-DMID can have value greater than the MAX value returned by atoi,
hence use atoll instead.
- This change is done to accommodate the change in atoi implementation
from Lolipop onwards.
Change-Id: I58b1ba4ff18642238eaa34f41dd4bd9eeb1a25a9
Add quad mic recording support for standalone audio recording
and for VoIP calls using audio path.
Change-Id: I10724b6350b9940b93971ac720747dee1a9936cb
add entry for 8939 skul sound card name to load
the mixer controls for 8939.
Change-Id: I62ec602ab1d86c22c6262ebd66c45e208ea7b962
Signed-off-by: Meng Wang <mwang@codeaurora.org>
* quic/LA.BR.1: (37 commits)
hal: Performance mode during recording usecase
audio: Support for compress offload recovery
audio: Fix for clip skip issue during SSR
hal: add checks for calibration buffer allocation failures
mm-audio: aenc-aac: fix integer overflow for encoded buffer timestamp calculation
hal: Add support for 8909 QRD skue
hal: Fix for FM mute issue on start of FM record
hal: Fix array length computation for backend id array.
hal : set Non SA+ app type for PCM Capture
hal : Fix for SA+ app type selection for PCM Capture usecase
hal : Fix for SA+ app type selection for PCM Capture usecase
hal: use 0 as default return value for out_get_render_position
audio: hal: Add support for pm8916 on msm8909
hal: Define a new combo device for AANC and Fluence
hal: Define a new combo device for AANC and Fluence
hal: use 0 as default return value for out_get_render_position
hal: Update DS2 implementation to support ACDB based license mechanism
hal : Overwrite the APP type for PCM RX and TX path
audio: HAL to support for peripheral manager
hal : Set the input device for VOIP calls using audio path
...
Conflicts:
hal/Android.mk
hal/audio_extn/audio_extn.c
hal/audio_extn/audio_extn.h
hal/audio_extn/dolby.c
hal/audio_extn/utils.c
hal/audio_hw.c
hal/msm8916/hw_info.c
hal/msm8916/platform.c
hal/msm8974/platform.c
hal/platform_api.h
Change-Id: Ibfa171e8f3af713dbb2cffbaf2ca2b0df3e8ae73
Currently, metadata such as min/max block size is sent only for first
stream in FLAC gapless playback. This causes incorrect configuration
and, subsequently, framedrops in decoding of second stream and onwards
Sending additional parameters
Change-Id: I74abd18e7e5ebdce5a35e81d6af4d6f8107e07c9
CRs-Fixed: 781837
Aligning PCM buffer size to 1024 bytes is not required in case
of PCM offloading. It actually creates problem when playing
6 ch, 96KHz, 16-bit stream PCM in offload mode, due to aligning
PCM data becomes unequal for the six channels.
Align the buffer to LCM of 2, 4, 6, 8 i.e. 24.
Change-Id: I0823ca2ccdd5fceaf23a182b763a1d69f91f602f
During device switch, all the use cases on a shared backend will
be derouted before disabling the current device and rerouted after
enabling the new sound device. Rerouting of voip use cases
was not taken care properly which is fixed now.
Change-Id: I394c0d34fe36d42ce615414ea0524a58c10a5f79
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.11 (GNU/Linux)
iEYEABECAAYFAlS+Y7EACgkQoUgPZYCpAfGhcQCgzFtFc1pVn2TUWr42hAfxqmcJ
KNMAn0ijSKfcIWXQfCXYaiv3W7khSvRT
=4cCU
-----END PGP SIGNATURE-----
Merge tag 'AU_LINUX_ANDROID_LA.BF64.1.05.00.00.021.095' into merge_branch
AU_LINUX_ANDROID_LA.BF64.1.05.00.00.021.095 based on quic/aosp/LA.BF64.1
* tag 'AU_LINUX_ANDROID_LA.BF64.1.05.00.00.021.095': (30 commits)
audio: Reject incall record request when mode is not IN_CALL
mm-audio: aenc-aac: fix integer overflow for encoded buffer timestamp calculation
hal: Add support for parsing hdmi edid information for msm8992
hal: Fix glitch in audio playback during device switch
hal: add support to query hpx state using getparameter
hal: add all compressed formats for app type selection
hal: Performance mode during recording usecase
hal: fix null pointer de-reference
hal : add 8992 for msm8992 audio
post_proc: Support Headphone:X in offload and non tunnel modes
post_proc: Add support for non-tunnel DSP audio effects
audio: Avoid mute during AFE PROXY Usecase
hal: Fix DTS HPX Get parameters
hal: bug fixes for PCM offload
hal: Fix progress bar issue during compress offload playback
hal: Fix setting stream channel map properly
hardware/qcom/audio: Integrate Eagle framework for HeadphoneX
hal: Remove setting DAP bypass /DAP on from HAL
hal: Avoid resetting device to speaker for HDMI passthrough case
hal: Audio HAL / APM changes for HDMI passthrough
...
