741 lines
25 KiB
C++
Raw Normal View History

/*
* Copyright (C) 2006-2007 The Android Open Source Project
*
* Licensed under the Apache License, Version 2.0 (the "License");
* you may not use this file except in compliance with the License.
* You may obtain a copy of the License at
*
* http://www.apache.org/licenses/LICENSE-2.0
*
* Unless required by applicable law or agreed to in writing, software
* distributed under the License is distributed on an "AS IS" BASIS,
* WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
* See the License for the specific language governing permissions and
* limitations under the License.
*/
#define LOG_TAG "AudioSystem"
//#define LOG_NDEBUG 0
#include <utils/Log.h>
#include <binder/IServiceManager.h>
#include <media/AudioSystem.h>
#include <media/IAudioPolicyService.h>
#include <math.h>
#include <system/audio.h>
// ----------------------------------------------------------------------------
namespace android {
// client singleton for AudioFlinger binder interface
Mutex AudioSystem::gLock;
sp<IAudioFlinger> AudioSystem::gAudioFlinger;
sp<AudioSystem::AudioFlingerClient> AudioSystem::gAudioFlingerClient;
audio_error_callback AudioSystem::gAudioErrorCallback = NULL;
// Cached values
DefaultKeyedVector<int, audio_io_handle_t> AudioSystem::gStreamOutputMap(0);
DefaultKeyedVector<audio_io_handle_t, AudioSystem::OutputDescriptor *> AudioSystem::gOutputs(0);
// Cached values for recording queries
uint32_t AudioSystem::gPrevInSamplingRate = 16000;
int AudioSystem::gPrevInFormat = AUDIO_FORMAT_PCM_16_BIT;
int AudioSystem::gPrevInChannelCount = 1;
size_t AudioSystem::gInBuffSize = 0;
// establish binder interface to AudioFlinger service
const sp<IAudioFlinger>& AudioSystem::get_audio_flinger()
{
Mutex::Autolock _l(gLock);
if (gAudioFlinger.get() == 0) {
sp<IServiceManager> sm = defaultServiceManager();
sp<IBinder> binder;
do {
binder = sm->getService(String16("media.audio_flinger"));
if (binder != 0)
break;
LOGW("AudioFlinger not published, waiting...");
usleep(500000); // 0.5 s
} while(true);
if (gAudioFlingerClient == NULL) {
gAudioFlingerClient = new AudioFlingerClient();
} else {
if (gAudioErrorCallback) {
gAudioErrorCallback(NO_ERROR);
}
}
binder->linkToDeath(gAudioFlingerClient);
gAudioFlinger = interface_cast<IAudioFlinger>(binder);
gAudioFlinger->registerClient(gAudioFlingerClient);
}
LOGE_IF(gAudioFlinger==0, "no AudioFlinger!?");
return gAudioFlinger;
}
status_t AudioSystem::muteMicrophone(bool state) {
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af == 0) return PERMISSION_DENIED;
return af->setMicMute(state);
}
status_t AudioSystem::isMicrophoneMuted(bool* state) {
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af == 0) return PERMISSION_DENIED;
*state = af->getMicMute();
return NO_ERROR;
}
status_t AudioSystem::setMasterVolume(float value)
{
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af == 0) return PERMISSION_DENIED;
af->setMasterVolume(value);
return NO_ERROR;
}
status_t AudioSystem::setMasterMute(bool mute)
{
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af == 0) return PERMISSION_DENIED;
af->setMasterMute(mute);
return NO_ERROR;
}
status_t AudioSystem::getMasterVolume(float* volume)
{
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af == 0) return PERMISSION_DENIED;
*volume = af->masterVolume();
return NO_ERROR;
}
status_t AudioSystem::getMasterMute(bool* mute)
{
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af == 0) return PERMISSION_DENIED;
*mute = af->masterMute();
return NO_ERROR;
}
