Age | Commit message (Collapse) | Author |
|
23b18ee115 am: 7115f08897 am: 32dfdab05e am: d228ea9525
am: cfb51307dd
Change-Id: Id18c0ac4a18d1c1deda5c7cb8db7c2c5dd34c3b0
|
|
am: 23b18ee115
Change-Id: Id58f5502b0bd970254557f2e4a4ebf9afc2ba168
|
|
am: 500d293743
Change-Id: I92532658dc53e5071ff8cd75dacca04703a2d62f
|
|
Removed offload visualizer library linked by mistake
Bug: 32917432
Test: cts security VisualizerEffectTest.java
Change-Id: I59a481e5ef2489e63d890a67fc4ec327507e460c
|
|
Do not hold a mutex when calling recognition callback
Bug: 28823662
Change-Id: I66b7b9c997b6b06c6ae13e39d44803e749ffbae9
|
|
Change-Id: I77d6eaf4ac31d969fd42e9a96418203bc682476f
|
|
Bug: 22594886
Change-Id: I1caa47d0b217de083a68483a1cea4b9addb1b170
|
|
Bug: 20226809
Change-Id: I864aa498c3a04545a9e61aa56a96adc949393777
|
|
On T132, IPI latency is observed to be 100's of microseconds. This
change and corresponding kernel change for APB DMA interrupts will
ensure that audio interrupt and waking thread, FastMixer, both run
on same core, thereby eliminating IPI latency altogether.
Bug: 20226809
Change-Id: I6d3b7006acfaee4d21ea7ac89ad2cce3494a99eb
|
|
Audio underruns can be handled either by stopping pcm device and
preparing it again or by playing silence frames till application
fills the buffer again. On average, stopping and preparing pcm device
takes 2-4 times more time than just playing silence frames and let
the application catch-up. So switching to that approach now.
Bug: 20226809
Bug: 20300203
Change-Id: I53a1d32879e594d15fa2b0142852649edb2b5835
|
|
Fix thread starvation issue where the capture or playback threads
running in FIFO priority would constantly acquire the stream mutex preventing
other threads to complete routing commands.
Bug: 21880828.
Change-Id: I36e9c609857dd1bc959809326350a04c3af92228
|
|
Change-Id: I29fdb1e5c57fb1fa555e323e506cb41ccb5f19e2
|
|
Change-Id: I09336551e8510e1c5be79f6e498da8b143c4421d
|
|
Change-Id: I5908318faea6d3e7c33e15310df20699120a89e6
|
|
set_parameters functions must return 0 in case
of successful operation.
Change-Id: I6b2f4d8b794fac888da8aaae10bf1491128d8814
|
|
This reverts commit 127585760e20910a81f8fdf06a1dd8eee2e94829.
|
|
issue""""
This reverts commit e4fd15511b972a2caa2d313c94143036b75395cc.
|
|
compress offload path""""
This reverts commit 83472e8f2be82832d36d4de9324624e6d0d417ed.
|
|
offload path"""
This reverts commit 6ea970a1e88c14219708b1f1ad0937d149167fa0.
|
|
This reverts commit ca371c457f68010f8d4a81ad844ad9819b16dd9d.
|
|
This reverts commit 5b10fbc0e8f651add1bc8b742b75237d1b56e800.
|
|
Bug: 18371513.
Change-Id: I9509237c05c053a2d0ec96a3b561641dccc5ad8f
|
|
This reverts commit 148e67547f7697e59b79a9268adb9ee9bd40ef7f.
Bug: 16683770
Change-Id: I3a2f286b4aa8b0cef22917ee31460e26ad9eddc1
|
|
path""
This reverts commit 5da021f1ae8f12ae6e23edeea44ea805960d6e53.
Bug: 16683770.
Change-Id: Ib8494c37725eb00dcb64ff490985cc1d0c0e4b8a
|
|
* commit '2697dc6996c6882887d742a6229955e64a1d3170':
Add missing includes.
|
|
Change-Id: Id2ee93ca30788d0ca26f89346632493b0d3bc205
|
|
Current SCO_PERIOD_SIZE will make FastCapture turned on for a call
initiated with BT-SCO. Increase it to apply AudioStreamIn instead.
