好久没写了，今天碰巧有个小伙伴问我关于音频流这一块的，久了还有点记不起来，我就顺便写一下，后面就不用又找一遍代码了，所谓好记性不如烂笔头。

所以，这里是关于如何从AudioTrack 写入数据到audioflinger，以及audioflinger如何写入到hal层的音频流处理流程，主要写一下audioflinger处理流程，和写一些细节。

获取音频流

1、client写入数据：

app client 通过创建AudioTrack后，在播放的时候会不断的调用audiotrack的write方法，不断的向audioflinger写数据。

//frameworks\av\media\libaudioclient\AudioTrack.cpp
ssize_t AudioTrack::write(const void* buffer, size_t userSize, bool blocking)
{if (mTransfer != TRANSFER_SYNC) {return INVALID_OPERATION;}if (isDirect()) {AutoMutex lock(mLock);int32_t flags = android_atomic_and(~(CBLK_UNDERRUN | CBLK_LOOP_CYCLE | CBLK_LOOP_FINAL | CBLK_BUFFER_END),&mCblk->mFlags);if (flags & CBLK_INVALID) {return DEAD_OBJECT;}}if (ssize_t(userSize) < 0 || (buffer == NULL && userSize != 0)) {// Sanity-check: user is most-likely passing an error code, and it would// make the return value ambiguous (actualSize vs error).ALOGE("AudioTrack::write(buffer=%p, size=%zu (%zd)", buffer, userSize, userSize);return BAD_VALUE;}size_t written = 0;Buffer audioBuffer;while (userSize >= mFrameSize) {audioBuffer.frameCount = userSize / mFrameSize;status_t err = obtainBuffer(&audioBuffer,blocking ? &ClientProxy::kForever : &ClientProxy::kNonBlocking);if (err < 0) {if (written > 0) {break;}if (err == TIMED_OUT || err == -EINTR) {err = WOULD_BLOCK;}return ssize_t(err);}size_t toWrite = audioBuffer.size;memcpy(audioBuffer.i8, buffer, toWrite);buffer = ((const char *) buffer) + toWrite;userSize -= toWrite;written += toWrite;releaseBuffer(&audioBuffer);}if (written > 0) {mFramesWritten += written / mFrameSize;}return written;
}

2、service端的写

作为service端，不断的接收并处理client写入的数据。

//frameworks/av/services/audioflinger/Tracks.cpp
bool AudioFlinger::PlaybackThread::OutputTrack::write(void* data, uint32_t frames)
{Buffer *pInBuffer;Buffer inBuffer;bool outputBufferFull = false;inBuffer.frameCount = frames;inBuffer.raw = data;//处理客户端write数据，这个就不在这里细说了，这里不打算细讲这个环形缓冲，有兴趣的可以自行阅读代码
}

AudioFlinger音频流处理

主要处理流程：
track预处理
混音处理
向HAL输出数据

bool AudioFlinger::PlaybackThread::threadLoop()
{           //1、track预处理mMixerStatus = prepareTracks_l(&tracksToRemove);//2、混音处理threadLoop_mix();if (mMixerBufferValid) {//void *buffer = mEffectBufferValid ? mEffectBuffer : mSinkBuffer;audio_format_t format = mEffectBufferValid ? mEffectBufferFormat : mFormat;// mono blend occurs for mixer threads only (not direct or offloaded)// and is handled here if we're going directly to the sink.if (requireMonoBlend() && !mEffectBufferValid) {mono_blend(mMixerBuffer, mMixerBufferFormat, mChannelCount, mNormalFrameCount,true /*limit*/);}//将mMixerBuffer 按hal 层格式 实际最终还是拷贝到 mSinkBuffer 对象上memcpy_by_audio_format(buffer, format, mMixerBuffer, mMixerBufferFormat,mNormalFrameCount * mChannelCount);}//3、向HAL输出数据ret = threadLoop_write();
}

track预处理

向mAudioMixer更新track对象等操作

//prepareTracks_l
AudioFlinger::PlaybackThread::mixer_state AudioFlinger::MixerThread::prepareTracks_l(Vector< sp<Track> > *tracksToRemove)
{.....for (size_t i=0 ; i<count ; i++) {const sp<Track> t = mActiveTracks[i];....// if an active track doesn't exist in the AudioMixer, create it.if (!mAudioMixer->exists(name)) {status_t status = mAudioMixer->create(name,track->mChannelMask,track->mFormat,track->mSessionId);if (status != OK) {ALOGW("%s: cannot create track name"" %d, mask %#x, format %#x, sessionId %d in AudioMixer",__func__, name, track->mChannelMask, track->mFormat, track->mSessionId);tracksToRemove->add(track);track->invalidate(); // consider it dead.continue;}}....}.....
}