Conflicts:
Android.mk
hal/Android.mk
hal/audio_extn/audio_extn.h
hal/audio_extn/utils.c
hal/audio_hw.c
hal/voice.c
Change-Id: Ia92b2002486894e8711a48480324479fe114aeb6
CLOCK_BOOTTIME is identical to CLOCK_MONOTONIC except it also includes
any time that the system is suspended.
Change-Id: I78e5136332a366a32b66a393ead29f2c5972f7d2
add check for calibration buffer allocations failures and
enhance debug messages in the calibration update funtions.
Change-Id: I08c48f57c4fe747af8c19d12c986a5b9279bca03
- FM audio is muted on start of record session when
playback record concurrency enabled
- Add support for FM playback device switch on
record concurrency enabled
Change-Id: I51795dedf0f32c41c6a84889bb7f62ee21713720
For Voip calls using audio path ,the input devices is not
getting selected when output device is changed.
Removed the source type check to select the input device
when there is an active input and output device switch.
Change-Id: Iab7a6fddb3dda045fab94cd72de8e658721d9237
CRs-fixed: 712950
Currently the deep buffer path uses 4 * 20msec buffer.
During device switch the AudioPolicyManager delays routing
command by 2 * output latency (80msec) to ensure that all
the data written to driver/DSP is played out on current
device itself.
The stream side buffering in the DSP in legacy PCM mode is
95msec and hence the depth of pipeline is 80 + 95 = 175msec
which exceeds the device switch delay 160msec (2 * 80). So
the tail (data written to driver before headset plugin) is
heard on headset and perceived as glitch.
Ensure that the buffering in the kernel is greater than or
equal to the buffering in the DSP to fix the issue.
Change-Id: I01a3862d63ce4c258056620693dee08761c7e83f
CRs-Fixed: 771446
When MULTI_VOICE_SESSIONS feature flag is not enabled, the state of
incall recording stream is returned incorrectly. Fix the issue by
moving the corresponding piece of code from voice_extn.c to voice.c.
Change-Id: Id350bad65d12a4f165e80ee11cda3a8f342d38ee
Encoded buffer timestamp is 64 bit wide and is derived by multiplying
two 32 bit values, while calculating timestamp, these values are
multiplied with result stored in intermediate storage of 32 bit used
by compiler internally, this was associated with overflow if the result
is larger than max supported by 32 bit storage.
Changing m_frame_count to 64 bit will fix possible overflow for frame
count as well as force compiler to use intermediate variable of
required size.
Change-Id: I5ab709094d1c6b984d27f98327d3d8afa179fc92
CRs-Fixed: 767185
- If application selects AUDIO_SOURCE_VOICE_CALL when voice call is
not active, audio HAL is selecting afe-proxy use case which is
wrong. afe-proxy usecase/device expected to be selected during
USB Voice call scenario only
- Fix is to fail opening the input stream with
AUDIO_DEVICE_IN_TELEPHONY_RX or AUDIO_DEVICE_IN_VOICE_CALL if
current mode is not IN_CALL
CRs-Fixed: 772386
Change-Id: I8ee3e72daec86e26d05b01a117d7a01f7f6b69a8
Encoded buffer timestamp is 64 bit wide and is derived by multiplying
two 32 bit values, while calculating timestamp, these values are
multiplied with result stored in intermediate storage of 32 bit used
by compiler internally, this was associated with overflow if the result
is larger than max supported by 32 bit storage.
Changing m_frame_count to 64 bit will fix possible overflow for frame
count as well as force compiler to use intermediate variable of
required size.
Change-Id: I5ab709094d1c6b984d27f98327d3d8afa179fc92
CRs-Fixed: 767185
Currently the deep buffer path uses 4 * 20msec buffer.
During device switch the AudioPolicyManager delays routing
command by 2 * output latency (80msec) to ensure that all
the data written to driver/DSP is played out on current
device itself.
The stream side buffering in the DSP in legacy PCM mode is
95msec and hence the depth of pipeline is 80 + 95 = 175msec
which exceeds the device switch delay 160msec (2 * 80). So
the tail (data written to driver before headset plugin) is
heard on headset and perceived as glitch.
Ensure that the buffering in the kernel is greater than or
equal to the buffering in the DSP to fix the issue.
Change-Id: I01a3862d63ce4c258056620693dee08761c7e83f
CRs-Fixed: 771446
Linux Build Service Account
Ashish Jain
Karthik Reddy Katta
Chaithanya Krishna Bacharaju
Meng Wang
Preetam Singh Ranawat
Amit Shekhar
Alexy Joseph
Narsinga Rao Chella
Shruthi Krishna
Fred Oh
Dhanalakshmi Siddani
Sidipotu Ashok
Naresh Tanniru
Ravi Kumar Alamanda
Apoorv Raghuvanshi
Pavan Chikkala
Banajit Goswami