OSDN Git Service
Marcel Holtmann [Fri, 27 Jun 2014 10:48:23 +0000 (12:48 +0200)]
doc: Add management command to get clock information
Luiz Augusto von Dentz [Fri, 27 Jun 2014 10:51:46 +0000 (13:51 +0300)]
unit/test-queue: Fix memory leak
This fixes the following memory leak:
32 bytes in 1 blocks are definitely lost in loss record 20 of 35
at 0x4C291D4: calloc (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
by 0x400DC2: queue_new (queue.c:65)
by 0x400AA5: test_foreach_remove_all (test-queue.c:92)
by 0x4E9E5E0: ??? (in /usr/lib64/libglib-2.0.so.0.3800.2)
by 0x4E9E7A5: ??? (in /usr/lib64/libglib-2.0.so.0.3800.2)
by 0x4E9EB1A: g_test_run_suite (in /usr/lib64/libglib-2.0.so.0.3800.2)
by 0x40083E: main (test-queue.c:109)
Luiz Augusto von Dentz [Fri, 27 Jun 2014 10:27:22 +0000 (13:27 +0300)]
shared/queue: Fix invalid read
If the queue elements are destroyed by queue_destroy the head will point
to freed memory causing the following error when unit/test-queue is run:
Invalid read of size 8
at 0x401040: queue_foreach (queue.c:194)
by 0x4E9E5E0: ??? (in /usr/lib64/libglib-2.0.so.0.3800.2)
by 0x4E9E7A5: ??? (in /usr/lib64/libglib-2.0.so.0.3800.2)
by 0x4E9EB1A: g_test_run_suite (in /usr/lib64/libglib-2.0.so.0.3800.2)
by 0x40083E: main (test-queue.c:109)
Address 0x7f65738 is 8 bytes inside a block of size 16 free'd
at 0x4C28577: free (in /usr/lib64/valgrind/vgpreload_memcheck-amd64-linux.so)
by 0x400E29: queue_destroy (queue.c:93)
by 0x40102C: queue_foreach (queue.c:219)
by 0x4E9E5E0: ??? (in /usr/lib64/libglib-2.0.so.0.3800.2)
by 0x4E9E7A5: ??? (in /usr/lib64/libglib-2.0.so.0.3800.2)
by 0x4E9EB1A: g_test_run_suite (in /usr/lib64/libglib-2.0.so.0.3800.2)
by 0x40083E: main (test-queue.c:109)
Marcel Holtmann [Fri, 27 Jun 2014 09:44:39 +0000 (11:44 +0200)]
doc: Add address type information to get connection info command
Jakub Tyszkowski [Thu, 26 Jun 2014 13:09:11 +0000 (15:09 +0200)]
android/pts: Update MCAP results
Jakub Tyszkowski [Thu, 26 Jun 2014 13:09:10 +0000 (15:09 +0200)]
android/mcaptest: Support MDL reconnection if MCL reconnected
This is needed in some PTS test cases.
Jakub Tyszkowski [Thu, 26 Jun 2014 13:09:09 +0000 (15:09 +0200)]
android/mcaptest: Support MDL and MCL disconnection
This adds support for MDL disconnection triggered with timeout after the
MDL is connected. It also adds MCL disconnection timeout after the MDL
has been disconnected. If less than 0 is passed as '-e' or '-f' option
parameter, disconnection is not triggered. This is needed to pass some
PTS test cases.
Jakub Tyszkowski [Thu, 26 Jun 2014 13:09:08 +0000 (15:09 +0200)]
android/mcaptest: Support incoming connections
This adds support for handling incoming Control Link (MCL) and Data
Link (MDL) connections.
Jakub Tyszkowski [Thu, 26 Jun 2014 13:09:07 +0000 (15:09 +0200)]
android/mcaptest: Support connect
This initiates the connection
Jakub Tyszkowski [Thu, 26 Jun 2014 13:09:06 +0000 (15:09 +0200)]
android/mcaptest: Add skeleton for mcap test tool
This adds skeleton for simple tool for testing mcap library. As mcap-lib
uses bt_io, this tool works in daemon-like mode, by utilizing Glib's main
loop.
