* Add a command line tool for MCAP related tests
* This tool can access the Bluetooth stack via a command line program
* Type "help" to obtain a list of available console commands in this
  Bluetooth
* Bluetooth must be disabled when this tool is running

Bug: 37867299
Test: make, PTS MCAP test
Change-Id: I106e6625ee8a29b7fdad88cdfeb99419054ef771

When two devices are bonded, even if the device is supposed to act just
as a GATT Server, service discovery is performed. Doing "read by type"
should not trigger connection timeout, or disconnection, as the server
should wait for the client to finish connection.

Please note that it does not GATT Client behaviour, as they manage the
connection timeout through L2CA_SendFixedChnlData called on each data
packet.

Bug: 34951749
Test: manual
Change-Id: If53cf63dc204238a60839497992c870b1b39aa60

* PTS tests requires MCAP APIs to be called at protocol level
* This CL creates a test interface to enable the above functionality

Bug: 37867299
Test: make, no user visible effect
Change-Id: I15cc6cc613ce8f7c57564296e45077ab877c269b

* Fix error response message for clock sync opcodes
* Fix error response message for CREATE_MDL opcode
* Fix logging in the MCAP stack

Bug: 37867299
Test: make, PTS MCAP test
Change-Id: I7a37ac04806bca20d732357f00b4761577b4be19

Bug: 37905488
Test: Code still compiles
Change-Id: I479fc82a15f977b7e3d50b07302241909efc8856

Bug: 37803501
Test: Code still compiles
Change-Id: I9326e58aff0d80e441cb92c41e0e4a7d70e32b1b

Bug: 38330724
Test: Manual: enable/disable Bluetooth
Change-Id: Iaece4cd608c67b28b436ded9fcc51c0bf28f4098

Make sure that all access happens on jni thread

Bug: 38298938
Test: Code compilation
Change-Id: I38246626d2c6390586b314b27889dd9ba905f406

The volume is erratic on the speakers when the speaker is reporting
linear increase, with decreases for higher volumes.

Test: check that absolute volume not used for Kinivo Speakers
Bug: 37872586
Change-Id: Ic5f409e8631bece66a75372b916faa9c104e2024

Bug: 38351937
Test: A2DP streaming to LDAC headset and switching MONO/STEREO in Settings
Change-Id: I20c5069fd5d104b9ebe767ae2728af7d7b2ae1d9

Bug: 38123054
Test: manual
Change-Id: Ic97a964ca2a25a84c6abcf41acb0f3fb583488a2

When SDP_MAX_RECORDS were reached in SDP db, which usually happen by
registering big number of GATT services, unregistration of a service was
causing a crash. This was because we moved from memory after the array
into last element.

Bug: 38265174
Test: manual
Change-Id: I7b0235213fb07c85dcc1245f1c840b41267521d5

Since Bluetooth 4.2 and errata ESR08 there may not be more than one
connection between two LE device addresses. Also the stack assumes
there is at maximum one connection to the same address. This patch
makes sure there are no connected devices in the white list when a
connection attempt is started. Since some (even 4.2) controllers don't
handle this correctly, currently this method is used regardless of
controller in this patch.

When the maximum L2CAP connections were reached and a new connection
was established to a device using auto connect, the stack hung and
would no longer create new connections until Bluetooth was restarted,
since the state change to BLE_CONN_IDLE was forgotten. This patch
resets the state correctly, and also never initiates a connection
unless there is space to avoid connect-disconnect loop.

There were also bugs in the background_connections hash map; memory was
not freed when an element was erased and an incorrect hash function
which used the pointer to a bd addr instead of the bd addr itself which
basically meant that elements were never removed. This patch removes
the dynamic memory allocation and uses a correct hash function.

There was a bug that might lead to that the white list was filled
beyond its maximum, due to the counter was updated on the HCI command
complete event, which might run too late. Now the space is instead
calculated based on what commands have been sent to the controller.

The address type of the address added to the white list must also be
tracked, otherwise it might be updated due to a BLE scan, and later the
wrong address is removed from the white list. This patch fixes this.
(Preferably 49-bit bd addrs should be used as identifier through the
whole stack but we're not there yet.)

There was a queue of size 10 with pending white list operations. That
queue got full if there was initially 10 devices in the white list,
then the 10 devices were removed and immediately after 10 other devices
were added. This patch removes the queue altogether by instead syncing
against the background_connections hash map.

Bug: https://code.google.com/p/android/issues/detail?id=219910
Test: stress-testing with a bunch of BLE devices and inspecting HCI log
Change-Id: I78de654ffbea5f4962a189caf984f7f2934e8fbe

If the A2DP Sink device sends "Set Configuration" with its preferred
SBC configuration, try using that configuration when selecting
the codec parameters instead of the A2DP Sink SBC capability.