混音处理

void AudioFlinger::MixerThread::threadLoop_mix()
{// mix buffers...mAudioMixer->process();mCurrentWriteLength = mSinkBufferSize;// increase sleep time progressively when application underrun condition clears.// Only increase sleep time if the mixer is ready for two consecutive times to avoid// that a steady state of alternating ready/not ready conditions keeps the sleep time// such that we would underrun the audio HAL.if ((mSleepTimeUs == 0) && (sleepTimeShift > 0)) {sleepTimeShift--;}mSleepTimeUs = 0;mStandbyTimeNs = systemTime() + mStandbyDelayNs;//TODO: delay standby when effects have a tail
}
//frameworks/av/media/libaudioprocessing/include/media/AudioMixerBase.hvoid process() {preProcess();//这里才是处理数据(this->*mHook)();postProcess();}

数据获取&混音处理 (this->*mHook)();
mHook 以process__genericResampling 为例：

//frameworks/av/media/libaudioprocessing/AudioMixerBase.cpp
// generic code with resampling
void AudioMixerBase::process__genericResampling()
{ALOGVV("process__genericResampling\n");int32_t * const outTemp = mOutputTemp.get(); // naked ptrsize_t numFrames = mFrameCount;for (const auto &pair : mGroups) {const auto &group = pair.second;const std::shared_ptr<TrackBase> &t1 = mTracks[group[0]];// clear temp buffermemset(outTemp, 0, sizeof(*outTemp) * t1->mMixerChannelCount * mFrameCount);for (const int name : group) {const std::shared_ptr<TrackBase> &t = mTracks[name];int32_t *aux = NULL;if (CC_UNLIKELY(t->needs & NEEDS_AUX)) {aux = t->auxBuffer;}// this is a little goofy, on the resampling case we don't// acquire/release the buffers because it's done by// the resampler.if (t->needs & NEEDS_RESAMPLE) {(t.get()->*t->hook)(outTemp, numFrames, mResampleTemp.get() /* naked ptr */, aux);} else {size_t outFrames = 0;while (outFrames < numFrames) {t->buffer.frameCount = numFrames - outFrames;//获取client写入的数据t->bufferProvider->getNextBuffer(&t->buffer);t->mIn = t->buffer.raw;// t->mIn == nullptr can happen if the track was flushed just after having// been enabled for mixing.if (t->mIn == nullptr) break;(t.get()->*t->hook)(outTemp + outFrames * t->mMixerChannelCount, t->buffer.frameCount,mResampleTemp.get() /* naked ptr */,aux != nullptr ? aux + outFrames : nullptr);outFrames += t->buffer.frameCount;//通知释放 通知client 有可以写入的buffer了t->bufferProvider->releaseBuffer(&t->buffer);}}}/**这个mainBuffer与 mOutput->write((char *)mSinkBuffer, 0); 中的mSinkBuffer 相关联//相关代码如下：//frameworks\av\services\audioflinger\Threads.cppAudioFlinger::PlaybackThread::mixer_state AudioFlinger::MixerThread::prepareTracks_l(Vector< sp<Track> > *tracksToRemove){........mAudioMixer->setParameter(name,AudioMixer::TRACK,AudioMixer::MIXER_FORMAT, (void *)mMixerBufferFormat);mAudioMixer->setParameter(name,AudioMixer::TRACK,AudioMixer::MAIN_BUFFER, (void *)mMixerBuffer);........}*///这里是将outTemp 数据按最终输出格式拷贝给到 mainBufferconvertMixerFormat(t1->mainBuffer, t1->mMixerFormat,outTemp, t1->mMixerInFormat, numFrames * t1->mMixerChannelCount);}
}

向HAL输出数据

  // frameworks\av\services\audioflinger\Threads.cppssize_t AudioFlinger::MixerThread::threadLoop_write()
{// FIXME we should only do one push per cycle; confirm this is true// Start the fast mixer if it's not already runningif (mFastMixer != 0) {FastMixerStateQueue *sq = mFastMixer->sq();FastMixerState *state = sq->begin();if (state->mCommand != FastMixerState::MIX_WRITE &&(kUseFastMixer != FastMixer_Dynamic || state->mTrackMask > 1)) {if (state->mCommand == FastMixerState::COLD_IDLE) {// FIXME workaround for first HAL write being CPU bound on some devicesATRACE_BEGIN("write");//AudioStreamOut  *mOutput;//致此，audioflinger的写入流程基本结束，剩下的就是写入到audiohal层，//最后在audiohal处理之后就会通过tinyalsa写入到驱动完成了声音流的输出过程了mOutput->write((char *)mSinkBuffer, 0);ATRACE_END();int32_t old = android_atomic_inc(&mFastMixerFutex);if (old == -1) {(void) syscall(__NR_futex, &mFastMixerFutex, FUTEX_WAKE_PRIVATE, 1);}
#ifdef AUDIO_WATCHDOGif (mAudioWatchdog != 0) {mAudioWatchdog->resume();}
#endif}state->mCommand = FastMixerState::MIX_WRITE;
#ifdef FAST_THREAD_STATISTICSmFastMixerDumpState.increaseSamplingN(mAudioFlinger->isLowRamDevice() ?FastThreadDumpState::kSamplingNforLowRamDevice : FastThreadDumpState::kSamplingN);
#endifsq->end();sq->push(FastMixerStateQueue::BLOCK_UNTIL_PUSHED);if (kUseFastMixer == FastMixer_Dynamic) {mNormalSink = mPipeSink;}} else {sq->end(false /*didModify*/);}}return PlaybackThread::threadLoop_write();
}