Jakub Tyszkowski [Thu, 26 Jun 2014 13:09:05 +0000 (15:09 +0200)]
android/pts: Change MCAP PIXIT value
Handling single data channel will simplify mcaptest tool.
This can be changed later when mcaptest will be improved.
Luiz Augusto von Dentz [Thu, 26 Jun 2014 10:14:07 +0000 (13:14 +0300)]
core: Fix bogus check to resume pending authorizations
After incoming pairing the code will attempt reverse service discovery
and queue any authorization request until the services are resolved, once
this process was completed the code should resume the authorization
process no matter how many are queued.
Szymon Janc [Thu, 26 Jun 2014 09:12:28 +0000 (11:12 +0200)]
android/gatt: Fix not sending notification on get included service
If there was and error notification with GATT error status should be
send.
Szymon Janc [Thu, 26 Jun 2014 08:19:27 +0000 (10:19 +0200)]
android/gatt: Fix not sending notification if no included services
Ravi kumar Veeramally [Thu, 26 Jun 2014 09:04:48 +0000 (12:04 +0300)]
android/health: Provide a helper funtion to search channel by id
Ravi kumar Veeramally [Wed, 25 Jun 2014 16:30:15 +0000 (19:30 +0300)]
android/hal-health: Return app_id and channel_id -1 in case of error
Right now returning only status of request but java layer expecting
-1 for app_id and channel_id in failure case.
Ravi kumar Veeramally [Wed, 25 Jun 2014 16:30:13 +0000 (19:30 +0300)]
android/health: Rename matching functions
Petri Gynther [Wed, 25 Jun 2014 17:38:29 +0000 (10:38 -0700)]
core: Fix adapter_connect_list_add()
adapter_connect_list_add() needs to ensure that trigger_passive_scanning() is
called even for existing devices on the connect list.
This issue can be seen with BLE HoG remotes trying to reconnect to BlueZ. If
the connection fails, att_error_cb() adds the failed device back to connect
list, but since the device is already on the list, no code ends up calling
trigger_passive_scanning(). After this point, passive scanning is disabled
and remains disabled for good, and BLE reconnects to BlueZ no longer work.
bluetoothd[931]: src/adapter.c:device_found_callback() hci0 addr ...
bluetoothd[931]: src/device.c:device_set_legacy() legacy 0
bluetoothd[931]: src/adapter.c:stop_passive_scanning()
bluetoothd[931]: src/adapter.c:stop_passive_scanning_complete() status 0x00 (Success)
bluetoothd[931]: src/device.c:device_connect_le() Connection attempt to: ...
bluetoothd[931]: src/adapter.c:discovering_callback() hci0 type 6 discovering 0
bluetoothd[931]: src/adapter.c:connect_failed_callback() hci0 ... status 2
bluetoothd[931]: plugins/policy.c:conn_fail_cb() status 2
bluetoothd[931]: src/adapter.c:bonding_attempt_complete() hci0 bdaddr ... type 1 status 0x2
bluetoothd[931]: src/device.c:device_bonding_complete() bonding (nil) status 0x02
bluetoothd[931]: src/device.c:device_bonding_failed() status 2
bluetoothd[931]: src/adapter.c:resume_discovery()
bluetoothd[931]: src/device.c:att_connect_cb() connect error: Transport endpoint is not connected (134)
bluetoothd[931]: src/device.c:att_error_cb() Enabling automatic connections
bluetoothd[931]: src/adapter.c:adapter_connect_list_add() ignoring already added device ...
... passive scan is disabled and remains disabled for good ...
Petri Gynther [Wed, 25 Jun 2014 17:38:17 +0000 (10:38 -0700)]
core: Fix discovering_callback()
discovering_callback() should not call trigger_passive_scanning() when
outgoing BLE connection is pending.
