OSDN Git Service

android-x86/system-bt.git
5 years agoAdd binder interfaces for metadata APIs
Ugo Yu [Sat, 12 Jan 2019 00:13:16 +0000 (08:13 +0800)]
Add binder interfaces for metadata APIs

Bug: 121051445
Test: build pass

Change-Id: Icd51b7cc0a4dec5a2bffec4ded67caffeedb126b

5 years agoMerge "Fix BT a bad/ out of bounds array read"
Sal Savage [Mon, 14 Jan 2019 21:43:55 +0000 (21:43 +0000)]
Merge "Fix BT a bad/ out of bounds array read"

5 years agoMerge "SMP: Fix invalid auth_req in LE Security Request handler"
Treehugger Robot [Mon, 14 Jan 2019 14:23:01 +0000 (14:23 +0000)]
Merge "SMP: Fix invalid auth_req in LE Security Request handler"

5 years agoDo not crash on LE Create Connection "Command Disallowed"
Jakub Pawlowski [Thu, 3 Jan 2019 14:13:11 +0000 (15:13 +0100)]
Do not crash on LE Create Connection "Command Disallowed"

Being in "Connecting" used to mean we are either doing direct
connection, or background connection procedure. Now we have just one
connection procedure. This means there is no ambiguity when the
controller responds with "Command Disallowed" - we know exactly which
state to go back to.

Bug: 110412113
Bug: 112827989
Test: there is no test, or reproduction scenarion to verify that
      Ran GattConnectTest for sanity

Change-Id: I9ed1cdc7ae4032cda4e8a166310eca0004c7f9dc

5 years agoSMP: Fix invalid auth_req in LE Security Request handler
Manoj Babulal [Thu, 27 Dec 2018 10:43:19 +0000 (02:43 -0800)]
SMP: Fix invalid auth_req in LE Security Request handler

Due to wrong pointer de-reference, auth_req value read
in smp_proc_sec_req (LE Security Request packet handler)
was invalid.

Logcat trace shows invalid auth_req value:
.. bt_smp  : smp_proc_sec_req: auth_req=0xf9
   ....
.. bt_smp  : smp_proc_sec_req: auth_req=0xe9

After the fix, auth_req value is correct:
.. bt_smp  : smp_proc_sec_req: auth_req=0x1
   ....
.. bt_smp  : smp_proc_sec_req: auth_req=0x5

This data flow can be traced in code on this path:
smp_data_received -> smp_sm_event -> smp_proc_sec_req

Issue seen after following change:
https://android-review.googlesource.com/c/platform/system/bt/+/492442

Issue was discovered during execution of PTS test
cases GAP/SEC/AUT/BV-13-C and GAP/SEC/AUT/BV-14-C.

Fixes: 122066779

Test: Verified logcat trace shows valid auth_req
value prints from smp_proc_sec_req() function

Change-Id: I072b4794fb56a05b7ac7f2f4ca606514d7524550

5 years agoFix BT a bad/ out of bounds array read
Sal Savage [Wed, 9 Jan 2019 20:50:07 +0000 (12:50 -0800)]
Fix BT a bad/ out of bounds array read

Bug: 121375525
Test: Used automotive hardware and verified that pairing and media
      streaming and control still work and crashes no longer happen.
Change-Id: I3476425131718aded980f57a8b9be92a8e9f9d3b

5 years agoFix potential NULL dereference in smp_proc_confirm
Jakub Pawlowski [Fri, 11 Jan 2019 12:43:33 +0000 (13:43 +0100)]
Fix potential NULL dereference in smp_proc_confirm

Bug: 122102910
Test: compilation
Change-Id: Idd16e1c1924165924847f1aad64ddd44a8a1b8ab

5 years agoMerge "Fix bt_trace.h #define"
Treehugger Robot [Fri, 11 Jan 2019 00:21:24 +0000 (00:21 +0000)]
Merge "Fix bt_trace.h #define"

5 years agoAdd host tests to TEST_MAPPING
Dan Shi [Thu, 10 Jan 2019 18:55:23 +0000 (10:55 -0800)]
Add host tests to TEST_MAPPING

Also added some missing test_suites setting.

Following tests don't have host_supported set to true in Android.bp:
net_test_audio_a2dp_hw
net_test_btif
net_test_btif_profile_queue
net_test_device
net_test_hci
net_test_stack
net_test_stack_ad_parser
net_test_stack_smp

Following tests failed to run in host with error "did not report any run":
net_test_performance
net_test_stack_rfcomm

Bug: 111921389
Test: system/bt$ atest --host :postsubmit
Change-Id: I5f55cead7a2ad1cadea48220f45781f4f74949dd

5 years agoFix bt_trace.h #define
Jack He [Thu, 10 Jan 2019 04:47:57 +0000 (20:47 -0800)]
Fix bt_trace.h #define

Test: make
Change-Id: Ie1177f2311db0b7b5dfb440f3f982d4a79aa460d

5 years agoRemove code related to Health Device Profile and MCAP Protocol
Jack He [Fri, 4 Jan 2019 01:11:24 +0000 (17:11 -0800)]
Remove code related to Health Device Profile and MCAP Protocol

* Health Device Profile (HDP) and MCAP protocol has been largely
  replaced by BLE. New applications should use Bluetooth Low Energy
  instead of legacy Bluetooth Health Device Profile

