Change dts notifier nodes path to /data/misc/dts/ to
satisfy seandroid policy.
Also remove AudioUtil source and header file from this project
as writing to route node is already moved to AudioPolicyManager.
Change-Id: I656e28077a6224b25e26d7476401bfb11fd141c8
During pause/resume, don't send DTS HPX fade-out and fade-in
to all active streams, we should only send to the one that is
pausing or resuming.
Change-Id: Ib527e0be29e81a38f702152b1263a140de773145
CRs-fixed: 780001
- On resuming the playback after headset removal, hal uses the
stream out device which is not updated one to get the snd out
device for sending acdb cal data.So it sends the wrong acdb id.
- Use the snd device associated with usecase to get the acdb id.
Change-Id: I4057b21e27e5d4836b4e8954cf71bf88c235ae7f
- Add support to use low lantency output as primary
output when AUDIO_USE_LL_AS_PRIMARY_OUTPUT is enabled
Change-Id: Ie7faf9f7af4c97b7afdda8084b7acf162abbc0b6
Written PCM frames count is not increased when HAL stays in SSR,
it can make mismatch with the written framecount in AudioFlinger,
so rendered frames count return from AudioFlinger is not correct.
Then A/V sync will be missed since the played audio duration is
calculated by rendered frames count.
Increase the written frames count even the frames are not written
to the PCM driver during SSR.
CRs-Fixed: 791738
Change-Id: I53cfeae9294ae1f5a6466a749da90d1f8ef297cb
Since small bufs uses blocking writes, a write will be blocked for
the default max poll time (20s) in the event of an SSR.Reduce the poll
time to observe and deal with SSR faster.
Change-Id: I8693a55596bc9578f9b5a04469d12155f191dcf1
- By default CLOCK_REALTIME is used.
- Change the clock type to CLOCK_MONOTONIC
to synchronize the framework and VOIP driver
clock and timestamps.
Change-Id: I8ab64e948cea5d027bf28c4e101ff517fda3558c
Offload latency is set to 96. Change
it to 50 to reflect the rendering latency more correctly
Modify pcm offload latency to return offload
latency rather than the latency based on buffering
since volume is applied immediately for offload use
cases. Also reduce log spam in offload use cases
Change-Id: If28d49d66e8f989074c4c7eaec523855e9f4290a
CRs-Fixed: 787448
Thulium target has support for HDMI. Enabled compilation of
file for HDMI EDID support on Thulium target.
Change-Id: I605f4c57a14ff9fcfa064a9ebda996cef349f91b
Add support for more audio devices for RX arbitration,
including speaker, headphones and the combo
speaker/headphones device.
Change-Id: I822fa0b70033ab067f50f7a019a21ef2e300a665
Add new rx and tx devices in order to route voice call to
afe proxy ports from where the data can be picked real time
using proxy playback and record usecases.
Change-Id: I0a26e4644588885934f36558312cbd846c38084f
Currently, ALIGN macro can align properly only if the buffer size
is a power of 2.
To have same PCM samples for all channels, the buffer size requires to
be multiple of (number of channels * bytes per sample).
For writes to succeed, the buffer must be written at address which is
multiple of 32.
Alignment of 96 satsfies both of the above requirements.
Change-Id: I5ab33ef198b5b8f18ae9f09c4361e5cff84e700e
CRs-Fixed: 795936
When compress offload playback is active, if the Visualizer effect
is enabled, decoded PCM audio is routed to AFE Proxy port and
read from it by the Visualizer wrapper library. When audio is routed
to proxy port, current output device specific calibration is also
being applied which is not desired.
Avoid this by sending default audio calibration i.e. no post-processing
to be applied the data.
Change-Id: Id576c4ed7bbb482683074e3e33aa5760b7597d37
PCM offload bit width is selected as 16 bit even for 24 bit pcm offload
playback streams if QTI_FLAC_DECODER flag is not defined. Fix this issue
by defining appropriate flag PCM_OFFLOAD_ENABLED_24.
Change-Id: I82169252770b764344bf4531fb994e2195244dde
Incall recording usecase should select non-fluence
device in audio path to avoid recorded data from
getting cancelled in audio path when fluence is
enabled.
Fix by adding missing break for the case statements.
Change-Id: I3832013423956683f403c8322d0547325c336c73
While sound trigger lab data being read, if CPE SSR
happens the pcm read returns -ENETRESET. Instead of
updating the sound card sate, mark the sound trigger
state as inactive to avoid reading the lab data from
driver. Marking the session as inactive handle both
CPE and ADSP SSR scenarios where HAL is prevented
to read from kernel.
Change-Id: Id82eed597c01265a7d26f17e5ac3c5cd0e17adbf
CRs-fixed: 790726
Mingming Yin
Linux Build Service Account
Dhanalakshmi Siddani
Preetam Singh Ranawat
Naresh Tanniru
Zhou Song
Weiyin Jiang
Banajit Goswami
Fred Oh
Pavan Chikkala
Alexy Joseph
Ravit Dennis
Amit Shekhar
Ravi Kumar Alamanda
Vidyakumar Athota
Shiv Maliyappanahalli
bait_dispatcher_monitor_system
Chaithanya Krishna Bacharaju
Bharath Ramachandramurthy