This issue can be seen with BLE HoG devices trying to reconnect to BlueZ:
bluetoothd[931]: src/adapter.c:device_found_callback() hci0 addr ...
bluetoothd[931]: src/device.c:device_set_legacy() legacy 0
bluetoothd[931]: src/device.c:device_set_rssi() rssi -75
bluetoothd[931]: src/adapter.c:stop_passive_scanning()
bluetoothd[931]: src/adapter.c:discovering_callback() hci0 type 6 discovering 0
bluetoothd[931]: src/adapter.c:trigger_passive_scanning()
bluetoothd[931]: src/adapter.c:stop_passive_scanning_complete() status 0x0b (Rejected)
bluetoothd[931]: Stopping passive scanning failed: Rejected
Petri Gynther [Wed, 25 Jun 2014 17:38:03 +0000 (10:38 -0700)]
core: Fix stop_passive_scanning_complete()
Consider MGMT_STATUS_REJECTED as success for stop_passive_scanning().
MGMT_STATUS_REJECTED may be returned from kernel because the passive
scan timer had expired in kernel and passive scan was disabled just
around the time we called stop_passive_scanning().
This issue can be seen with BLE HoG devices trying to reconnect to BlueZ:
bluetoothd[931]: src/adapter.c:device_found_callback() hci0 addr ...
bluetoothd[931]: src/device.c:device_set_legacy() legacy 0
bluetoothd[931]: src/device.c:device_set_rssi() rssi -75
bluetoothd[931]: src/adapter.c:stop_passive_scanning()
bluetoothd[931]: src/adapter.c:discovering_callback() hci0 type 6 discovering 0
bluetoothd[931]: src/adapter.c:trigger_passive_scanning()
bluetoothd[931]: src/adapter.c:stop_passive_scanning_complete() status 0x0b (Rejected)
bluetoothd[931]: Stopping passive scanning failed: Rejected
Luiz Augusto von Dentz [Wed, 25 Jun 2014 11:49:07 +0000 (14:49 +0300)]
android/hog: Fix using invalid range for external report
Andrei Emeltchenko [Thu, 26 Jun 2014 07:14:52 +0000 (10:14 +0300)]
android/health: Fix setting wrong PSM for SDP record
Luiz Augusto von Dentz [Wed, 25 Jun 2014 09:09:26 +0000 (12:09 +0300)]
android/hog: Fix code style
Luiz Augusto von Dentz [Wed, 25 Jun 2014 09:09:25 +0000 (12:09 +0300)]
android/hidhost: Add hid_device_new helper function
Andrei Emeltchenko [Tue, 24 Jun 2014 12:12:02 +0000 (15:12 +0300)]
android/health: Add setting mcl callbacks for incoming connection
For incoming connection NULL is passed as argument since application is
not known yet.
Andrei Emeltchenko [Tue, 24 Jun 2014 12:12:01 +0000 (15:12 +0300)]
android/health: Refactor create_device
create_device() should not search for apps again since we have searched
already, and we do not have always app_id.
Andrei Emeltchenko [Tue, 24 Jun 2014 12:12:00 +0000 (15:12 +0300)]
android/health: Add actual error message to error print
Andrei Emeltchenko [Tue, 24 Jun 2014 12:11:59 +0000 (15:11 +0300)]
android/health: Refactor setting mcl callbacks to function
Function will be used in other places as well.
Andrei Emeltchenko [Tue, 24 Jun 2014 12:11:58 +0000 (15:11 +0300)]
android/health: trivial: Fix missing space
Andrei Emeltchenko [Tue, 24 Jun 2014 12:11:57 +0000 (15:11 +0300)]
android/health: Remove extra check
gerr is always set on error condition.
