- Widevine sends ADTS data, set ADTS based on the parameter
- Set offload fragment size 2k for all video streaming
Change-Id: I4c2376408e2998354db583a9f54649476a880f01
str_parms_to_str() requires the caller to free() the
returned string. Free the memory pointed to by "str"
before overwriting the "str" with the new pointer which
is returned by strdup() function.
Change-Id: I602ae37cce684e0c62b0ec310080bb0cd675e35a
CRs-fixed: 608384
- Modem & Audio uses same DSP which will limit
Audio usecases during call mode
- Route all audio output sessions to ULL path
during in call mode
Route all non music sessions to ULL path if WFD
session is active(to disable multiple post processing on ADSP)
Change-Id: I0c2d124cf0cef6ff2064405795dc2f16eddbe540
Due to changes in machine driver pcm device id has changed.
Change updates pcm device id for vifeedback usecase.
Change-Id: I75323f80b07fda9cf21d9aef09e4849d767c11b9
CRs-fixed: 602026
Compress voip input requires input source type to be of
AUDIO_SOURCE_VOICE_COMMUNICATION. Move the opening of compress
voip input to in_set_parameters() function which is the first
place in HAL where the source type information is available.
Change-Id: I3758eba6049ccd4f1a4465ef264420d9afb6a795
Move platform_parser to root hal directory. Rename
platform_parser to platform_info. Change name of
XML file read from platform_info.xml to
audio_platform_info.xml. The xml now only needs
information for ACDB ID's that you want overwritten.
Names in the XML now match sound device enums in
platform.c.
Change-Id: I6f9bfcf4149f08d88cb2a95ba021bdf7296aaf4d
Audio HAL set corresponding preset id into adsp, but doesn't disabe
reverb explicitly when 'None' preset is chosen. However, preset=0
means custom preset in adsp, so adsp won't stop reverb processing.
Change-Id: I0c42dd13c09579a6bd3bef2141b857477ec65f2e
CRs-Fixed: 604982
- When strategy is STRATEGY_ENFORCED_AUDIBLE, do not select
unsupported combo devices Proxy+Speaker, HDMI+Speaker,
and FM_TX + Speaker.
Change-Id: Icd4f6de74807a6efc5f6eb8258c020ad990aaf9b
Crs-fixed: 600799
-During switch to tunnel playback, offload effects-flags remain
disabled. This stops effects being applied on tunnel playback.
-Send effects enable params to DSP to start effects during switch.
Change-Id: Ibe1e50b41ca25b51537e4a0ad80dbd5184379133
CRs-Fixed: 593509
-Add check_and_set_gapless_mode and get_offload_buffer_size functions
at the begining of hal file.
Change-Id: I134dffac4aa81ed52fc6a2668cb53fbda60103ca
-Deep buffer period count in the driver is 8. This adds to latency
of 160ms. This also results in the common block to have 8 buffers.
-The total latency used for AV sync is 320ms. AV sync issues are
seen due to this big latency
-Reduce the deep buffer cound from 8 to 4. This would reduce the
latency.
CRs-Fixed: 598792
Change-Id: I1c057ee9d7c8a8cb747661e3dc62c9f44bf9ee79
-Partial drain is send to the driver after next track ioctl.
-If next track times out or returns any error then call drain.
-Partial drain will not be executed if next track fails.
CRs-Fixed: 594977
Change-Id: I0665f21a5ea01f7776408b5155344f542021c299
Use acdb_loader_init_v2 instead of acdb_loader_init_ACDB.
The new v2 api allows the codec name to be passed to
acdb-loader allowing the calibration files to be selected
based on the codec name.
Change-Id: Ib596111902b5e873fe9c770ced36dadd4d001232
Add XML parser which parses the platform_info.xml
on the device. That xml contains ACDB ID information
and is populated from the device project folder to
the /etc folder on the device. It is used to overwrite
hardcoded ACDB ID's in platform.c.
Change-Id: I86419bf0f48bcf7f0125da58626adab1d23fa50a
This change is needed to fix a compilation warning in Compress VoIP
code when the feature flag "AUDIO_FEATURE_DISABLED_MULTI_VOICE_SESSIONS"
is defined.
Change-Id: Ie19f542dc71705bc27a7278c83236904782f3a5e
CRs-Fixed: 590219
Add support for voip_out_stream_count in get_parameters which
will be queried by Audio Track to decide whether to open direct
output or not in case of VoIP.
Change-Id: Ic0d058dff32d9002cd452689636f342e748a7fe5
HAL doesnt let voice calls start if the device is set to speaker
as voice calls generally start on devices other than speaker.
However QCHAT calls start by calling setForceUse and setting the
device to speaker.
Fix by updating a flag whenever device routing happens when
AUDIO_MODE_IN_CALL. Reset the flag when mode is AUDIO_MODE_NORMAL.
CRs-fixed: 596074
Change-Id: I546959d2b0123828562dba1f6439aa494a365c3f
- Issues: camera shutter sound is not heard on
speaker when wired headset is connected.
- Rootcause: combo output devices are not updated
properly for voice call use cases.
- Fix: update combo devices before voice call use
cases in platform_get_output_snd_device.
CRs-Fixed: 591822
Change-Id: Ic2e75a4ce25f77499c07b4a8a32af774cdea16c2
Linux Build Service Account
ApurupaPattapu
Narsinga Rao Chella
Naresh Tanniru
Gopikrishnaiah Anandan
Ben Romberger
wjiang
Apoorv Raghuvanshi
Amit Shekhar
Krishnankutty Kolathappilly
Dhananjay Kumar
Vidyakumar Athota
Subhash Chandra Bose Naripeddy
Shiv Maliyappanahalli
Mingming Yin