status_t AudioSystem::setStreamVolume(int stream, float value, int output)
{
if (uint32_t(stream) >= AUDIO_STREAM_CNT) return BAD_VALUE;
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af == 0) return PERMISSION_DENIED;
af->setStreamVolume(stream, value, output);
return NO_ERROR;
}
status_t AudioSystem::setStreamMute(int stream, bool mute)
{
if (uint32_t(stream) >= AUDIO_STREAM_CNT) return BAD_VALUE;
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af == 0) return PERMISSION_DENIED;
af->setStreamMute(stream, mute);
return NO_ERROR;
}
status_t AudioSystem::getStreamVolume(int stream, float* volume, int output)
{
if (uint32_t(stream) >= AUDIO_STREAM_CNT) return BAD_VALUE;
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af == 0) return PERMISSION_DENIED;
*volume = af->streamVolume(stream, output);
return NO_ERROR;
}
status_t AudioSystem::getStreamMute(int stream, bool* mute)
{
if (uint32_t(stream) >= AUDIO_STREAM_CNT) return BAD_VALUE;
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af == 0) return PERMISSION_DENIED;
*mute = af->streamMute(stream);
return NO_ERROR;
}
status_t AudioSystem::setMode(int mode)
{
if (mode >= AUDIO_MODE_CNT) return BAD_VALUE;
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af == 0) return PERMISSION_DENIED;
return af->setMode(mode);
}
status_t AudioSystem::setParameters(audio_io_handle_t ioHandle, const String8& keyValuePairs) {
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af == 0) return PERMISSION_DENIED;
return af->setParameters(ioHandle, keyValuePairs);
}
String8 AudioSystem::getParameters(audio_io_handle_t ioHandle, const String8& keys) {
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
String8 result = String8("");
if (af == 0) return result;
result = af->getParameters(ioHandle, keys);
return result;
}
// convert volume steps to natural log scale
// change this value to change volume scaling
static const float dBPerStep = 0.5f;
// shouldn't need to touch these
static const float dBConvert = -dBPerStep * 2.302585093f / 20.0f;
static const float dBConvertInverse = 1.0f / dBConvert;
float AudioSystem::linearToLog(int volume)
{
// float v = volume ? exp(float(100 - volume) * dBConvert) : 0;
// LOGD("linearToLog(%d)=%f", volume, v);
// return v;
return volume ? exp(float(100 - volume) * dBConvert) : 0;
}
int AudioSystem::logToLinear(float volume)
{
// int v = volume ? 100 - int(dBConvertInverse * log(volume) + 0.5) : 0;
// LOGD("logTolinear(%d)=%f", v, volume);
// return v;
return volume ? 100 - int(dBConvertInverse * log(volume) + 0.5) : 0;
}
status_t AudioSystem::getOutputSamplingRate(int* samplingRate, int streamType)
{
OutputDescriptor *outputDesc;
audio_io_handle_t output;
if (streamType == AUDIO_STREAM_DEFAULT) {
streamType = AUDIO_STREAM_MUSIC;
}
output = getOutput((audio_stream_type_t)streamType);
if (output == 0) {
return PERMISSION_DENIED;
}
gLock.lock();
outputDesc = AudioSystem::gOutputs.valueFor(output);
if (outputDesc == 0) {
LOGV("getOutputSamplingRate() no output descriptor for output %d in gOutputs", output);
gLock.unlock();
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af == 0) return PERMISSION_DENIED;
*samplingRate = af->sampleRate(output);
} else {
LOGV("getOutputSamplingRate() reading from output desc");
*samplingRate = outputDesc->samplingRate;
gLock.unlock();
}
LOGV("getOutputSamplingRate() streamType %d, output %d, sampling rate %d", streamType, output, *samplingRate);
return NO_ERROR;
}
status_t AudioSystem::getOutputFrameCount(int* frameCount, int streamType)
{
OutputDescriptor *outputDesc;
audio_io_handle_t output;