Andrei Emeltchenko [Tue, 24 Jun 2014 12:11:56 +0000 (15:11 +0300)]
android/health: trivial: Remove empty line
Andrei Emeltchenko [Tue, 24 Jun 2014 12:11:55 +0000 (15:11 +0300)]
android/health: Fix missing error check
Andrei Emeltchenko [Tue, 24 Jun 2014 12:11:54 +0000 (15:11 +0300)]
android/health: Fix possible overflow
Andrei Emeltchenko [Tue, 24 Jun 2014 12:11:53 +0000 (15:11 +0300)]
android/health: Remove unneeded check
In case of error err is always set so the check is not needed.
Andrei Emeltchenko [Tue, 24 Jun 2014 12:11:52 +0000 (15:11 +0300)]
android/hal-health: trivial: Correct coding style
Andrei Emeltchenko [Thu, 19 Jun 2014 10:28:21 +0000 (13:28 +0300)]
android/gatt: Remove dead code
The code apparently left after wrong rebase as it duplicates code
several lines above.
Andrei Emeltchenko [Thu, 19 Jun 2014 10:36:19 +0000 (13:36 +0300)]
android/gatt: Fix wrong size argument
Instead of sizeof of (uint8_t *) pass real length, though for that
function it doesn't matter at this point.
Szymon Janc [Wed, 25 Jun 2014 14:51:54 +0000 (16:51 +0200)]
android/cts: Update tests results with CTS 4.4 R3
No LE GATT tests were run yet.
Andrei Emeltchenko [Tue, 24 Jun 2014 12:07:46 +0000 (15:07 +0300)]
HDP: Fix NULL check
Improves readability of the code using similar check for not NULL.
Andrei Emeltchenko [Tue, 24 Jun 2014 12:07:45 +0000 (15:07 +0300)]
HDP: trivial: Remove empty lines
Andrei Emeltchenko [Tue, 24 Jun 2014 12:07:44 +0000 (15:07 +0300)]
HDP: Fix possible memory leak
Andrei Emeltchenko [Tue, 24 Jun 2014 12:07:43 +0000 (15:07 +0300)]
mcap: Fix possible overflow
Andrei Emeltchenko [Tue, 24 Jun 2014 12:07:42 +0000 (15:07 +0300)]
HDP: Fix checking always constant error code
Function sdp_set_add_access_protos() always returns 0, so there is no
sense to check for error code.
Andrei Emeltchenko [Tue, 24 Jun 2014 12:07:41 +0000 (15:07 +0300)]
HDP: Fix checking always constant error code
sdp_set_access_protos() always returns 0, there is no sense to check for
error code. Fixes compile warnings.
Andrei Emeltchenko [Tue, 24 Jun 2014 12:07:40 +0000 (15:07 +0300)]
HDP: Remove extra debug statements
DBG() includes function name by default, so including it is not needed.
Szymon Janc [Mon, 23 Jun 2014 14:28:34 +0000 (16:28 +0200)]
android/bluetooth: Log error if loading LTKs or IRKs failed
Szymon Janc [Mon, 23 Jun 2014 14:28:33 +0000 (16:28 +0200)]
android/gatt: Use proper address when connecting
If device was using RPA but address was resolved IDA must be used.
Szymon Janc [Mon, 23 Jun 2014 14:28:32 +0000 (16:28 +0200)]
android/bluetooth: Map RPA to Identity address on New IRK event
This patch allows to pair with devices using RPA. There is no API
to notify HAL about new Identity Address. When New IRK event is
received ID address is stored in device and map to RPA is done.
This allows to use RPA after address was resolved making this
transparent to Android Framework. When daemon is restarted device
is advertised to Framework with ID address.
Szymon Janc [Mon, 23 Jun 2014 14:28:31 +0000 (16:28 +0200)]
android/bluetooth: Add get_device_android_addr helper
This will be used to fake address for Android Framework.
