This was causing stack stitching problems where a one-way call with
violations followed by a two-way call without violations was getting
the previous one-way call's violation stack stitched on to the second
caller's stack.
The solution is a little more indirect than I would've liked
(preserving the binder's onTransact flags until enforceInterface) but
was seemingly necessary to work without changing the AIDL compiler.
It should also be sufficiently cheap, since no new calls to
thread-local IPCThreadState lookups were required. The additional
work is just same-thread getter/setters on the existing
IPCThreadState.
Change-Id: I4b6db1d445c56e868e6d0d7be3ba6849f4ef23ae
Merge commit 'b8fd047311e329f2b8dbe3d228488ba844718ee1' into gingerbread-plus-aosp
* commit 'b8fd047311e329f2b8dbe3d228488ba844718ee1':
Fix for 512 limit in assetManager.list
Also replaced all doThrow by jniThrow.
OutOfMemory after string creation were removed: should have been thrown before.
Bug http://b/issue?id=2949164
Change-Id: Idea8e27fdedeb43e3976776c477766e4dcdebcf8
This change is the result of three cherry-picks:
- Add dalvik.vm.gc.preverify dalvik.vm.gc.postverify properties.
git cherry-pick --no-commit 0ef82fcf
- Add the property dalvik.vm.gc.verifycardtable to set the -Xgc:-Xgc:verifycardtable option for the vm.
git cherry-pick --no-commit 8b4faf54
- Eliminate short JIT debugging properties
git cherry-pick --no-commit 57a673fc3db9d84908467ae6d245fd60d4637b2f
Change-Id: I5f8002ed1e431344570add02f58e2641c8fae549
Merge commit '9f7257f0a14cd17d8b55bc3bd1584dac2f9dd1b2'
* commit '9f7257f0a14cd17d8b55bc3bd1584dac2f9dd1b2':
Fix a bug, where one thread is using JNIEnv associated with another thread.
Before this change, all framework assets would be decoded at drawing time
outside of zygote. This was forcing all apps to re-decode the assets and
zygote to keep an in-memory copy of each asset. This behavior is now
opt-in by setting the inPurgeable flag on BitmapFactory.Options.
Change-Id: Ief823139163d8071b8ee1267746622faf52eb8ec
Before this change, all framework assets would be decoded at drawing time
outside of zygote. This was forcing all apps to re-decode the assets and
zygote to keep an in-memory copy of each asset. This behavior is now
opt-in by setting the inPurgeable flag on BitmapFactory.Options.
Change-Id: Ic703f57adb26b2a701ecff0a653d35a93e26d47c
We see abort messages like this when using JavaPixelAllocator and JavaMemoryUsageReporter.
W/dalvikvm( 680): JNI WARNING: threadid=2 using env from threadid=10
W/dalvikvm( 680): in Landroid/graphics/LargeBitmap;.nativeClean (I)V (CallVoidMethodV)
To fix it, we keep JavaVM, rather than JNIEnv, in JavaPixelAllocator and JavaMemoryUsageReporter,
because JavaVM allows us to get the JNIEnv corresponds to the current thread.
Change-Id: Ibd4b394a53dd3fdccc5a442eeb0dedf836479575
* Add flags field in OBB footer to support overlays.
* Remove unused 'crypto' and 'filesystem' fields in obbtool (could
later be supported in the "flags" field of the OBB footer).
* Add notes to document OBB classes before shipping.
Change-Id: I386b43c32c5edef55210acb5d3322639c08010ba
This depends on a kernel patch that implements read(2)
in the ashmem driver.
Bug http://b/issue?id=2595601
Change-Id: Ie3b10aa471aada21812b35e63954c1b2f0a7b042
Previously, the input dispatcher assumed that the input channel's
receive pipe file descriptor was a sufficiently unique identifier for
looking up input channels in its various tables. However, it can happen
that an input channel is disposed and then a new input channel is
immediately created that reuses the same file descriptor. Ordinarily
this is not a problem, however there is a small opportunity for a race
to arise in InputQueue.
When InputQueue receives an input event from the dispatcher, it
generates a finishedToken that encodes the channel's receive pipe fd,
and a sequence number. The finishedToken is used by the ViewRoot
as a handle for the event so that it can tell the InputQueue when
the event has finished being processed.
Here is the race:
1. InputQueue receives an input event, assigns a new finishedToken.
2. ViewRoot begins processing the input event.
3. During processing, ViewRoot unregisters the InputChannel.
4. A new InputChannel is created and is registered with the Input Queue.
This InputChannel happens to have the same receive pipe fd as
the one previously registered.
5. ViewRoot tells the InputQueue that it has finished processing the
input event, passing along the original finishedToken.
6. InputQueue throws an exception because the finishedToken's receive
pipe fd is registered but the sequence number is incorrect so it
assumes that the client has called finish spuriously.
The fix is to include a unique connection id within the finishedToken so
that the InputQueue can accurately confirm that the token belongs to
the currently registered InputChannel rather than to an old one that
happened to have the same receive pipe fd. When it notices this, it
ignores the spurious finish.
I've also made a couple of other small changes to avoid similar races
elsewhere.
This patch set also includes a fix to synthesize a finished signal
when the input channel is unregistered on the client side to
help keep the server and client in sync.
Bug: 2834068
Change-Id: I1de34a36249ab74c359c2c67a57e333543400f7b
Add extension to WifiLock to allow apps to operate
in high performance mode (high power & disable suspend
optimizations for battery consumption).
Bug: 2834260
Change-Id: I8b33d307f3d569bc92ba2139b9ed224ffc147547
When the driver was configured to run with power save mode disabled the
power save mode incorrectly got reverted back to AUTO mode right after
DHCP response. The power save mode value is now saved so that the device
properly reverts back to a previous mode after DHCP response.
Change-Id: Ie68cd107872d233bf422e24130a1eb9f6432db91
Bug: 2834260
This change adds the following blending modes for shaders and color filters:
Add
Multiply
Screen
Overlay
Darken
Lighten
Change-Id: Iff22f5ce6041b43c71b1857d73013f5010ab3413
This introduces basic infrastructure that should allow content
providers holding complex data to perform on-demand conversion
of their data to streams of various types. It is achieved through
two new content provider APIs, one to interrogate the possible
stream MIME types the provider can return, and the other to
request a stream of data in a particular MIME type.
Because implementations of this will often need to do on-demand
data conversion, there is also a utility intoduced in ContentProvider
for subclasses to easily run a function to write data into a
pipe that is read by the client.
This feature is mostly intended for cut and paste and drag and
drop, as the complex data interchange allowing the source and
destination to negotiate data types and copy (possible large)
data between them. However because it is fundamental facility
of ContentProvider, it can be used in other places, such as for
more advanced GET_CONTENT data exchanges.
An example implementation of this would be in ContactsProvider,
which can now provider a data stream when a client opens certain
pieces of it data, to return data as flat text, a vcard, or other
format.
Change-Id: I58627ea4ed359aa7cf2c66274adb18306c209cb2
The makefile variable USE_OPENGL_RENDERER must be set to true to compile
libhwui and the related code in the JNI layer.
This change also removes obsolete APIs from Canvas that must not be used
and would be confusing if left in. These APIs were remnants of our first
attempt at an OpenGL renderer for the view hierarchy and had not been
taken out before Android 1.0 was released.
Change-Id: I2475ff1307212bab26c926724f3c508681c7dae1
