- 10 Jul, 2018 13 commits
-
-
Jamie Madill authored
Bug: angleproject:2717 Change-Id: I542f3b0f2de21857d7fea0267f07d2d0eec78a8c Reviewed-on: https://chromium-review.googlesource.com/1131567Reviewed-by:
Geoff Lang <geofflang@chromium.org> Reviewed-by:
Luc Ferron <lucferron@chromium.org> Commit-Queue: Jamie Madill <jmadill@chromium.org>
-
Jamie Madill authored
This will make it easier to insert internal uniform layouts in Vulkan. Bug: angleproject:2717 Change-Id: Ic1a76848337cc67a57698913c584d8596bf4f27e Reviewed-on: https://chromium-review.googlesource.com/1127300Reviewed-by:
-
Tobin Ehlis authored
Update Vulkan Layers/Loader/Tools/Header to latest versions as of 7/10. The layers were updated to correctly comprehend which device extensions are enabled which should fix a false error being flaged on negative Viewports in ANGLE. Bug: angleproject:2704 Change-Id: Ie51660f539beb249c3b11dc7e0d553ff7cabfde3 Reviewed-on: https://chromium-review.googlesource.com/1131919 Commit-Queue: Tobin Ehlis <tobine@google.com> Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by:
Jamie Madill <jmadill@chromium.org>
-
Jamie Madill authored
This was introduced in a prior patch. Bug: angleproject:2717 Change-Id: I8f34377cbb9cbede4237efb1610400a8ed20ae2a Reviewed-on: https://chromium-review.googlesource.com/1131566Reviewed-by:
-
Luc Ferron authored
- Fixes all deqp functional_fragment_ops_depth_stencil_* with Y flipping enabled. Bug: angleproject:2673 Change-Id: I94a4225dec8adf9113309e8b8b2c8aa61f6a2bb9 Reviewed-on: https://chromium-review.googlesource.com/1129857 Commit-Queue: Luc Ferron <lucferron@chromium.org> Reviewed-by:
-
Jamie Madill authored
This cleans up another use of the proxy context. Bug: angleproject:2714 Change-Id: Icba5bf76a3e9f811ee571529de16fd8162c76b3a Reviewed-on: https://chromium-review.googlesource.com/1128928Reviewed-by:
-
Jamie Madill authored
This cleans up another usage of getProxyContext. Bug: angleproject:2714 Change-Id: I5337e30cdd96e282e355c727bb40cd30e049ec1a Reviewed-on: https://chromium-review.googlesource.com/1128926Reviewed-by:
Frank Henigman <fjhenigman@chromium.org> Commit-Queue: Jamie Madill <jmadill@chromium.org>
-
Jamie Madill authored
The BindingPointer pattern isn't necessary. Every time we delete a Texture we call ReleaseTexImage internally. There shouldn't be any time that we keep an orphaned Texture is Surface. This cleans up one place where we were using the ProxyContext. Bug: angleproject:2714 Change-Id: I3b0fd2125d02ea7545922ec6da7f487451bed871 Reviewed-on: https://chromium-review.googlesource.com/1128925 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by:
-
Luc Ferron authored
- Validated by running: -functional_texture_specification_basic_copytexsubimage2d_* -functional_texture_specification_basic_copyteximage2d_* With the flip flag enabled. Bug: angleproject:2673 Change-Id: I5a3041af79f9316256b0650ab7e3fd0e086e46e3 Reviewed-on: https://chromium-review.googlesource.com/1129820 Commit-Queue: Luc Ferron <lucferron@chromium.org> Reviewed-by:
-
Luc Ferron authored
- The depth/stencil related tests are now disabled since I found out that the tests are too simple to find issues if there is a flip since they use only a single color and no gradient / checked board. This is left to implement in the Vulkan backend later on. Bug: angleproject:2673 Change-Id: I8f7091d4b9c8c3bec0353ebab28304b6209ea350 Reviewed-on: https://chromium-review.googlesource.com/1129629Reviewed-by:
-
Jamie Madill authored
These methods can mutate the display resources. Necessary for getting rid of ProxyContext and for the Vulkan error refactor. Bug: angleproject:2714 Bug: angleproject:2713 Change-Id: Ibffb1a382ecb064daaa7c664f9fc65cbcf927b37 Reviewed-on: https://chromium-review.googlesource.com/1128927Reviewed-by:
-
Jacek Caban authored