Bug: 111562841
Test: make, unit test, use Bluetooth
Change-Id: I3da1173a8a928af4b26a29e71746241c22af2236
Merged-In: I3da1173a8a928af4b26a29e71746241c22af2236
(cherry picked from commit dc13609b58b2f6c6185951f650036f5098e19e92)

5 years agoMerge "HCI: Simplify btu_hcif_hdl_command_status()"
Treehugger Robot [Wed, 9 Jan 2019 21:59:39 +0000 (21:59 +0000)]
Merge "HCI: Simplify btu_hcif_hdl_command_status()"

5 years agoMerge "Group all Hearing Aids device information together"
Treehugger Robot [Wed, 9 Jan 2019 17:36:40 +0000 (17:36 +0000)]
Merge "Group all Hearing Aids device information together"

5 years agoHCI: Simplify btu_hcif_hdl_command_status()
Jack He [Wed, 9 Jan 2019 01:48:20 +0000 (17:48 -0800)]
HCI: Simplify btu_hcif_hdl_command_status()

* Move 2nd layer switch statements into top layer
* Check if p_cmd is nullptr at the beginning of the function so
  that each case does not need to check it individually. p_cmd
  is never null.
* Skip parameter total length at beginning of the function

Test: make, unit test, use Bluetooth for music and calls
Change-Id: I23cf983a1ade44bcea0b1009860efdfaf3b8bca8

5 years agoHCI: Fix command status handling of HCI_QOS_SETUP
Jack He [Wed, 9 Jan 2019 01:53:31 +0000 (17:53 -0800)]
HCI: Fix command status handling of HCI_QOS_SETUP

* QOS setup complete event was handled instead of QOS setup command

Test: make
Change-Id: I894099f369675bd25397ee0005c6838f39220eed

5 years agoGroup all Hearing Aids device information together
Stanley Tng [Fri, 21 Dec 2018 00:44:15 +0000 (16:44 -0800)]
Group all Hearing Aids device information together

A small refactor where the various information about a Hearing Aids
device is gathered into a class. Instead of passing a long list
of function parameters, it is cleaner to group them inside a single
class. It also makes it easier to add new elements in the future.

Test: Manual pair and audio stream with Hearing Aids
Bug: 119617521
Change-Id: I193b5cc915a6c36e7e36d1f06a5adcda7c17ab1c

5 years agoRemove BTM_SCO_WAKE_PARKED_LINK flag
Jack He [Sat, 5 Jan 2019 00:18:12 +0000 (16:18 -0800)]
Remove BTM_SCO_WAKE_PARKED_LINK flag

* Under majority of configurations, SCO connection can only be created
  if in non-parked mode since SCO is a circuit switch link

Bug: 122375953
Test: make
Change-Id: I58569025b435953b964e5aded64b3ace27d4e73c

5 years agoClean up BLE connection state state machine
Jakub Pawlowski [Wed, 2 Jan 2019 23:41:53 +0000 (00:41 +0100)]
Clean up BLE connection state state machine

* get rid of BLE_DIR_CONN - we never go into this state any more
* BLE_BG_CONN -> BLE_CONNECTING
* use btm_ble_get_conn_st instead of directly accessing control block

Test: compilation
Bug: 112827989
Change-Id: I4aa09a8b558c6f2161806a6d27bee0560fd19fd8

5 years agoMove all HCI LE Create Connection management into single file
Jakub Pawlowski [Mon, 31 Dec 2018 11:54:04 +0000 (12:54 +0100)]
Move all HCI LE Create Connection management into single file

Currently handling of LE Create Connection, LE Create Connection
Cancel, LE Connection Complete and other HCI commands related to LE
connection establishment is in two separate files.
Move all this code into single file.

Bug: 112827989
Test: compilation
Change-Id: Ibc2f6ca2324f4df8a23c2722762d0e5297782b33

5 years agoRemove BTM_SCO_INCLUDED flag
Jack He [Thu, 3 Jan 2019 01:38:53 +0000 (17:38 -0800)]
Remove BTM_SCO_INCLUDED flag

* Instead of removing SCO related code during compilation, platforms
  should chose not to call SCO related methods when SCO is not supported
* For example, platforms can choose to not initialize HFP to avoid using
  SCO commands
* This CL removes BTM_SCO_INCLUDED compile time flag to simplify code
  flow and improve readability

Bug: 122279647
Test: mm -j40; run Bluetooth stack and make phone call
Change-Id: I67d290839091f66c289a428496304c8b3e1dbe21

5 years agoRemove connection_manager from gatt namespace
Jakub Pawlowski [Fri, 28 Dec 2018 18:03:09 +0000 (19:03 +0100)]
Remove connection_manager from gatt namespace

connection_manager handles more than GATT. Celanup naming.

Bug: 112827989
Test: compilation
Change-Id: Id5fdfacb12921cff04204cf2589f73a7c6a0d01f

5 years agoGet rid of is_ble_connecting and ble_connecting_bda
Jakub Pawlowski [Thu, 27 Dec 2018 20:02:13 +0000 (21:02 +0100)]
Get rid of is_ble_connecting and ble_connecting_bda

These variables make no sense any more, as we might be attempting
multiple connections at the same time.

