OSDN Git Service

android-x86/system-bt.git
9 years agoFree Avrcp Controller callbacks on Cleanup
Anubhav Gupta [Mon, 8 Sep 2014 10:30:22 +0000 (16:00 +0530)]
Free Avrcp Controller callbacks on Cleanup

Free AVRCP Controller callbacks on cleanup so that callbacks
are initialized properly on next initialization.

Change-Id: Ieaf4c1989a7c40c591adfb7cb47912532ceee46e

9 years agoProtect reset of out stream with mutex
Nitin Srivastava [Mon, 22 Dec 2014 14:24:09 +0000 (19:54 +0530)]
Protect reset of out stream with mutex

Protect reset of out stream with mutex so that the access
of the same is properly synchronized in multi-threaded environment.

Change-Id: I6dd81d8c7fa219ee8d1e4d82aa1e7ec52dcecec8

9 years agoLE Privacy 1.2 and LE secure connections
Satya Calloji [Fri, 6 Mar 2015 18:38:22 +0000 (10:38 -0800)]
LE Privacy 1.2 and LE secure connections

Bug: 19816438
Original author: Chaojing Sun <cjsun@broadcom.com>
Change-Id: I5951f4d4e038f8348a62aa6d19b2111bae0b3ecc

9 years agoL2CAP and SDP Search API for BT profiles (2/2)
Kim Schulz [Wed, 25 Mar 2015 09:39:40 +0000 (10:39 +0100)]
L2CAP and SDP Search API for BT profiles (2/2)

Added support for exposing L2CAP to Java such that OBEX over L2CAP
is made possible.

Added support to create SDP records as a seperate step.(as opposed to
creating a SDP record when a BluetoothSocket is created).
This allows both a RFCOMM channel and a L2CAP PSM to be included in a
SDP record. (Additionally the content of the SDP record is set by the
profile in Java, in stead of beeing hardcoded in the socket layer.)
This completes the L2CAP channel exposure to Java.

Change-Id: Iaf68a07d910145cdd33e940d73cd680f79164100

9 years agoUpdate Service Availability.
Mallikarjuna GB [Wed, 5 Nov 2014 09:48:50 +0000 (15:18 +0530)]
Update Service Availability.

 - service availability is updated on CIND response
   during SLC set-up.
 - update current service availability on +CIEV response from
   remote device.

Change-Id: I57643a6ac3aad5339d686cf941902a8bf88d99fb

9 years agoAvoid duplicate disconnection callbacks for fixed channel connections
Satya Calloji [Tue, 17 Mar 2015 20:12:01 +0000 (13:12 -0700)]
Avoid duplicate disconnection callbacks for fixed channel connections

Fixed channel connections are getting a disconnect callback when the
HCI command is issued and when the HCI disconnect completes. This
causes problems if the upper layer trying to reconnect at link
disconnecting state.
Triggering the disconnect callback only once solves this problem.

Original author: Chaojing Sun <cjsun@broadcom.com>

Bug: 19816438
Change-Id: Ib661c968e586975a7fc7244e2d0745f71d52e3e9

9 years agoAdd a section on assertions to the style guide.
Sharvil Nanavati [Wed, 8 Apr 2015 06:22:08 +0000 (23:22 -0700)]
Add a section on assertions to the style guide.

Change-Id: I5627b5b2d6b25f0472c53b8150b14bab79b200c3

9 years agoMerge "Avoid null-pointer dereference in l2c_release_lcb()" into m-wireless-dev
Andre Eisenbach [Wed, 8 Apr 2015 18:11:54 +0000 (18:11 +0000)]
Merge "Avoid null-pointer dereference in l2c_release_lcb()" into m-wireless-dev

9 years agoAvoid null-pointer dereference in l2c_release_lcb()
Andre Eisenbach [Wed, 8 Apr 2015 17:01:56 +0000 (10:01 -0700)]
Avoid null-pointer dereference in l2c_release_lcb()

Change-Id: Id46d678e45797c4d3d3c8e024f42cfd06b1d2da2

9 years agoAdd a 'Bluedroid conventions' section to the style guide.
Sharvil Nanavati [Wed, 8 Apr 2015 06:10:59 +0000 (23:10 -0700)]
Add a 'Bluedroid conventions' section to the style guide.

Also introduce a memory management section underneath 'Bluedroid
conventions' that describes common ownership semantics.

Change-Id: I9f1f148437fe21788618bcdc884b2eea5cbff141

9 years agoAdd a section to the style guide on variable declarations.
Sharvil Nanavati [Wed, 8 Apr 2015 05:45:21 +0000 (22:45 -0700)]
Add a section to the style guide on variable declarations.

Change-Id: I0eaa435378b56953c1c228c795d0b4bf9bd8ce6c

9 years agoMinor cleanup of formatting in style guide.
Sharvil Nanavati [Wed, 8 Apr 2015 05:41:01 +0000 (22:41 -0700)]
Minor cleanup of formatting in style guide.

Remove trailing whitespace on a line and add inline code quotations.

