GitLab

Test: mma -j46
Change-Id: I8f534276d972e13d61ab84c7fdab9f6687c38f88

Usecase
1) Pair with carkit and connect HF and A2DP/AVRCP.
2) Carkit sends SAP connect request. Reject this request from DUT.
3) Power off/on carkit.
4) Carkit sends SAP connection request to DUT.

Expectation:
SAP connection is rejected, but carkit can reconnect HF and A2DP/AVRCP

Observation:
Unable to connect any profiles after power cycling carkit.

Root Cause:
Even thought DUT rejects incoming SAP RFCOMM connection request whose access
was previously rejected by DUT, RFCOMM connection is not closed from DUT until
time out of 30 seconds. This prevents carkit from sending new connect request
for previously connected profiles like HFP, A2DP/AVRCP.

Fix:
Close RFCOMM connection in case request is from remote device whose access is
already denied by DUT in previous attempt.

Change-Id: I31bd2dfd862f7f69e46b8012924942e08e850c95

Use case :
1. Send large file to dut (Above 4 GB)
Failure:
Crash in com.android.bluetooth.
Process: com.android.bluetooth
java.lang.IllegalArgumentException: Length must be
   between 0 and 0xFFFFFFFF
at javax.obex.HeaderSet.setHeader(HeaderSet.java:333)
at com.android.bluetooth.opp.BluetoothOppObexClientSession$ClientThread
   .sendFile(BluetoothOppObexClientSession.java:411)
at com.android.bluetooth.opp.BluetoothOppObexClientSession$ClientThread
    .doSend(BluetoothOppObexClientSession.java:337)

Root cause :
Largest length in OBEX message header is unsigned 4 bytes, so actual
file length (4 GB) cannot be written into Obex header which leads
to exception in OBEX code.

Fix :
Add length check before push file and skip sendfile if length is above
4GB to prevent crash.

Change-Id: I9066fb7456d85fdc197fe21fed26fb2bb89ff09e

When scanner died, it should be unregistered as scanner, not as GATT
client. Otherwise app will not be able to start scan when it's
restarted.

This bug was introduced when separating GATT client from LE scan
clients.

Bug: 33556286
Test: start scanning, kill app, restart the app
Change-Id: If8da5512081e1986bbee7362984091f298500f4e

Test: sl4a FilterTest
Change-Id: Ic16e523524022803107b5e0653a35e3f2cbc5186

* Add one extra item in the A2DPSession message to keep track of audio
  playback time for A2DP source

Bug: 33694310
Test: code compilation
Change-Id: Iee1e47997bf9cf9ba3f7f812c965eab36f19c0f4

When the app acts as a pointing device, it (according to the
specification) generates ~80 reports per second.
Removing 5 log lines per report makes logs shorter by 400 lines
per second.

Test: build, run, logcat
Change-Id: Ief2300294b7c7ed56cb67bcf1b77f70187291b8e

Test: run a bugreport, check in "APP SERVICES" section
Bug: 29356402
Change-Id: I51a2ab0ec51d4607d39eaca225584d3d539fbfd6

* Extended the Bluetooth A2DP AIDL interface to get/set the current
  codec configuration

* If the current codec config is changed such that the audio feeding
  parameters have changed, then inform the Audio Manager so it can
  reset accordingly the audio feeding parameters in the Audio HAL to
  the Bluetooth stack.

Also, fix the BT_BOND_STATE_* values (unused) in the
AbstractionLayer.java class so they match the corresponding
BT_BOND_STATE_* values in
hardware/libhardware/include/hardware/bluetooth.h

Test: A2DP streaming to headsets, TestPlans/71390
Bug: 30958229
Change-Id: Ic3e35154d9dbc32b66522322cd5d2a96bc5a977c

MediaBrowserService can reject the connection from a MediaBrowser.

Test packages beforehand so we don't advertise browsing for a player we
can't browse to.

Refactor and cleanup some code around building media player lists and
remove some dead code.

Bug: 33390650
Bug: 32382082
Test: use carkit to browse to different players
Change-Id: I5a4c18ab77a24f3cf2aa34d1c15e3213edfed61d

Test: make
Change-Id: I1322eb7fb5d253c0968a8bbe11a19cca8f1834f9

This change reflects the API update, and also fixes a couple
issues with the code (non-initialized, wrongly compared and
otherwise unused mHidDevice field).

It keeps the current logic of using a single device at a time,
yet introduces checks that the application is actually
addressing the right device.

Test: make
Change-Id: I088417046907c8310692cf9205075d6da0cd81e6

* changes:
  Rename the Bluetooth profile classes for HID Device role.
  HIDD: Add support for HID Device Role

BluetoothManagerService will dump basic state about Bluetooth now for
standard dumpsys arguments. Move it from here and track which apps enable /
disable Bluetooth.

Test: enable/disable a few times, check that apps are recorded correctly
Bug: 33692282
Change-Id: I7505ba25a5cafda061fa23774cc31a6b83909d34

This CL reflects the changes from frameworks/base.

Test: make
Change-Id: I0e3a3335572f6ab4683e8a3b5335f3b51afba2f5

This patch adds support for HID Device Role which enables phone being
used as a Bluetooth keyboad or mouse.

Change-Id: I931867442111ad997a34a166c7b2ec1daf317ddd

Get rid of BTA state machine states for BLE scanning.

Test: sl4a FilteringTest
Bug: 30622771
Change-Id: If616c1100f8d1153568c2b5487a0b117e735930b

Bug: 30622771
Test: sl4a ConcurrentBleAdvertisingTest
Change-Id: I1cfdd29075a41057ae99b5b458876a4a9916e845

A bug was introduced in commit
d828b42e which causes normal
GATT clients (neither scanners nor advertisers) not being freed
properly when remote process terminates, leading to out of space
of available GATT clients after some time.

This patch adds back a ClientDeathRecipient to handle this.

Test: kill an app that has gatt client, restart it and create new
gatt client, verify the client id is same as before in logcat

Change-Id: I318cb1693984444a172cb81be2023a064dbc02a4

Bug: 30622771
Test: sl4a BleScanTest
Change-Id: Ic35017fc62f64c5ebdbca43b6b6e8e8b16062b07

This patch converts btgatt_scanner_interface_t struct into
BleScannerInterface class. It also refactors three most important
methods from this interface: RegisterAdvertiser, Scan, and Unregister.
Rest of this interface will be updated in following patches.

Bug: 30622771
Test: sl4a BleScanApiTest
Change-Id: I0f0ad31dab88efbcd74a11d1908b80fcb78a5c6c