Test: sl4a GattConnectTest, manual bonding test
Bug: 112827989
Change-Id: I7e1d01e8d80383b49ff9c20fb3758cef03c52882

5 years agoUse conneciton_manager for all LE connection establishment
Jakub Pawlowski [Wed, 19 Dec 2018 21:23:47 +0000 (22:23 +0100)]
Use conneciton_manager for all LE connection establishment

Test: sl4a GattConnectTest, manual bonding test
Bug: 112827989
Change-Id: I55878e03d268c43002e22ce4ebfbf0e334523882

5 years agoGet rid of direct connect queueing logic
Jakub Pawlowski [Mon, 17 Dec 2018 12:59:54 +0000 (13:59 +0100)]
Get rid of direct connect queueing logic

Instead of queueing requests, we should use whitelist to process the
requests simulteanously.
This patch does remove the queueing logic, succeeding patch will add the
logic to schedule through whitelist.

Test: compilation
Bug: 112827989
Change-Id: Ia7132812f704658ecb6919e52c70b291e6874660

5 years agoCall btm_ble_set_conn_st before on_connection_complete
Jakub Pawlowski [Fri, 21 Dec 2018 16:16:01 +0000 (17:16 +0100)]
Call btm_ble_set_conn_st before on_connection_complete

on_connection_complete can trigger BTM_WhiteListRemove, which sends
"LE Create Connection Cancel" if the connection is pending.

Make sure connection state is updated before calling on_connection_complete.
Otherwise we send an unnecessary "LE Create Connection Cancel" request.

Bug: 112827989
Test: establish Direct connection with at least one device in the
      background connection procedure. Verify HCI snoop log content

Change-Id: I1b8532f4e58d5ab62b471ce0bde6c48640c83ff7

5 years agoMerge code in scanner/advertiser connection complete handler
Jakub Pawlowski [Fri, 14 Dec 2018 18:53:59 +0000 (19:53 +0100)]
Merge code in scanner/advertiser connection complete handler

Test: sl4a GattConnectTest
Bug: 112827989
Change-Id: I860693c11c263643858cfa0b4e4deb6a9f231cdd

5 years agoUnify code in scanner/advertiser connection complete handler.
Jakub Pawlowski [Fri, 14 Dec 2018 18:38:44 +0000 (19:38 +0100)]
Unify code in scanner/advertiser connection complete handler.

btm_ble_set_conn_st is called in every possible scenario of
l2cble_scanner_conn_comp. Call it at start, instead of before each
return.

Test: sl4a GattConnectTest
Bug: 112827989
Change-Id: I969fc59b934dc1c7a52be00741fdf38402545b26

5 years agoSplit l2cu_create_conn into two separate, transport-specific functions
Jakub Pawlowski [Fri, 14 Dec 2018 14:15:54 +0000 (15:15 +0100)]
Split l2cu_create_conn into two separate, transport-specific functions

There is no common logic between LE and BR/EDR paths.

Test: compilation
Bug: 112827989
Change-Id: I507d3a0d8efa0e4936b22722a9c1cc846e0be3bd

5 years agoMerge "Fix mtu assignment with correct value"
Ted Wang [Thu, 20 Dec 2018 06:16:47 +0000 (06:16 +0000)]
Merge "Fix mtu assignment with correct value"

5 years agoMerge "use epoll_create1"
Treehugger Robot [Wed, 19 Dec 2018 23:51:28 +0000 (23:51 +0000)]
Merge "use epoll_create1"

5 years agoEnable BLE background connection when stack is started
Hansong Zhang [Tue, 17 Jul 2018 23:24:24 +0000 (16:24 -0700)]
Enable BLE background connection when stack is started

Remove variable btm_cb.ble_ctr_cb.bg_conn_type. We can enable the
background mode when the Bluetooth stack is started and don't have
to wait until the first background connection request from HID
Device/Hearing Aid/GATT. As soon as we add the first device to white
list, the stack will send HCI Create Connection (White List) command.

Test: manual. Bond to a BLE mouse and turn off/on Bluetooth. Also test
with initiating another direct/white list connection while another white
list connection is pending
Bug: 111562702
Bug: 112827989
Change-Id: I82a476489d181d17aa1c1da8202df1be76928d77

5 years agoRevert "Remove redundant GATT_StartIf"
Jakub Pawlowski [Wed, 19 Dec 2018 12:43:03 +0000 (13:43 +0100)]
Revert "Remove redundant GATT_StartIf"

This change broke initial HID connection.

This reverts commit cb45b1e380126f300d9f7d89a8382438c5d06cd3.