Tested with Firefox build, upstream bug: https://bugzilla.mozilla.org/show_bug.cgi?id=1471632 Change-Id: I4b44847dfc69ee26cf2215b0a0b7573becfd369d Reviewed-on: https://chromium-review.googlesource.com/1117187 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by:
-
Olli Etuaho authored
This makes the compiler a few kilobytes smaller, and prepares getting rid of TString altogether. BUG=angleproject:2267 TEST=angle_unittests Change-Id: I93a003fe27b99bef72f872fa1066e2e108f934c5 Reviewed-on: https://chromium-review.googlesource.com/1107713Reviewed-by:
Corentin Wallez <cwallez@chromium.org> Commit-Queue: Olli Etuaho <oetuaho@nvidia.com>
-
- 09 Jul, 2018 12 commits
-
-
Omar El Sheikh authored
Updated ScopedVkLoaderEnvironment destructor to delete an environment variable that gets set during initialization of the renderer to load the Mock/Null driver for tests that request it. Bug: angleproject:2698 Change-Id: Ibbac795b6315971b303d97a55d24565a403d056c Reviewed-on: https://chromium-review.googlesource.com/1120940Reviewed-by:
Tobin Ehlis <tobine@google.com> Commit-Queue: Jamie Madill <jmadill@chromium.org>
-
Tobin Ehlis authored
Update Vulkan Layers/Loader/Tools/Header to latest versions as of 7/3. This updates the ICD build to use inputs from the Header repo instead of copies of the file from its own repo. Those copies have been deleted and some of the generation scripts in Tools repo have been renamed. Also put script dependencies from Vulkan-Headers repo into a shared var that's imported by the other repos to minimize repeat code. Bug: angleproject:2704 Change-Id: I98a18be055482dccdad8468115b4045d09342bfb Reviewed-on: https://chromium-review.googlesource.com/1124966Reviewed-by:
-
Jamie Madill authored
This will be used to specify the applied Vulkan. We will use this to implement OpenGL line rasterization rules. Bug: angleproject:2717 Change-Id: I3395bf620a01c4b84b19a00037d05f148e5523f3 Reviewed-on: https://chromium-review.googlesource.com/1120151 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by:
-
Jamie Madill authored
This fixes a bug in SPIRV-Tools validation. vec3s were not being packed properly. Affected some of the more complex dEQP shader tests. Unblocks better Debug error reporting. Bug: angleproject:2552 Change-Id: I83447813a4ae3347157321a30a91b5dee55ff6c1 Reviewed-on: https://chromium-review.googlesource.com/1129824 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by:
-
Jamie Madill authored
This completes a prior refactor. DynamicDescriptorPool couldn't robustly handle multiple pools so only accept one as a parameter. We were already only using one at a time in practice. This removes the need to look up the pool index from the descriptor type which became problematic later on. Bug: angleproject:2717 Change-Id: I7fb92d5ab55bbe1f35b6b9cfcd8981cc1ca358f1 Reviewed-on: https://chromium-review.googlesource.com/1127157 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by:
-
Geoff Lang authored
This is a simpler way of flipping when the client is aware of the extension. This allows Chrome to render with the correct orientation. BUG=angleproject:2709 Change-Id: I52216b765a42930f2be043a07fe441a9f875a14d Reviewed-on: https://chromium-review.googlesource.com/1127342Reviewed-by:
-
Jamie Madill authored
This will allow the Vulkan back-end to call Vulkan commands that return VkResult within syncState. Bug: angleproject:2717 Change-Id: I05e1379e71acaac56bcc9493c8561e82da000934 Reviewed-on: https://chromium-review.googlesource.com/1120150 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by:
-
Tobin Ehlis authored