Szymon Janc [Mon, 23 Jun 2014 14:28:30 +0000 (16:28 +0200)]
android/bluetooth: Fix passing address type to SSP replies
Szymon Janc [Mon, 23 Jun 2014 14:28:29 +0000 (16:28 +0200)]
android/bluetooth: Pass device to send_bond_state_change function
Szymon Janc [Mon, 23 Jun 2014 14:28:28 +0000 (16:28 +0200)]
android/bluetooth: Add helper for finding device with android address
Grzegorz Kolodziejczyk [Mon, 23 Jun 2014 14:18:17 +0000 (16:18 +0200)]
android/hidhost: Fix adding hidhost device on devices list
Create hog connection cb functionality tried to operate on stored on
list hidhost device. Adding hidhost device to list should be done
before operating on it.
It caused sending bogus connecting state change.
Ravi kumar Veeramally [Mon, 23 Jun 2014 15:06:04 +0000 (18:06 +0300)]
android/health: Update PTS results
Update test results for TC_SRC_CC_BV_03_C, TC_SRC_CC_BV_07_C,
TC_SNK_CC_BV_04_C, TC_SNK_HCT_BV_01_I, TC_SNK_CC_BV_08_C and
TC_SRC_HCT_BV_01_I.
Ravi kumar Veeramally [Mon, 23 Jun 2014 15:06:03 +0000 (18:06 +0300)]
android/client: Write dummy data on fd on health source connect
PTS expects dummy data on fd when it connects in source role.
Ravi kumar Veeramally [Mon, 23 Jun 2014 15:06:02 +0000 (18:06 +0300)]
android/client: Fix coding style
Ravi kumar Veeramally [Mon, 23 Jun 2014 15:06:01 +0000 (18:06 +0300)]
android/health: Fix error messages
Ravi kumar Veeramally [Mon, 23 Jun 2014 15:06:00 +0000 (18:06 +0300)]
android/health: Fix wrong mdep config index
MDEP Id starts from 1, (0 value for Echo). HAL api sends mdep cfg index
which starts from 0.
Luiz Augusto von Dentz [Mon, 23 Jun 2014 12:46:17 +0000 (15:46 +0300)]
android/pics: Do not set Suspend feature in HOGP
This requires integration with Android power management so leave it out
for now.
Luiz Augusto von Dentz [Mon, 23 Jun 2014 12:46:16 +0000 (15:46 +0300)]
android/pics: Fix having both types of discover set in HOGP
Currently the code discover by service handle and not by UUID.
Szymon Janc [Mon, 23 Jun 2014 13:24:09 +0000 (15:24 +0200)]
android/client: Fix closing wrong FD in health HAL
Andrei Emeltchenko [Mon, 23 Jun 2014 11:19:35 +0000 (14:19 +0300)]
core: Fix using address of array instead of size
The dev->name variable is an array, so the address of it can never be
NULL.
Ravi kumar Veeramally [Sun, 22 Jun 2014 11:59:01 +0000 (14:59 +0300)]
anrdroid/client/health: Cache fd and close on channel disconnection
Cache fd and close them on channel disconnect or destroy notification.
When running PTS tests it is expecting to close all data channels before
exiting test case.
Ravi kumar Veeramally [Sun, 22 Jun 2014 11:59:00 +0000 (14:59 +0300)]
android/health: Implement mcap_mdl_deleted_cb
mcap_mdl_deleted_cb will be called if remote device sends MDL_DELETE_REQ.
Free channel data and notify channel is destroyed.
Ravi kumar Veeramally [Sun, 22 Jun 2014 11:58:59 +0000 (14:58 +0300)]
android/health: Create and connect MDL
Request for md_create_mdl and on successful response
connect mdl and pass fd. First data channel should be reliable
data channel.
Ravi kumar Veeramally [Sun, 22 Jun 2014 11:58:58 +0000 (14:58 +0300)]
android/health: Notify channel status on channel destroy call
Ravi kumar Veeramally [Sun, 22 Jun 2014 11:58:57 +0000 (14:58 +0300)]
android/health: Cache remote mdep id on channel connect request
Remote mdep is required to initiate MD_CREATE_MDL_REQ request.