Test: A2DP streaming, PTS AVDTP Tests: TC_ACP_SRC_SIG_SMG_BV_18_C,
      TC_ACP_SRC_SIG_SMG_BV_20_C, TC_ACP_SRC_SIG_SMG_BV_22_C
      PTS A2DP Tests: A2DP/SRC/SET/BV-04-I, A2DP/SRC/SET/BV-06-I
Bug: 37723311
Bug: 35664023
Change-Id: I48e2aa664b899c1cda5df4d1f86f9cfa3a9a67f6

Add a mechanism to set the RTP Header Marker bit per codec.
Currently, the Marker bit is set only for AAC.

Test: A2DP streaming, PTS AVDTP tests: TC_ACP_SRC_TRA_BTR_BV_01_C,
      TC_INT_SRC_SIG_SMG_BV_17_C, TC_INT_SRC_TRA_BTR_BV_01_C
Bug: 37723283
Change-Id: I9b8e77d44c750746c169df9628d5539ad2406b92

- use references instead of pointers where possible. Thanks to it we can
  remove some reduntant null checks and simplify the code.
- use directly allocated memory instead of own buffers. Thanks to it the
  stack uses around 12kb less of memory after startup.
- use list and vector from std library, instead of some hand-written
  implementations.

This patch is a prerequisite for further server refactoring, that will
further reduce the space used, make unit-testing possible, and enable
proper handling of PTS GATT tests.

Test: sl4a GattReadTest
Bug: 38225928
Change-Id: I1620be682259ccb8f0c02754806e355e3f1ad0c1

Update include files to include bt_target.h as it has the necessary
defines to correctly allow bdroid_buildcfg.h to enable and disable
dumping to output_sample.pcm.

Bug: 38180732
Test: Stream Bluetooth audio to device and observe that
output_sample.pcm does not get written or appended to.

Change-Id: Ia7926f93a47489686d315e3c9abd1618cb8adda3

* changes:
  eSCO: Format changes to original vendor patch (4/5)
  eSCO: Additional fix to get eSCO working for HFP and HSP (3/5)
  eSCO: Remove BTM_WBS_INCLUDED from the stack (2/5)
  eSCO: BT 4.1 Enhanced SCO command (1/5)

The A2DP_GetTrackBitsPerSample() API was replaced in the past
by A2dpCodecConfig::getAudioBitsPerSample() .
 * Removed A2DP_GetTrackBitsPerSample() and related codec-specific APIs:
   - A2DP_GetTrackBitsPerSampleSbc
   - A2DP_GetTrackBitsPerSampleAac
   - A2DP_GetTrackBitsPerSampleAptx
   - A2DP_GetTrackBitsPerSampleAptxHd
   - A2DP_GetTrackBitsPerSampleLdac
     NOTE: The removed A2DP_GetTrackBitsPerSampleLdac implementation
     had a bug: a2dp_ldac_caps.bits_per_sample -> ldac_cie.bits_per_sample
     This bug itself was uncovered when switching to clang-3960126
   - A2DP_VendorGetTrackBitsPerSample()
 * Replaced few leftover usages of A2DP_GetTrackBitsPerSample()
   A2dpCodecConfig::getAudioBitsPerSample()
 * Updated the unit tests

Test: Code compilation, unit tests, A2DP streaming
Bug: 37752547
Change-Id: I7feeda0f32922b8ef888b157673e3e2ade98b179

* Format changes to original vendor patch
* Re-order part of SCO creation code to improve readability
* Make log more concise and usefull

Bug: 19540029
Test: make, HFP PTS test, testplans/86884, testplans/87103
Change-Id: Ie0199519afb30f56568c04fef728fce44a172063

* Removed vendor specific call after codec negotiation and before eSCO
  command is sent to Bluetooth chip
* Removed vendor specific call after a SCO connection is requested by
  the peer
* Removed vendor specific save audio state methods

Bug: 19540029
Test: make, HFP PTS test, testplans/86884, testplans/87103
Change-Id: I899405a8eaa2f38ee0de04949c5288f9eefa74df

All of our platforms support WBS.

Bug: 19540029
Test: make, HFP PTS test, testplans/86884, testplans/87103
Change-Id: I384d49f73935ebb35b17132e71c85174797c435e

Added support for BT 4.1 enhanced SCO feature on the stack.
This feature allows the stack to create a SCO connection with
remote device by using Hci_Enhanced_Setup_Synchronous_Connection
command after checking the controller (4.1) support for
enhanced SCO command.
Added the command parameters to use the command in both wide band
speech(WBS) and narrow band speech(NBS) scenarios.

Number of Broadcom vendor specific commands(VSCs)that are needed
to be sent to Broadcom controllers along with this command have
also been updated accordingly

NOTE: This change would also need a firmware patch for
the feature to work on Broadcom contollers which will
be delivered to Google in a separate change set.