5 years agoMerge "Add stng@google.com to the OWNERS list"
Treehugger Robot [Tue, 18 Dec 2018 21:11:28 +0000 (21:11 +0000)]
Merge "Add stng@google.com to the OWNERS list"

5 years agoMerge "Add a host test for net_test_avrcp"
Treehugger Robot [Tue, 18 Dec 2018 19:05:15 +0000 (19:05 +0000)]
Merge "Add a host test for net_test_avrcp"

5 years agoMerge "Make sure Hearing Aids' Connection Intervals is unchanged"
Treehugger Robot [Tue, 18 Dec 2018 17:04:26 +0000 (17:04 +0000)]
Merge "Make sure Hearing Aids' Connection Intervals is unchanged"

5 years agoFix mtu assignment with correct value
Ted Wang [Tue, 18 Dec 2018 02:31:06 +0000 (10:31 +0800)]
Fix mtu assignment with correct value

Assigned mtu value with correspond avct channel configuration

Bug: 120524683
Test: manual
Change-Id: Idb70c678cef1d704e232aef8706082a48b073ef8

5 years agoAdd stng@google.com to the OWNERS list
Stanley Tng [Tue, 28 Aug 2018 23:06:55 +0000 (16:06 -0700)]
Add stng@google.com to the OWNERS list

Test: N/A
Change-Id: Ic164086fc97fa72aa68ad751b35f339fa71233bb

5 years agouse epoll_create1
Nick Kralevich [Mon, 17 Dec 2018 17:09:43 +0000 (09:09 -0800)]
use epoll_create1

epoll_create leaks file descriptors. Use epoll_create1(EPOLL_CLOEXEC)
instead.

Bug: 120983106
Test: compiles and boots
Change-Id: I40abaaea6dad641dc03d141adb29e963e853edc4

5 years agoMake sure Hearing Aids' Connection Intervals is unchanged
Stanley Tng [Fri, 14 Dec 2018 01:31:31 +0000 (17:31 -0800)]
Make sure Hearing Aids' Connection Intervals is unchanged

Make sure that the connection interval of the LE Connection remains the
desired one and reset it if this interval gets changed.

Test: Manual with Hearing Aids device and repeated
connection/disconnection
Bug: 117886712
Change-Id: Id11d1d2e7901bfbbc9c452b94aba968afbc52ee0

5 years agoAdd a host test for net_test_avrcp
Dan Shi [Sat, 15 Dec 2018 01:17:40 +0000 (17:17 -0800)]
Add a host test for net_test_avrcp

Bug: 111921389
Test: atest
Change-Id: I26e4036bb76271ac6181429c904efe0f8b438256

5 years agoMerge "List test_config_template to android platform only."
Dan Shi [Sat, 15 Dec 2018 01:03:29 +0000 (01:03 +0000)]
Merge "List test_config_template to android platform only."

5 years agoList test_config_template to android platform only.
Dan Shi [Fri, 14 Dec 2018 17:56:36 +0000 (09:56 -0800)]
List test_config_template to android platform only.

This allows host side test config to be auto-generated from the right
template.

Bug: 120993548
Test: m -j net_test_avrcp && atest net_test_avrcp --host
Change-Id: I003713ea3c1c4c1c0686551828d817227ec19019

5 years agoUse White List for direct connection establishment to multiple devices
Jakub Pawlowski [Wed, 14 Nov 2018 15:15:04 +0000 (16:15 +0100)]
Use White List for direct connection establishment to multiple devices

Currently, we implement GAP Direct Connect procedure exactly as the spec
says: we attempt connection to just one device.

This means that an app doing Direct Connection can take over the ability
to establish connections, and if it repeats the request, can cause other
devices to not connect.

From now on, we will use White List for both Background and Direct
connections. When there is at least one direct connection, we will
increase the scan parameters used for connection.

Bug: 112827989
Test: added unittests, also sl4a GattConnectTest
Change-Id: Ic0b2f631e3c84d5e8e0f4683b0148f031bfd5639

5 years agoRemove unused battery level support
Jakub Pawlowski [Thu, 13 Dec 2018 23:27:11 +0000 (00:27 +0100)]
Remove unused battery level support

Change-Id: I88239a835562a5e3c32b830e4c2c31fed1f107b3

5 years agoMerge "Remove redundant GATT_StartIf"
Jakub Pawlowski [Fri, 14 Dec 2018 08:40:04 +0000 (08:40 +0000)]
Merge "Remove redundant GATT_StartIf"

5 years agoRemove redundant GATT_StartIf
Jakub Pawlowski [Thu, 13 Dec 2018 16:53:15 +0000 (17:53 +0100)]
Remove redundant GATT_StartIf

GATT_StartIf is just sending connect callback about all currently
connected devices, even ones that this app is not interested in.

For btif/java apps, we will trigger connection attempt to just one
device right after registration, which would trigger connect callback
anyway.

For all bta layer apps, they do guard inside connect callback to make
sure code is triggered for already known device, but the devices are
considered as known only after they already connected, and some event
triggered app to create control block.

Thus, GATT_StartIf is surplus feature that should be removed.

Test: run all sl4a GATT tests
Change-Id: I8c45b92dbbf80894c32bb870eb392e766063c934

5 years agoMerge "Pass package name as a part of startDiscovery"
Treehugger Robot [Thu, 13 Dec 2018 12:42:58 +0000 (12:42 +0000)]
Merge "Pass package name as a part of startDiscovery"

5 years agoMerge "DO NOT MERGE - Merge pie-platform-release (PPRL.181205.001) into master"
Treehugger Robot [Thu, 13 Dec 2018 05:45:20 +0000 (05:45 +0000)]
Merge "DO NOT MERGE - Merge pie-platform-release (PPRL.181205.001) into master"

5 years agoPass package name as a part of startDiscovery
Zach Johnson [Thu, 13 Dec 2018 01:13:29 +0000 (17:13 -0800)]
Pass package name as a part of startDiscovery