Change-Id: I2f85ddf95f0dcfb0d9fdb1e6663315be15efeb62

9 years agoOnFound/OnLost feature changes based on Google spec v0.90
Satya Calloji [Tue, 10 Feb 2015 01:40:52 +0000 (17:40 -0800)]
OnFound/OnLost feature changes based on Google spec v0.90

Bug: 19609671
Change-Id: If9e61b405a141cb77fb65741d08227284cb22ccd

9 years agoStatic code analysis cleanups and minor bug fixes
VenkatRaghavan VijayaRaghavan [Thu, 5 Feb 2015 12:40:47 +0000 (04:40 -0800)]
Static code analysis cleanups and minor bug fixes

- Fix for initializing all the un-initialized variables
- Handling system call failure like socket,fcntl etc...
- In btpan_tap_close validate tap_if_down and close the socket.
- Invalid memset size, which could have potential memory issue
- Change -1 to INVALID_FD where appropriate

Bug: 19417758
Change-Id: Id31832f8678b2d72c80740c29b946a94e7ae0197

9 years agoSet dual-mode bits correctly in multi-adv instance flags
Satya Calloji [Thu, 5 Feb 2015 21:12:15 +0000 (13:12 -0800)]
Set dual-mode bits correctly in multi-adv instance flags

Bug: 19369962
Change-Id: I081f9db56d52e85cff50f668752322b8ab8adf42

9 years agoMerge "Fix pthread_t confusion." into m-wireless-dev
Ian Coolidge [Fri, 3 Apr 2015 06:02:48 +0000 (06:02 +0000)]
Merge "Fix pthread_t confusion." into m-wireless-dev

9 years agoDon't truncate the top 16 bits of the LE advertisement data mask.
Ian Coolidge [Wed, 4 Mar 2015 01:00:17 +0000 (17:00 -0800)]
Don't truncate the top 16 bits of the LE advertisement data mask.

This is stored elsewhere as unsigned 32-bit integer.
It should be stored the same way here.

Bug: 19592970
(cherry picked from commit 91f6987ecca884b0995b77fea014046b8a56385a)

Change-Id: I21dbe07e4a1095e3e8b38a56b9673a1cebb32b08

9 years agoFix pthread_t confusion.
Ian Coolidge [Tue, 3 Feb 2015 04:41:11 +0000 (20:41 -0800)]
Fix pthread_t confusion.

Don't assign pthread_t to pid_t or UINT32.
Just use pthread_t throughout.

(cherry picked from commit df3459935a4c12744a9a78812157890b60ccb77d)

Change-Id: I1ab086a547a731358003471be867cfb189f61c56

9 years agoPrevent slave link block removal at cancel connect
Nitin Arora [Wed, 19 Nov 2014 22:24:08 +0000 (14:24 -0800)]
Prevent slave link block removal at cancel connect

This change prevents the cancellation of direct LE connection
from removing the lcb, if there is already an LE connection up
with DUT as peripheral

Change-Id: If11759e40d19f428959b28025dde2d65b6ed7cd6

9 years agoosi: add compat code for glibc
Scott James Remnant [Thu, 2 Apr 2015 22:22:29 +0000 (15:22 -0700)]
osi: add compat code for glibc

glibc doesn't declare strlcpy(), strlcat() or gettid(). Add a header
of compatibility routines that only get included when compiling on this.

Change-Id: I889ac7f1d22cf7056757890fcff0403107488d62

9 years agofixed missing {} around if scope
Kim Schulz [Mon, 30 Sep 2013 07:52:55 +0000 (09:52 +0200)]
fixed missing {} around if scope

a range of lines was supposed to be part of an if-true scope, but was not (due to missing {})

Bug: 10905622
Change-Id: I7adad37216cc78403f2d187d3404565f527d67c1

9 years agoAdd missing header for APPL_TRACE_EVENT
Scott James Remnant [Thu, 2 Apr 2015 22:22:27 +0000 (15:22 -0700)]
Add missing header for APPL_TRACE_EVENT

Change-Id: Iebe9d899dc73867549c40a569094ff9b5c340808

9 years agoUse int64_t instead of non-standard __int64
Scott James Remnant [Thu, 2 Apr 2015 22:22:25 +0000 (15:22 -0700)]
Use int64_t instead of non-standard __int64

Change-Id: I482b861364531d27ef89aab6f5617a060176de83

9 years agoMake atomic functions static, as well as inline
Scott James Remnant [Thu, 2 Apr 2015 22:22:22 +0000 (15:22 -0700)]
Make atomic functions static, as well as inline

'inline' is only a hint for the compiler, and if ignored this results
in extern functions being created. Add 'static' so that the resulting
functions are also only local to the compilation unit.

Change-Id: I947201b8170c8c73f15e7ccdd68cfc7f07cde1e9

9 years agoFix non-standard pthread mutex interactions.
Scott James Remnant [Thu, 2 Apr 2015 22:23:22 +0000 (15:23 -0700)]
Fix non-standard pthread mutex interactions.

pthread_mutex_t is intended to be opaque.
pthread_attr_t is intended to be opaque.

