Floss needs this field in order to emulate how metrics work in
BlueZ. This CL calls the SDP/GATT function in bta_dm_act and pass
the result to btif, where it is stored and also passed further to
the upper layers.

Bug: 240782193
Tag: #floss
Test: Verify VendorId field in /var/lib/bluetooth/bt_config.conf
      after pairing with an LE and Classic devices.
Change-Id: I16f60aaf0de546b0eab3af705abb790610161489

Floss needs this field from the EIR in order to emulate how metrics
work in BlueZ. Therefore, store the appearance into storage, and pass
it with the callback for Floss only.

Bug: 240782193
Tag: #floss
Test: Verify Appearance field in /var/lib/bluetooth/bt_config.conf
      after pairing with an LE mouse.

Change-Id: Iac4f2fea7f1f4f711b3f99a7d416799831b1f948

Test: HOGP PTS tests (next CL)
Bug: 242326310
Tag: #feature
Change-Id: Iabf59cc20d1392bf341f925dabe58421629d8917

As per offline discussion, all pairing/bonding related interfaces for
both classic + LE will be moved out of Host (since Host is getting huge)
and put in a separate interface. We can't call it SM since we have
classic stuff in here too.

Test: existing PTS
Bug: 242326310
Tag: #refactor
Change-Id: I2704098edd72e50d53366b210a2bf425f11d3388

* changes:
  Fix admin enable/disable of Bluetooth file sharing, part 2
  Fix admin enable/disable of Bluetooth file sharing

BluetoothOppLauncherActivity & BluetoothOppBtEnableActivity
'enabled' state should be updated wether Bluetooth is ON or
OFF. Previously, the state of these Activities was set with
resource config, but since Bluetooth started replacing them
by sysprops, we had to enable / disable these Activities in
BluetoothOppService. However, the bt sharing tile should be
available even when Bluetooth is OFF, and as their state is
already updated in BluetoothManagerService when there is an
user change, we can use that instead but with the condition
that BluetoothOppBtEnableActivity is enabled too. All other
Bluetooth OPP Activities enabled / disabled state are still
handled in BluetoothOppService as they require Bluetooth to
be ON to work anyway, and don't impact UI tiles. This fixes
both an issue where the bt sharing tile doesn't appear when
Bluetooth has been turned OFF, and an issue where Bluetooth
would crash if it hadn't been turned ON before using the bt
sharing tile.

Test: manual
Tag: #feature
Bug: 242963176
Merged-In: I67840dbd9f4cfc42026a14a1fc467fc95aa3a587
Change-Id: I67840dbd9f4cfc42026a14a1fc467fc95aa3a587

This fixes the bluetooth package detection functionality in the case
when *all* components are disabled. This occurs in the CTS test, but did
not occur in manual testing of ag/19219300.

Also did manual testing to make sure we don't regress (toggled the no_bluetooth_sharing option, verified that it appeared/disappeared)

Bug: 239372439
Test: atest CtsDevicePolicyManagerTestCases:DeviceOwnerTest
Tag: #stability
Merged-In: Ib9530a8829a946ea00a21111e459d697086f2ef8
Change-Id: Ib9530a8829a946ea00a21111e459d697086f2ef8

This is an extremely partial fix for a mostly-broken feature. For
fully-managed devices, this fixes the toggle for the file sharing
option, by detecting which Bluetooth package should be used rather than
just guessing.

However, on work profiles, the file sharing option does not really work
anyway, it just infinitely loads. In addition, it is not correctly
disabled when Bluetooth stops, so the icon just remains but doesn't
do anything.

A full fix would add the enable/disable sequence for *work* UserHandles
to BluetoothOppManager, so the LAUNCHER_ACTIVITY is properly enabled and
disabled. In addition, it would fix the issue described in the linked
bug to prevent infinite loading (dedicated bug: b/112625123).

Tag: #stability
Bug: 237389774
Test: Manual, verifying the *badged* BT file sharing icon appears/disappears as needed
Merged-In: I070987bc884947246c372c65691c4866ebd1e7c7
Change-Id: I070987bc884947246c372c65691c4866ebd1e7c7

This fixes a bug in the GD LE scanning manager where the own address type is always set to use the random address on startup. This bug causes the BT process to crash when we try to scan while the sysprop
bluetooth.core.gap.le.privacy.enabled is set to false.