Test: manual
Bug: 118347252
Change-Id: I52adf73009a9766ad9aabf41657187343a2d0f41

5 years agoFix performance-for-range-copy warnings
Chih-Hung Hsieh [Wed, 12 Dec 2018 22:30:52 +0000 (14:30 -0800)]
Fix performance-for-range-copy warnings

Bug: 30413223
Test: make with WITH_TIDY=1 DEFAULT_GLOBAL_TIDY_CHECKS=-*,performance*
Change-Id: I7d5537c24c39d5c42ac17c8a67646e4b9321d6bb

5 years agoTurn stack/gatt/conneciton_manager.cc into text file
Jakub Pawlowski [Tue, 11 Dec 2018 20:53:58 +0000 (21:53 +0100)]
Turn stack/gatt/conneciton_manager.cc into text file

Change-Id: I031bfaa1745de55aeb0a0dbde3a7a67be0f9c982

5 years agoMerge changes I7f8b5cfd,I68d65028
Treehugger Robot [Wed, 12 Dec 2018 05:05:27 +0000 (05:05 +0000)]
Merge changes I7f8b5cfd,I68d65028

* changes:
  test_vendor: Add PacketBuilder classes
  test_vendor: Add PacketView for copyless parsing

5 years agoDO NOT MERGE - Merge pie-platform-release (PPRL.181205.001) into master
Xin Li [Tue, 11 Dec 2018 22:45:08 +0000 (14:45 -0800)]
DO NOT MERGE - Merge pie-platform-release (PPRL.181205.001) into master

Bug: 120502534
Change-Id: Icdcf2220fd28f92b182da73bee207151e80db295

5 years agoMerge "OWNERS -eisenbach, +zachoverflow"
Zach Johnson [Tue, 11 Dec 2018 04:21:52 +0000 (04:21 +0000)]
Merge "OWNERS -eisenbach, +zachoverflow"

5 years agotest_vendor: Add PacketBuilder classes
Myles Watson [Wed, 14 Nov 2018 00:56:38 +0000 (16:56 -0800)]
test_vendor: Add PacketBuilder classes

Templated PacketBuilder class for inserting fixed width
types with the correct endianness.

Test: rootcanal-packets_test_host --gtest_filter=*PacketBuilderEndian*
Change-Id: I7f8b5cfd225ebae35f310b44fc0782b940a23007

5 years agotest_vendor: Add PacketView for copyless parsing
Myles Watson [Wed, 7 Nov 2018 20:10:46 +0000 (12:10 -0800)]
test_vendor: Add PacketView for copyless parsing

A View is holds bounds and a shared pointer.
A PacketView is a forward list of View, templated for endianness.
A BaseIterator iterates over PacketViews.
An Iterator is a BaseIterator with extract, which is templated for endianness.

Test: rootcanal-packets_test_host
Change-Id: I68d6502861be61f411c0f5f2da33ead3d2392694

5 years agoOWNERS -eisenbach, +zachoverflow
Andre Eisenbach [Tue, 4 Sep 2018 17:17:20 +0000 (10:17 -0700)]
OWNERS -eisenbach, +zachoverflow

5 years agoFix GATT not working due to address mismatch
Jakub Pawlowski [Thu, 6 Dec 2018 18:32:37 +0000 (19:32 +0100)]
Fix GATT not working due to address mismatch

If the initial connection is established when using RPA, device can
later decide to use it's Identity Address for reconnections, if it
doesn't care about privacy.
When we receive such Identity Address, we must translate it into pseudo
address before passing it further down the stack.

Test: Connect to device using RPA, try to reconnect when it advertise
      it's Public address
Change-Id: Icb8eed25e1aad7003f1a3e0fed684168a8344544

5 years agoConfig: Reduce LOG(WARNING) to VLOG(1) when config key is not found
Jack He [Fri, 7 Dec 2018 21:58:12 +0000 (13:58 -0800)]
Config: Reduce LOG(WARNING) to VLOG(1) when config key is not found

Fixes: 120495903
Test: make
Change-Id: I724b8883ea400ed3c79fe31b75f1e2aa65b45b4d

5 years agoMerge "Preserve x18 while calling aptX encoder libraries."
Peter Collingbourne [Fri, 7 Dec 2018 00:47:06 +0000 (00:47 +0000)]
Merge "Preserve x18 while calling aptX encoder libraries."

5 years agoPreserve x18 while calling aptX encoder libraries.
Peter Collingbourne [Mon, 19 Nov 2018 20:58:18 +0000 (12:58 -0800)]
Preserve x18 while calling aptX encoder libraries.

Shadow call stack (SCS) is a security mitigation that uses a
separate stack (the SCS) for return addresses. In versions of
Android newer than P, the compiler normally cooperates with the
system to ensure that the SCS address is always stored in register
x18. This requires that everything in a process that uses SCS is
built with -ffixed-x18. However, two libraries (libaptX_encoder.so
and libaptXHD_encoder.so) are supplied in binary format and were
not compiled with -ffixed-x18. We are working with the vendor to get
these libraries rebuilt with -ffixed-x18, but until then, we need to
prevent this library from clobbering x18 so that we can turn on SCS
in the Bluetooth process.