Change-Id: Ie0f5558a96f8e5fc8b2f59763acbfe5340fa81a4

9 years agobtif_pan: Grab interface declarations from net/if.h only.
Scott James Remnant [Thu, 2 Apr 2015 22:23:16 +0000 (15:23 -0700)]
btif_pan: Grab interface declarations from net/if.h only.

This is the libc interface file.
Using linux/if.h produces definition collisions on glibc.

Change-Id: I8744258ba0c4698d8f9f5b1215bb3cb1d86618fe

9 years agoFix missing #includes needed for glibc
Scott James Remnant [Thu, 2 Apr 2015 22:22:14 +0000 (15:22 -0700)]
Fix missing #includes needed for glibc

Some standard library functions are used while relying on bionic headers
including the headers the functions were actually declared on. Add those
missing #includes so that bluetooth.default.so will compile on glibc.

Change-Id: Ied9f89ce5a05911fca63f6bfe1b8cc8196ab3b1c

9 years agoMerge "Simultanious LE Central/Peripheral operation not possible" into m-wireless-dev
Andre Eisenbach [Tue, 31 Mar 2015 22:29:40 +0000 (22:29 +0000)]
Merge "Simultanious LE Central/Peripheral operation not possible" into m-wireless-dev

9 years agoBT: fix possible out of bounds read
kschulz [Wed, 11 Feb 2015 13:41:51 +0000 (14:41 +0100)]
BT: fix possible out of bounds read

In btm_ble_bgconn.c, function btm_update_bg_conn_list() if the last element in
"p_cb->bg_dev_list" is removed, the line:

"p_next = p_bg_dev + 1;"

will get "p_next" to point at an entry outside the bounds of the array.
There check for "p_next->in_use" may render a TRUE result depending on what
data exist in the memory pointed to by "p_next", which is unknown.
If the check passes, the sentence:

" memcpy(p_cur, p_next, sizeof(tBTM_LE_BG_CONN_DEV));"

will copy random data into the entry pointed at by "p_cur".
This can potentially corrupt the white list filling it with entries that
contain garbage and not allowing to add new valid devices to the list.

Bug: 18748601
Change-Id: I98b4be4e9f4568ea77f89146eae1e38f19fa1ee3

9 years agoSDP parameter checks too loose
Kim Schulz [Tue, 24 Sep 2013 07:52:52 +0000 (09:52 +0200)]
SDP parameter checks too loose

SDP Server's client req handler currently only check that the parameter length field does not indicate the parameters to excite the packet length. Sending SDP PDUs with
- fixed sdp_server_client_req to also check for length filed

If an invalid SDP Service Search Request Packet has the continuation info length field set to SDP_CONTINUATION_LEN, but does not contain any continuation information, the SDP server should answer with SDP_INVALID_CONT_STATE.
- fixed process_service_search so the length field is correctly checked
- fixed possible access to 2 bytes of memory outside the packet boundry

Bug: 10904535
Change-Id: I41fd021327fbf4f810ee4319bc95fa0dba0b48df

9 years agoInclude log library to make bdtool
Nitin Arora [Mon, 30 Mar 2015 19:22:30 +0000 (12:22 -0700)]
Include log library to make bdtool

This change adds the liblog library in the
make file for bdtool to provide for separate
compilation of the tool without causing link
errors while importing libosi

Change-Id: I87752ef7e1026080bd56fb377c73b457775f12a3

9 years agoUsing generic format specifier for logging pointers
Nitin Arora [Wed, 25 Mar 2015 00:15:50 +0000 (17:15 -0700)]
Using generic format specifier for logging pointers

This change uses generic format specifier for logging the
pointers which can be 64 bit or 32 bit depending upon the
system used for compilation.
Also address of operator is used before casting a int variable
to void pointer.

Change-Id: Ie9dec408f3a5f674079f77c047d8192f65a004ca

9 years agoRemove BLE_PERIPHERAL_ADV_NAME flag.
Wei Wang [Fri, 27 Mar 2015 18:28:54 +0000 (11:28 -0700)]
Remove BLE_PERIPHERAL_ADV_NAME flag.

BLE_PERIPHERAL_ADV_NAME forces inclusion of device name
in advertise data. Since setIncludeDeviceName is already
available in advertise API to control the presence of
device name the flag should be removed.

Change-Id: I8637529b2d809bb9829783f60bd7e9889784c8b7

9 years agoMake A2DP audio HAL a shared library again
Andre Eisenbach [Sat, 28 Mar 2015 00:12:19 +0000 (17:12 -0700)]
Make A2DP audio HAL a shared library again

The audio HAL is loaded separately from the Bluetooth stack and thus
needs to stay separate.

Change-Id: I8d8601a6e753099f4f058dc0192b93c53be2ffb1

9 years agoMerge "Fix macro for maximum allowed HOGP devices" into m-wireless-dev
Andre Eisenbach [Fri, 27 Mar 2015 01:01:34 +0000 (01:01 +0000)]
Merge "Fix macro for maximum allowed HOGP devices" into m-wireless-dev

9 years agoFix macro for maximum allowed HOGP devices
Nitin Arora [Mon, 10 Mar 2014 20:25:27 +0000 (13:25 -0700)]
Fix macro for maximum allowed HOGP devices

This checks HOGP dev handle count less than or equal to maximum
allowed HOGP devices.