Bug 18710410
Change-Id: Idfe641a96c137d8a7973a7b9e638d9911e5cb0f9
|
|
This reverts commit 476a97240a37b5ad79f2db295a7f1eb58959982a.
|
|
This reverts commit 4402bf721e148e1cf97821063192fcd8b1bbb000.
|
|
Fix offload spk no sound issue. (mr1)
Bug: 16683770
Change-Id: Idfff572e69516f1f6ed26f41752f2edada4a6325
Signed-off-by: andy_su <andy_su@htc.com>
|
|
Bug: 16683770
Change-Id: I6fe056db6ea1ed3a7be7341873b8b2c3876fb935
|
|
Fix codec doesn't close during pause condition (mr1)
Bug: 16683770
Change-Id: Idff3ad7dc942021d300a23b70560e9039b284879
Signed-off-by: andy_su <andy_su@htc.com>
|
|
Add support for offload volume control.
Bug: 16683770
Change-Id: I6ceb89b6c3fa016c4f1ff00e7573a1337f158415
Signed-off-by: andy_su <andy_su@htc.com>
|
|
Pass directly the sample rate value to struct snd_codec
instead of the ALSA enum.
Bug: 17398311.
Change-Id: I88e828c36743413fbcc75eb35538c72751fa7768
|
|
Tfa9895 need pcm output continously during fw loading. Reopen pcm
with larger stop_threshold only for the case.
Bug 18024252
Change-Id: I93b0610d44129c2cc4b52b3188f4b79121c0c8df
|
|
Buganizer#: 18024252
Change-Id: I3bbec1a1ce56b04d65cad438b5b6dea4054dffb8
Signed-off-by: andy_su <andy_su@htc.com>
|
|
Bug: 17360889
Change-Id: I3b078c89c9429e01f580fbca81e3e631efdff2ac
Signed-off-by: Dmitry Shmidt <dimitrysh@google.com>
|
|
out_set_parameters()" into lmp-dev
* commit '6337d893bc851a3f1e0ef84885a88e99509f79e4':
audio: fix assertion in out_set_parameters()
|
|
adev->lock_inputs should be locked when modifying
adev->active_intput.
Bug: 18011058.
Change-Id: Ifcd28f260899c7307781249610198179f7645110
|
|
Bug: 17918918
Change-Id: I85f149f2d800e478b2acb569f75bdd38ffd3f4f7
Signed-off-by: Dmitry Shmidt <dimitrysh@google.com>
|
|
Bug: 17459931
Change-Id: I654c9034281e024b8ce82f65c50eeea26ada6db6
Signed-off-by: andy_su <andy_su@htc.com>
|
|
There is a problem with mutex acquisition order when removing or
adding an echo reference from do_input_standby_l() or start_input_stream():
the output stream mutex is acquired after the audio device mutex creating
a potential cross deadlock as the required order is output stream mutex
then audio device mutex.
Bug: 17830596.
Change-Id: I465f4eeb1f1f2e76d6cc3d685d96e4758c6bc473
|
|
Can't guarantee I2S signal in previous way.
Enlarge stop_threshold for spk to avoid I2S drain in unexpected cases.
Also Avoid I2S turned off when selecting a new device.
Bug: 17490341
Change-Id: I322076c7643cacf6e8878ce454dea5a0a0f93309
Signed-off-by: wesley_wang <wesley_wang@htc.com>
|
|
Allow capture from DSP when headset mic is specified
as capture device.
Bug: 17325425.
Change-Id: Ie9ca236e722be2789403a462bf325129fd49752a
|
|
Bug: 17490341
Change-Id: Ic30a976ee76a02db849471db3442e722b6267ee4
Signed-off-by: Dmitry Shmidt <dimitrysh@google.com>
|
|
If dummybufthread fails to write data to pcm device then stop_threshold will
automatically trigger stop on pcm device which disables i2s fifo. If we use
large value for stop_threshold then this automatic trigger can be avoided
which is needed incase of dummybufthread for hardware configuration.
Bug: 17379121
Change-Id: I0e83a3f6f8febe900589f38dc2e9f6cf63d94033
Signed-off-by: Dmitry Shmidt <dimitrysh@google.com>
|
|
|
|
Bug: 17325425
Change-Id: I67372152146f735d5c976df892509572c9f3a3ec
Signed-off-by: Dmitry Shmidt <dimitrysh@google.com>
|
|
Bug: 17129712
Change-Id: Ib0336e2b7f4c15f999d3e5984b1d87bfde884a65
|
|
|