GitLab

am 977acabd: (-s ours) am eebc8ab9: DO NOT MERGE Change pairing_cb to assume temporary pairing by default

* commit '977acabd':
  DO NOT MERGE Change pairing_cb to assume temporary pairing by default

* commit 'eebc8ab9':
  DO NOT MERGE Change pairing_cb to assume temporary pairing by default

When pairing takes place, the pairing_cb.is_temp flag indicates whether
a pairing is temporary or permanent. Link keys are not stored for
temporary pairings. Since this is a "positive" flag, resetting the
pairing_cb control block (ex. memset to 0), it will assume persistent
pairing by default. Under certain circumstances, this can lead to a link
key being stored for temporarily secured connection.

This patch reverses the flag to be a "negative" flag. Renamed to
"persistent_bond", the default 0 meaning is now used to indicate a
temporary bond. If the lag is not properly set now, it will default to a
temporary bond and will not save the link key erronously.

Bug: 18345373
Change-Id: I06b1ba9331a70ebc29f4437bf836164658dec5ae

When pairing takes place, the pairing_cb.is_temp flag indicates whether
a pairing is temporary or permanent. Link keys are not stored for
temporary pairings. Since this is a "positive" flag, resetting the
pairing_cb control block (ex. memset to 0), it will assume persistent
pairing by default. Under certain circumstances, this can lead to a link
key being stored for temporarily secured connection.

This patch reverses the flag to be a "negative" flag. Renamed to
"persistent_bond", the default 0 meaning is now used to indicate a
temporary bond. If the lag is not properly set now, it will default to a
temporary bond and will not save the link key erronously.

Bug: 18345373
Change-Id: I8558abc9eace609b1ba02485b0ff775fb6f3c6af

If the encryption times out during the re-connect process for a HOGP
device, the device is removed from the whitelist and thus will not
reconnect until it is added back to the list. On platforms that do not
have a way to reconnect to a device manually, this means the device will
never reconnect again until the stack is restarted.

This patch checks the reason for the encryption failure and does not
remove the device from the whitelist if the encryption failed to a non
authentication/key related error.

Bug: 18233442
Change-Id: Ifaec4a9224ff9281956715b38d34c2d7c3fecb01

DO NOT MERGE Do not set sig_id in gerneral reject message per AVDTP SPEC 1.2 automerge: 595bcb4c  -s ours
automerge: 23699356  -s ours

* commit '23699356':

automerge: f81ffdb1

* commit 'f81ffdb1':

automerge: 8403c6d2

* commit '8403c6d2':

automerge: f9336d34  -s ours

* commit 'f9336d34':

automerge: 0e74efe5

* commit '0e74efe5':

automerge: 595bcb4c  -s ours

* commit '595bcb4c':
  DO NOT MERGE Do not set sig_id in gerneral reject message per AVDTP SPEC 1.2

automerge: 98ad505f

* commit '98ad505f':
  Increment seq before adding to packet in avdt_scb_hdl_write_req_no_frag

With new link management of controller buffers between
different ACL connections, it is possible that a ACL
conn would get 0 buffers - thus no profile can run.
This affects classic and LE ACL connections.

This work around allows a floor of 1 buffer as the
floor for all connection, irrespective of controller
capability.

Bug: 17976311
Change-Id: I7cd0d34af47c441c5ea7f2b836b591a4fec62ad1

If the encryption fails during the re-connection process with a
previously paired LE device, the pairing is removed. Thus, a single
connection failure could lead to the device be unpaired.

This patch prevents unpairing when encryption fails.

Additionally, the IRK is restored from NV ram to fix reconnection to
dual-mode devices.

Bug: 18072300
Change-Id: I077a16d0715fdaa07b95b1c664c18bd839754896

HID reports from a bonded device can be missed when Bluetooth is
restarted HOGP report discovery is still active. Caching the
report information and loading it back when the stack resets
will ensure incoming HID reports can be processed immediately.

Bug: 17999991
Change-Id: I4608935f8749537d6b05625b894445a21f844ee0

Some car kits (notably Renesas car kits used in Nissan/Infiniti models)
will throttle incoming A2DP frames if too many frames are sent at once.
This patch lowers the maximum number of frames sent per tick to not
overwhelm the remote device.

Bug: 17880920
Change-Id: I7bfbba625405c96159d1c94eebd8dc498f100daa

Implements a "blacklist" identifying troublesome remote devices which
may require multiple pairing attempts to succeed.