Change-Id: I63e6c3bc64a8c305b739b5eade0e846315e1ad1d

9 years agoAdd option for schedule_next_instance to force rescheduling
Zach Johnson [Thu, 26 Mar 2015 22:07:43 +0000 (15:07 -0700)]
Add option for schedule_next_instance to force rescheduling

If we're rescheduling a periodic timer in the context of an alarm
expiration we want to force rescheduling to happen.

The alarm at the front of the list after scheduling the next
instance may be a different alarm. The old code would assume it was
already scheduled correctly, meaning nothing would ever get scheduled
after that.

Change-Id: I8accf5d004e69dfd6477ab7ec529eea42b3a88f8

9 years agoDemote bta_av_link_role_ok error level to info
Chris Manton [Thu, 26 Mar 2015 21:49:25 +0000 (14:49 -0700)]
Demote bta_av_link_role_ok error level to info

Change-Id: I4399432c9071614fc610d9de084028e48aeaf8fd

9 years agoRemove unused file hci/include/bt_hci_lib.h
Chris Manton [Thu, 26 Mar 2015 21:08:38 +0000 (14:08 -0700)]
Remove unused file hci/include/bt_hci_lib.h

Change-Id: I60d87da42fa28dd57d50384396e30504103d1ae5

9 years agoEnsure alarms are called back when they expire in the past
Zach Johnson [Thu, 26 Mar 2015 06:32:11 +0000 (23:32 -0700)]
Ensure alarms are called back when they expire in the past

Turns out the posix timers we're dealing with aren't well behaved.

If the timer is TIMER_ABSTIME the following is supposed to be true:
"If the specified time has already passed, the function shall succeed
and the expiration notification shall be made."

But alas, this is not the case. If the expiration time happens to be
in the past (e.g. very short timer which gets hit with a context switch
before the timer can be set) the timer decides to disarm itself. This
means no more callbacks happen, and no more alarms are processed ever.

sadness.

But thankfully, we can use timer_gettime to check the state of the timer
after timer_settime. If timer_gettime tells us the timer is disarmed
right after we armed it, we want to perform the alarm callback
ourselves.

Put all timer callbacks on the same thread (as would be the case in the
underlying timer implementation already) and used a sempahore to signal
when an alarm expires in the normal posix timer callback and also in the
timer didn't set case.

Change-Id: I39854b241369b2da52afbaaba0c01d3167a47e32

9 years agoRewrite alarms to use a single posix timer instance
Zach Johnson [Tue, 24 Mar 2015 01:07:46 +0000 (18:07 -0700)]
Rewrite alarms to use a single posix timer instance

This means we lose the ability to send the intended alarm
as context with the posix timer, but this can be handled
gracefully enough by making sure we actually have an
expired alarm at the front of the list when the timer
callback occurs.

Change-Id: I909195b5c3155075096328c529655eeb5b7d80d0

9 years agoChange logtag to conform to rest of module
Chris Manton [Thu, 26 Mar 2015 20:28:51 +0000 (13:28 -0700)]
Change logtag to conform to rest of module

Change-Id: I9b99111f50dc4f1c581ef96e78b36d14b51a6866

9 years agoMove unused include file out of hci_audio.h
Andre Eisenbach [Thu, 26 Mar 2015 21:14:40 +0000 (14:14 -0700)]
Move unused include file out of hci_audio.h

This fixes include path issues for libbt.

Change-Id: I416f563c0056216e86898d8e9cc111b600c73ce0

9 years agoResolved hardware error observed during SCO Connection setup
Mudumba Ananth [Wed, 25 Mar 2015 12:20:02 +0000 (05:20 -0700)]
Resolved hardware error observed during SCO Connection setup

Hardware error was caused due to a faulty HCI command formed in the
process of vendor specific pre-SCO setup in the stack(set_audio_state)

Fixed the above problem and also added back the vendor (interface)
mapping for the set_audio_state functionality to facilitate the
sending of the pre-SCO vendor specific commands.

Made common vendor library audio setting by moving to hci subsystem.

General cleanup around this functionality.

Bug: 19923226
Change-Id: I4a743f6725459f360bd2a90e0a46f08fcca2292d

9 years agoRandom address does not get written properly
Satya Calloji [Tue, 3 Feb 2015 22:41:49 +0000 (14:41 -0800)]
Random address does not get written properly

The random address not getting written with correct
instance ID, as per Google spec.

Bug: 19370739
Change-Id: I7e5fc8d274bad5f127bd791469a46e2f257eaddd

9 years agoBuilds server with service, characteristic, and descriptor UUIDs. Then starts and...
Steve Gu [Thu, 6 Nov 2014 19:16:41 +0000 (11:16 -0800)]
Builds server with service, characteristic, and descriptor UUIDs. Then starts and stops the server.

Change-Id: I22e269010c5e4d96d9562155e03903c3ea957c97

9 years agoAdds tests for BLE advertising and BLE server register. Also fixes random seed.
Steve Gu [Wed, 29 Oct 2014 01:38:30 +0000 (18:38 -0700)]
Adds tests for BLE advertising and BLE server register. Also fixes random seed.

