Fix a idmap leak in AssetManager::addSystemOverlays.
And, The fix could also prevent fd leak of idmap.
Test: none
Bug: 32691930
Signed-off-by: Hyangseok Chae <neo.chae@lge.com>
(cherry picked from commit 6a742a38509693f8b39ee9a5ad2803fca12688bf)
Change-Id: Idc4af77db2b0cb739bd6b009b6af0f9123be1aac
from AOSP: https://android-review.googlesource.com/#/c/183305/
On the critical path of the cold launch of applications
the main thread of the started application tells the RenderThread
to create a surface. This process is synchronous and blocks
the main thread of the application until the creation
of the EGLContext is complete.
As a consequence the launch time of the application is delayed
by time spent allocating the EGL Context in the RenderThread.
With this optimization the launch time of any application
is improved (for example settings by 20 to 40 ms).
Change-Id: Ibf47aaa0abb8dedf7aa00693073db3785d9d6b08
Signed-off-by: Thomas Buhot <thomas.buhot@intel.com>
Signed-off-by: Zhiquan Liu <zhiquan.liu@intel.com>
The name 'sku' does not fit well as we recommend to use the same
value for different SKUs if vendors wish to have the same colors
on those SKUs.
Bug: 32268656
Change-Id: Ib5e5a3386676453dafeb13a6a6bf91f81bb48a11
A race could occur when printing the list of Asset allocations for
debugging purposes.
Each Asset object would insert themselves into a global linked list
on construction and remove themselves on destruction. Iterating the list
and the insertion/remove operations all acquire a global lock.
The race occurs after the Asset subclass destructor runs but before the Asset
base class destructor runs, which performs the actual removal from the list.
The vtable of the object being destroyed ends up pointing at the base Asset class'
vtable, and during the iteration of the global list, a pure virtual method is called
leading to an abort, since the wrong vtable is dereferenced.
This change moves the insertion/removal of the Asset object into the global list
to the concrete class, which adds some maintenance overhead but solves the problem.
Bug:31113965
Test: make libandroidfw_tests
Change-Id: I1a620897e5e04a8519ee247883bba0719b1fa6f3
(cherry picked from commit 0358efe4f76f42d9eea91600202a5ab0831d9cef)
idmap --scan is executed as a part of the pre-loading in ZygoteInit.
The pre loading is executed in parallel for each supported architecture
(32/64 bit). This will cause a race condition in the creation of the
overlays.list file and the idmap files for the system overlays.
Apply flock on overlays.list to prevent the file from being thrown away
and recreated when it is in use by another Zygote.
Cherry-picked from AOSP.
Bug: 28032298
Test: manual
Change-Id: I51d39f121d207b11181340b68b164b60020f0c61
This CL fixes a SkShader leak in VD when applying local matrix
to the shader. Specifically, the usage of newWithLocalMatrix(...)
increments the shader's ref count in every draw() call for
Gradient VectorDrawable, whereas there's no balancing call to
decrement the ref count in draw(). In this CL, we assume
the ownership of the shader returned from newWithLocalMatrix(...)
to ensure the correct ref count management.
Also, add test to verify that shader is no longer being leaked
BUG: 32067647
Test: this CL
Change-Id: Ic15fe46cde06a73d81b44e2d3c56b51907344cc0
This CL fixed an issue where calling start() from onAnimationFinished()
caused AnimationListenerBridge::onAnimationFinished(...) to be unsafely
re-entered and the new start listener was (incorrectly) reset to null.
BUG: 31971397
Test: test apk in the bug linked above
Change-Id: Ica809ef2dab884950b93b54f2d0cb4b81e9830f1
The SharedZip's Asset is not thread-safety,the getResourceTableAsset()
and setResourceTableAsset(Asset* asset) function of the SharedZip is
not sync with a same lock.
Consider the following sequence of events:
Thread A calls setResourceTableAsset(Asset* asset),it will set
mResourceTableAsset = asset; then to calls getBuffer() of the asset.
Thread B calls getResourceTableAsset(),which return mResourceTableAsset,
then to calls getBuffer() of the mResourceTableAsset,the asset and
mResourceTableAsset is same one object.
Thread A to delete mZipInflater in getBuffer().
Thread B to delete mZipInflater in getBuffer().
It will cause crash becuase double delete mZipInflater in getBuffer().
https://code.google.com/p/android/issues/detail?id=211941
Bug:31734545
Change-Id: I5a7d67fdf64c4aa03f505b37a2fa840f4443d158
Signed-off-by: songjinshi <songjinshi@xiaomi.com>
See go/sku-colors. This changes the directory to search for framework
overlays if the right system property is defined. This allows
OEMs to specify different resources based on device SKUs.
Bug: 31692079
Change-Id: I9cb121b286b7f52aa26de1757fde1f3110cd47fd
Currently PointerController starts listening to display events
immediately (in its constructor) and never explicitly removes
the callback. The reference dangling from the looper
prevents the PointerController instance from being deleted
when all the clients have released their references.
As a result, when USB or BT mouse is disconnected,
the mouse stays frozen on screen and only goes away
after a 15 sec inactivity timeout.
This change introduces an intermediary LooperCallback
which holds only a weak reference to PointerController.
The pointer now disappears immediately upon mouse
disconnect.
Bug: 30824220
Change-Id: I5f7208dbfa381b3e21f248cc0da402f307faa184
bug:30895941
Prevents a race where frame work could interleave between frames,
causing SurfaceView position updates to be delivered out of order.
Change-Id: I01e4cc557b69dcf33e877a0e16c0d115ec95e4cc
Bug: 30442298
We already do this for initialize(), fix
it so that update() is parallel with the
UI thread as well.
Shaves ~7ms off of the 99th percentile on
NotificationShade open & close
Change-Id: I1791df495453fb9e1e12362c68e3d20e837e62be
bug:30537130
BeginLayerOps were being rejected in a way that allowed the associated
EndLayerOps to still be recorded. This was a violation of DisplayList
content expectations, and caused crashes in FrameBuilder when trying to
play these DisplayLists back.
Change-Id: I531b840aa5c4ffb1ee458da3f4b366978eaeafbe
This CL fixed an integer overflow error caused by using integer literal
when it should be a float literal that defines a double.
BUG:30739855
Change-Id: If2bdb0eb7a2e93d28434bfe72ec11e8f1f2a1eda
