Test: Build + Th + boot | This should be a no-op change
Bug: 277168419
Change-Id: Iae5fe93b721528af57578009c117e8e5ae3becbe
Merged-In: Iae5fe93b721528af57578009c117e8e5ae3becbe

Test: Build + Th + boot | This should be a no-op change
Bug: 277168419
Change-Id: I0ddc4496ed6fbf785613841eb85a71a4f2a92425
Merged-In: I0ddc4496ed6fbf785613841eb85a71a4f2a92425

Without array that are causing conflict with internal branches
Conflicting array are clean in follow-up

Test: Build + Th + boot | This should be a no-op change
Bug: 277168419
Change-Id: I555f87510b48cd5704ce9349ccda4060ea42db4b

Problem: While bluetooth is streaming media (with appropriate audio
focus), any other media/audio app can acquire transient audio focus
i.e. AUDIOFOCUS_GAIN_TRANSIENT and AUDIOFOCUS_GAIN_TRANSIENT_MAY_DUCK
This results in bluetooth process to receive the respective LOSS
events. When it receives the AUDIOFOCUS_LOSS_TRANSIENT it sends a
pause command to the remote device. In addition, it also mutes the
stream in case the remote device rejects the pause command to ensure
no sound is made by the bluetooth process. The logic to set the gain
was broken post the lib AAudio migration resulting in bluetooth
process making additional sound on top of the other audio app.
Similarly, when other media/audio app had requested for gain_may_duck
focus, bluetooth process wasn't ducking the media sound resulting in
overlap of both the audio.

Solution: Updated the business logic to update the stream buffer
based on the gain value set by the bluetooth process. On different
transient audio focus events, different gain (in range of [0.0f,1.0f])
is set by the handler. On focus gain event, it sets back the gain to
1.0f.

Bug:254724838
Tag:#stability
Test: Tested by setting the different gain value in the code and
validating on device.

Change-Id: Iafedf759949df6e54cad7f21d64dc247712fd380

The following commands are optional in the specification,
but Android is actually using them (without validation) and
they are required to enable LL privacy.

/!\ This change just implements the commands, they are not
    enabled in the Supported Commands mask.

Bug: 276513139
Test: m root-canal
Change-Id: Ia07ee5d6592ed4a1f5574d773dbb78deadea0292

Bug: 277105543
Tag: #floss
Test: manual
Change-Id: Ia33ae040e7488aaf1d34d7861600527e8b861a26

framework

Tag: #feature
Bug: 265077412
Test: atest DatabaseManagerTest
Change-Id: I5e2b2e2cd76feca010735cd08c749b5ccd82f17e

Test: atest BumbleBluetoothTests
Bug: 266873249
Change-Id: Iacb1a99cbe6ebcd8cb5ee9478d2fbd1f9357cb2b

Rather than depending on libbluetooth-core, this exchanges the order
so that libbluetooth-core depends on libbluetooth_core_rs.

But since they are static libraries, circular symbol dependencies are
allowed, but header-only dependencies can only flow one way. This lets
C++ access types defined only in Rust, while the reverse direction
still works.

Bug: 272572974
Test: mm
Change-Id: I64d063520104c611374c467a66a9b516f36f334f

Bug: 271343657
Merged-In: I1be2cd3b5daa08e24ef44ace7901b3312d2f14ce
Change-Id: Ifd404a6976fab0b0ca632ef928d621a00fe8c4bf

- Preparation for migrating RootCanal to github,
  removing the final dependencies to the bluetooth stack

- Will allow editing of the hci_packets.pdl source
  without incurring large changes in gabeldorsche

Bug: 256013143
Test: mmm packages/modules/Bluetooth
Change-Id: If79872ca4857786d8bd7a33f5674aa6b7ea15746

We want to have non connected set members in the allow list when group
is streaming.

On the event that one device got disconnected we do as follow
1. If this is connection timeout - device goes to allow list when group
   is streaming
2. If this is disconnection initiated by remote side don't know if both
   devices are going to be disconnected or it is just single earbud put
   back to the charing case. With this patch, when earbud got disconnected
   by remote side, we schedule a function to check after 3sec if group is
   still streaming - if so device goes to allow list.

Bug: 275753723
Test: manual (streaming, put one bud to the charging case, wait for
screen off, take it back)
Test: atest BluetoothInstrumentationTests
Tag: #feature

Change-Id: I26dc62ec75c8642fe9b5c686563cc0dec57adb38

Also handles after merge comment in
aosp/2520603

Bug: 276693370
Test: manual
Tag: #feature
Change-Id: I28513860646737ce65e9331d2da598131bb9ae34

* changes:
  PTS: Add GAP/SEC/AUT/BV-13-C
  Floss: Set le sec_flags in btm_sec_encrypt_change

The INIT_classic_discovery_only flag starts inquiry with both the BR and
LE mode flags set (so that LE results are parsed properly for inquiry).
However, we need to make sure to complete both BR and LE modes when we
complete so that we end discovery correctly.

Bug: 277192596
Tag: #floss
Test: Start discovery, let it end naturally and confirm state cb
Change-Id: I15edfc324b8372602748a8935a2fbe9269930aac

Pend the volume update event until the device is fully connected if not
yet.

BUG: 255700075
Tag: #floss
Test: ./build.py and test with a2dp and hfp device

Change-Id: I557df78d90f6bd55569c27da444b3e39e40100dc

Test: Build + Th + boot | This should be a no-op change
Bug: 277168419
Change-Id: I09623a2256170a3315ed18d6f9070b428a3cd904

Tag: #feature
Bug: 265077412
Test: Manual
Change-Id: I485e255e2d11309670025f3b45dfac807ef4586f

framework

Tag: #feature
Bug: 265077412
Test: Manual
Change-Id: I8f57a6a628c4dc6a1f35a889d2e6d921d0a6758c