Tag: #stability
Bug: 243862787
Test: system/gd/cert/run --clean --sl4a_sl4a IrkRotationTest or turn on BT after setting the sysprop to false
Change-Id: Ifd93743f8fd8e4c6b412cb3747a67a35554e97b2

Fix: 212644286
Test: QA testing
Change-Id: Ic9f96affbee64382873e12bc766254ea8260f0b1

The OnPair interface (as well as potential others in the future) are
bidirectional streaming rpcs. This CL adds a utility in Kotlin to
emulate the native-Kotlin gRPC interface on top of the Java interface
provided (Flow -> Flow rather than callback-based).

In Python, the native interface consumes an iterator and supplies a new
one. So to interleave rx and tx messages, we need multiple threads. This
CL also adds a client-side utility to simplify usage, by making it mimic
channels (which are the grpc.aio interface, as well as what Go gRPC
does). It means we consume one thread for every active streaming gRPC
call, but this is unavoidable unless we use asyncio channels, which
would require replacing all the mmi2grpc calls with async calls.

Bug: 242326310
Test: next CL
Tag: #feature
Change-Id: I526982d3163120689797919382a7dff8ff33378f

Needed for HID PTS tests. We just "fake" the link policy and our current
role by tracking them locally, rather than using LMP to coordinate with
the peer. This can be changed at a future date if a test depends on this
working.

Bug: 239986609
Test: HID PTS tests
Tag: #feature
Change-Id: I3b2b97a200b0279251fc68e37e23b8d3cfb8bdeb

Same type configuration would be shared between multiple context types.
This will limit internal configuration handovers and stream
reconfigurations while changing context type.

Bug: 240145073
Bug: 233961821
Tag: #feature
Sponsor: jpawlowski@
Test: atest bluetooth_le_audio_test
Test: atest bluetooth_le_audio_client_test
Change-Id: Ie0c19a0a4d1958ea6e641e7a9c3297cf611a5851

Merge "[PTS-Bot] Add 7 AVCTP test cases AVCTP/TG/CCM/BV-03-C AVCTP/TG/CCM/BV-04-C AVCTP/CT/CCM/BV-02-C AVCTP/CT/CCM/BV-03-C AVCTP/TG/NFR/BV-03-C AVCTP/TG/NFR/BI-01-C AVCTP/TG/FRA/BV-03-C"

- Read characteristics using UUID now returns a list of
read characteristics. Each characteristics that has a
matching UUID and handle contained in the start handle
and end handle of the request is read.
- MMIs don't return "No" anymore, instead either we
assert that the result is correct or raise an exception if
the check does not pass
- Ran the ktfmt on Gatt.kt and GattInstance.kt
- Fixed SDP tests that checked the wrong value

Bug: 239103399
Test: atest pts-bot:GATT/CL/GAR --rerun-until-failure
Test: atest pts-bot:GATT/CL/GAD --rerun-until-failure
Tag: #feature
Change-Id: Ife8d43545c77c2eb52d0a0ca57643a6213c0103c

Add check for string size before creating slice. Add unit tests.

Bug: 243426979
Tag: #floss
Test: ./build.py --target test
Change-Id: Iae1cc0dcf48951ca41d652c181da84a940d342ea

AVCTP/TG/CCM/BV-03-C
AVCTP/TG/CCM/BV-04-C
AVCTP/CT/CCM/BV-02-C
AVCTP/CT/CCM/BV-03-C
AVCTP/TG/NFR/BV-03-C
AVCTP/TG/NFR/BI-01-C
AVCTP/TG/FRA/BV-03-C

Test: atest pts-bot:AVCTP -v
Change-Id: Ide34f01b8c22d9077ec5d45e92854979d40ac08e

The only exception is hearing aid device,
which isn't overriden by another profile
of another device.

Bug: 237496193
Test: atest ActiveDeviceManagerTest
Tag: #feature
Change-Id: Ia9c3b255a3613d0e10fb362dca5b113adbd92a19

btif_hh_connect() can be called through main thread and it lead to crash
with following message. "JNI DETECTED ERROR IN APPLICATION: a thread
xxxx is making JNI calls without being attached"
So we fix to switch callback to jni thread.

Bug: 243753956
Tag: #stability
Test: reconnect some mouse after BT turning on
Change-Id: Ib3576bca9026e7fdf164c972f2da22471005e01f