if (streamType == AUDIO_STREAM_DEFAULT) {
streamType = AUDIO_STREAM_MUSIC;
}
output = getOutput((audio_stream_type_t)streamType);
if (output == 0) {
return PERMISSION_DENIED;
}
gLock.lock();
outputDesc = AudioSystem::gOutputs.valueFor(output);
if (outputDesc == 0) {
gLock.unlock();
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af == 0) return PERMISSION_DENIED;
*frameCount = af->frameCount(output);
} else {
*frameCount = outputDesc->frameCount;
gLock.unlock();
}
LOGV("getOutputFrameCount() streamType %d, output %d, frameCount %d", streamType, output, *frameCount);
return NO_ERROR;
}
status_t AudioSystem::getOutputLatency(uint32_t* latency, int streamType)
{
OutputDescriptor *outputDesc;
audio_io_handle_t output;
if (streamType == AUDIO_STREAM_DEFAULT) {
streamType = AUDIO_STREAM_MUSIC;
}
output = getOutput((audio_stream_type_t)streamType);
if (output == 0) {
return PERMISSION_DENIED;
}
gLock.lock();
outputDesc = AudioSystem::gOutputs.valueFor(output);
if (outputDesc == 0) {
gLock.unlock();
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af == 0) return PERMISSION_DENIED;
*latency = af->latency(output);
} else {
*latency = outputDesc->latency;
gLock.unlock();
}
LOGV("getOutputLatency() streamType %d, output %d, latency %d", streamType, output, *latency);
return NO_ERROR;
}
status_t AudioSystem::getInputBufferSize(uint32_t sampleRate, int format, int channelCount,
size_t* buffSize)
{
// Do we have a stale gInBufferSize or are we requesting the input buffer size for new values
if ((gInBuffSize == 0) || (sampleRate != gPrevInSamplingRate) || (format != gPrevInFormat)
|| (channelCount != gPrevInChannelCount)) {
// save the request params
gPrevInSamplingRate = sampleRate;
gPrevInFormat = format;
gPrevInChannelCount = channelCount;
gInBuffSize = 0;
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af == 0) {
return PERMISSION_DENIED;
}
gInBuffSize = af->getInputBufferSize(sampleRate, format, channelCount);
}
*buffSize = gInBuffSize;
return NO_ERROR;
}
status_t AudioSystem::setVoiceVolume(float value)
{
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af == 0) return PERMISSION_DENIED;
return af->setVoiceVolume(value);
}
status_t AudioSystem::getRenderPosition(uint32_t *halFrames, uint32_t *dspFrames, int stream)
{
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af == 0) return PERMISSION_DENIED;
if (stream == AUDIO_STREAM_DEFAULT) {
stream = AUDIO_STREAM_MUSIC;
}
return af->getRenderPosition(halFrames, dspFrames, getOutput((audio_stream_type_t)stream));
}
unsigned int AudioSystem::getInputFramesLost(audio_io_handle_t ioHandle) {
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
unsigned int result = 0;
if (af == 0) return result;
if (ioHandle == 0) return result;
result = af->getInputFramesLost(ioHandle);
return result;
}
Issue 2667801: [Audio Effect Framework] AudioFlinger, AudioMixer AudioTrack modifications. First drop of audio framework modifications for audio effects support. - AudioTrack/AudioRecord: Added support for auxiliary effects in AudioTrack Added support for audio sessions Fixed left right channel inversion in setVolume() - IAudioFlinger: Added interface methods for effect enumeraiton and instantiation Added support for audio sessions. - IAudioTrack: Added method to attach auxiliary effect. - AudioFlinger Created new classes to control effect engines in effect library and manage effect connections to tracks or output mix: EffectModule: wrapper object controlling the effect engine implementation in the effect library. There is one EffectModule per instance of an