GitLab

Bug: 25467621
Change-Id: Ifdb52295fa5b8c5a39524a722dc044227c735240

Some devices (ex. Nexus Player remote) do not take well to having the
remote version queried while a remote feature request is also sent and
subsequently fail encryption.

This patch serializes the query of the remote version and requires both
remote features and the remote version to be read before calling a
connection established.

Bug: 25357767
Change-Id: Ie70eea241b47b81778d8e0df47f3a8f376316932

Request remote version information for LE links when connecting and
ensure LMP version is >= 8 (Bluetooth 4.2) before setting the LinkKey
bits in the pairing request.

This is to ensure older remote devices do not get confused by the
additional bit(s) in the pairing request.

Bug: 25277879
Change-Id: I25424e541fa5244d8f522dffc62cdedc24f4f35b

Bug: 17901099
Change-Id: Ib61113b19c9c9798b9ea992fad0ce997dcc2e2f8

Whenever a device has more than 1 ACL link active and transferring
data on one of its links via PAN, HH or JV (RFCOMM) the sys busy/idle
state toggles frequently. To avoid triggering role switches for each
of these events we filter this out and let the other SYS events handle
it in combination with other role policy management code. Ideally we
should revert the toggling to properly reflect the busy/idle state of
each profile but to limit risk of side effects at this stage we will
make this intermittent change.

This also affects audio streaming in certain cases.

Bug: 24570959, 25129209
Change-Id: I141e17ee069c82624e153fd8de5db90ae93724b9

Since alarms [/callbacks] are used in the critical audio path, raising
the thread priority is required to ensure timely encoding/transmission
of audio packets and avoiding thread pre-emption.

This patch indiscriminatly raises the (single) callback thread priority,
affecting all alarm callbacks. In the future, thread priotiry should be
adjusted based on the alarms pending in the queue.

Bug: 24570959
Change-Id: I64928fd7325f1d6f76815a185dbccf1e01ae30d2

128-bit UUIDs for GATT services discovered using SDP are byte-ordered
incorrectly. This patch fixes both the SDP discovery as well as the SDP
record creation code.

Bug: 24344122
Change-Id: Iee4bf33dcbbc5ee0f2380b48330107232bd0401b

Sometimes, during result-heavy LE scans, the UART byte stream can become
corrupted, leading to assertions caused by mis-interpreting the bytes
following the corruption.

This workaround looks for tell-tale signs of a BLE event and attempts to
skip the correct amount of bytes in the stream to re-synchronize onto a
packet boundary.

Bug: 23934838
Change-Id: Ifadaecf8077cb1defc7134c59c97302fca660f81

Remove the absolute volume control whitelist to enable absolute volume
control for all compatible devices.

Bug: 25081918
Change-Id: Iab0e4b4c4fd3dac0fefdf62ec80419dd64b56c0c

Bug: 24737963
Change-Id: If3b59437439305b490c5fc260a7080cee04f09c5

The audio sub-system and audio related tasks run at elevated thread
priorities in order to ensure timely handling of audio packets. This
patch raises the thread priority of lower layer Bluetooth tasks involved
in delivering audio packets to avoid thread pre-emption and subsequent
audio skipping.

Bug: 24570959
Change-Id: I5e19ee4590207df753f7b8bfc22174959722ec2d

* commit '74dad515':
  Disable opening network debug ports for security reasons

By default, we open up to three TCP ports that are used
for debugging purpose:

 * TCP port 8872 - used for forwarding btsnoop logs at real time
   Note: the port is open only if "Bluetooth HCI snoop log" is enabled
   in the Developer options
 * TCP port 8873 - used for HCI debugging
 * TCP port 8879 - used for debugging the Bluetooth counters

Those ports are disabled by default.
To enable, the following #define should be added at the top of the
corresponding file(s): btcore/src/counter.c hci/src/btsnoop_net.c
hci/src/hci_inject.c

   #define BT_NET_DEBUG TRUE

Bug: 24371736

Change-Id: I5cb43af1a5d29c331eb5ef61a24dccbe95df6f40

By default, we open up to three TCP ports that are used
for debugging purpose:

 * TCP port 8872 - used for forwarding btsnoop logs at real time
   Note: the port is open only if "Bluetooth HCI snoop log" is enabled
   in the Developer options
 * TCP port 8873 - used for HCI debugging
 * TCP port 8879 - used for debugging the Bluetooth counters

Those ports are disabled by default.
To enable, the following #define should be added at the top of the
corresponding file(s): btcore/src/counter.c hci/src/btsnoop_net.c
hci/src/hci_inject.c

   #define BT_NET_DEBUG TRUE

Bug: 24371736

Change-Id: I5cb43af1a5d29c331eb5ef61a24dccbe95df6f40

A case, where tap read thread is always exist, Sometimes
data packets get recieved on older fd, which is not available.
which is causing assert due to fd mismatch in race condition.
when next pan connection is immediately available. If last pan
connection gets disconnected, then tap_read_thread should be
destroyed in btif context to fix this issue.