Ravi kumar Veeramally [Sun, 22 Jun 2014 11:58:56 +0000 (14:58 +0300)]
android/health: Check if device and channel already exists or not
On channel connect request, check if device is already exists or not.
Also check if channel is already created for remote device or not.
Marcel Holtmann [Mon, 23 Jun 2014 08:37:11 +0000 (10:37 +0200)]
AUTHORS: Mention Arman's contributions
Marcel Holtmann [Mon, 23 Jun 2014 08:36:23 +0000 (10:36 +0200)]
shared/att: Fix a few compiler warnings
Arman Uguray [Wed, 18 Jun 2014 20:25:43 +0000 (13:25 -0700)]
shared/att: Handle incoming response PDUs
This patch adds handling for incoming ATT protocol response PDUs. Basic
incoming PDU handling logic added to the io read handler.
Arman Uguray [Wed, 18 Jun 2014 20:25:42 +0000 (13:25 -0700)]
shared/att: Implement write handler and bt_att_send.
This patch implements the write handler logic, including the way send
operations are process from the various internal queues. Added PDU
encoding for the Exchange MTU request.
Arman Uguray [Wed, 18 Jun 2014 20:25:41 +0000 (13:25 -0700)]
shared/att: Implement basic boilerplate.
This patch implements the getters, setters, creation, ref, and unref
functions for struct bt_att. Also added is a simple table for
determining the ATT op type given an opcode and the io read handler that
currently does nothing.
Arman Uguray [Wed, 18 Jun 2014 20:25:40 +0000 (13:25 -0700)]
shared/att: Introduce struct bt_att.
This patch introduces struct bt_att, which handles the transport and
encoding/decoding for the ATT protocol. The structure of the code
follows that of src/shared/mgmt and lib/mgmt.h, where individual
parameter structures are defined for all ATT protocol requests, responses,
commands, indications, and notifications. The serialization and
endianness conversion for all parameters are handled by bt_att.
struct bt_att is based around struct io and operates on a raw file
descriptor.
Andrei Emeltchenko [Thu, 19 Jun 2014 13:55:35 +0000 (16:55 +0300)]
hciconfig: Fix adding missing break
Andrei Emeltchenko [Thu, 19 Jun 2014 13:55:34 +0000 (16:55 +0300)]
hcitool: Fix adding missing break
Andrei Emeltchenko [Thu, 19 Jun 2014 13:55:33 +0000 (16:55 +0300)]
bnep: Fix redundant check
Andrei Emeltchenko [Thu, 19 Jun 2014 13:55:32 +0000 (16:55 +0300)]
monitor: Fix checking boolean return value
Since btsnoop_read_hci() is used from src/shared.
Andrei Emeltchenko [Thu, 19 Jun 2014 13:55:31 +0000 (16:55 +0300)]
tools/csr: Fix wrong error check
Andrei Emeltchenko [Thu, 19 Jun 2014 13:55:30 +0000 (16:55 +0300)]
queue: Fix unreachable code
I assume passing function == NULL means using direct_match() so make it
reachable.
Andrei Emeltchenko [Thu, 19 Jun 2014 13:55:29 +0000 (16:55 +0300)]
tools/mpris-player: Fix overflow before type widening
Expression is evaluated using 32-bit arithmetic before conversion to
64-bit.
Andrei Emeltchenko [Thu, 19 Jun 2014 13:55:28 +0000 (16:55 +0300)]
android/ipc-tester: Fix possible double close
I case of error we may close fd and pipe twice.
