Calling into IBluetoothLeAudio will call the LeAudioService from
Bluetooth app that does not live in the systemServer.
Therefore, exception should be rethrown as runtimeException

Bug: 298264617
Test: atest CtsBluetoothTestCases | Cts does not handled failure in the
      binder nor any of the Test framework we currently have
Change-Id: I6adb47ceb1d075f001ff972dded6e44716368363

Calling into IBluetoothVolumeControl will call the VolumeControlService
from Bluetooth app that does not live in the systemServer.
Therefore, exception should be rethrown as runtimeException

Keeping the only call into the systemServer as a rethrow from
systemServer

Bug: 298264617
Test: atest CtsBluetoothTestCases | Cts does not handled failure in the
      binder nor any of the Test framework we currently have
Change-Id: Iaa23917a29833671c3e21a4a89d2b2fd92c808a9

Calling into IBluetoothCsipSetCoordinator will call the
CsipSetCoordinatorService from Bluetooth app that does not live in the
systemServer.
Therefore, exception should be rethrown as runtimeException

Bug: 298264617
Test: atest CtsBluetoothTestCases | Cts does not handled failure in the
      binder nor any of the Test framework we currently have
Change-Id: Id89076adf63c0cbaa6d536b04ae1421ae5464745

Calling into IBluetoothLeAudio will call the LeAudioService from
Bluetooth app that does not live in the systemServer.
Therefore, exception should be rethrown as runtimeException

Bug: 298264617
Test: atest CtsBluetoothTestCases | Cts does not handled failure in the
      binder nor any of the Test framework we currently have
Change-Id: Idfb7ae662fcadf33358bfd728554fcdb25e92a4c

Calling into IBluetoothHeadset will call the HeadsetService from
Bluetooth app that does not live in the systemServer.
Therefore, exception should be rethrown as runtimeException

Bug: 298264617
Test: atest CtsBluetoothTestCases | Cts does not handled failure in the
      binder nor any of the Test framework we currently have
Change-Id: I617c28686fc3af639b2ca980129974633403ff2f

Calling into mService(IBluetoothGatt) will call the GattService from
Bluetooth app that does not live in the systemServer.
Therefore, exception should be rethrown as runtimeException

Bug: 298264617
Test: atest CtsBluetoothTestCases | Cts does not handled failure in the
      binder nor any of the Test framework we currently have
Change-Id: Ie2ed38e5076242c1001e231f70105890c86ea116

Calling into mService(IBluetoothGatt) will call the GattService from
Bluetooth app that does not live in the systemServer.
Therefore, exception should be rethrown as runtimeException

Bug: 298264617
Test: atest CtsBluetoothTestCases | Cts does not handled failure in the
      binder nor any of the Test framework we currently have
Change-Id: Ia75654e3d5a6060a72d4031499ac98f8c89da697

Merge "Make Phone Policy changes on connection state changes handled via direct calls routed via adapter service." into main

Change the include header and BUILD.gn libs.

Bug: 294165109
Tag: #floss
Test: m - None.

Change-Id: I62e354678591affeb43feb1b3ac2da3f01bf8752

Merge "Pandora: Refactor the helpers for disconnecting and reconnecting rootcanal devices" into main

Make Phone Policy changes on connection state changes handled via direct calls routed via adapter service.

Tag: #refactor
Bug: 296932947
Test: atest BluetoothInstrumentationTests
Change-Id: Ib4ad2f08d047e4a63bc01ecdd467b72f1bb78218

Directly ask the app what profiles are connected.
This removes the need to store profiles and register listeners.

Bug: 271963349
Test: atest BluetoothAirplaneModeListenerTest
Change-Id: I4c31e5ca855fc2007cf6d9a35537eb8e0eb8e78e

- fix issues with device and phy ids
- simplify the API to be more representative of the expected behaviour

Bug: 281083061
Test: atest pts-bot
Change-Id: I3f2916923c867cb68b2cdad0ada228dcfa33656e

* changes:
  Protect system APIs based on feedback.
  Apply delivery group policies to ACTION_DISCOVERY_FINISHED.

API council highlighted that system APIs should be protected with
permissions, which this change adds.  It also cleans up the
interactive signal to be on BroadcastOptions, since broadcasts are
the only client.