effect in a given audio session EffectChain: group of effects associated to one audio session. There is one EffectChain per audio session. EffectChain for session 0 is for output mix effects, other chains are attached to audio tracks with same session ID. Each chain contains a variable number of EffectModules EffectHandle: implements the IEffect interface. There is one EffectHandle object for each application controlling (or using) an effect module. THe EffectModule maintians a list of EffectHandles. Added support for effect modules and effect chains creation in PlaybackThread. modified mixer thread loop to allow track volume control by effect modules and call effect processing. -AudioMixer Each track now specifies its output buffer used by mixer for accumulation Modified mixer process functions to process tracks by groups of tracks with same buffer Modified track process functions to support accumulation to auxiliary channel Change-Id: I26d5f7c9e070a89bdd383e1a659f8b7ca150379c
2010-06-01 23:49:17 -07:00
int AudioSystem::newAudioSessionId() {
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af == 0) return 0;
return af->newAudioSessionId();
}
void AudioSystem::acquireAudioSessionId(int audioSession) {
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af != 0) {
af->acquireAudioSessionId(audioSession);
}
}
void AudioSystem::releaseAudioSessionId(int audioSession) {
const sp<IAudioFlinger>& af = AudioSystem::get_audio_flinger();
if (af != 0) {
af->releaseAudioSessionId(audioSession);
}
}
// ---------------------------------------------------------------------------
void AudioSystem::AudioFlingerClient::binderDied(const wp<IBinder>& who) {
Mutex::Autolock _l(AudioSystem::gLock);
AudioSystem::gAudioFlinger.clear();
// clear output handles and stream to output map caches
AudioSystem::gStreamOutputMap.clear();
AudioSystem::gOutputs.clear();
if (gAudioErrorCallback) {
gAudioErrorCallback(DEAD_OBJECT);
}
LOGW("AudioFlinger server died!");
}
void AudioSystem::AudioFlingerClient::ioConfigChanged(int event, int ioHandle, void *param2) {
LOGV("ioConfigChanged() event %d", event);
OutputDescriptor *desc;
uint32_t stream;
if (ioHandle == 0) return;
Mutex::Autolock _l(AudioSystem::gLock);
switch (event) {
case STREAM_CONFIG_CHANGED:
if (param2 == 0) break;
stream = *(uint32_t *)param2;
LOGV("ioConfigChanged() STREAM_CONFIG_CHANGED stream %d, output %d", stream, ioHandle);
if (gStreamOutputMap.indexOfKey(stream) >= 0) {
gStreamOutputMap.replaceValueFor(stream, ioHandle);
}
break;
case OUTPUT_OPENED: {
if (gOutputs.indexOfKey(ioHandle) >= 0) {
LOGV("ioConfigChanged() opening already existing output! %d", ioHandle);
break;
}
if (param2 == 0) break;
desc = (OutputDescriptor *)param2;
OutputDescriptor *outputDesc = new OutputDescriptor(*desc);
gOutputs.add(ioHandle, outputDesc);
LOGV("ioConfigChanged() new output samplingRate %d, format %d channels %d frameCount %d latency %d",
outputDesc->samplingRate, outputDesc->format, outputDesc->channels, outputDesc->frameCount, outputDesc->latency);
} break;
case OUTPUT_CLOSED: {
if (gOutputs.indexOfKey(ioHandle) < 0) {
LOGW("ioConfigChanged() closing unknow output! %d", ioHandle);
break;
}
LOGV("ioConfigChanged() output %d closed", ioHandle);
gOutputs.removeItem(ioHandle);
for (int i = gStreamOutputMap.size() - 1; i >= 0 ; i--) {
if (gStreamOutputMap.valueAt(i) == ioHandle) {
gStreamOutputMap.removeItemsAt(i);
}
}
} break;
case OUTPUT_CONFIG_CHANGED: {
int index = gOutputs.indexOfKey(ioHandle);
if (index < 0) {
LOGW("ioConfigChanged() modifying unknow output! %d", ioHandle);