Marcel Holtmann [Mon, 23 Jun 2014 06:56:15 +0000 (08:56 +0200)]
doc: Add comment for single mode controllers without public address
Johan Hedberg [Mon, 23 Jun 2014 05:05:13 +0000 (08:05 +0300)]
doc/mgmt-api: Fix a couple of typos
Marcel Holtmann [Sun, 22 Jun 2014 20:05:51 +0000 (22:05 +0200)]
doc: Add extra mode for a reduced LE privacy mode
Marcel Holtmann [Sun, 22 Jun 2014 20:00:42 +0000 (22:00 +0200)]
doc: Add management support for handling connection parameters
Marcel Holtmann [Sat, 21 Jun 2014 02:21:08 +0000 (04:21 +0200)]
doc: Add mode for enabling generation of debug keys
Johan Hedberg [Sun, 22 Jun 2014 07:22:05 +0000 (10:22 +0300)]
client: Wait for D-Bus before processing stdin
This should make it possible to do some scripting with bluetoothctl.
Otherwise stdin is processed before we're connected to D-Bus and most
commands will fail as a consequence.
Ravi kumar Veeramally [Fri, 20 Jun 2014 12:23:30 +0000 (15:23 +0300)]
android/health: Fix queue creation for mdeps and devices
Create queue for mdeps, devices and channels when creating app
and device struct. It is simpler to read code than on demand
creation.
Marcel Holtmann [Fri, 20 Jun 2014 18:32:08 +0000 (20:32 +0200)]
Release 5.20
Marcel Holtmann [Fri, 20 Jun 2014 18:26:51 +0000 (20:26 +0200)]
build: Update library version
Lukasz Rymanowski [Thu, 19 Jun 2014 22:50:34 +0000 (00:50 +0200)]
android/pts: Update GAP PTS status
After enabling privacy for peripheral one more test gets enabled in PTS
Lukasz Rymanowski [Thu, 19 Jun 2014 22:50:33 +0000 (00:50 +0200)]
android/pics: Enable privacy for peripheral
Szymon Janc [Fri, 20 Jun 2014 14:33:34 +0000 (16:33 +0200)]
android/bluetooth: Fix loading device address type from storage
Address type is stored as "AddressType" key. Also this is valid only
for LE capable devices.
Szymon Janc [Fri, 20 Jun 2014 15:22:15 +0000 (17:22 +0200)]
android/bluetooth: Pass device to send_device_property function
Szymon Janc [Wed, 18 Jun 2014 17:50:11 +0000 (19:50 +0200)]
android/socket: Convert service UUID to 16 or 32 bits if possible
Jakub Tyszkowski [Wed, 18 Jun 2014 15:29:46 +0000 (17:29 +0200)]
android/pts: Update GAP results
Jakub Tyszkowski [Wed, 18 Jun 2014 14:09:21 +0000 (16:09 +0200)]
android/pts: Update GAP results
Jakub Tyszkowski [Wed, 18 Jun 2014 12:58:54 +0000 (14:58 +0200)]
android/gatt: Minor coding style fix
Jakub Tyszkowski [Wed, 18 Jun 2014 12:58:53 +0000 (14:58 +0200)]
android/gatt: Fix not freeing discover_srvc_data
This fixes the following issue:
32 bytes in 1 blocks are definitely lost in loss record 92 of 142
==19365== at 0x4C29DB4: calloc (in
/usr/lib/valgrind/vgpreload_memcheck-amd64-linux.so)
==19365== by 0x43275A: search_dev_for_srvc (gatt.c:1302)
==19365== by 0x4354C9: connect_cb (gatt.c:1424)
==19365== by 0x448AD3: accept_cb (btio.c:203)
==19365== by 0x4E79D12: g_main_context_dispatch (in
/lib/x86_64-linux-gnu/libglib-2.0.so.0.3200.4)
==19365== by 0x4E7A05F: ??? (in
/lib/x86_64-linux-gnu/libglib-2.0.so.0.3200.4)
==19365== by 0x4E7A459: g_main_loop_run (in
/lib/x86_64-linux-gnu/libglib-2.0.so.0.3200.4)
==19365== by 0x40459D: main (main.c:538)