Age | Commit message (Collapse) | Author |
|
reduce the number of the retries to avoid hitting timecheck timeout 5s.
Bug: 172888755
Test: usb audio function simple test
Change-Id: I0a30b3246e13e440acf89b6b9f05480ca796259b
|
|
am: 4d21b82ae9 am: fbc7cb969a
Original change: https://android-review.googlesource.com/c/platform/hardware/qcom/audio/+/1588196
MUST ONLY BE SUBMITTED BY AUTOMERGER
Change-Id: I14b8c1b072a1d34de18da7acbb081bec2a9e8093
|
|
am: 4d21b82ae9
Original change: https://android-review.googlesource.com/c/platform/hardware/qcom/audio/+/1588196
MUST ONLY BE SUBMITTED BY AUTOMERGER
Change-Id: Ie0b85bfc15859b4a4bbb3b5d4937d6641e62c1a0
|
|
Added SPDX-license-identifier-Apache-2.0 to:
hal/Android.mk
post_proc/Android.mk
visualizer/Android.mk
voice_processing/Android.mk
Added SPDX-license-identifier-Apache-2.0 SPDX-license-identifier-BSD to:
legacy/alsa_sound/Android.mk
Added SPDX-license-identifier-Apache-2.0 SPDX-license-identifier-BSD
SPDX-license-identifier-LGPL
to:
legacy/libalsa-intf/Android.mk
Bug: 68860345
Bug: 151177513
Bug: 151953481
Test: m all
Exempt-From-Owner-Approval: janitorial work
Change-Id: I2ffc382d8a925e559275dc2c0a4359d54fa87e09
|
|
2651b854c7 am: fc76f54175
Original change: https://android-review.googlesource.com/c/platform/hardware/qcom/audio/+/1589292
MUST ONLY BE SUBMITTED BY AUTOMERGER
Change-Id: I77be964eea368259509f6766ce18da714fd87125
|
|
2651b854c7
Original change: https://android-review.googlesource.com/c/platform/hardware/qcom/audio/+/1589292
MUST ONLY BE SUBMITTED BY AUTOMERGER
Change-Id: I9ed596fa1f74d99d55efaa9f63e88fab067374ce
|
|
Bug: 150578172
Test: m
Change-Id: I9714cd4416fe0ce23e57cd7db162bfb45b0f84d5
|
|
frames.
Update get input buffer size logic for record use cases
such that the size is also a multiple of bytes per sample
period. For 1 channel 24 bit recording the computed
value was not a multiple of bytes per sample and this
mismatches with the computed value by ALSA causing
buffer invalid errors because the use case pointers
difference was reaching stop threshold.
Bug: 152727483
Test: repro steps in the bug.
Merged-In: I953366e09d98352d30b82f40c18a261d2355cf2c
Change-Id: I953366e09d98352d30b82f40c18a261d2355cf2c
|
|
frames.
Update get input buffer size logic for record use cases
such that the size is also a multiple of bytes per sample
period. For 1 channel 24 bit recording the computed
value was not a multiple of bytes per sample and this
mismatches with the computed value by ALSA causing
buffer invalid errors because the use case pointers
difference was reaching stop threshold.
Bug: 152727483
Test: repro steps in the bug.
Change-Id: I953366e09d98352d30b82f40c18a261d2355cf2c
|
|
compress-offload am: 9e85e402af am: c2510cab63
Original change: https://googleplex-android-review.googlesource.com/c/platform/hardware/qcom/audio/+/12628350
Change-Id: If972db65bbf33d132cb14b0ba4f8ed5fcaa68c3f
|
|
compress-offload
Remove the standby patch which was workaround for b/37551929 but impacting basic usecases.
Bug: 162394465
Test: Test music playback with and without usb headset.
Change-Id: I0c5e0ff162475f3958de79fa851018669edb4998
|
|
16240d402b am: cdc9b25709
Original change: https://googleplex-android-review.googlesource.com/c/platform/hardware/qcom/audio/+/12516726
Change-Id: Icc4114fac95fb2cbbbd31f149ee5fb81baee14be
|
|
When audio-playback-voip use case device shares the same
backend as the new device, audio-playback-voip device is switched,
and voip volume must be re-applied.
Bug: 167288716
Test: repro steps in bug.
Change-Id: I61692d7d5acf5d3147c32fc22761d1ed75c487a2
|
|
Check if a2dp path start failed and do retry from
out_write to recover the path is possible, which
can avoid blocking write if path set up failed.
Bug: 148926518
Test: manual test
Change-Id: If0df472ca0fb6454588fe2b58f1b0cc53bb6e650
Signed-off-by: Eric Huang <ericsphuang@google.com>
|
|
efff023555 am: b286e42949
Original change: https://googleplex-android-review.googlesource.com/c/platform/hardware/qcom/audio/+/11690442
Change-Id: I0c2ecf8f248e1ea9ef3185927aeecb714e027817
|
|
-Set the property-vendor.audio.offload.gapless.enabled to true to enable
gapless playback by default.
