Bug: 301661850
Test: mma packages/modules/Bluetooth
Change-Id: I55c60f4a91ae0a2ca1cda2045b48d960af756b90

Bug: 301661850
Test: mma packages/modules/Bluetooth
Change-Id: Ia3230e88c944a935bd5f62dd6d15db35b75004a9

Bug: 301661850
Test: mma packages/modules/Bluetooth
Change-Id: I71d0af48dbf6ce62b9f523e5ec6c4ae70dabadef

Bug: 301661850
Test: mma packages/modules/Bluetooth
Change-Id: I3a0e12a21ae434ef3811fc48306f34ddf03c7c54

* changes:
  Remove unused includes from mock .h files
  Remove unused includes from mock .cc files
  Remove UNUSED_ATTR from mocks

Bug: 305971273
Test: mma -j32
Change-Id: Iaf587f8f1b6776575811433e49b36c62f398dcbf

Bug: 305971273
Test: mma -j32
Change-Id: I115c6dfd3397881500dcba341b10c6fa1e511e78

Bug: 305971273
Test: mma -j32
Change-Id: I4b323bbaf974de70af1a2a2e6083428e9d9372bc

Bug: 305971273
Test: mma -j32
Change-Id: Ia74a8c70b3f7dcf5c74b4d452a894018af6be7d6

Module dumper works *with* modules but is not a module

Module should have no knowledge of wakelocks

Bug: 304519090
Test: m .

Change-Id: I8122c1f47d204e107d410760c6b1f726025a3976

Allow latency to be same as sdu interval.

Bug: 300216034
Test: atest bluetooth_le_audio_test
Tag: #feature
Change-Id: I47bff899acd3af5b074dc35327be24d4e555ea7c

When the LocalSink is resumed when stream is already created,
the context logic would incorrecty changed metadata to LIVE.
This patch fixes that.

Bug: 302481144
Test: atest bluetooth_le_audio_client_test
Tag: #feature

Change-Id: I8025dcf388217303e6a4c315b03fdfd19278dbea

Bug: 298582813
Test: atest VolumeControlService
Tag: #feature
Change-Id: I50b9f3a82e543bcb79f9c0fc0004704fb4e84b32

* changes:
  Floss: Remove automatic SCO call start
  Floss: Call IBluetoothTelephonyCallback method on UHid Open/Close events
  Floss: Add UHID Mute support
  Floss: Add CallSource to CallInfo
  Floss: Add set_mps_qualification_enabled flag
  Floss: Add UHID device creation

Bug: 304299980
Test: mma packages/modules/Bluetooth
Change-Id: Iad317aff366ac3ca15851cc49412e0674f50786e

* changes:
  floss: hcidoc: Add SecurityMode3 tag
  floss: hcidoc: fix the disconnection due to host power off
  floss: hcidoc: Add tags for remote feature request
  floss: hcidoc: Refactor OddDisconnectionRule

Security mode 3 is a legacy security mode for classic connection.
Currently it doesn't work on floss, and it is reported that they
don't work well on BlueZ and iOS either.

Bug: 293397988
Test: m -j
Test: Run hcidoc to analyze report 91211217971
Change-Id: I12e25fdd43e4eeba84479beab9965bc0054c4aba

If BT is powered off, it is possible that the controller doesn't
wait until the peer device reply the disconnection request.
Therefore in aosp/2651792 we treat those disconnection request as
a special case. However, sometimes the event will still be replied,
thus we process the disconnection twice.

This CL fixes that by first checking whether the handle was
disconnected due to host power off.

Bug: 293397988
Tag: #floss
Test: Verify report 91229383170 doesn't report disconnection for
      nonexistent connection.
Test: m -j
Change-Id: I1db61463d1d93dd0120951c3a7b514d0167befd6

Sometimes we see disconnections due to receiving errors (due to unknown
handle, etc.) when requesting remote features, and sometimes even no
replies. Use the OddDisconnectionsRule to detect this behavior and
output an appropriate tag.

Bug: 293397988
Tag: #floss
Test: m -j
Test: Run hcidoc on report 91216966534, 91199447662
Change-Id: I1261212f327fefab8e30ac2f30698dec7a25c498

Most rules detect the type of command or event in the process()
function, then pass them (or their parameters) to a smaller, more
specific functions. However, OddDisconnectionRule mixes these
responsibilities so it's hard to follow what one part is doing.

Refactor the rule to follow the general style.

Bug: 293397988
Tag: #floss
Test: m -j
Change-Id: I41958f63e602b9b788835fd89b1221f7515b3d4e

Originally, btmanagerd only restores the enabled state on the
ReadIndexList Complete Event. However, the first ReadIndexList after
reboot is usually empty and the HciDevicePresence is caused by the later
IndexAdded event.

Instead of sending StartBluetooth message on MGMT event, this patch load
the previous state from disk on a new HCI device presents. The handler
of HciDevicePresence shall then start the adapter if floss is enabled.

This patch also fixes an issue that start_hci_if_floss_enabled is passed
a real HCI index while it actually wants virtual.

Bug: 302495893
Test: mmm -j packages/modules/Bluetooth
Test: Verified btadapterd is started on IndexAdded MGMT event on reboot
Change-Id: I43f73dc45ab822614e072f9555e290a7cb5f4e6f

Bug: 301661850
Test: mma packages/modules/Bluetooth

Change-Id: I5ed6c1ef897432277e995646f88a71ba2c1fb6b3

Bug: 287651018
Test: mma .

Change-Id: I0777e5cf887abddf5bba140e60acf1269b3aa6b2

Bug: 287651018
Test: mma .
No-Typo-Check: legacy code

Change-Id: I87f46240d658d73d898b05559a2ff4492da9f6c8

We currently keep track of scn usage in both BTM and BTA.  Remove the
scn array from bta to prepare for the btm counterpart to move to bta.

Bug: 299158235
Test: m Bluetooth
Change-Id: Ic209c62da9137dfe103565017fa49eab158a12d4

* changes:
  AutoConnect hfp: Improve test behavior
  if mostRecentlyActiveA2DPDevice is null, auto connect to mostRecentlyActiveHfpDevice device
  MetadataDatabase: fix format
  Prepare database for mostRecentlyActiveHfpDevice
  Metadata: Add builder