Change-Id: Ic15360baefb133b761ff077d27946c0ea7fa3edb

9 years agoAdds BLE tests to bdtest.
Steve Gu [Mon, 27 Oct 2014 19:59:02 +0000 (12:59 -0700)]
Adds BLE tests to bdtest.

Change-Id: Idcdc01339f0e8a14b6fedf27545566e7efdf53c2

9 years agoMerge "Add section on booleans and bitfields to the style guide" into m-wireless-dev
Andre Eisenbach [Thu, 26 Mar 2015 00:04:14 +0000 (00:04 +0000)]
Merge "Add section on booleans and bitfields to the style guide" into m-wireless-dev

9 years agoAdd section on booleans and bitfields to the style guide
Zach Johnson [Tue, 24 Mar 2015 17:05:15 +0000 (10:05 -0700)]
Add section on booleans and bitfields to the style guide

Change-Id: I75df8b638d71bfe2ba1dc98d3eb148d8a61f5b9c

9 years agoRectify allocation of GATT control block
Nitin Arora [Fri, 20 Mar 2015 00:00:35 +0000 (17:00 -0700)]
Rectify allocation of GATT control block

This change ensures that a valid value is returned while
requesting available control block or allocating memory
for a new control block for saving GATT discovery handles.
The bug if not fixed causes memory corruption and thus
modifying some callback handles.

Change-Id: I148ee82d2ea9fe77bebd34fd345b4fb6a6eee3cd

9 years agoRevert "Resolved hardware error observed during SCO Connection setup"
Chris Manton [Wed, 25 Mar 2015 18:01:11 +0000 (18:01 +0000)]
Revert "Resolved hardware error observed during SCO Connection setup"

This reverts commit 4d8ac4ff015ff87c4e985ac0d9a5f3c2a6b0d9fa.

Change-Id: Ie1433c82014b6ddce78c5363100ab109ae0fecde

9 years agoResolved hardware error observed during SCO Connection setup
Mudumba Ananth [Wed, 25 Mar 2015 12:20:02 +0000 (05:20 -0700)]
Resolved hardware error observed during SCO Connection setup

Hardware error was caused due to a faulty HCI command formed in the
process of vendor specific pre-SCO setup in the stack(set_audio_state)

Fixed the above problem and also added back the vendor (interface)
mapping for the set_audio_state functionality to facilitate the
sending of the pre-SCO vendor specific commands.

Bug: 19923226
Change-Id: I03c6a1c048c21549b7326a783b70588686283eb3

9 years agoSimultanious LE Central/Peripheral operation not possible
Kim Schulz [Wed, 25 Mar 2015 08:08:01 +0000 (09:08 +0100)]
Simultanious LE Central/Peripheral operation not possible

Bug: 18925497
Change-Id: I5a20e38ffebce4b251e4728944fce73db50e18f8

9 years agoAllow dual-mode devices to be processed by BR/EDR scan
Nitin Arora [Wed, 22 Jan 2014 18:33:26 +0000 (10:33 -0800)]
Allow dual-mode devices to be processed by BR/EDR scan

This change lets a Dual mode device which is already processed
as LE device for the current inquiry, to be also processed as
a BR/EDR device in case an extended inquiry result is recvd.
This helps setting the cod correctly.

Change-Id: I0a0c598ebdafe2b6153015c3cda5dde8f7a15873

9 years agoRewrite alarm to avoid drift in the periodic version
Zach Johnson [Mon, 23 Mar 2015 20:59:10 +0000 (13:59 -0700)]
Rewrite alarm to avoid drift in the periodic version

Also improve names of some functions.

Change-Id: Ia5aeebc8226ce5c36abdd372ea4de5195314e273

9 years agoMerge "Remove unused remote_name_* variables" into m-wireless-dev
Andre Eisenbach [Tue, 24 Mar 2015 01:11:01 +0000 (01:11 +0000)]
Merge "Remove unused remote_name_* variables" into m-wireless-dev

9 years agoRemove unused remote_name_* variables
Andre Eisenbach [Tue, 24 Mar 2015 01:07:54 +0000 (18:07 -0700)]
Remove unused remote_name_* variables

Change-Id: Ib0db95646564439a9458542db9b1891ea818c5d7

9 years agoMerge "Style guide: Use C99 style flexible arrays" into m-wireless-dev
Andre Eisenbach [Tue, 24 Mar 2015 00:44:36 +0000 (00:44 +0000)]
Merge "Style guide: Use C99 style flexible arrays" into m-wireless-dev

9 years agoStyle guide: Use C99 style flexible arrays
Andre Eisenbach [Thu, 19 Mar 2015 22:38:34 +0000 (15:38 -0700)]
Style guide: Use C99 style flexible arrays

Change-Id: I012d9f4195833e2c6c9a5c953e71b7a7b5847673

9 years agoDelete 0 index LE scan filter at inquiry cancel
Nitin Arora [Fri, 26 Sep 2014 21:05:24 +0000 (14:05 -0700)]
Delete 0 index LE scan filter at inquiry cancel

This patch ensures that the 0 index filter is deleted when the
inquiry is cancelled due to initiation of a direct connection
from inquiry or due to return from the Settings menu screen

Change-Id: I281c6ace15b17fb7e413c42d36e016b8f82d2ace

9 years agoIgnore LE security request if BR/EDR pairing is ongoing
Nitin Arora [Sat, 22 Mar 2014 01:28:04 +0000 (18:28 -0700)]
Ignore LE security request if BR/EDR pairing is ongoing

This change checks the btm state to see if a BR/EDR pairing
is ongoing when a security request is received from a remote
device.