-Set mixer control to the driver to decide if gapless playback needs
to be used or not
-If gapless playback is not enabled then partial drain from HAL is
treated as full drain.
Bug: 157046086
Test: Test with Play Music
Change-Id: If010b7f730cc70dd55c2d4ff77ac272e60f3c4e2
Signed-off-by: Robert Lee <lerobert@google.com>
|
|
d620d11473 am: 3a77e5cd6a am: aaad8cf5af
Change-Id: I91428621f2d3430131a27f307b67f3279f9ea7e5
|
|
When A2DP sink is disconnected, the A2DP offload path may be
disabled before the audio framework pauses the output stream, in
which case the HAL forces the audio route to speaker.
When this happens, the front end must be muted to avoid audio
leaking over speaker path potentially at full volume if
A2DP absolute volume is enabled.
Bug: 156044862
Test: repro steps in the bug.
Change-Id: I90de8729ac862ca794dbdcf85fc1fa9dfabbf23b
|
|
Bug: 155495129
Test: built crosshatch_svelte successfully.
Change-Id: I00c6932865f773b0a7393addc45d0acf149a9d44
|
|
Add delay from xml, add platform capture delay
with audio source for AudioRecord, and add more delay in
platform_render_latency for playback usecases.
Bug: 137600762
Test: Manual test with Clarity test apk
Change-Id: I95088fc1826cbb0ecaf2409d6e184dbaaaae7c12
Signed-off-by: millerliang <millerliang@google.com>
|
|
|
|
Bug: 154338876
Test: build
Change-Id: I93b858a1991b18e281ca488351d2526d5f5dd4d8
|
|
- To select correct sound device for recording with
AUDIO_SOURCE_VOICE_RECOGNITION and AUDIO_SOURCE_UNPROCESSED, adding more
conditions to determine sound device.
Bug: 153704905
Bug: 153854533
Test: solotester for unprocessed with ch mask 2,3,4
Test: solotester for voice_recognition with ch mask 2,3,4
Signed-off-by: Jasmine Cha <chajasmine@google.com>
Change-Id: I87ffaacd54ca2a5a5b9918948d73a9583e74b0dd
|
|
Right now AIDL doesn't provide an API for vendor code to query if the
AIDL HAL declared. This will lead to 1) long blocking call for device
doesn't have the AIDL HAL and also 2) potential problem if the service
starts late after 5s timeout.
This is a workaround for addressing 1) on devices having HIDL, but still
affects device that doesn't have HIDL service.
Bug: 149797408
Test: build
Change-Id: I60aada1ab1a044db965987bca5f6eee5b6262b20
|
|
Bug: 148798433
Test: Build
Change-Id: Ie14d90bf17faad68beb16a9916e2fab9627b1eb4
|
|
Change-Id: Ie5298d1d83de872c5a2eb64c35ca2a8fcfac0c75
|
|
Free and assign NULL to global static device pointer instead of local
pointer to avoid free after use issue.
Bug: 144583303
Signed-off-by: Harrison Lingren <hlingren@google.com>
Change-Id: Idfdef719320efcd792c7d2ebd7ec2dfe5d3fbfbd
|
|
am: 7a6a02c8e2
Change-Id: I598b7bf1837c9203b3a6701df105dc1cebe123d5
|
|
Fix error codes returned by stream pause and resume
methods for the offload case.
Fix mmap buffer size check.
Bug: 144575694
Test: atest VtsHalAudioV6_0TargetTest
Change-Id: Iead2ed737a1994fd36f7b0f69ad107afcf14ddd6
Merged-In: Iead2ed737a1994fd36f7b0f69ad107afcf14ddd6
|
|
am: 73d6b59977
Change-Id: I44632c224c6ee9edca73fda24593362379220fff
|
|
From: Preetam Singh Ranawat <apranawat@codeaurora.org>
Date: Mon, 24 Jun 2019 15:11:28 +0530
Subject: audio: free and assign NULL to global static device pointer
-free and assign NULL to global static device pointer instead of local
pointer to avoid use after free issue.
Bug: 142267478
Test: manual
Change-Id: I6f64fe0f6034844279c9a481726426dc5b989b41
|
|
am: 32bbb9a379
Change-Id: Iadbf380c97bcd6e0cd3b30222a724f1c93989b9a
|
|
Since these were combined into libhidlbase.
Bug: 135686713
Test: build only (libhwbinder/libhidltransport are empty)
Change-Id: If2f4ffb73967cfdba61d0aeecb75a6c5d70e7fde
|
|
Starting from Q-launching devices, sysprops defined by vendor should
start with vendor. or ro.vendor. Therefore, renaming
ro.qc.sdk.audio.fluencetype to ro.vendor.audio.sdk.fluencetype. For
legacy compatibility, the HAL reads both sysprops; first the new one and
then falls back to the old one.
Bug: 139108926
Test: making a video call via Duo, make sure sound works
Test: vol adjust during a phone call works
Change-Id: I54babf96de702a142cbe860f59d6b5b011bfa374
|
|
This reverts commit 25054b7a665f02f441169353514116fa9d772953.