Bug: 19540029
Test: make, HFP PTS test, testplans/86884, testplans/87103
Change-Id: I1014d81be5cbe91078a4484dd072ac3957bfdfe4

If there are no attributes available, it's possible to return zero
attributes from a GetItemAttributes request.

This is triggered by some carkits who are unhappy when they only get
titles back (the only required item) and ask for the rest of the
attributes (which we don't have).

Test: connect to honda carkit and get snoop logs to confirm
Bug: 36055995
Bug: 35956792
Change-Id: I5a31e206565d212e456111d8c6b542c7a1569e5a

Characteristic and descriptor can't use UUID equal to GATT Attribute
Types. Using such UUID would result in invalid GATT database, and
crashes during service discovery.

Test: manual
Bug: 38134693
Change-Id: Ide2c21109f885cbc79287452b1dabd3f532de385

Provides an interface for application to send Set Idle
and Get Idle commands to remote HID Device. Support for these
two commands was missing from existing code, so existing code
design is reused to add support for these two commands.

Without this support following mandatory PTS test cases for HID 1.0
cannot be passed, TC_HOS_HID_BV_05/06.

Test: Executed PTS tests TC_HOS_HID_BV_05/06 and confirmed if they can
pass

Bug: 34344715
Change-Id: I14bc2dc349354fd87ef8aba25357514398532e87

* Generic linux does not have property_get_int32. Instead,
  osi_property_get_int32() is created to handle OS_GENERIC cases
* Some linux header have sigevent.sigev_notify_attributes typed as
  (pthread_attr_t *) whereas others typed it as (void *), as any pointer
  can be implicitly casted to (void *), the current casting to (void *)
  is unncessary and will break build on systems using (pthread_attr_t *)

Test: make, unit test, no user visible effect
Change-Id: I24b33da453dc9d40656168a3bcd900d9c99219ce

Swapping out the queues with base::MessageLoop allows for an easier refactor
of alarms as we can just replace them with an instance of base::Timer. Also
refactored out the data_dispatcher as most of the code wasn't being used.

Bug: 37245162
Test: Bluetooth sanity test and listening to music
      added net_test_btu_message_loop
Change-Id: I73c359f07a225733dc89f51422a6a24ce807c593

This is for PTS test TP/LE/CFC/BV-20-C.

The scenario is there is already one COC between DUT and PTS.
Then PTS tries to establish one more COC with same cid which
should be rejected.

Change-Id: I094a4058ac0e7bb9d7344e66d9feafee17bc3d53
Signed-off-by: Jaganath Kanakkassery <jaganathx.kanakkassery@intel.com>

SM/MAS/SCJW/BI-01-C [Just Works, IUT Initiator, Secure Connections -Pairing Failed
For Lower tester to handle step 15.
Lower Tester responds with a Pairing Failed command in phase 2 with reason code
0x0C (Numeric Comparison Failed).

Bug: 37707831
Change-Id: Ia7103330d1d1ecb06e39502c2a0976585b203ba5
Signed-off-by: Ravishankar Srivatsa <Ravishankar.Srivatsa@intel.com>
Test: manual

After we reach 7 ACL connections we have 7 peer device records.
Once we reach this then we lose a connection and shrink the list we
copy garbage into the last peer device record.  This can cause the flag
remove_dev_pending to be set to something > 0 thus true which causes the
stack to remove the device and its bonding information on the next
acl change event.

ACL Connections can happen from any connection (pairing, SDP, etc...)

Bug: 36598959
Test: Manual
Change-Id: Ifbaa4098edba442274ffde183960ef53169988e7

Before this patch, if there was a message on the command queue and
there were no command credits, the thread reactor would spin trying
to process the message on the command queue and would continue until
a credit was received. This led to a bug where upon switching users,
hci_thread would spin and try to use 100% of the CPU. This is fixed
by moving over to a message loop and queue system. The message loop
processes all the messages. If there aren't enough command credits,
command messages are deferred to the command queue and popped off
whenever more credits are aquired. The deferred queue has priority
to credits over recently posted messages.

Bug: 37733903
Test: Swap users with the real time scheduling patch applied, and
      general Bluetooth usage.
      TestTracker: 86249
Change-Id: Ib775e47f6d4810d3d7d8af5b3ba84adc4ada3da5
(cherry picked from commit 417ffed2)

* merge BTM_CheckAdvData and BTM_CheckEirData into GetFieldByType
* check wether AD data is properly formatted when the packet is
  received. Some controllers were returning malformed data that
  resulted in stack crashes for BT5 packets.
* add tests to verify the helpers work as expected

Test: added net_test_stack_ad_parser
Bug: 37671082
Change-Id: I49e43d7cf7d0a8ace1ee45d9b14b2b8440096b05

BT spec limits the length of device name to 248 bytes. With the new
LE advertising packets, that can go up to 1650 bytes long, we must check
the length of device name received.

Bug: 37671082
Test: manual
Change-Id: Iad309d638003c2391014c9764605d84ed5717cb1