Bug: 112907825
Bug: 120621517
Change-Id: Id27829ed5696903c42b4aeb75c3b3880c97a3a36

5 years agoMerge "Hearing Aid: Add clear conn_id and close ctrl UIPC when BT turn off"
Treehugger Robot [Thu, 6 Dec 2018 02:43:24 +0000 (02:43 +0000)]
Merge "Hearing Aid: Add clear conn_id and close ctrl UIPC when BT turn off"

5 years agoMerge "Enable bounds sanitization in Bluetooth."
Ivan Lozano [Wed, 5 Dec 2018 23:38:16 +0000 (23:38 +0000)]
Merge "Enable bounds sanitization in Bluetooth."

5 years agoMerge "Group BR-EDR related SDP variables under transport"
Treehugger Robot [Wed, 5 Dec 2018 22:04:09 +0000 (22:04 +0000)]
Merge "Group BR-EDR related SDP variables under transport"

5 years agoMetrics: Fix a typo in log
Jack He [Tue, 4 Dec 2018 22:30:28 +0000 (14:30 -0800)]
Metrics: Fix a typo in log

Test: make
Change-Id: I92a49c62ebf13d8da7f13b39c426d0eb14a83db6

5 years agoMerge PQ1A.181205.006 from Pi-QPR1-Release into ToT for Pi-Platform release.\n\nBUG...
Bill Rassieur [Tue, 4 Dec 2018 17:00:02 +0000 (17:00 +0000)]
Merge PQ1A.181205.006 from Pi-QPR1-Release into ToT for Pi-Platform release.\n\nBUG: 120448245

Change-Id: I39b32c98eac813e96ba19087a27c97d0da02ab5e

5 years agoGroup BR-EDR related SDP variables under transport
Srinu Jella [Thu, 11 Feb 2016 14:17:39 +0000 (19:47 +0530)]
Group BR-EDR related SDP variables under transport

Use Case: Random RNR initiated after HOGP Keyboard disconnection

steps:
1. keep madcatz in advertisement mode and scan, pair from DUT.
2. keep dailog Keyboard in advertisement mode and scan, pair from DUT.
3. Take Keyboard out of range to initiate disconnection from remote
or initiate disconnection from remote.

Failure: DUT trying to do RNR to random device.

Root Cause: when the disconnection happened for LE device, it sets
the wait_disc to true, that is causing random RNR to go for BR-EDR
device.

Fix: Group all BR-EDR related SDP variables under transport check
so that if some other operation on LE will not corrupt these BR-EDR
variables to mislead.

Test: code compilation.
Bug: 69031582
Change-Id: Ic9003f2b85bbaaa30b201b8a639a9d9b52055ac5

5 years agoSplit Timer into OnceTimer and RepeatingTimer
Jakub Pawlowski [Mon, 3 Dec 2018 15:50:40 +0000 (16:50 +0100)]
Split Timer into OnceTimer and RepeatingTimer

Follow the naming convention in libbase, where Callback was split into
OnceCallback, and RepeatingCallback, to make it obvious how many times
callback is expected to be called.

Test: unit tests were split from TimerTest into OnceTimerTest and
RepeatingTimerTest
Bug: 110303473
Change-Id: Ibbb9bb74a2e6e1a0f0f74c6ec44f63c5a4a1537f

5 years agoHearing Aid: Add clear conn_id and close ctrl UIPC when BT turn off
weichinweng [Tue, 27 Nov 2018 05:47:25 +0000 (13:47 +0800)]
Hearing Aid: Add clear conn_id and close ctrl UIPC when BT turn off

* On some condition, BT will stay on BLE_ON state when BT turn off. For
the case, Hearing Aid should be correctly disconnected and ctrl UIPC
close until next BT turn on.
* This CL add clear BtaGATTQueue and Close GATTC to make sure Queue
clean and HA enter disconnected state. And don't restart ctrl UIPC if
HearingAid is shutdown.

Bug: 118803637
Test: make, BT off/on when play A2DP and an app registered.
Change-Id: I1c52307637ee6b34222de9308dcfe1392ad170e8

5 years agoTest: Save SSR dump after test
Jack He [Fri, 30 Nov 2018 00:51:22 +0000 (16:51 -0800)]
Test: Save SSR dump after test

* Save SSR dump from /data/vendor/ssrdump after test invocation

Bug: 120098494
Test: atest :postsubmit
Change-Id: Iafc1666c8cc9fbc7f69b909be630b3d0c1445fe4

5 years agoMerge "Metrics: Add Bluetooth address obfuscator"
Treehugger Robot [Fri, 30 Nov 2018 07:59:20 +0000 (07:59 +0000)]
Merge "Metrics: Add Bluetooth address obfuscator"

5 years agoMerge "Assert if alarm callback is null"
Treehugger Robot [Fri, 30 Nov 2018 01:58:17 +0000 (01:58 +0000)]
Merge "Assert if alarm callback is null"

5 years agoMetrics: Add Bluetooth address obfuscator
Jack He [Fri, 16 Nov 2018 10:59:43 +0000 (02:59 -0800)]
Metrics: Add Bluetooth address obfuscator

* Add an address obfuscator that takes a 256-bit randomly generated
  salt and hashes a Bluetooth MAC address using HMAC-SHA256 algorithm
* The salt is generated per Android device during the first time
  Fluoride native layer is initialized after a connectivity reset
* The output of the above operation is an unique identifier for a
  Bluetooth device on the same Android device, but yet remains
  anonymous across different Android devices.
* Add a logging method loghex(std::array) that allows dumping integral
  arrays into a hex string
* Add more error logging to btif_config_get_bin() so that we know why it
  fails in the future
* Change mutex in btif_config to recursive_mutex to avoid deadlock when
  calling btif_config API functions with API function

Bug: 112969790
Test: mm -j40, bluetooth_test_common
Change-Id: I11f01cc82fe805c97fd7c86ad83b3d324ee5af53

5 years agoAssert if alarm callback is null
Stanley Tng [Mon, 26 Nov 2018 22:44:38 +0000 (14:44 -0800)]
Assert if alarm callback is null

Crash the system when callback is null. This will help to identify and debug the problem. Previously, we got a CFI crash which is hard to isolate.

