- May 03, 2024
-
-
Oriol Prieto Gasco authored
Set the container field of aconfig flags Test: printflags --format='{package}:{container}:' | grep "::" Bug: 312769710 Flag: NONE Change-Id: I84fda0d5e9813940609436d825738ff786908ff0
-
- Apr 23, 2024
-
-
Ryan Prichard authored
Upgrading to clang-r522817's new libc++ requires that <mutex> be included here. Bug: 333165689 Test: treehugger Flag: EXEMPT, fix build error Change-Id: I14ff396b21feb4b569a3c8d967c8295a04005beb
-
- Mar 13, 2024
-
-
Aditya Kumar authored
``` In file included from external/libcxx/include/algorithm:644: external/libcxx/include/memory:2335:19: error: invalid application of 'sizeof' to an incomplete type 'android::FileBlobCache' 2335 | static_assert(sizeof(_Tp) > 0, | ^~~~~~~~~~~ external/libcxx/include/memory:2652:7: note: in instantiation of member function 'std::default_delete<android::FileBlobCache>::operator()' requested here 2652 | __ptr_.second()(__tmp); | ^ external/libcxx/include/memory:2606:19: note: in instantiation of member function 'std::unique_ptr<android::FileBlobCache>::reset' requested here 2606 | ~unique_ptr() { reset(); } | ^ frameworks/base/libs/hwui/pipeline/skia/ShaderCache.h:40:7: note: in instantiation of member function 'std::unique_ptr<android::FileBlobCache>::~unique_ptr' requested here 40 | class ShaderCache : public GrContextOptions::PersistentCache { | ^ frameworks/base/libs/hwui/pipeline/skia/ShaderCache.h:35:7: note: forward declaration of 'android::FileBlobCache' 35 | class FileBlobCache; | ^ 1 error generated. ``` Test: Build Fixes: b/328705924 Change-Id: I6396f5fafef1941201c05d0b107f48e6d28a6631
-
- Feb 05, 2024
-
-
Liangliang Sui authored
1. When |frame->pendingDirty.isEmpty()| is executed there, it must not be empty. Because the judgment has been made at the beginning of the applyRenderNodeTransform method. 2. When |dirty->isEmpty()|, |dirty->setIWH(frame.width(), frame.height());| can be returned directly after execution. There is no need to judge bufferAge, and subsequent steps are redundant. Because the maximum range of dirty is [0, 0, frame.width(), frame.height()] Change-Id: Iccb3d39e4139c26ed5e7e24c39d1e141cb380dab Signed-off-by:
Liangliang Sui <coolsui.coding@gmail.com>
-
- Jan 25, 2024
-
-
Jihoon Kang authored
java_api_contribution modules are dynamically generated from droidstubs modules or java_sdk_library_import, and is not a user-facing module. Defining this module is a .bp file is a no-op. Test: m nothing Bug: 322242515 Change-Id: I5a6c925f94017f4f92b6abb0c93079d8c85c3eeb
-
- Jan 22, 2024
-
-
Kevin Lubick authored
Change-Id: Iacacba4e4cf03b1c9fd02ffef39ca99e4c597160
-
qiubowen authored
1. mDamageGenerationId is a random value when initialized. When this random value happens to be equal to info.damageGenerationId, the dirty range will reach the maximum DIRTY_MAX. 2. When info.damageGenerationId is equal to 0, which is when the first frame of the APP is drawn, there is no need to reach the maximum DIRTY_MAX in the dirty range. Bug: 321612817 Test: build Change-Id: Iab6a04170926d10cdf2a62ae0fbf9cbe5eeb8bbf Signed-off-by:
qiubowen <qiubowen@xiaomi.com>
-
- Jan 18, 2024
-
-
Jiang Tian authored
Making the scope more accurate that only acquire the lock when trying to access frame info in FrameInfoVisualizer, then make it irrelevant to the real draw operation. Bug: 317995179 Test: 1.going to developer options 2. swapping the "profile hwui" option from "none" to "bars" and back a couple times, no crash Change-Id: I069a28a7e847c0c3fca94fd9c43e95382f501b80 Merged-In: I069a28a7e847c0c3fca94fd9c43e95382f501b80
-
- Dec 11, 2023
-
-
Yi Kong authored
Since https://r.android.com/2856020, eng builds no longer perform costly link time optimizations. The workaround that disables LTO build speed is no longer necessary. Test: presubmit Change-Id: Ic76b7bb5866e8df4b19793aa06253910e5c5d384
-
- Dec 07, 2023
-
-
Dichen Zhang authored
