- When hdmi device supports multi channel PCM formats,
  framework picks compress passthrough profile as
  playback path. Avoid it by forcing routing to
  deep buffer based on system property and format type.

Change-Id: I7cca7f60a38e0d300037b0cae76d9f2c9bb0f4a1

Remove AAudio MMAP/NOIRQ properties from qssi product
config as these are vendor specific properties already defined in
target specific audio config files.
Enabing here makes it a default product property where
AAudio MMAP/NOIRQ data path is not supported for certain QSSI targets
across AOSP upgrades.

Change-Id: I35dd160c5c9657d8e7c5bb3b8417afd3fb9a75a3

Change-Id: I73c3ce823f4979b46ef145f52d3e49e5f78acdc4

This reverts commit 3664e304.

Change-Id: I92342c44a5907a80a7b56e10461643952af46490

replace PRODUCT_PROPERTY_OVERRIDES with
PRODUCT_PRODUCT_PROPERTIES. system side
PRODUCT_PROPERTY_OVERRIDES variable is not
effecting the properties into the build.prop file.

Also, remove all vendor properties from here.

Change-Id: Ib6744c19e76c58ca4a74a1240bfcf0e60b4c58e4

-Add HDR pre proc libraries
-Add HDR pre proc config files

Change-Id: I1c44ec7f12006be7bd7be88893ddba16ec708b79

as part of audio self-containment, created a
audio_system_product.mk file where audio related 
product system defs can be configured in this file.

the softlink for this will be created in
vendor/qcom/defs/product-defs/system/ .

Change-Id: Ief3a0a4544511ed3106c333fe4cdb7eb105a0138

Change-Id: I23e876c5936680ea256c857b6b421c6daaaec6df

Sound Trigger detection is not supported on USB headset, so
the capture streams should not start on USB headset. This
change prevents USB headset device selection for Sound
Trigger capture streams.

Change-Id: I238ab26a7a9d206d6b2f8e92b35be74b051b31ed

When offload output starts with FLAG_DIRECT, then setup a
voip call, VOIP_RX flag will be overwritten to FLAG_NONE.
This causes voip usecase isn't set for voip call.

Fix by adding VOIP_RX flag check to avoid settting 
FLAG_NONE for voip call.

Change-Id: Iee89d53d7529e6235b31903376b30e3a23657e3f

Stereo output is started with setting voip_rx flag since
voip_rx profile supports Stereo config.

Change-Id: Ib79102cef1f5e4af4cfa8efb106c3297d6f9f8ed

Update custom policy for SCO volume against below AOSP update.

audiopolicy: keep BT SCO vol at 0 dB on VoIP

The BT SCO volume of VoIP is set as the value computed with the volume
index, while that of voice call keeps constant (0 dB). It causes the
output sound of VoIP being multiplied by a gain computed with the
volume index, then the sound hears different if the volume index has
been changed before. Therefore, the volume of VoIP on BT SCO should keep
at 0 dB as well.

Bug:78206701
commit: da7581eca3523a21142a3d259531c938f8dc717c
Change-Id: I888b2f7c0869902dd5e0adf77f711e7b9fd7e108

CRs-Fixed: 2703382
Change-Id: Ie4f8bd1ac5defdadf7501ba84026daa46cda842f

AudioPolicyManager: Ensure direct output matches exact format
Otherwise AudioFlinger will fail on createTrack.

Change-Id: I8ad8e9a800465b9b167df205f488f1a48b27e527

Restructure Voip flag override logic to avoid doing any override when
device used is AUDIO_DEVICE_OUT_TELEPHONY_TX.
Overriding flag for this device leads to incorrect output selection
for uplink voice usecase resulting in muted uplink data.

Change-Id: I361a321633b7d53fd1ee8334c723471e374530fd

Don't check for reuse and don't close previous output if multiple
concurrent offload decode is supported and new offload request comes.

Change-Id: I8f428ba3e6dfa4783589e9da997bdf3e35e64ae0

Keep the logic to be the same as default policy manager.
This is to address volume unexpected applied for in call
music delivery.

Change-Id: Id46fd45457015bf9432532ff7aba5f67b29b6206

Guard system properties with odm flag.

Change-Id: I748006a53906203664c74a91eedaaf916866fc0d

During voice call, when headset remove, new device of primary
output is first set to handset, and voice volume is also
calculated in handset. Later other outputs loop to select new
device and caculate volume in new device. Due to inactive output
selectes AUDIO_DEVICE_NONE, it uses previous device and set wrong
voice volume.

Fix it by add check to avoid applying volume for inactive output.

Change-Id: Icc0dd9e1f00cafd6548d20c678c882e770effc0d

- add dumpsys support.
- usage dumpsys media.audio_policy

Change-Id: Ic3a80d0d06278f83fd30b103b69bec9f13f5bba9

During voice call, when device switch from speaker to headset,
first new device of primary output is set to headset, voice
volume is calculated in headset, and then it is applied for
voice call correctly. Later, other outputs loop to select new
device and caculate volume in new device. Due to inactive output
selectes AUIOD_DEVICE_NONE, it causes wrong voice volume is
calculated to set.

Fix it by add check to avoid applying volume for inactive output.

Change-Id: I092d870fe5e32c324c953e595cfbc5fce5696321

When selecting output for haptics, channelmask is needed.

Change-Id: Id58cd2f90615bfb008a3eebbaace12328c41e945

enable 3D audio feature for qssi.

Change-Id: Idbe942398f8aae1d6665f21ad92a8bf09c789b88

libnbaio is not installed for system/lib/ for HY11
builds.
Add to PRODUCT_PACKAGES explicitly to acheive this.

Disable audio hal extention feature flag to disable config store
which is not required now.

Change-Id: Ie06d18961e17ee8125e9b7016fbe43094d994395

The AudioPatch API changed upstream to make the handle a private member
variable.
https://keystone-qcom.googlesource.com/platform/frameworks/av/+/afd4cea46c12f55a7ccd47271314d59d833aa5c0

Bug: 149114520
Change-Id: Ic9092b003b35ed51d640b9349c4dfa2dc3980d79

Change-Id: I2959eff9d3177be881b19e7bbc24746827240d9f

libsoundtrigger was removed upstream and a new middleware service is now
in use. Reflect those changes here.

Bug: 147608992
Change-Id: I2bdc5711a3430b9fd125582570e042da1696d6e4