[Description]
If bt controller device is usb interface,
sco loopback single pkt and bandwidth test can not pass
because pkt size is limited accroding to core spec
[Root Cause]
USB sco packet size is limited in spec
[Solution]
If is bt usb chip, continuous receive sco packets
from controller until size is enough
[Test Report]
Test Pass
Bug: 286355871
Change-Id: Idb150d8a72149f7e1dfaccfd7bc82a91710da59c
This reverts commit 1adfae9364.
Reason for revert: Causes Sabrina/Boreal bt stacks to timeout on startup
Bug: 247053367
Change-Id: Ibef97f054510c826920bb30725fed168c755cc14
[Description]
If aidl open HCI fail at the first time, the aidl state
can not change to enable for the next time to initialize.
[Root Cause]
Bluetooth AIDL State is always equal to INITIALIZING after
open hci fail at the first time.
[Solution]
When open HCI fail at the first time, initialize the mstate to ready
Bug: 285517442
Test: pass
Change-Id: I19756d6602ba931242d0f95942db55c505a210bb
This reduces the code size by not generating unnecessary code for
non-root elements.
Bug: 283700992
Test: m
Change-Id: I23bd5c2a0ccc40fded3ca08155ead5f4680c2ed9
The "provider" is managed with a shared_ptr but we do not hold it and
instead are giving the raw inner pointer as binderDiedCallbackAidl.
This can randomly generate crash as the provider may be freed outside of
this code.
Replacing the provider with a context that we can manually allocate and
deallocate.
Setup AIBinder_DeathRecipient_setOnUnlinked to clean the data allocated
Bug: 245009140
Test: m android.hardware.bluetooth.audio-impl and start / stop session +
manually kill bluetooth process during audio play
Change-Id: I0c14c062a8bde7e532ff02f01991d66da33ec569
Call SetTerminalRaw on the fd that was opened, since mFD has not been set. If SetTerminalRaw fails, log it but continue, since some platforms do not need this call.
Bug: 279870989
Test: start Bluetooth on a device that doesn't need SetTerminalRaw
Change-Id: I97c92c407b7c80442a0ba0812a630525219670a0
Issue found from the test H4ProtocolAsyncTest#TestMultiplePackets.
The packetizer does not correctly handle a packet termination
at the beginning of a chunk.
Test: atest --host bluetooth-vendor-interface-hci-test
Bug: 276652301
Change-Id: I1b61efcfa5f817fdb8b5870e81d660be5279fdea
The requirements have been added to RootCanal:
- LE Periodic Advertising
- LL Privacy
- LE 2M Phy
Test: atest
Bug: 275847929
Change-Id: Ifa49494091727bf162554a658629bc4ec3c3cd63
The default HAL only uses one thread, which is 0 more than
the default of 1.
Bug: 205758693
Test: VtsHalBluetoothTargetTest
Change-Id: I8828cab2b4eb4d2a28ecc41cb75e3eb470cd0da6
* Enforce that LMP version must be greater than or equal to HCI version
* Enforce that minimum advertising set is 16 when HCI version is 5.0 and
above
* Enforce Link Layer Security, 2M Phy, Coded PHY requirements at BT 5
* Enforce minimum 8 resolving list entries at BT 5
Bug: 263257831
Test: vts
Change-Id: I53a6713563c0b9a5ba27eecf2a69caff0a57f30b
AptxAdaptive LE/LEX codec and AAC ABR AIDL side changes
Bug: 254567319
Tag: #feature
Test: m android.hardware.bluetooth.audio-freeze-api &&
m android.hardware.bluetooth.audio-update-api &&
atest VtsHalBluetoothAudioTargetTest
Change-Id: I74a2df64ba6ef2ae02e08850b0f6c7784b24b02b
Added two new latency modes for Audio HAL to enable Dynamic Spatial
audio with LE audio, and the path through which the head tracking data
has to be routed.
Bug: 268568163
Test: m android.hardware.bluetooth.audio-update-api
Tag: #feature
Change-Id: I92b529e4eee3c862934d46164117450975e017ef
This field is used to indicate the PCM stream source for given stream
handle
Tag: #feature
Bug: 264692800
Test: m android.hardware.bluetooth.audio-update-api
Change-Id: I3df9066e98ca552329bb58ae5fbc5346d1021b10
Merged-In: I3df9066e98ca552329bb58ae5fbc5346d1021b10
This test isn't a performance test. In rare cases it
takes longer than 100ms (usually around 5ms).
Bug: 268408355
Test: bluetooth-btlinux-hci-test
Change-Id: I6d2f23eb05541bc0f9676cbfd8b097340a131e09
Sending a reset and discarding extra packets
helps ensure that tests are hermetic for devices
that do not power off the chip with rfkill.
Add a mutex to make sure that close only gets
called once.
Bug: 266221125
Test: atest VtsHalBluetoothTargetTest
Change-Id: Ifb259a675202aa5399a8c7570ba8c4df2083e10e