Change-Id: Ia3f9c9db6c1aed944857174528c14ed66f184152

9 years agoMerge "Extend Bluetooth startup timeout from 3s to 8s" into m-wireless-dev
Andre Eisenbach [Mon, 23 Mar 2015 20:26:23 +0000 (20:26 +0000)]
Merge "Extend Bluetooth startup timeout from 3s to 8s" into m-wireless-dev

9 years agoMerge "Reset LE connecting flag on disconnect only if BDA matches" into m-wireless-dev
Andre Eisenbach [Mon, 23 Mar 2015 20:21:42 +0000 (20:21 +0000)]
Merge "Reset LE connecting flag on disconnect only if BDA matches" into m-wireless-dev

9 years agoMerge "Remove ACL connection reference counting" into m-wireless-dev
Andre Eisenbach [Mon, 23 Mar 2015 20:12:53 +0000 (20:12 +0000)]
Merge "Remove ACL connection reference counting" into m-wireless-dev

9 years agoReset LE connecting flag on disconnect only if BDA matches
Nitin Arora [Fri, 6 Jun 2014 22:21:25 +0000 (15:21 -0700)]
Reset LE connecting flag on disconnect only if BDA matches

This change will make sure that the flag representing that BLE device
is connecting is made false only if a disconnection of the same
LE device is received, whose connection is in progress. This is ensured
by matching the bda of the disconnecting device and the LE device which
is in connecting state. This is to prevent a BR/EDR disconnection from
cancelling the connection timeout of an LE device

Change-Id: Ia5b47681479dc0be217cdfc3e15af9ca5323575c

9 years agoSend GATT discovery complete only if discovery is active
Nitin Arora [Tue, 18 Feb 2014 23:28:59 +0000 (15:28 -0800)]
Send GATT discovery complete only if discovery is active

This change forces a discovery complete event at gatt close
event only if service discovery is in progress and not when
inquiry is on going. Otherwise, a false discovery complete
is sent to the host, which cause a never ending inquiry.

Change-Id: I7ca207382b04ed233c149ca50040dab993067870

9 years agoExtend Bluetooth startup timeout from 3s to 8s
Andre Eisenbach [Mon, 23 Mar 2015 16:46:08 +0000 (09:46 -0700)]
Extend Bluetooth startup timeout from 3s to 8s

Change-Id: I0e8af541bf9c5837d92b5ccc047b3dd9d3765767

9 years agoRemove ACL connection reference counting
Andre Eisenbach [Thu, 19 Mar 2015 22:15:05 +0000 (15:15 -0700)]
Remove ACL connection reference counting

In addition to maintaining ACL link control blocks, which carry a
"in_use" flag, ACL links are reference counted in the num_acl links
variable.

The reference counting is thrown off when a SM connection initiated for
pairing is followed up by a GATT connection after pairing completes. The
2nd connection is counted against num_acl, even so other parts of the
code recognize the connection as a duplicate ACL connection to the same
BDA.

This patch removes the unecessary reference counting and relies on the
control block status instead, which is correctly maintained.

Bug: 19019189
Change-Id: I0a5f4d279fc634689d0e8c0ef4ad639d0ce5d687

9 years agoDemote, cleanup and extend observed logging
Chris Manton [Thu, 12 Mar 2015 16:22:48 +0000 (09:22 -0700)]
Demote, cleanup and extend observed logging

9 years agoCode formatting and language feature additions to the style guide.
Sharvil Nanavati [Thu, 12 Mar 2015 18:53:44 +0000 (11:53 -0700)]
Code formatting and language feature additions to the style guide.

9 years agoUse fully qualified path for btcore includes.
Sharvil Nanavati [Thu, 12 Mar 2015 22:55:21 +0000 (15:55 -0700)]
Use fully qualified path for btcore includes.

9 years agoUse fully qualified path for OSI includes.
Sharvil Nanavati [Thu, 12 Mar 2015 22:42:50 +0000 (15:42 -0700)]
Use fully qualified path for OSI includes.

9 years agoStart putting together a style guide for bluedroid.
Sharvil Nanavati [Wed, 11 Mar 2015 04:14:40 +0000 (21:14 -0700)]
Start putting together a style guide for bluedroid.

It's far from complete but provides a starting point for codifying
best practices and guidelines for developing in this codebase.

9 years agoAtomically update config files.
Sharvil Nanavati [Thu, 12 Mar 2015 06:50:42 +0000 (23:50 -0700)]
Atomically update config files.

There's a race condition on shutdown between Java and native code.
A stack shutdown is issued from Java and in the meantime the
calling thread calls System.exit(0). If that line is reached before
the native code has had a chance to shut down cleanly, the system
could be stuck in a bad state.