Bug: 267646347, 264907403
Test: TH
Ignore-AOSP-First: platform API not in AOSP
(cherry picked from https://googleplex-android-review.googlesource.com/q/commit:9247108c2c6a5c5a6b8e370f209d0f6ff1c85108)
Merged-In: I34b1a3121334fb14b0292c48d6a589a1bb73e629
Change-Id: I34b1a3121334fb14b0292c48d6a589a1bb73e629

- The "set-defer-until-active" policy is applied so that the
broadcast targeted to apps in the Cached state is deferred until
they come out of that state.
- The "deliver-most-recent" policy is applied so that if there are
already pending broadcasts waiting to be delivered when a new
broadcast is sent, the old ones are discarded.

Bug: 266474854
Test: atest CtsBluetoothTestCases
(cherry picked from https://googleplex-android-review.googlesource.com/q/commit:a44cb9326a83c85340fb96fd967486b5168fa0f7)
Merged-In: Iaa505809fdb2f67a440a9a568fd44a753814b0e2
Change-Id: Iaa505809fdb2f67a440a9a568fd44a753814b0e2

In conversational mode, when we disable 32k bidirectional dual mic,
we select other configs like 16k bidirectional dual mic.

However, when users connect to only one bud while they put one earbud
back into cases, we select single device on 32k, which is not suitable.

This is because CIG is already configured for 16k but when there is
only one device connected, we choose 32k for single device, and this
will make us removing CIG and create CIG again.

The user experience is not good as the audio will stop for some time.
This patch handles this case.

Also, not insert config to the mpa when it is empty.

Bug: 298946279
Test: check early exit and later join on conversational streaming
Test: atest -c bluetooth_le_audio_test bluetooth_le_audio_client_test
Change-Id: I50a0ccc0906a2c3edefcfa553224498e87a54267

Merge "Remove all ACTION_CONNECTION_STATE_CHANGED for A2DP, A2DP Sink, HFP, HFP Client, LE Audio in Scan Manager." into main

If device got disconnected while LeAudio service is connected, it could
stay in the state, that handles are known and stored, but LeAudio group
id is not yet know.
In such case, device will stay in wrong state.

This patch fixes it.
Bug: 296761332
Test: atest bluetooth_le_audio_client_test
Tag: #feature

Change-Id: I8f58fa65c8cd7a7f334605f73c8dd75666877e58

This reverts commit 87367338.

Bug: 257375445
Test: atest BluetoothInstrumentationTests
Reason for revert: The tests are not flaky anymore
(cherry picked from https://android-review.googlesource.com/q/commit:a5fd48440610185ee229c798e49610bed8589e71)
Merged-In: I175110efe6b126c47377cd607c31f895087db353
Change-Id: I175110efe6b126c47377cd607c31f895087db353

This specific commands lets a tester change the configuration
of a connected device at runtime.

This requirement originates from the PTS-bot test runner
version 8.4.1, which requires different, non compatible dongles
for different tests (e.g. LE only v5.4+ for LE audio tests and
Dual v5.0- for A2DP tests).

The alternate solution would be to open multiple HCI ports
for all required configurations, but it does add complexity
to the test configuration whereby set_device_configuration
can be invoked prior to a test execution.

Bug: 282818191
Test: m root-canal && manual
Change-Id: I1d7570223679f9ab432d6681bbcd6c47bb42c03a

Remove all ACTION_CONNECTION_STATE_CHANGED for A2DP, A2DP Sink, HFP, HFP Client, LE Audio in Scan Manager.

Bug: 296932947
Test: atest BluetoothInstrumentationTests
Tag: #refactor
Change-Id: Iaa7881e122f4f59d9be43c502578928e93abafc0

Bug: 298708986
Test: m .

Change-Id: Iee38a7f0c4abf159473ee37910691e4dca041abd

The HAL crashing is an unrecoverable failure. Sometimes, the system
kills the HAL as part of a system shutdown, in which case Bluetooth
will also be killed. Sleeping for one second after being notified that
the HAL died is a way to make sure we don't report deaths that are a
result of the system killing the HAL out from under the Bluetooth
stack.

Bug: 294022818
Test: mma -j32
Test: Kill the hal, then Bluetooth with different timings
Change-Id: I9c17f4637639852965fdb529687a43ab26fde776