- 01 Sep, 2016 1 commit
-
-
Jakub Pawlowski authored
This patch fixes crashes on reconnection to HID device. All those issues are already fixed in AOSP. Bug: 31221398 Change-Id: I43b64d0b11b1f548cb373332777ea92d8a075466
-
- 29 Aug, 2016 8 commits
-
-
Andre Eisenbach authored
am: 88cc389d Change-Id: I6c325fa710adfefbbe3e448ebc28a0777253d5d3
-
Andre Eisenbach authored
am: f7d632a2 Change-Id: I6cf3f7c2579e0ef395142aa3165c1e608e7e2fbc
-
Andre Eisenbach authored
am: fb309b3c Change-Id: Ib7494cc47ff0e93f7f20a12dc79695697d2eecf0
-
Andre Eisenbach authored
am: d2a37eff Change-Id: I87c0ba767d7dcf91cfd6cb4903fea4efd6df392b
-
TreeHugger Robot authored
-
- 23 Aug, 2016 4 commits
-
-
Jakub Pawlowski authored
am: 970d9d18 Change-Id: Iaff1375e619409afe9570b6fb79867070a339902
-
Jakub Pawlowski authored
am: aab757d3 Change-Id: Id9397aa59aae3570821c9838d428541cfb6a878a
-
Jakub Pawlowski authored
This patch creates new connection mode - opportunistic connection. When such connection is made, no call to gatt_update_app_hold_link_status will be made when it's started. This means that connecting and disconnecting in this mode won't trigger disconnect timer. When other, app creates regular connection and then disconnects, the physical connection might get disconnected. Opportunistic connection is used only for code setting CCC right now. Bug: 30186455 Change-Id: Ia5be7682b6c3dcb4993060f818dee603aef3e19e
-
Jakub Pawlowski authored
Right now, in GATT_Deregister we make a call to gatt_update_app_use_link_flag, which might result in a call to GATT_SetIddleTimeout on links that this client was never connected to. In order to fix that, we must check if gatt_update_app_hold_link_status made any changes to p_tcb->app_hold_link. If not, then skip updating the timer. Bug: 30186455 Change-Id: I6f18ec7fe4e80e048adc9e360cae703b7d64d62a
-
- 19 Aug, 2016 1 commit
-
-
TreeHugger Robot authored
-
- 18 Aug, 2016 2 commits
-
-
Jakub Pawlowski authored
Right now, in GATT_Deregister we make a call to gatt_update_app_use_link_flag, which might result in a call to GATT_SetIddleTimeout on links that this client was never connected to. In order to fix that, we must check if gatt_update_app_hold_link_status made any changes to p_tcb->app_hold_link. If not, then skip updating the timer. Change-Id: I6f18ec7fe4e80e048adc9e360cae703b7d64d62a
-
Jakub Pawlowski authored
This patch creates new connection mode - opportunistic connection. When such connection is made, no call to gatt_update_app_hold_link_status will be made when it's started. This means that connecting and disconnecting in this mode won't trigger disconnect timer. When other, app creates regular connection and then disconnects, the physical connection might get disconnected. Opportunistic connection is used only for code setting CCC right now. Bug: 30186455 Change-Id: Ia5be7682b6c3dcb4993060f818dee603aef3e19e
-
- 09 Aug, 2016 1 commit
-
-
Satomi authored
This reverts commit 5f0e23d5 . Instead of reusing service discovery data, gatt profile initiates discovery by service uuid before regular service discovery is started, and both discovery procedures interfere with each other. As a result, link is dropped. Bug: 30462943 Bug: 30481143 Change-Id: Ic63339f4f46eda70e575a4103974766d54a9d624 Signed-off-by:
Satomi <michitsuta@casio.co.jp>
-
- 05 Aug, 2016 1 commit
-
-
Jakub Pawlowski authored
When read RSSI timeouts, callback is called with NULL result. This was causing crash. Bug: 30562545 Change-Id: If379c2de4ceab48c021601851a8c83f995440460
-
- 28 Jul, 2016 2 commits
- 27 Jul, 2016 1 commit
-
-
Ayan Ghosh authored
Use case: 1. Pair to Remote 2. Turn OFF/ON BT on DUT 3. After DUT's BT is turned ON, Remote would reconnect to DUT Failure: Bluetooth settings UI continues to display Connecting when BT was Turned OFF and ON. Root cause: - This is a connection collision case where remote is not responding to DUT initiated SDP attribute fetch request for AV. AV holds outgoing connection for sometime and meanwhile remote starts AV connection from its end. Then DUT disconnects the SDP channel after a second without waiting to fetch requested info, and this confuses DUT's AV state machine. The outgoing connection could not proceed as incoming AV L2CAP connects, because DUT's AV state machine is not equipped with handling this scenario. On incoming connection timer expiry, it does not start AV media channel as it finds AV Open has not been done from upper layer after incoming L2CAP got connected. - When the incoming path 2s timer expires, the DUT checks whether Open has been called after incoming state is set. This is not the case here, because Open was triggered earlier as as part of the outgoing connection. As a result, A2DP cannot connect at all. Fix: - Properly set collision flags for handling AV connection collision. - Update BTA AV handle even when BTIF AV state machine is yet to be built. Bug: 30362987 Change-Id: I02a3adb62479b0f762bc792a5727d06e11eaaa52 (cherry picked from commit db1e3a5fbb41cd1faf572c5f89ec9212c5ff2986)
-
- 25 Jul, 2016 1 commit
-
-
Tom Turney authored