If a draw does not have enough vertices for its primitive count to produce a primitive then skip it. If an instanced draw has 0 instances or not have enough vertices for its primitive count to produce a primitive then skip it. This means a Point with 0 vertices, a Line w/ 0-1 vertices, or a tri with 0-2 primitives. Updated some redundant code in the D3D11 backend. Draws below the minDrawCount will no longer be passed to the backend so updated the associated state in StateManager11 to only track the case where all primitives should be culled due to GL state settings. BUG=angleproject:2568 TEST=functional_transform_feedback_basic_types_interleaved_lines_lowp_int Change-Id: I9faa767c12004fcdec923ec70a8ee5615d789813 Reviewed-on: https://chromium-review.googlesource.com/1120849 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by:
-
Luc Ferron authored
- This change fixes the tests that were failing in angle_end2end_tests - MaxTextureSizeTest.* Bug: angleproject:2673 Change-Id: I9c671b8be0e2360a80f6298897ec298c6e998863 Reviewed-on: https://chromium-review.googlesource.com/1128811 Commit-Queue: Luc Ferron <lucferron@chromium.org> Reviewed-by:
-
Luc Ferron authored
- This fixes all ViewportTest.* and MipmapTest.* - Tests left to fix in end2end: - BlitFramebuffer* - MaxTextureSizeTest.* - PointSpritesTest.* Every other test in angle_end2end_tests are working. Bug: angleproject:2673 Change-Id: I162083bc847c15fa5490ab524ad4c22747d232ea Reviewed-on: https://chromium-review.googlesource.com/1126333 Commit-Queue: Luc Ferron <lucferron@chromium.org> Reviewed-by:
-
Jamie Madill authored
By making mData public we can use aggregate initialization. The syntax is the same as for initializing a std::array. It might be possible to also use std::map initialization syntax using a list of pairs. That is left for a follow-up. Bug: angleproject:2568 Change-Id: I30a6f280172cb197208d14b1a53b9cc12b7834d8 Reviewed-on: https://chromium-review.googlesource.com/1127181 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by:
-
Olli Etuaho authored
This code is analoguous to the code in TextureFunctionHLSL and is now implemented in a similar manner. BUG=angleproject:2267 TEST=angle_unittests, angle_end2end_tests Change-Id: Ie3503766217dad4f3848f2d4b2fc3f62b3edce0c Reviewed-on: https://chromium-review.googlesource.com/1110366Reviewed-by:
-
- 06 Jul, 2018 1 commit
-
-
Geoff Lang authored
In the future this can be implemented with a flush. BUG=angleproject:2707 Change-Id: I1dd4eae9551f7797865e6e17e6caa77f6ec5a9b2 Reviewed-on: https://chromium-review.googlesource.com/1126457 Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by:
-
- 05 Jul, 2018 7 commits
-
-
Jamie Madill authored
In a few places we initialize resource on the stack in the Vulkan back-end. When we do this we should use a RAII wrapper so that if we generate an error and unwind the stack we can auto-release the object. This fixes a bug that was being triggered by an unexpected failure on Present with the Vulkan back-end on Intel. There are several other places in the code that should be modified to use the new wrapper. Bug: angleproject:2690 Change-Id: I49a1c5516756f8b7dba833aca65926aa7b8bd5c6 Reviewed-on: https://chromium-review.googlesource.com/1118610 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by:
-
Geoff Lang authored
BUG=angleproject:2707 Change-Id: Ibd70437e3a40a9a63b78ac9a4546c97611a97ba5 Reviewed-on: https://chromium-review.googlesource.com/1126458 Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by:
-
Geoff Lang authored
any_input_dirty would return early when finding a dirty input and not finish filling the new_hashes map. This would require multiple runs of the code generation script to fully generate all the outputs. BUG=angleproject:2695 Change-Id: Ie62190efe2765df432b0a535fb8d33ed2ffa66a7 Reviewed-on: https://chromium-review.googlesource.com/1127439 Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by:
-
Geoff Lang authored