Bug: 313791883 Change-Id: I1afd802a6c3958fa369c03bfef6a7cb379795bc4 Test: build
-
- Dec 06, 2023
-
-
Jorge Betancourt authored
Test: none Bug: 280116960 Change-Id: I76e1ce2c4fcd81ba08bb29c98a4319a6d5e4d4ac
-
- Dec 05, 2023
-
-
John Reck authored
Test: enable flag && silkfx Bug: 314810174 Change-Id: Ic69b633d6042145a537c7d39e713f804faff6600
-
- Nov 30, 2023
-
-
Harish Mahendrakar authored
Bug: 313791883 Test: treehugger Change-Id: Ia3347a0fb30ac4f2e5fddb8eded4ec023b0e3f59
-
mashixing authored
The original code contained assertions that checked `fICount` and `fIOffset`, which seem to be mistakenly copied from the Skia codebase. This commit corrects the assertions to properly reflect the variables used in the AOSP codebase, updating them to `mIndexCount` and `mIndexOffset` as intended. The correction ensures accurate runtime checks for the expected conditions and prevents compilation error due to undeclared identifier. Bug: 313566801 Change-Id: I1432598be14a6960d0dbe373e1ac997f5f5f8ee6 Signed-off-by:
Shixing Ma <mashixing@xiaomi.corp-partner.google.com>
-
- Nov 29, 2023
-
-
Nader Jawad authored
Fixed issue where profile canvas always leverage the transform from the vulkan surface instead of leveraging the transform from the hardwarebuffer if one was provided. Fixes: 307978217 Test: manual Change-Id: If27e24828f84977bde321a7635f1531550518491
-
- Nov 28, 2023
-
-
Zhao Qin authored
Gpu resource is not purged as expected in new implementation when app stays in foreground. Test: manual, adb shell dumpsys gfxinfo Change-Id: I15deb71693e726661c8bae70c7ae30cff2dcfa6b
-
- Nov 22, 2023
-
-
Tyler Freeman authored
fix(force invert): fix white-on-white text when text is drawn against a fill rect in the same RenderNode This fixes issues with custom Views that draw background and text in the same View/RenderNode. Instead of trying to force the individual paints to be dark, we simply invert the whole view. Adds a new hasFill() member to DisplayList and Canvas so that we record when the RenderNode has a call that fills the background. Also adds a new Container usage hint type, for this exact situation. Bug: 282821643 Bug: 293883260 Test: RenderNodeTests.cpp: mmm -j8 frameworks/base/libs/hwui && adb push $ANDROID_PRODUCT_OUT/data/nativetest/hwui_unit_tests/hwui_unit_tests /data/nativetest/hwui_unit_tests/hwui_unit_tests && adb shell /data/nativetest/hwui_unit_tests/hwui_unit_tests Change-Id: Ia0fe9c739bf07c1f3a8508d3f295a1dc54ee48f9
-
- Nov 21, 2023
-
-
John Reck authored
Fixes: 243390200 Test: make Change-Id: I0150737e744016d56f8f4c4464d90b29b8505344
-
John Reck authored
Bug: 303549484 Test: monkey runs Change-Id: I6843010151c3b7bd3933c7f22c706f36e11cd5a5
-
Kevin Lubick authored
Change-Id: I8943f5f12d3d61a15d7fc1fe7091a8685eeb7bdd
-
- Nov 17, 2023
-
-
Kevin Lubick authored
Android Renderengine doesn't draw any text, but hwui needs to turn bytes into SkTypefaces and it currently uses Skia's "Empty Custom" SkFontMgr, which is really just a wrapper around FreeType. This makes that connection explicit instead of relying on SkFontMgr::RefDefault(), which is going away soon. Change-Id: Ida96538c76e069ec86e8dc078ef329cfaeac5d78 Bug: b/305780908
-
- Nov 16, 2023
-
-
John Reck authored
Skip using the gainmap shader when drawing to a bitmap with a known SDR colorspace as it'll be a no-op in rendering results since W=1.0, but it'll be very expensive in runtime due to CPU evaluation of the SKSL shader. Before: createScaledBitmapWithGainmap_median (ns): 179,455,037 After: createScaledBitmapWithGainmap_median (ns): 15,643,742 Bug: 311085927 Test: benchmark in CL Change-Id: I3f2506a2bd86bbe29c6e0837d31e4a231a7c182a
-
- Nov 15, 2023
-
-
John Reck authored
Skip using the gainmap shader when drawing to a bitmap with a known SDR colorspace as it'll be a no-op in rendering results since W=1.0, but it'll be very expensive in runtime due to CPU evaluation of the SKSL shader. Before: createScaledBitmapWithGainmap_median (ns): 179,455,037 After: createScaledBitmapWithGainmap_median (ns): 15,643,742 Bug: 311085927 Test: benchmark in CL Change-Id: I3f2506a2bd86bbe29c6e0837d31e4a231a7c182a