This reverts commit 1e7afc82. Revert Reason: This CL causes regressions b/30367872, b/30368041, b/30368763. Change-Id: I48da5caac1e9fe32462c236385f768b896fa6542
-
- 24 Jul, 2016 1 commit
-
-
Andre Eisenbach authored
The various position dependent lookup tables in the AT command parser were out of sync, causing invalid responses to the AT+CBPS command for example. This patch gets rid of positionally dependent enums for simple lookup tables that correlate all the values for easier, less error prone maintenance of the related tables. Change-Id: I84f7d4f4a803e6e52bf2021c6c74917e90bbc846 Fixes: 29978908
-
- 19 Jul, 2016 1 commit
-
-
ugo_yu authored
Precondition: 1. Pair with specific HOGP mouse. Steps: 1. Pair with HOGP mouse (iBuffalo BSMBB09DS) 2. Unpair the mouse. 3. Pair the mouse again. Failure: Pairing is successful, but the mouse won't connect. Root Cause: Some HOGP mice send notifications before the stack writes the client characteristic configuration descriptor. The Bluetooth stack doesn't deregister these notifiactions when disconnecting. Fix: Deregister notifications when the GATT connection closes. Fixes: 29111689 Change-Id: I520faeaa5fc396381cb248d3a2c49efbffa12261
-
- 16 Jul, 2016 1 commit
-
-
Andre Eisenbach authored
Bug: 29615817 Change-Id: I5256b31a746433f10c2e7de1629952f668c99d16
-
- 15 Jul, 2016 1 commit
-
-
Subramanian Srinivasan authored
The Gatt server multi response queue is made NULL after each Gatt server response is sent by BT stack in response to the remote Gatt client's request. But the multi response queue is not initialized again to process subsequent read multiple Gatt server responses. Hence, the BT stack does not send Gatt server responses for the subsequent read multiple requests from remote device which inturn leads to disconnection. This change would process read multiple responses by initializing the queue when it is null. Some of Bluetooth certification test cases involve testing Gatt server's capability to process and send read multiple response. Test case: TC_GAR_SR_BV_05_C This change also cleans up unnecessary initialization and freeing of multi rsp queue. This change makes sure that the stack initializes multi rsp queue only when it is necessary(ie only when the Gatt multiple read response is being processed). Bug: 29011255 Change-Id: Ia3e2e1569ee16cac9c518f95501945257b9fb7fd (cherry picked from commit 3737592a7846766083041e81d72a2fab0b27fb38)
-
- 12 Jul, 2016 1 commit
-
-
Ajay Panicker authored
Change-Id: I02b213913685d831bfe00e1938baef953428e048
-
- 11 Jul, 2016 1 commit
-
-
Jacky Cheung authored
On BTA_DM_AUTH_SMP_UNKNOWN_ERR and BTA_DM_AUTH_SMP_CONN_TOUT, bond state status will be notified as BT_STATUS_AUTH_FAILURE. This is observed during the pairing flow with iOS devices. Plus minor cleanup on SMP errors and fail reasons. Bug: 29779689 Change-Id: Ia32fe6bb84b07272d5159409f65217a06a2e5f8e
-
- 07 Jul, 2016 2 commits
-
-
Andre Eisenbach authored
Fixes: 30004157 Change-Id: Id170856fbee5ea345f02f55d3a6e3376d846c2de
-
Andre Eisenbach authored
Fixes: 30004157 Change-Id: Id170856fbee5ea345f02f55d3a6e3376d846c2de (cherry picked from commit e8b81851c33095fe3b3ac78329073309be5421e1)
-
- 06 Jul, 2016 1 commit
-
-
Jacky Cheung authored
These events are used to track HCI timeout and configuration file usage respectively. Change-Id: I5c8b6ef20b61f03b9ccb74f3b1c9784e4653b770
-
- 02 Jul, 2016 1 commit
-
-
Jacky Cheung authored
Delay authorization complete during Bond Pending state in SMP. Instead of automatically advancing the Bond Pending state, create a 500ms window to allow the slave to stop pairing by sending over the Pairing Failed command during the Bond Pending state. BUG: 28475887 Change-Id: Id6c30247d15258cd18eb34827694b6cd79ca753a
-
- 29 Jun, 2016 1 commit
-
-
Andre Eisenbach authored
To add another datapoint and potentially narrow down *why* we had to flush the buffers, add RSSI logging to keep track of signal strength during audio drop-out events. Change-Id: Ie52f56290951130165a9f375a898a9bacf459d55
-
- 27 Jun, 2016 2 commits
-
-
Andre Eisenbach authored
am: d1387007 Change-Id: Iff3e3643f4cc36c093e5a7d0d47af8e3aa79fe9c
-
Andre Eisenbach authored
Bug: 29157087 Change-Id: If227963b939a9b0ed55e6d900364c71932d3cfe8
-
- 24 Jun, 2016 5 commits
-
-
Andre Eisenbach authored
am: d881c4b4 Change-Id: Ie00432fcec7e869c212a04374d6fcd58094eb643
-
Andre Eisenbach authored
Currently, our TX queue does not allow for quick recovery of minor congestion. As we only clear as much room in the queue as is needed to fix the news frames needed for a given timer tick, any minor congestion on the TX path can result in single frames to be dropped over an excessively long amount of time. From the "just rip the band-aid off" school of thought, drop the whole queue instead to cause a single drop-out on congestion and catch up to the live stream immediately, hoping that the stream recovers... Bug: 29601962 Change-Id: I482c3cfef6d8b51143e54733eb2da5f70663636f
-
Andre Eisenbach authored
am: 6b142ed4 Change-Id: I7dd99e2af5ec4b09a9cba9830df9cf0c677ca46b
-
Andre Eisenbach authored
am: ffdb33b5 Change-Id: I1df2268b7ddd28220f69d4eeeaee30097a976e7c
-
Andre Eisenbach authored
-