Dead lock sequence:
1. called onSetA2dpConnectionState()
1-1. synchronized(mConnectedDevices)
2. called onServiceDisconnected()
2-1. synchronized(mA2DPAvrcpLock)
3. waiting to lock(mConnectedDevices) in onServiceDisconnected()
4. waiting to lock(mA2DPAvrcpLock) in onSetA2dpConnectionState()
========================================================================================
- watchdog issue
----- pid 3306 at 2014-10-14 16:15:12 -----
Cmd line: system_server
"main" prio=5 tid=1 MONITOR
- waiting to lock <0x4343e9b0> (a java.util.HashMap) held by tid=46 (AudioService)
at android.bluetooth.BluetoothA2dp$2.onServiceDisconnected(BluetoothA2dp.java:529)
"AudioService" prio=5 tid=46 MONITOR
- waiting to lock <0x42ab7a58> (a java.lang.Object) held by tid=1 (main)
========================================================================================
Signed-off-by: Seunghwan Hong <seunghwan.hong@lge.com>
Change-Id: I99e061c07be01aabcd26786ef2ebb71f46717b93
CallStaticVoidMethod is varargs function, and calling it with
a literal 0 like CallStaticVoidMethod(..., 0) will treat the
argument as a 4 byte int in both 32 and 64 bit processes.
This is incorrect for pointer arguments where NULL should be
used instead.
Reviewed-by: Liao, Bruce <bruce.liao@intel.com>
Signed-off-by: Yong Yao <yong.yao@intel.com>
Change-Id: I9d700d3790a80dbee6826f64baf9ef5d81ca390f
Adjust format strings to not produce Clang warnings in both 32-bit and
64-bit builds
Change-Id: I76c29d8d5d0fb4b5e9d9518077652370ffe9e871
Signed-off-by: Bernhard Rosenkränzer <Bernhard.Rosenkranzer@linaro.org>
- allow Media.Image.setCropRect(null) to unset cropRect
- use full image if cropRect is null in MediaCodec$Image
Bug: 18124320
Change-Id: I7abb6175402a1fe913e16c0a682fabee79c55741
Don't refer to classes as structs and view versa -- causes warnings in
clang (in code built with -Werror)
Change-Id: I4d7f461bcf6fe4e43a14ad2f28ee0e0ddea95abc
Signed-off-by: Bernhard Rosenkränzer <Bernhard.Rosenkranzer@linaro.org>
Some JNI functions ignore the JNI environment and class information, but
still take the parameters -- causing a build failure with clang (and gcc
with -Wextra enabled). Ignore this.
Change-Id: I049fcf65991b19d2416fce105699311803b43cfc
Signed-off-by: Bernhard Rosenkränzer <Bernhard.Rosenkranzer@linaro.org>
Remove unused variables and static functions clang complains about,
disable warnings about unused parameters (needed for clang and for gcc
with -Wextra enabled)
Change-Id: I76a22cd0158b3c7375c54e3d4d15bc1ac448591e
Signed-off-by: Bernhard Rosenkränzer <Bernhard.Rosenkranzer@linaro.org>
C++11 defines a real char16_t, which is not implicitly convertible to
uint16_t (and by extension jchar). Add casts as needed.
Bug: 18300613
Change-Id: I00752002ef2e938bdb57f70947e8fd53ec103293
There is race between SoundPoolThread and SoundPool / AudioManager
threads during releasing SoundPool.
AudioManager deletes a global reference before setting SoundPool
callback to NULL. If, at that time, a call to the SoundPool::notify
fuction happens then mCallback is valid but mUserData is not.
The following log will show up to indicate the problem:
JNI ERROR (app bug): accessed deleted global reference 0xXXXXXXXX
This fix is to clear the SoundPool's callback before releasing global
reference.
Change-Id: I5e6d647edc0444340db879428048e2c0a068a8b4
We weren't setting the ratings in the new MediaMetadata from RCC.
This adds some code to set them correctly.
bug:18051829
Change-Id: I358eeb2b25b3fb9bd9a5d207eb3314bb2cdff25e
Apps can end up in priority mode by setting ringer-mode = silent.
Now they can leave priority mode by setting ringer-mode = non-silent.
(normal or vibrate)
Bug: 17884168
Change-Id: I54c853885f4ae9ee618041dd7ac6ab0663fc7b37
Current implmentation only sends rotation updates on orientation
changes, so does not handle direct 0<->180 or 90<->270 transitions.
Update rotation based on an OrientationEventListener instead of
Intent.ACTION_CONFIGURATION_CHANGED
Bug 17606902
Change-Id: I01dfcd1c587f5b2e8a96365c2389782ad77936ef
If an AudioRecord is created with the "fixedVolume" tag
when recording from REMOTE_SUBMIX, treat the
device DEVICE_OUT_REMOTE_SUBMIX as a fixed full volume
device during the recording. Also register a death
handler during the recording.
Otherwise this is a no-op.
Bug 17635294
Change-Id: I8d26fe777047126f34308e1e1b7ac28ba269ad89
Found a very simple fix. Just set the cached metadata so we have
it to copy over when editMetadata(false) is called.
bug:17796693
Change-Id: Ib27f0c3d28e7f2a3c7d9495697f36c8045e2bcf6