Bug: 24093456
Change-Id: Ic1053200a7be4c2091d6c394634831ca3fbd61df

Bug: 23740164
Change-Id: Ib26a64b624c711443201adc4fde6b041ecb0dde1

Bug: 23978964
Change-Id: I36b6d05bb82eb0283eed3b465038d65890c4c245

On some platforms, the logs indicate that the BTU task is stalled by
UHID driver when writing the HID descriptor to the kernel. This patch
converts the UHID socket to non-blocking to avoid stalling the main
BTU task.

Bug: 23978964
Change-Id: I5a7a3e106fb2c967d68f077faedcd4fe62bbd912

A case where DUT is acting as PANU device on BR/EDR link & also
acting as central on LE link. As LE link & PANU link need not to
be master & it's not a scatter-net scenario, because there is only
a BR link. So checking for BR device count before initiating role
switch to avoid continuous loop of role switches.

Cherry-picked from: https://partner-android-review.googlesource.com/#/c/376058/

Change-Id: I70e4041a71ba965f1f9e25729c3cfd20ae03d6b4
Bug: 23290136

And complete implementation of out_get_render_position()

Bug: 21199150
Bug: 24042067
Change-Id: I95038ad68433e401b096a0159282a2ae372570e1

BT Inquiry is disallowed when service discovery
is in progress. Innitially when GATT discovery
is in progress (crosskey pairing), it resulted
in inquiry requests being rejected. During an
insecure connection like OPP, bonding link key
is not stored. In an earlier patch (b820aec7),
BR/EDR discovery was bypassed if the bonding was
temporary. However, this was not done for LE.
Applying the same logic to LE, resolved the
inquiry issue reported in this bug.

Bug: 23468169
Change-Id: I6111c80a420aeb0bc36a0393eb215c112512d251

If there are errors when processing GATT related events (e.g.,
configuring the MTU), don't use recursive calls into
bta_gattc_sm_execute(), because it breaks the free-ing of some
of the memory.

Bug: 23756301
Change-Id: I3c685170e868ffbf4e488d2bb5a31904e3f7b39d

Sometimes allow_device_sleep() is called before wake_assert, this result in transmit_done
does not start idle timer. This is because the "wake_state" does not match while
allow_device_sleep() is called before wake_assert() is called. So start idle timer
anyway when transmit_done is called.

Bug: 23754156
Change-Id: I5437aea0abc4b7f5cf847701e2d0a889c6823fcb

When calculating the appropriate LPM's idle timeout,
the local chip name was NULL. This is fixed by calculating
LMP's idle timeout when the chip is enabling.

Bug: 23248798
Change-Id: I8ee2b28fb4363359e035e242fdcff8a736ee448c

* commit '24923939':
  LE Multi-Advertising State Change Sub-event Wrong Variable Type

Wrong connection handle type in multi-adv event parsing,
causing random address (confirm value) mismatch and pairing failure.

Bug id: 23201007

Change-Id: I13ce231360937e711f61eb0777805b07bcde7074

* commit 'dcfafe9a':
  Fix incorrect SetConnScanParam function mapping

BTA_DM_API_BLE_CONN_SCAN_PARAM_EVT is defined before
BTA_DM_API_BLE_SCAN_PARAM_EVT in the enum but the corresponding
entries in bta_dm_action were inverted. This resulted in incorrect
invocation of set_conn_scan_params while trying to set parameters
for scanning. The call was hence failing here and was not
resulting in setting of the new scan parameters in the Controller
irrespective of the scan mode selected from the App. This would
essentially result in the breakage of ScanSettings.SCAN_MODE_<xx>
Android SDK APIs.

Bug: 23602042
Change-Id: I17e1b14a32250e3ccb7631a67690ec2e0a6bc321

* commit 'b7460434':
  Fix memory corruption of BLE whitelist hashmap

When adding devices to the whitelist for LE background connections, a
local variable is used as the hashmap key, preventing any successful
lookups going forward. Thus the device will repeatedly add the same
device to the whitelist and preventing successful connections going
forward.

Bug: 23423602
Change-Id: I3d2590c0a1dd66c6e1864ea53f875a713660b645

* commit 'aa41fec0':
  Check the return value when reading HCI type byte

Add missing return value check when reading the HCI type byte.
This check is needed as a safeguard. For example, function
event_uart_has_bytes() could be called (indirectly)
within the run_reactor() loop not only when there are bytes to read,
but also if there is an error (e.g., EPOLLHUP | EPOLLRDHUP | EPOLLERR).

Bug: 23105107
Change-Id: Ic3b6e4d656406949e384c8106b0c607f7c221759

* commit 'd2199cbb':
  Disable remote TCP connections

For security reasons, TCP sockets now listen on the loopback
IPv4 address 127.0.0.1 for incoming TCP connections.

Bug: 23272146
Change-Id: I88523f643f305f2281740575d7011b6077bf0843