This should avoid generating diffs when running the code generation with different source directories or on different platforms. BUG=angleproject:2695 Change-Id: I67776883bdbeb867a49bea00f16998c04f7857b4 Reviewed-on: https://chromium-review.googlesource.com/1127355 Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by:
-
Geoff Lang authored
BUG=angleproject:2695 Change-Id: I9a083bf67b17ca0dc1ee213b75caa4ecece62f00 Reviewed-on: https://chromium-review.googlesource.com/1127354 Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by:
-
Geoff Lang authored
BUG=angleproject:2691 Change-Id: Iafb1b4828fa0d54974bfb9d11a436ccaa910b942 Reviewed-on: https://chromium-review.googlesource.com/1127356Reviewed-by:
-
Geoff Lang authored
Pass the VkPhysicalDeviceProperties to GenerateDefaultConfig to initialize the pbuffer size limitations in a platform independent way. BUG=angleproject:2622 Change-Id: Id99bc505a1965cc037c5dcd65af031c1c655d424 Reviewed-on: https://chromium-review.googlesource.com/1126406 Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by:
-
- 04 Jul, 2018 4 commits
-
-
Luc Ferron authored
- Hide the implementation behind a feature flag, currently disabled permanently as I'm working on fixing the different failures. - SimpleOperationTest.* passing Bug: angleproject:2673 Change-Id: Ic86520c3cc478d62bebbaeaf4c6b33c439a67b0f Reviewed-on: https://chromium-review.googlesource.com/1119089 Commit-Queue: Luc Ferron <lucferron@chromium.org> Reviewed-by:
-
Geoff Lang authored
-Add a new renderer and context type to own native EGL contexts and handle destruction. -Track the current EGL surface and context per-thread. -Support unvirtualized contexts by creating a new context for every client context. BUG=angleproject:2464 Change-Id: Ib2efa1d88c771b4a78625e0e3546f6ed95678c91 Reviewed-on: https://chromium-review.googlesource.com/1110943 Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by:
Yuly Novikov <ynovikov@chromium.org>
-
Geoff Lang authored
BUG=angleproject:2464 Change-Id: If1834cf88aed0ffae12bb584d6936c6c09a296a0 Reviewed-on: https://chromium-review.googlesource.com/1117022 Commit-Queue: Geoff Lang <geofflang@chromium.org> Reviewed-by:
-
Luc Ferron authored
- We weren't keeping track of mSize correctly, causing some very specific index buffer bindings to fail. The size returned when allocating the buffer can be a bit different than the size requested, and that extra space between mSize and the allocated size was not meant to be used, causing errors in the validation layers. Bug: angleproject:2580 Change-Id: I47eb7b8de6f4f657de14385b77ba6a459add599b Reviewed-on: https://chromium-review.googlesource.com/1118607Reviewed-by:
-
- 03 Jul, 2018 2 commits
-
-
jchen10 authored
According to the spec(section 18.1), all copy commands are treated as transfer operations for the purposes of synchronization barriers. Some tests are flaky on Intel platform without such barriers. Bug: angleproject:2663 Bug: angleproject:2664 Change-Id: Ic8bc9a0eb000670342c0df0449257324f04ad1f8 Reviewed-on: https://chromium-review.googlesource.com/1124103 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by:
-
Jeff Gilbert authored
Bug: angleproject:2699 Change-Id: I234e0aee2526750a442ad74b5c57f7c536310786 Reviewed-on: https://chromium-review.googlesource.com/1123880 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by:
-
- 29 Jun, 2018 1 commit
-
-
Jamie Madill authored
Using mtime is fundamentally flawed when working with git. Replace these flaky checks with hash comparisons. The hashes are stored in an autogenerated json file that will be stored in the repository. This makes the run_code_generation script robust against any and all input changes. It also removes the need to track script outputs as dependencies. Bug: angleproject:2697 Change-Id: I60f2a87a8680b1f775ad678b05112f5b16c7dde7 Reviewed-on: https://chromium-review.googlesource.com/1120159 Commit-Queue: Jamie Madill <jmadill@chromium.org> Reviewed-by:
-