break;
}
if (param2 == 0) break;
desc = (OutputDescriptor *)param2;
LOGV("ioConfigChanged() new config for output %d samplingRate %d, format %d channels %d frameCount %d latency %d",
ioHandle, desc->samplingRate, desc->format,
desc->channels, desc->frameCount, desc->latency);
OutputDescriptor *outputDesc = gOutputs.valueAt(index);
delete outputDesc;
outputDesc = new OutputDescriptor(*desc);
gOutputs.replaceValueFor(ioHandle, outputDesc);
} break;
case INPUT_OPENED:
case INPUT_CLOSED:
case INPUT_CONFIG_CHANGED:
break;
}
}
void AudioSystem::setErrorCallback(audio_error_callback cb) {
Mutex::Autolock _l(gLock);
gAudioErrorCallback = cb;
}
bool AudioSystem::routedToA2dpOutput(int streamType) {
switch(streamType) {
case AUDIO_STREAM_MUSIC:
case AUDIO_STREAM_VOICE_CALL:
case AUDIO_STREAM_BLUETOOTH_SCO:
case AUDIO_STREAM_SYSTEM:
return true;
default:
return false;
}
}
// client singleton for AudioPolicyService binder interface
sp<IAudioPolicyService> AudioSystem::gAudioPolicyService;
sp<AudioSystem::AudioPolicyServiceClient> AudioSystem::gAudioPolicyServiceClient;
// establish binder interface to AudioFlinger service
const sp<IAudioPolicyService>& AudioSystem::get_audio_policy_service()
{
gLock.lock();
if (gAudioPolicyService.get() == 0) {
sp<IServiceManager> sm = defaultServiceManager();
sp<IBinder> binder;
do {
binder = sm->getService(String16("media.audio_policy"));
if (binder != 0)
break;
LOGW("AudioPolicyService not published, waiting...");
usleep(500000); // 0.5 s
} while(true);
if (gAudioPolicyServiceClient == NULL) {
gAudioPolicyServiceClient = new AudioPolicyServiceClient();
}
binder->linkToDeath(gAudioPolicyServiceClient);
gAudioPolicyService = interface_cast<IAudioPolicyService>(binder);
gLock.unlock();
} else {
gLock.unlock();
}
return gAudioPolicyService;
}
status_t AudioSystem::setDeviceConnectionState(audio_devices_t device,
audio_policy_dev_state_t state,
const char *device_address)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return PERMISSION_DENIED;
return aps->setDeviceConnectionState(device, state, device_address);
}
audio_policy_dev_state_t AudioSystem::getDeviceConnectionState(audio_devices_t device,
const char *device_address)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return AUDIO_POLICY_DEVICE_STATE_UNAVAILABLE;
return aps->getDeviceConnectionState(device, device_address);
}
status_t AudioSystem::setPhoneState(int state)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return PERMISSION_DENIED;
return aps->setPhoneState(state);
}
status_t AudioSystem::setRingerMode(uint32_t mode, uint32_t mask)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return PERMISSION_DENIED;
return aps->setRingerMode(mode, mask);
}
status_t AudioSystem::setForceUse(audio_policy_force_use_t usage, audio_policy_forced_cfg_t config)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return PERMISSION_DENIED;
return aps->setForceUse(usage, config);
}
audio_policy_forced_cfg_t AudioSystem::getForceUse(audio_policy_force_use_t usage)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return AUDIO_POLICY_FORCE_NONE;
return aps->getForceUse(usage);
}
audio_io_handle_t AudioSystem::getOutput(audio_stream_type_t stream,
uint32_t samplingRate,
uint32_t format,
uint32_t channels,
audio_policy_output_flags_t flags)
{
audio_io_handle_t output = 0;
// Do not use stream to output map cache if the direct output
// flag is set or if we are likely to use a direct output
// (e.g voice call stream @ 8kHz could use BT SCO device and be routed to
// a direct output on some platforms).