Bug: 139108926
Test: making a video call via Duo, make sure sound works
Test: vol adjust during a phone call works
Change-Id: Iea2371c79aa8061009a4f26f0574f1e1ea5742c9
|
|
This reverts commit 83fde0ec25e15f4aa94deae1e53738c2fe039775.
Reason for revert: No vol adjust on calls
No audio in duo vid calls.
Bug: 139108926
Change-Id: I21a6f81d1880251fd5bf9269363fbfda59674fd3
|
|
Starting from Q-launching devices, all vendor-defined sysprops should be
prefixed with "vendor.". To support such new devices as well as old
devices where the renamining isn't required, audio HAL is using both
properties: vendor.audio_hal.* is read first and then fall backs to
audio_hal.*.
Bug: b/138278883
Test: play some audio
Test: run VtsTrebleSysProp
Change-Id: I15158467f81ac99c36c40f6ca0a6c00cbca7fb24
|
|
|
|
The file descriptor for MMAP shared memory was created
and passed to AudioFlinger but never closed.
Now the FD is closed when the stream is goes to standby.
Bug: 134381208
Test: Get HAL pid.
Test: adb shell ps | grep audio
Test: adb shell ls -l /proc/{halpid}/fd | wc
Test: Run Oboetester or other app that uses MMAP
Test: Completely close the app.
Test: adb shell ls -l /proc/{halpid}/fd | wc
Test: Count should NOT grow each time.
Change-Id: Ieaaf1c6bdc96e7ecf01cee23215fb39f79662111
|
|
The sound_trigger session might have been paused while
soundtrigger client starts to record.
In this corner case, this input won't be detected as
soundtrigger input in audio HAL since the st_session handle
has been removed because the st_session is in pause state.
Then soundtrigger input will try to open pcm driver and
act as normal audio record, it will cause race condition
on pcm node.
To fix this, set the input with AUDIO_INPUT_FLAG_HW_HOTWORD
flag to be always soundtrigger input.
Bug: 135059114
Test: hotword/music detection works
Change-Id: I4475bfb1d3e2c9b998c583e9244c0aa4f07b2134
Signed-off-by: Carter Hsu <carterhsu@google.com>
|
|
Add support for incall-music-uplink2 usecase to enable Dual Sim
Dual Standby (DSDS) voice scenarios
Bug: 132080107
Test: Local test
Change-Id: I26206b1887f002bb58fcfb3378600ef81e111662
Signed-off-by: BubbleFang <bubblefang@google.com>
|
|
Android Q supports concurrent capture and will
create more than one AEC/NS effects if app creates
over 1 voice record tracks.
Refine the AEC/NS control to support mutilple
AEC/NS effects.
Also correct the voice-call stream type in STHAL.
Bug: 129111371
Bug: 128456220
Test: manual
Change-Id: Icd0863f54e17cd6a5ee765f9bb5fe1b4580743a4
Signed-off-by: Carter Hsu <carterhsu@google.com>
|
|
1. Audio HAL should notified STHAL before sending
audio calibration of audio usecase to kernel.
2. Fix the condition logic of pcm_capture, the
stream.out and stream.in are union, so we can't
check if it's NULL.
Bug: 129111371
Test: manual
Change-Id: I39676410555f7f528c0705059c312a1250d489f0
Signed-off-by: Carter Hsu <carterhsu@google.com>
|
|
sound" into qt-dev
|
|
|
|
HAL routes A2DP stream to speaker on releaseAudioPatch
from APM until the stream is put to standby from framework.
APM may releaseAudioPatch for an output where A2DP is
suspended and SCO is active leading to HAL switching all
streams to speaker. Fix this behavior by routing streams
to speaker only if both SCO and A2DP are not active.
Bug: 126848701
Test: repro steps in bug
Change-Id: Ic315b16480fe6d0c81c45716534a42b800e2d9da
|
|
|
|
Report only index channel masks for USB output profiles
supporting more than two channels.
Bug: 120947396
Test: play multichannel audio over USB
Change-Id: I10f6d4751a5a17674d24fb0d074f1dd6a71a06e1
|
|
Bug: 130443032
Test: voice call test.
Change-Id: Icb9b92a1f6be23cbf93970d6b10777856127dcc5
|
|
There are three usecases, the first is low-latency-playback(touch sound),
the second is audio-playback-voip(end call tone), the third is
deep-buffer-playback or compress-offload-playback(Music).
When low-latency-playback stop, it will trigger reroute to handset for
audio-playback-voip first then reroute to speaker for deep-buffer-playback
or compress-offload-playback.
The fix is to trigger reroute only when audio-playback-voip stop.
Bug: 130128135
Test: Music resume smoothly after press end call button w/ touch sound
Change-Id: Id5e8cefd53a852d0ce98a9f5ce095bee5832cb5a
Signed-off-by: Robert Lee <lerobert@google.com>
|