- 13 Oct, 2016 1 commit
-
-
Treehugger Robot authored
-
- 12 Oct, 2016 5 commits
-
-
Pavlin Radoslavov authored
Removed file execute permissions for: - bta/include/bta_hh_co.h - embdrv/sbc/decoder/Android.mk - embdrv/sbc/Android.mk - embdrv/Android.mk - btif/co/bta_ag_co.cc Test: visual inspection of file permissions Change-Id: I746debb1e8ceaa1cc14f867641b25cb42c04c7be
-
Marie Janssen authored
Test: mma in system/bt Change-Id: I357eb44a97f3935262252cf97c13cef078a95e4d
-
Marie Janssen authored
-
Pavlin Radoslavov authored
Test: code compilation Change-Id: Ib61126d7921d088b045dca6c5b69a5379603fcc3
-
Marie Janssen authored
Test: run unit tests / sanity connection to devices Change-Id: I293c3600affd229fea67cdd6624eba7f186cbcb7
-
- 11 Oct, 2016 3 commits
-
-
Stephen Li authored
automerger BUG:32069275 Change-Id: I60cc50b21fe0d5a5464bd3acc5897ba7c4e05b73
-
Marie Janssen authored
Test: compiles Change-Id: I34f68c2b166a28622155358ff75478af8d18bed0
-
Ajay Panicker authored
Test: Manually tested Bug: 32060762 Change-Id: Ia6ca97c119d421aef1136f97afccb3f0c126334d
-
- 10 Oct, 2016 12 commits
-
-
George Burgess IV authored
- sprintf() does not limit the length of the character string when writing to a buffer and may result in buffer overflow - snprintf() requires the maximum write length as a parameter. When the maximum length supported is smaller than the reserved buffer length, the call will not result in buffer overflow Bug: 31859081 Test: TestTracker/64195/3975 Change-Id: I519f8ef7b9b162fd79094f89148250d783c734c0
-
Treehugger Robot authored
-
Andre Eisenbach authored
Firmware fixes are available for affected headsets (Thanks, Jabra!!). Change-Id: I60b42366d1e8155895cc58fff69f6404415a0406
-
Jakub Pawlowski authored
Bug: 30622771 Test: all related tests were updated Change-Id: I9695284ae249fa427ae4b2f3e9649da951102c6c
-
Jakub Pawlowski authored
The return value of BTM_VendorSpecificCommand is currently just an information wether the callback was passed to it or not, but many places through stack use this value as error indication. Remove it. Test: unit tests pass Change-Id: Idf73fc49fb25e52c3aeb64f2305b4498ecacd9a4
-
Treehugger Robot authored
-
Andre Eisenbach authored
Reverting for now after talking to Jacky because this causes circular dependencies as lower layer (stack/) functions are calling higher layer APIs (btif/) directly. To restore a change like this, the data should be collected at a lower layer in the stack and reported through polling or a callback. This reverts commit 8ec8ca4a. Change-Id: I748843864ae7198ea021fe70d2643a62097ab029
-
Treehugger Robot authored
-
Jakub Pawlowski authored
The return value of BTM_VendorSpecificCommand is currently just an information wether the callback was passed to it or not, but many places through stack use this value as error indication. Remove it. Test: unit tests pass Change-Id: Idf73fc49fb25e52c3aeb64f2305b4498ecacd9a4
-
Jakub Pawlowski authored
btsnd_hcic_* functions are always returning true, so this return value is not really useful. Test: unit tests pass Change-Id: I06ced65392d180ceed565e45b7831de6aede283c
-
Jakub Pawlowski authored
btsnd_hcic_* functions return only true. This patch makes sure that no logic is waiting for "false" value. Few methods, that had their return value dependent on result of btsnd_hcic_* were made void. Test: unit tests pass Change-Id: I43cff62f461ff35f4bd84e5338f1a587b6a05c58
-
Joseph Pirozzo authored
Bug: 32054308 Test: manual, UPF 55 Change-Id: Ib89f7557fa390d60691afe8399a768e8127e52cb (cherry picked from commit a1fa1bb0d638d5f444bf959fb839e824d69de31d)
-
- 07 Oct, 2016 3 commits
-
-
Treehugger Robot authored
-
Ben YoungTae Kim authored
Incorrect size check for UHID_OUTPUT_EV for incoming data from UHID driver of kernel was leading to exiting of poll thread. This was causing IOT issues with Apple Magic Mouse that require output report to be sent for proper functioning of scroll functionality. Bug: 28942565 Change-Id: I2eaba1f70b7cd51fb21933031d439543eb141fd2
-
Gurpreet Ghai authored
Use case: BT will get crash in SNS testing. steps: SNS testing BT ON/OFF. Failure: Crash is observed. Root Cause: If the number of supported services increased, larger service record array is required. Smaller array is resulting in out of record error causing L2CAP deregister. Fix: Check added to avoid null pointer exception if null is returned while fetching sock from id. Also cleanup of SDP and security records done during BT turning Off. Change-Id: I2e77127262119cabf2c372f894427dcdf3350e80
