Android 7.0 Audio ：通话中的音频调用接口

对于Audio在通话中的处理，根据不同的功能，有不同的调用路径，

 

 

 

1) 通话中 打开speaker 流程   enableSpeaker

 

对于在通话中打开扬声器，走图示1的调用路径，

 

 

在telephony部分，调用流程如下，

InCallAudioManager. enableSpeaker--telecomAdapter.setAudioRoute(CallAudioState.ROUTE_SPEAKER)/binderIPC-- mInCallService.setAudioRoute(route);-- mPhone.setAudioRoute(route);-- mInCallAdapter.setAudioRoute(route);--mAdapter.setAudioRoute(route);/binder IPC--…--CallsManager.

setAudioRoute --mCallAudioManager.setAudioRoute(route)--CallAudioRouteStateMachine.sendMessageWithSessionInfo / CallAudioRouteStateMachine.USER_SWITCH_SPEAKER/--通过ActiveSpeakerRoute向下..—setSpeakerphoneOn-- mAudioManager.setSpeakerphoneOn(on)

 

这样，就到了Audio系统了，根据之前的章节，就知道接下来就是AM/AService/ASystem了，

AudioManager.setSpeakerphoneOn/ binder IPC--setSpeakerphoneOn --/ mForcedUseForComm = AudioSystem.FORCE_SPEAKER / sendMsg(mAudioHandler,MSG_SET_FORCE_USE, SENDMSG_QUEUE,

               AudioSystem.FOR_COMMUNICATION, mForcedUseForComm, null, 0) /内部handler--setForceUse --setForceUseInt_SyncDevices -- AudioSystem.setForceUse/jni-AudioSystem::setForceUse

 

在native部分的调用流程如下，

AudioSystem::setForceUse--AudioPolicyService::setForceUse（AudioPolicyInterfaceImpl.cpp）--AudioPolicyManager::setForceUse

APM处理部分如下，

 

void AudioPolicyManager::setForceUse(audio_policy_force_use_t usage,                                          audio_policy_forced_cfg_t config) {       if (mEngine->setForceUse(usage, config) != NO_ERROR) {         ALOGW("setForceUse() could not set force cfg %d for usage %d", config, usage);         return; }       checkA2dpSuspend();     checkOutputForAllStrategies();     updateDevicesAndOutputs();       if (mEngine->getPhoneState() == AUDIO_MODE_IN_CALL && hasPrimaryOutput()) {         audio_devices_t newDevice = getNewOutputDevice(mPrimaryOutput, true /*fromCache*/);         updateCallRouting(newDevice);     }     for (size_t i = 0; i < mOutputs.size(); i++) {         sp<SwAudioOutputDescriptor> outputDesc = mOutputs.valueAt(i);         audio_devices_t newDevice = getNewOutputDevice(outputDesc, true /*fromCache*/);         if ((mEngine->getPhoneState() != AUDIO_MODE_IN_CALL) || (outputDesc != mPrimaryOutput)) {             setOutputDevice(outputDesc, newDevice, (newDevice != AUDIO_DEVICE_NONE));         }         if (forceVolumeReeval && (newDevice != AUDIO_DEVICE_NONE)) {             applyStreamVolumes(outputDesc, newDevice, 0, true);         }     }       audio_io_handle_t activeInput = mInputs.getActiveInput();     if (activeInput != 0) {         sp<AudioInputDescriptor> activeDesc = mInputs.valueFor(activeInput);         audio_devices_t newDevice = getNewInputDevice(activeInput);         // Force new input selection if the new device can not be reached via current input         if (activeDesc->mProfile->getSupportedDevices().types() & (newDevice & ~AUDIO_DEVICE_BIT_IN)) {             setInputDevice(activeInput, newDevice);         } else {             closeInput(activeInput);         }     } }

 

 

a）mEngine->setForceUse

 

实现在Engine.cpp里面，主要是判断当前对设备设置是否合逻辑，并给mForceUse的成员赋值。

 

status_t Engine::setForceUse(audio_policy_force_use_t usage, audio_policy_forced_cfg_t config) {     switch(usage) {     case AUDIO_POLICY_FORCE_FOR_COMMUNICATION:         if (config != AUDIO_POLICY_FORCE_SPEAKER && config != AUDIO_POLICY_FORCE_BT_SCO &&             config != AUDIO_POLICY_FORCE_NONE) {             ALOGW("setForceUse() invalid config %d for FOR_COMMUNICATION", config);             return BAD_VALUE;         }         mForceUse[usage] = config;         break;

 

 

 

b）checkA2dpSuspend

checkA2dpSuspend--AudioPolicyService::AudioPolicyClient::suspendOutput--AudioFlinger::suspendOutput--thread->suspend