-
- Nov 14, 2023
-
-
Dichen Zhang authored
Bug: 299202809 Test: YuvImageTest.java Change-Id: Ifb1500114c9c212eb145c8538a74b46066fd5db7
-
- Nov 13, 2023
-
-
John Reck authored
Fixes: 309635621 Test: atest android.uirendering.cts.testclasses.SurfaceViewTests Change-Id: Ie38c491c5a87f19b5352774164187e713bdf5849
-
Seigo Nonaka authored
Bug: 283193133 Test: minikin_tests Change-Id: I214aaa5c94742a4ce84222fb540851d488157ef6
-
- Nov 10, 2023
-
-
Dichen Zhang authored
Bug: b/299202809 Test: YuvImageTest.java Change-Id: If1aa9598f75062e7d0684d5d0f4b60f1e4a19f4d
-
John Reck authored
Test: atest CtsUiRenderingTestCases passes on hwasan build Fixes: 310244900 Change-Id: I9ce160410801c6b7078ecd80eba59e14aac4ef51
-
- Nov 09, 2023
-
-
John Reck authored
Avoids stashing a semaphore in a field use unique_fd to ensure an unused present fence isn't leaked Test: sample app ran in a loop for 3 hours, FD usage remained constant Fixes: 295332012 Change-Id: Iee38e86587805ef3596e5f7efea3ca3b5c7758b3
-
- Nov 08, 2023
-
-
Minkyoung Kim authored
Make active Context valid while buildLayer() When RecordingCanvas::drawPoints() is executed through buildLayer(), active Context is null. It leads to crash Change-Id: Iefe7fa0e056e82f07cd54fa3bfd74180f89ae726
-
- Nov 07, 2023
-
-
Tyler Freeman authored
Bug: 282821643 Test: manual?? Change-Id: I0a3907ede2ad5d0c895ae70d3d1ec47b0902e3b4
-
Tyler Freeman authored
This allows the RenderNode to differentiate between normal force dark and force force-dark (force invert color). In the next CL we will treat force invert slightly differently, not allowing devs to opt-out of force dark in individual Views/RenderNodes. This also sets us up to specify a "force light" mode for force invert, if we ever need it. Bug: 282821643 Test: atest ViewRootImplTest Change-Id: I8cc0fe099cccabdd09ea072aca1e70527e91e1a8
-
Leon Scroggins III authored
Pixel has been using Vulkan in libhwui for several years now. We strongly recommended using it in the Android 14 (U) CDD, and we plan to make it MUST in Android 15 (V). Update the default to match. Bug: 293371537 Bug: 301558076 Test: CtsUiRenderingTestCases, CtsGraphicsTestCases Change-Id: I106381fadf6a686127cb6120afbaef53db5c4aa8
-
Dongya Jiang authored
support visual effects for backdrop contents of a View or RenderNode. Test: added unit test & hwui_unit passes Test: added BackdropBlurActivity in HwAccelerationTest, build & run it Signed-off-by:
Dongya Jiang <jiangdongya@coolpad.com> Change-Id: If1ac1b8aee53667f175e8fa80ecfc7bdfa28173d Merged-In: If1ac1b8aee53667f175e8fa80ecfc7bdfa28173d
-
- Nov 06, 2023
-
-
Seigo Nonaka authored
Revert submission 25211873-font_feature_cache Reason for revert: accessing feature flag in native causes the peformance regression. Reverted changes: /q/submissionid:25211873-font_feature_cache Change-Id: I1f4b619c827d2968fd8b75aa05f65a2ed89b62a2
-
- Nov 01, 2023
-
-
Kevin Lubick authored
Change-Id: I003aa8fcc4de2f6a4e34663a26ebc21d207dcd94 Bug: b/305780908
-
- Oct 28, 2023
-
-
Seigo Nonaka authored
Bug: 283193133 Test: minikin_tests Change-Id: Iae16bbada5f3076b08db9e67d2546eb5014a1fbf
-
- Oct 27, 2023
-
-
John Reck authored
Clip SV to its ancestor clipping bounds. This enables Z-above SurfaceView + scrolling containers to work more naturally Replaces the hidden API of setEnableSurfaceClipping Fixes: 298621623 Test: Sample app Change-Id: Iaa862598e37065677f5ba163a5ac7a6fab2739ea
-
Trevor David Black authored
Bug: 292545615 Test: Build Change-Id: I9a0b42984c3cc8284eb184f412b092a2201a6e4a
-
- Oct 25, 2023
-
-
Seigo Nonaka authored
As a first step of deprecating UI fonts, make setElegantTextFlag as true by default if the target SDK is set to Android 15 or later.. Bug: 279646685 Test: atest PaintDefaultsTest Change-Id: If8e40c075122cafd2f89ae703bf6a5f07ebc7bd0
-