-
- 06 Oct, 2016 1 commit
-
-
Jakub Pawlowski authored
Currently we have some unnecessary complicated code for handling advertising data. Get rid of it. Bug: 30622771 Test: Tests making sure flags and tx power are filled are added Change-Id: I807cdf11577c357c116c54cab50bbf3261e1ee16
-
- 05 Oct, 2016 3 commits
-
-
Pavlin Radoslavov authored
Use the (new) A2D_IsPeerSinkCodecValid() API call as appropriate to check whether the codec information of a Sink peer is valid. Previously, A2D_IsSourceCodecSupported() was used, and that didn't match the original code (before the refactoring). A2D_IsSourceCodecSupported() has extra checks, including min/max bitpool oundaries, and those shouldn't be used for this initial check - the bitpool boundaries can be adjusted later. Similarly, use the new A2D_IsPeerSourceCodecValid() API call instead of A2D_IsSinkCodecSupported(). Also: * Replaced A2D_IsValidCodec() with A2D_IsSourceCodecValid(), A2D_IsSinkCodecValid() A2D_IsPeerSourceCodecValid(), A2D_IsPeerSinkCodecValid() and added the appropriate unit tests. * Added extra debug messages to help identify similar issues in the future Bug: 31749230 Test: manual test with a carkit, and unit tests Change-Id: Iafaeb82744df9758e686194c91624992a0c55bdf
-
Jakub Pawlowski authored
Change-Id: I6e68f80d0d97081a4fc3879e6c11342923cfef7b
-
Jakub Pawlowski authored
This patch turns the code that manages LE multi-advertising into the BleAdvertisingManager class. All HCI calls made by this code were abstracted into BleAdvertiserHciInterface. Thanks to those changes, we can mock BleAdvertiserHciInterface and test the advertising manager. The BleAdvertisingManagerTest, still has some dependencies on other code from the "stack" module. These have to be implemented on top of the test file. In the future, this dependencies should be replaced by mockable classes. Test: Tests covering the change are added in this patch. Bug: 30622771 Change-Id: I4a363678815c235ee01255f462db2a671ef89610
-
- 04 Oct, 2016 3 commits
-
-
Myles Watson authored
cd vendor_libs/test_vendor_lib/ clang-format ../linux/bt_vendor_linux.c -style=file -i clang-format include/* test/* src/* -style=file -i Update the style guide since clang-format help uses the -style=file option. Test: mm -j32 Change-Id: I166adfb2bb9a87fefa1242c91e5f180011caa2ad
-
Mark Salyzyn authored
Test: compile Bug: 26552300 Bug: 31289077 Change-Id: I8fb392150441ca5b6bc43dc43a10a3009e22202b
-
Jakub Pawlowski authored
-
- 03 Oct, 2016 2 commits
-
-
Mark Salyzyn authored
Actually, remove reference, android logging is not used here. Test: compile Bug: 26552300 Bug: 31289077 Change-Id: I37731dc37f61c129f3440e5d9ff1d0e71fe193d0
-
Jakub Pawlowski authored
If the GATT service cache was successfully loaded from a file, the state must be reset to idle. Bug: 31175159 Change-Id: I92ebf1bec9d2e7467d7412b41923614ec5a13b6d
-
- 30 Sep, 2016 7 commits
-
-
Jorge E. Moreira authored
Moved all controller actions out of read callbacks to task callbacks to avoid race conditions when modifying internal controller state. Change-Id: Ie8bb762992755db2dddb6b7bb6a4fe52ab2d0020
-
Jack He authored
Bug: 31274290 Test: Manual Change-Id: I5a4bff5e25b08a32709cdc81bac86a641d80f113
-
Treehugger Robot authored
-
Treehugger Robot authored
-
Kim Schulz authored
if an LE device is found during an LE scan, the BT host will try to fetch its name (creating an LE connection) even though the advertising report received indicates that it is a non-connectable advertising. This is pointless and should not happen. Change-Id: I246e4795b4f4e33f3202908ebff3724e8e89523c
-
Patrick Tsai authored
Change-Id: I482f1096e87966f33ba8584ea60188abab3bcc42
-
Sebastien Griffoul authored
If the string argument of send_at_cmd is a null pointer, then the function should not crash: indeed this is a valid parameters which can be used to send an AT cmd which doesn't require any extra parameter. Test: trivial fix. Existing unit tests still pass. Change-Id: I3d83eebed660d703cd5a93a19dc73f4f354ab7bf Signed-off-by:
Sebastien Griffoul <sebastien.griffoul@intel.com>
-