This change uses rename(2) to replace an existing config file
atomically instead of truncating, writing, and then flushing. In
case of a crash or the race condition above, the config file will
still be valid.

There's an implementation gotcha. The config saving code uses ".new"
as a suffix on the original filename to write to a temporary file.
If a file by that name already exists, it will be overwritten.

9 years agoAdd ability to set periodic alarms
Zach Johnson [Wed, 11 Mar 2015 08:55:27 +0000 (01:55 -0700)]
Add ability to set periodic alarms

Adds alarm_set_periodic so the alarm code can have more
contextual information when rescheduling alarms.

Problem: A2DP would stream for a few seconds and then
stop working.

Cause: The Java garbage collector. Bluedroid reaches out
to javaland to acquire and release the wake lock. Alarm was
always reaching out to get the wake lock when it scheduled a
short timeout. If GC kicked in during that call out to make
sure we have the wake lock, it could take more than 100ms to
get back us. That would screw over the alarm implementation
particularly for small 20ms timers.

So now if the wake lock was already acquired, we don't try to
reacquire it.

Cool. But we still have thrashing. Why? Because the alarm code
doesn't know the alarm is actually being used in a periodic way.

Here's what used to happen:

alarm expires
alarm is removed
reschedule
alarm callback is called
alarm callback sets the alarm again
alarm is added
reschedule

The problem is the first reschedule will get rid of the wake lock
if the next alarm is too far out or doesn't exist, meaning the next
reschedule needs to get the wake lock again.

With the extra periodicity information we can eliminate the
unnecessary intermediate reschedule, meaning no thrashing on the
wake lock. yay!

9 years agoAllocate a2d list properly
Chris Manton [Fri, 6 Mar 2015 22:41:32 +0000 (14:41 -0800)]
Allocate a2d list properly

Some other minor cosmetic improvements

9 years agoRemove dead code checks for BLE support
Zach Johnson [Thu, 5 Mar 2015 23:24:40 +0000 (15:24 -0800)]
Remove dead code checks for BLE support

On some build configurations (looking at you, hammerhead) the controller
hasn't come up before BTU init stuff happens.

These BLE support checks were triggering the assert that ensures the
controller module has loaded when you try to access it.

However, these BLE support checks were dead code because the outcome was
the same regardless in either case (returning from the function). Removed.

It's a good thing these asserts were in place, because if these checks
were actually important to the correctness of the code we'd be reading
bad state (BLE supported would always be false during this stage) and
we would have no idea. Yay asserts!

9 years agoRejig hci upwards dispatch
Zach Johnson [Wed, 4 Mar 2015 22:47:00 +0000 (14:47 -0800)]
Rejig hci upwards dispatch

Split events and acl data in hci dispatch, and dispatch events by event code

Future code will now be able to dynamically tie into hci events, instead
of relying on a hard coded routing.

9 years agoReinstate btsnoop last log saving
Zach Johnson [Wed, 4 Mar 2015 22:06:13 +0000 (14:06 -0800)]
Reinstate btsnoop last log saving

Waited for the merge to complete before reworking
the feature into the stack.

9 years agoPrevent Memory overbound and initialize variables
VenkatRaghavan VijayaRaghavan [Fri, 6 Feb 2015 06:20:39 +0000 (22:20 -0800)]
Prevent Memory overbound and initialize variables

1. Fix for initializing all the un-initialized variables
2. Potential strncpy overbound

Bug: 19417758

9 years agoResolved trace level issue for AVDTP
VenkatRaghavan VijayaRaghavan [Tue, 10 Feb 2015 15:35:10 +0000 (07:35 -0800)]
Resolved trace level issue for AVDTP

Root Cause: Trace messages are getting logged

When setting TRC_AVDT=5 the trace level doesn't take effect. Due to
multiple initialization points the bt_stack.conf value was overwritten

Bug: 19417758

9 years agoMove list function comments to the header file
Zach Johnson [Mon, 2 Mar 2015 22:42:02 +0000 (14:42 -0800)]
Move list function comments to the header file

9 years agoMove hashmap function comments to the header file
Zach Johnson [Mon, 2 Mar 2015 22:28:54 +0000 (14:28 -0800)]
Move hashmap function comments to the header file

9 years agoRemove more GKI code.
Sharvil Nanavati [Sun, 28 Dec 2014 10:25:02 +0000 (02:25 -0800)]
Remove more GKI code.

This change removes all remaining GKI task remnants as well as
a few timer and buffer bits.

9 years agoRemove dependency on GKI tasks from BTIF media task.
Sharvil Nanavati [Mon, 2 Mar 2015 23:34:21 +0000 (15:34 -0800)]
Remove dependency on GKI tasks from BTIF media task.

The previous code started a thread and posted a long-running function
to the thread's event loop. In essence, it created a thread and ran
a custom thread main that did its own event handling using GKI_wait.

The new code doesn't use GKI_wait or GKI_send_event at all. Instead,
it uses thread_post and the thread's reactor to enqueue and dispatch
events.