Change-Id: I11b6b7da7230e37fbd5ec4c07cd38b30dc12b40e

When trying to reconnect a previously paired device, notifications may
be sent by the remote device before the characteristic handles have been
discovered. Thus notifications will be lost.

This patch adds a non-volatile attribute cache that allows attribute
handles to be restored before notifications are received.

Bug: 17999991
Change-Id: I97faefbc6a2fed86cbce7f64d620ed03944d89b2

* commit '595bcb4c':
  DO NOT MERGE Do not set sig_id in gerneral reject message per AVDTP SPEC 1.2

AVDTP SPEC 1.3 set sig_id in gerneral reject message. But current stack
is on AVDTP 1.2, we need set the field to be 0.
Setting sig_id would fail PTS test TP/SIG/SMG/BI-28-C

Bug: 17892440
Change-Id: I2037aae299eab15864fc7b53ada7f49c0f34b9c5

* commit '98ad505f':
  Increment seq before adding to packet in avdt_scb_hdl_write_req_no_frag

Bluetooth controllers do not allow multiple connection parameter
updates where there is one outstanding. This requires serializing
all connection parameter update requests and send the latest
request when the pending one is finished.

Bug: 17979666
Change-Id: Ic49bd89a4ded28014d74b40e67f2a3ea490b9ef7

Bluetooth Certification Test case TP/TRA/BTR/BV-01-C for Basic
Transport Service Support fails. The failure is because the PTS
is expecting the first RTP Media Frame to have a sequence number
of 1. But metallica sends it as 0.
bug 17893538

Change-Id: Idfe7ae382c3764616dab4733cdf83b1627b8b3c9

Running GKI timer engine via JAVA alarm service was causing severe
timer preemption at times which affects A2DP media task timer. This in
turn was stressing the compensation logic towards UIPC resulting in
very bursty reads of up to 40 UIPC frames at a time. This led to
instabilities on the UIPC pipe and trickled upwards in audio framework.
Inclusion of java alarm thread in A2DP high priority thread list results
in a much smoother data flow for entire A2DP data path and improved
audio quality.

Bug: 17520043
Change-Id: I7cea6ec64aa5b973b5fff767389a8827e681619e

This reverts commit a934f018.

This patch caused various headset audio break issues and slow
reconnections on headsets like the Bose AE2W.

Based on the original bug under which this patch was merged
(b/10648781) it is unlikely that this bug even fixed the
issue it was meant to address.

Bug: 17520043
Change-Id: I39bb9abbf5735abff25b5257425c34972287dc27

* commit 'ba2f6d99':
  Revert "anthias:17614185:Bluetooth: Getbuf returning invalid buffer pointer"

* commit 'ba2f6d99':
  Revert "anthias:17614185:Bluetooth: Getbuf returning invalid buffer pointer"

This reverts commit c1ccb0d9.

This change looks like a likely candidate for the culprit behind
a wide range of BT issues we're seeing across many devices.
Reverting this in order to test.

* commit '46bfe63d':
  DO NOT MERGE AVRCP: Proper handling of Fragmented AVCTP packet

* commit 'c1ccb0d9':
  anthias:17614185:Bluetooth: Getbuf returning invalid buffer pointer

The latest discovered address is not used when reconnecting to devices
using resolvable private addresses, if RPA resolution offloadin is not
supported in the controller.
This patch also address the exception case where BT off is memseting
all control value while the IRK list is still being cleaned up.

Bug: 17187661
Change-Id: Ifeb5b8297e7eadb0e34ec1afbd41f46bdd836cf3
Author: Chaojing Sun <cjsun@broadcom.com>

* Ensure all a2dp audiopath threads are configured to high
  priority when actively streaming.
* For now set bt hc worker thread always to URGENT_AUDIO
  until new thread api can adjust thread priority dynamically.

Bug: 17520043
Change-Id: I10e314085329278bcfcc4e8bc5ce899b03d22e2b

* commit '46bfe63d':
  DO NOT MERGE AVRCP: Proper handling of Fragmented AVCTP packet

- AVCTP : Allocate bigger buffer for reassembly
- AVRCP : Handle getelimentattr to eliminate duplicates
          and invalid attributes, this also fixes BOF.

Bug: 17669579
CRs-Fixed: 576831
Change-Id: I778c3df1386339ff8a00a755ffc9f24881b5167d