Test: Run with Hearing Aids but did not simulate the error condition
Bug: 119533256
Change-Id: I9bc2af7c3c405bea3a57c5feb9b213614484507f

5 years agoMerge pi-qpr1-release PQ1A.181105.017.A1 to pi-platform-release
Bill Yi [Thu, 29 Nov 2018 02:35:03 +0000 (18:35 -0800)]
Merge pi-qpr1-release PQ1A.181105.017.A1 to pi-platform-release

Change-Id: I2a349df7ad02b6b57044fe77d92b329e1a5afda0

5 years agoAllow Resolvable Private Address into LE White List
Jakub Pawlowski [Thu, 15 Nov 2018 22:22:21 +0000 (23:22 +0100)]
Allow Resolvable Private Address into LE White List

When doing background connection, we don't want to put RPA into the
white list, as it might expire in 15 minutes.
For direct connect procedure, RPA is ok - the request times out after 30
seconds.

This patch moves address type check from the White List to the background
connection procedure. It is prepearation for handling direct connect
with white list.

Bug: 112827989
Test: sl4a GattConnectTest
Change-Id: I907ad44d1d255c9212ed58112bb8b99464e46f43

5 years agoCentralize call to gatt_update_app_use_link_flag for GATT_Connect
Jakub Pawlowski [Tue, 13 Nov 2018 17:31:08 +0000 (18:31 +0100)]
Centralize call to gatt_update_app_use_link_flag for GATT_Connect

Currently both code paths, doing direct and background connection, call
gatt_update_app_use_link_flag. Move this call to common code.

Bug: 112827989
Test: sl4a GattConnectTest
Change-Id: I5d55f6467ffca78ff785ecdad02dd049ea5481bf

5 years agoMerge "Prevent potential deadlock in Hearing Aids"
Treehugger Robot [Wed, 21 Nov 2018 20:52:18 +0000 (20:52 +0000)]
Merge "Prevent potential deadlock in Hearing Aids"

5 years agoMerge "HFP: Send caller ID in +CLIP command (3/4)"
Treehugger Robot [Wed, 21 Nov 2018 20:21:19 +0000 (20:21 +0000)]
Merge "HFP: Send caller ID in +CLIP command (3/4)"

5 years agoMerge "btsock_signaled: provide debugging info for crash"
Treehugger Robot [Wed, 21 Nov 2018 19:51:32 +0000 (19:51 +0000)]
Merge "btsock_signaled: provide debugging info for crash"

5 years agoMerge "btif_av: remove dead code"
Treehugger Robot [Wed, 21 Nov 2018 19:04:36 +0000 (19:04 +0000)]
Merge "btif_av: remove dead code"

5 years agobtif_av: remove dead code
Hansong Zhang [Wed, 21 Nov 2018 00:04:18 +0000 (16:04 -0800)]
btif_av: remove dead code

Test: manual
Change-Id: I8b172ec2235ae5847f073f8d50ad71761bc37d02

5 years agoPrevent potential deadlock in Hearing Aids
Stanley Tng [Fri, 16 Nov 2018 19:34:28 +0000 (11:34 -0800)]
Prevent potential deadlock in Hearing Aids

Prevent a potential deadlock when calling the main thread but the main
thread is not running.

Bug: 117619792
Test: Manual test to make sure no regression
Change-Id: I6e001533144827b17b88547671765f790ed771bd

5 years agoMerge "Remove unused BTM_BleSetAdvParams"
Treehugger Robot [Tue, 20 Nov 2018 21:13:35 +0000 (21:13 +0000)]
Merge "Remove unused BTM_BleSetAdvParams"

5 years agoRemove unused BTM_BleSetAdvParams
Jakub Pawlowski [Tue, 20 Nov 2018 11:14:10 +0000 (12:14 +0100)]
Remove unused BTM_BleSetAdvParams

Test: compilation
Change-Id: I5797a16b3fd92c5ffa1ad1fd048699a591fd1e4c

5 years agoLinux tests build fix
Jakub Pawlowski [Tue, 20 Nov 2018 18:27:11 +0000 (19:27 +0100)]
Linux tests build fix

Test: compile on Linux using ninja
Change-Id: I180679a7a6f12abe83db0480d9007ebdfdf1d451

5 years agoFix an error in pure virtual function in earlier CL
Stanley Tng [Wed, 7 Nov 2018 20:50:37 +0000 (12:50 -0800)]
Fix an error in pure virtual function in earlier CL

The earlier CL "Fix Spurious Hearing Aid Start and Stop Cmds" is missing
this "=0" to make the virtual function pure. This CL fixes this
oversight.

