 Reviewed-on: https://chromium-review.googlesource.com/1147607Reviewed-by:
-
Geoff Lang authored
BUG=angleproject:2507 Change-Id: I0e5879acfbb7ba1b9d711f518205cebad2edb643 Reviewed-on: https://chromium-review.googlesource.com/1148657 Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by:
-
Yuly Novikov authored
Bug: 850060 Change-Id: I9672e6c93e92941e61763d77180bef1548f5fbde Reviewed-on: https://chromium-review.googlesource.com/1148897Reviewed-by:
-
Jamie Madill authored
Bug: angleproject:2743 Change-Id: I1693a5b301fa10a92263aab0cec14829851fe21f Reviewed-on: https://chromium-review.googlesource.com/1148461Reviewed-by:
-
Olli Etuaho authored
Instead of having separate main() and gl_main() functions in HLSL shaders, add initializing outputs and inputs directly to the main function that's in the AST. This works around some HLSL bugs and should not introduce name conflicts inside main() since all the user-defined variables are prefixed. BUG=angleproject:2325 TEST=angle_end2end_tests Change-Id: I5b000c96aac8f321cefe50b6a893008498eac0d5 Reviewed-on: https://chromium-review.googlesource.com/1146647Reviewed-by:
-