// TODO: the output cache and stream to output mapping implementation needs to
// be reworked for proper operation with direct outputs. This code is too specific
// to the first use case we want to cover (Voice Recognition and Voice Dialer over
// Bluetooth SCO
if ((flags & AUDIO_POLICY_OUTPUT_FLAG_DIRECT) == 0 &&
((stream != AUDIO_STREAM_VOICE_CALL && stream != AUDIO_STREAM_BLUETOOTH_SCO) ||
channels != AUDIO_CHANNEL_OUT_MONO ||
(samplingRate != 8000 && samplingRate != 16000))) {
Mutex::Autolock _l(gLock);
output = AudioSystem::gStreamOutputMap.valueFor(stream);
LOGV_IF((output != 0), "getOutput() read %d from cache for stream %d", output, stream);
}
if (output == 0) {
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return 0;
output = aps->getOutput(stream, samplingRate, format, channels, flags);
if ((flags & AUDIO_POLICY_OUTPUT_FLAG_DIRECT) == 0) {
Mutex::Autolock _l(gLock);
AudioSystem::gStreamOutputMap.add(stream, output);
}
}
return output;
}
status_t AudioSystem::startOutput(audio_io_handle_t output,
audio_stream_type_t stream,
int session)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return PERMISSION_DENIED;
return aps->startOutput(output, stream, session);
}
status_t AudioSystem::stopOutput(audio_io_handle_t output,
audio_stream_type_t stream,
int session)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return PERMISSION_DENIED;
return aps->stopOutput(output, stream, session);
}
void AudioSystem::releaseOutput(audio_io_handle_t output)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return;
aps->releaseOutput(output);
}
audio_io_handle_t AudioSystem::getInput(int inputSource,
uint32_t samplingRate,
uint32_t format,
uint32_t channels,
audio_in_acoustics_t acoustics,
int sessionId)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return 0;
return aps->getInput(inputSource, samplingRate, format, channels, acoustics, sessionId);
}
status_t AudioSystem::startInput(audio_io_handle_t input)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return PERMISSION_DENIED;
return aps->startInput(input);
}
status_t AudioSystem::stopInput(audio_io_handle_t input)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return PERMISSION_DENIED;
return aps->stopInput(input);
}
void AudioSystem::releaseInput(audio_io_handle_t input)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return;
aps->releaseInput(input);
}
status_t AudioSystem::initStreamVolume(audio_stream_type_t stream,
int indexMin,
int indexMax)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return PERMISSION_DENIED;
return aps->initStreamVolume(stream, indexMin, indexMax);
}
status_t AudioSystem::setStreamVolumeIndex(audio_stream_type_t stream, int index)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return PERMISSION_DENIED;
return aps->setStreamVolumeIndex(stream, index);
}
status_t AudioSystem::getStreamVolumeIndex(audio_stream_type_t stream, int *index)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return PERMISSION_DENIED;
return aps->getStreamVolumeIndex(stream, index);
}
uint32_t AudioSystem::getStrategyForStream(audio_stream_type_t stream)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return 0;
return aps->getStrategyForStream(stream);
}
uint32_t AudioSystem::getDevicesForStream(audio_stream_type_t stream)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return 0;
return aps->getDevicesForStream(stream);
}
audio_io_handle_t AudioSystem::getOutputForEffect(effect_descriptor_t *desc)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return PERMISSION_DENIED;
return aps->getOutputForEffect(desc);
}
status_t AudioSystem::registerEffect(effect_descriptor_t *desc,
audio_io_handle_t io,
uint32_t strategy,
int session,
int id)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return PERMISSION_DENIED;
return aps->registerEffect(desc, io, strategy, session, id);
}
status_t AudioSystem::unregisterEffect(int id)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return PERMISSION_DENIED;
return aps->unregisterEffect(id);
}
status_t AudioSystem::setEffectEnabled(int id, bool enabled)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return PERMISSION_DENIED;
return aps->setEffectEnabled(id, enabled);
}
status_t AudioSystem::isStreamActive(int stream, bool* state, uint32_t inPastMs)
{
const sp<IAudioPolicyService>& aps = AudioSystem::get_audio_policy_service();
if (aps == 0) return PERMISSION_DENIED;
if (state == NULL) return BAD_VALUE;
*state = aps->isStreamActive(stream, inPastMs);
return NO_ERROR;
}
// ---------------------------------------------------------------------------
void AudioSystem::AudioPolicyServiceClient::binderDied(const wp<IBinder>& who) {
Mutex::Autolock _l(AudioSystem::gLock);
AudioSystem::gAudioPolicyService.clear();
LOGW("AudioPolicyService server died!");
}
}; // namespace android