Test: Compile
Bug: 117619792

Change-Id: Id9e72b8d93f9671c08a613ecb43b4121662770a0

5 years agobtsock_signaled: provide debugging info for crash
Hansong Zhang [Thu, 15 Nov 2018 23:37:12 +0000 (15:37 -0800)]
btsock_signaled: provide debugging info for crash

Test: manual
Bug: 68108403
Change-Id: I89cfc15787f51594535e1e0c00e7a18a317a5d81

5 years agoDump gatt::connection_manager state into bugreport
Jakub Pawlowski [Tue, 13 Nov 2018 14:41:42 +0000 (15:41 +0100)]
Dump gatt::connection_manager state into bugreport

Knowing what's in the white list is useful for debugging.

Bug: 112827989
Test: adb bugreport, verify content.
Change-Id: I195d1ee2fe277f0f0ef534c73f7a964c5f5e61b6

5 years agoUse map instead of list for keeping apps interested in connection
Jakub Pawlowski [Sat, 10 Nov 2018 15:45:35 +0000 (16:45 +0100)]
Use map instead of list for keeping apps interested in connection

Bug: 112827989
Test: BleGattConnectionManager
Change-Id: I7b7ea608f5e27a9f17141ccec0c5d5c06bb22e60

5 years agoFirst tests for connection management
Jakub Pawlowski [Fri, 9 Nov 2018 13:41:01 +0000 (14:41 +0100)]
First tests for connection management

Test: added in this CL
Change-Id: Id3937530e1b0a16c056ae48face78b18c6ec32f0

5 years agointroduce gatt::connection_manager::get_apps_connecting_to
Jakub Pawlowski [Fri, 9 Nov 2018 17:40:04 +0000 (18:40 +0100)]
introduce gatt::connection_manager::get_apps_connecting_to

This is to clean up the interface, and don't expose internal control
structures.

Test: compilation
Change-Id: Id2afc5efb4c14be0b787cfabbe21fc1961381cb7

5 years agoNaming cleanup in stack/gatt/connection_manager.*
Jakub Pawlowski [Fri, 9 Nov 2018 11:42:32 +0000 (12:42 +0100)]
Naming cleanup in stack/gatt/connection_manager.*

Test: compilation
Change-Id: Ic73632cb258515856b9ae5616ec47881b4b28c61

5 years agogatt_utils_white_list -> connection_manager
Jakub Pawlowski [Fri, 9 Nov 2018 10:53:11 +0000 (11:53 +0100)]
gatt_utils_white_list -> connection_manager

Test: compilation
Change-Id: Ib6741ddd8332ff70efd47a8ede0176d887743ff4

5 years agoMerge "AVRCP: Change notification variable when the interim response is sent"
Treehugger Robot [Fri, 9 Nov 2018 02:57:12 +0000 (02:57 +0000)]
Merge "AVRCP: Change notification variable when the interim response is sent"

5 years agoEnable bounds sanitization in Bluetooth.
Ivan Lozano [Fri, 28 Sep 2018 18:23:51 +0000 (11:23 -0700)]
Enable bounds sanitization in Bluetooth.

This enables UBSan bounds sanitization in our bluetooth modules.
This adds run-time bounds checks where the compiler can statically
determine the array bounds.

Bug: 116517182
Test: Pairing devices, playing audio, sharing files.
Change-Id: I9a3ac98aaa6092ac0ad8d1ba080a76b3203bdb26

5 years agoAVRCP: Change notification variable when the interim response is sent
Ajay Panicker [Tue, 26 Jun 2018 21:52:56 +0000 (14:52 -0700)]
AVRCP: Change notification variable when the interim response is sent

This avoids the possibility of having a changed notification be sent
before an interim response is sent when the AVRCP service has to wait
for a response from the media layer for information regarding the
notification.

Also fix the PlayPosition test for AVRCP as it was testing for play
state correctness.

Bug: 109588762
Test: Run host test net_test_avrcp
Change-Id: Icb7bb1068191ecf2a2e390854084b6b9b47537fa
(cherry picked from commit 2d87908fe8825245dcc2e1dbde027331f1e2371d)

5 years agoMinor style fixes in gatt_main.cc
Jakub Pawlowski [Thu, 8 Nov 2018 18:54:49 +0000 (19:54 +0100)]
Minor style fixes in gatt_main.cc

Test: compilation
Change-Id: I446b9e83603e7224abe1eddd030b4502cc7dceb5

5 years agoMerge "Add code coverage report support to Bluetooth"
Treehugger Robot [Thu, 8 Nov 2018 22:02:48 +0000 (22:02 +0000)]
Merge "Add code coverage report support to Bluetooth"

5 years agoHFP: Send caller ID in +CLIP command (3/4)
Benson Li [Tue, 17 Jul 2018 10:05:16 +0000 (18:05 +0800)]
HFP: Send caller ID in +CLIP command (3/4)

* Append caller display name in ag_reslut data if there is name in
  CALL_STATE_INCOMING state
* Process all clip data and store the formatted string to ag_result
  data

Bug: 111378831
Test: runtest bluetooth, runtest -x BluetoothPhoneServiceTest.java
Change-Id: Ief48ed9cb5e0b32274624a90799ab9922dbd656e