///                void        suspend() { (void) android_atomic_inc(&mSuspended); }

 

最终是给&mSuspended赋值。

 

 

c）checkOutputForStrategy 根据策略类型和当前设置获取设备类型，并最input和output进行处理，最后invalidateStream。

其中，

AudioPolicyManager::checkOutputForStrategy--audio_devices_tAudioPolicyManager::getDeviceForStrategy--mEngine->getDeviceForStrategy，

Engine.cpp里，

根据strategy为STRATEGY_PHONE，以及刚刚设置的mForceUse[AUDIO_POLICY_FORCE_FOR_COMMUNICATION]为AUDIO_POLICY_FORCE_SPEAKER，匹配到正确的设备类型，刚好为AUDIO_DEVICE_OUT_SPEAKER，

audio_devices_t Engine::getDeviceForStrategyInt(routing_strategy strategy,                                                 DeviceVector availableOutputDevices,                                                 DeviceVector availableInputDevices,                                                 const SwAudioOutputCollection &outputs) const {     uint32_t device = AUDIO_DEVICE_NONE;     uint32_t availableOutputDevicesType = availableOutputDevices.types();       switch (strategy) {     case STRATEGY_PHONE:           switch (mForceUse[AUDIO_POLICY_FORCE_FOR_COMMUNICATION]) {           case AUDIO_POLICY_FORCE_SPEAKER:             // when not in a phone call, phone strategy should route STREAM_VOICE_CALL to             // A2DP speaker when forcing to speaker output               device = availableOutputDevicesType & AUDIO_DEVICE_OUT_SPEAKER;             break;         }     break;

 

 

d)之后是updateCallRouting，设置device的操作。

 

 

 

2）dtmf音

 

在telephony模块，调用流程（这里只分析了声音播放，并不包含信令处理）

CallsManager.playDtmfTone --DtmfLocalTonePlayer .playTone--mToneGenerator.startTone

 

 

在Audio模块，调用流程

 

ToneGenerator.startTone--jni--android_media_ToneGenerator_startTone--ToneGenerator::startTone

 

ToneGenerator则直接使用了AT，先initAudioTrack创建AT实例，prepareWave生成PCM数据，再mpAudioTrack->start()开始，

 

bool ToneGenerator::startTone(tone_type toneType, int durationMs) {   toneType = getToneForRegion(toneType);       if (mState == TONE_IDLE) {         ALOGV("startTone: try to re-init AudioTrack");         if (!initAudioTrack()) {             return lResult;         }     }       if (mState == TONE_INIT) {         if (prepareWave()) {             ALOGV("Immediate start, time %d", (unsigned int)(systemTime()/1000000));             lResult = true;             mState = TONE_STARTING;             if (clock_gettime(CLOCK_MONOTONIC, &mStartTime) != 0) {                 mStartTime.tv_sec = 0;             }             mLock.unlock();             mpAudioTrack->start(); …     }     mLock.unlock(); }

 

 

 

可以看到，initAudioTrack创建AT实例时，使用了AUDIO_OUTPUT_FLAG_FAST的方式播放，通过audioCallback回调完成数据填充。

bool ToneGenerator::initAudioTrack() {       // Open audio track in mono, PCM 16bit, default sampling rate, default buffer size     mpAudioTrack = new AudioTrack();     ALOGV("Create Track: %p", mpAudioTrack.get());       mpAudioTrack->set(mStreamType,                       0,    // sampleRate                       AUDIO_FORMAT_PCM_16_BIT,                       AUDIO_CHANNEL_OUT_MONO,                       0,    // frameCount                       AUDIO_OUTPUT_FLAG_FAST,                       audioCallback,                       this, // user                       0,    // notificationFrames                       0,    // sharedBuffer                       mThreadCanCallJava,                       AUDIO_SESSION_ALLOCATE,                       AudioTrack::TRANSFER_CALLBACK);       if (mpAudioTrack->initCheck() != NO_ERROR) {         ALOGE("AudioTrack->initCheck failed");         goto initAudioTrack_exit;     }       mpAudioTrack->setVolume(mVolume);       mState = TONE_INIT;       return true;   initAudioTrack_exit:       ALOGV("Init failed: %p", mpAudioTrack.get());       // Cleanup     mpAudioTrack.clear();       return false; }

 

这条通道走的是AF的路径，并不在上图的三个path之中。

 

 

 

3）标记3的场景有，Telephony向Audio设置当前是通话状态，这个会用setMode (MODE_INCALL)；向Audio设置当前是VOICE通话还是LTE通话，这个会用AudioSystem::setParameters()，

 

vendor\qcom\proprietary\qcril\qcril_qmi\qcril_am.cc