9 years agoDon't call GKI_send_event on BTU task since it's not a GKI task anymore.
Sharvil Nanavati [Mon, 2 Mar 2015 23:02:39 +0000 (15:02 -0800)]
Don't call GKI_send_event on BTU task since it's not a GKI task anymore.

9 years agoRemove GKI timers
Sharvil Nanavati [Mon, 2 Mar 2015 22:56:24 +0000 (14:56 -0800)]
Remove GKI timers

9 years agoAdding some documentation for the power management flow.
June R. Tate-Gans [Thu, 9 Oct 2014 21:16:25 +0000 (14:16 -0700)]
Adding some documentation for the power management flow.

9 years agoAdd sniff changed log statement on the mode changed event
Zach Johnson [Wed, 28 Jan 2015 19:16:15 +0000 (11:16 -0800)]
Add sniff changed log statement on the mode changed event

This will ensure we log accordingly when the other end
decides to put the connection into sniff mode.

9 years agoCollapse BNEP #defines that don't need to be configurable.
Sharvil Nanavati [Wed, 21 Jan 2015 07:12:31 +0000 (23:12 -0800)]
Collapse BNEP #defines that don't need to be configurable.

In some instances, defining these macros to FALSE would result in
broken runtime behavior (e.g. UUID fields containing uninitialized
values). This change collapses the following defines:

BNEP_SUPPORTS_PROT_FILTERS
BNEP_SUPPORTS_MULTI_FILTERS
BNEP_SUPPORTS_DEBUG_DUMP
BNEP_SUPPORTS_ALL_UUID_LENGTHS

9 years agoAdd a new l2cap_client class.
Sharvil Nanavati [Mon, 19 Jan 2015 18:03:08 +0000 (10:03 -0800)]
Add a new l2cap_client class.

The code to negotiate an L2CAP connection is currently scattered
and duplicated throughout the codebase. Each profile that uses
L2CAP has its own explicit or implicit state machine to go establish
a connection and go through the handshake. This class is intended
to consolidate that duplicated code and provide a usable API for
L2CAP client connections.

9 years agoAdd connection-specific function L2CA_SetConnectionCallbacks.
Sharvil Nanavati [Sun, 18 Jan 2015 06:12:38 +0000 (22:12 -0800)]
Add connection-specific function L2CA_SetConnectionCallbacks.

This function allows a client to specify callback routines per-
connection instead of per-PSM.

9 years agoproperty api unification, naming and testing
Chris Manton [Thu, 8 Jan 2015 18:47:59 +0000 (10:47 -0800)]
property api unification, naming and testing

9 years agoAdd a reference counted buffer implementation.
Sharvil Nanavati [Mon, 15 Dec 2014 09:37:59 +0000 (01:37 -0800)]
Add a reference counted buffer implementation.

This implementation introduces the notion of a 'slice'. A slice is
simply a sub-buffer of the original buffer. Slices refer to their
parent buffer so no memory is copied and slice creation becomes an
O(1) operation. Although I'm introducing a new concept, in practice,
a slice is indistinguishable from a buffer (they share a type and
all operations).

I expect slices to be used heavily during packet construction. For
example, an RFCOMM implementation would request a buffer from the
L2CAP layer (which may in turn request a buffer from the HCI layer).
The L2CAP layer would reserve some space in the buffer for its header
and return a slice to the RFCOMM layer.

9 years agoCollapse more feature flags.
Sharvil Nanavati [Mon, 29 Dec 2014 18:24:15 +0000 (10:24 -0800)]
Collapse more feature flags.

Start reviewing in include/bt_target.h.

9 years agoCollapse RFCOMM_INCLUDED and GAP_INCLUDED guards.
Sharvil Nanavati [Mon, 29 Dec 2014 06:29:42 +0000 (22:29 -0800)]
Collapse RFCOMM_INCLUDED and GAP_INCLUDED guards.

Both of these features are always available in bluedroid.

9 years agoDelete unused Broadcom-specific extensions API.
Sharvil Nanavati [Mon, 29 Dec 2014 12:51:27 +0000 (04:51 -0800)]
Delete unused Broadcom-specific extensions API.

9 years agoEliminate unusable wcassert code in favor of plain old C asssert.
Sharvil Nanavati [Mon, 29 Dec 2014 12:23:30 +0000 (04:23 -0800)]
Eliminate unusable wcassert code in favor of plain old C asssert.

The WC_* macros only expand to something meaningful when _DEBUG is
also defined. However, on enabling _DEBUG, bluedroid fails to build
because the wc_assert function is not defined anywhere. We can get
what I imagine is the equivalent behavior by switching over to the
standard C assert macro.

9 years agoDrop duplicate profile connection requests
Zach Johnson [Mon, 12 Jan 2015 21:56:46 +0000 (13:56 -0800)]
Drop duplicate profile connection requests

If profile connection requests come in too quickly
for us to handle, clients can effectively DoS the bluetooth
stack.

We used to see out of buffer errors because the requests
were batched up and sitting on buffers the rest of the stack
needed. With the switch to pure allocation we didn't see that
specific problem anymore, but we were still running into the
sanity assert.

Bug: 18323666