补充

每个client 创建audiotrack的时候都是要通过AudioFlinger来创建的。audioflinger交给Threads.cpp这边创建，创建完毕后会保存在Threads 的mTracks 对象里面，当每次loop的时候会先调用prepareTracks_l 来准备数据，此时会及时将mTracks 同步到AudioMixer（不是将对象直接赋值给AudioMixer对象，AudioMixer会同步创建一个AudioMixer::Track来与mTracks对应，这个AudioMixer::Track然后再包含Track，最后实际获取数据的时候是通过这个Track来获取到client数据的）。

sp<IAudioTrack> AudioFlinger::createTrack(const CreateTrackInput& input,CreateTrackOutput& output,status_t *status)
{......track = thread->createTrack_l(client, streamType, input.attr, &output.sampleRate,input.config.format, input.config.channel_mask,&output.frameCount, &output.notificationFrameCount,input.notificationsPerBuffer, input.speed,input.sharedBuffer, sessionId, &output.flags,input.clientInfo.clientTid, clientUid, &lStatus, portId);
......
}sp<AudioFlinger::PlaybackThread::Track> AudioFlinger::PlaybackThread::createTrack_l(const sp<AudioFlinger::Client>& client,audio_stream_type_t streamType,const audio_attributes_t& attr,uint32_t *pSampleRate,audio_format_t format,audio_channel_mask_t channelMask,size_t *pFrameCount,size_t *pNotificationFrameCount,uint32_t notificationsPerBuffer,float speed,const sp<IMemory>& sharedBuffer,audio_session_t sessionId,audio_output_flags_t *flags,pid_t tid,uid_t uid,status_t *status,audio_port_handle_t portId)
{......track = new Track(this, client, streamType, attr, sampleRate, format,channelMask, frameCount,nullptr /* buffer */, (size_t)0 /* bufferSize */, sharedBuffer,sessionId, uid, *flags, TrackBase::TYPE_DEFAULT, portId);lStatus = track != 0 ? track->initCheck() : (status_t) NO_MEMORY;if (lStatus != NO_ERROR) {ALOGE("createTrack_l() initCheck failed %d; no control block?", lStatus);// track must be cleared from the caller as the caller has the AF lockgoto Exit;}mTracks.add(track);sp<EffectChain> chain = getEffectChain_l(sessionId);if (chain != 0) {ALOGV("createTrack_l() setting main buffer %p", chain->inBuffer());track->setMainBuffer(chain->inBuffer());chain->setStrategy(AudioSystem::getStrategyForStream(track->streamType()));chain->incTrackCnt();}
......
}

最后

sourceinside阅读有些跳转不了，有点恼火。

Android 9 Audio系统笔记：AudioFlinger音频流处理流程相关推荐

Android 9 Audio系统笔记：音频路由实现——从AudioTrack到audiohal
目录一.动态路由的初始化 1.获取路由策略 2.向AudioPolicyManager注册路由策略二.动态路由的路由流程,以AudioTrack创建为例创建AudioTrack的路由选择如何定 ...
【初学音频】Android的Audio系统之AudioTrack
目录前言 1. AudioTrack 2. 用例介绍 2.1 过程 2.2 数据加载模式 2.3 音频流的类型 2.4 Buffer分配和Frame的概念 3. AudioTrack (Java空间 ...
Android8.0 Audio系统之AudioFlinger
继上一篇AudioTrack的分析,本篇我们来看AudioFlinger,AF主要承担音频混合输出,是Audio系统的核心,从AudioTrack来的数据最终都会在这里处理,并被写入到Audio的HA ...
Android 4.0系统打电话和接电话系统流程时序图详解
鄙人初学Android系统源码还不到半载,之前一段时间由于工作原因,稍微研究了一下Android4.0系统打电话和接电话的大致流程.现将自己的研究结果以时序图奉上,由于经验有限,若有错误还请见谅. 打 ...
[深入理解Android卷一全文-第七章]深入理解Audio系统
由于<深入理解Android 卷一>和<深入理解Android卷二>不再出版,而知识的传播不应该因为纸质媒介的问题而中断,所以我将在CSDN博客中全文转发这两本书的全部内容. ...
Android 音频(Audio)架构
一.概述 Android 的音频硬件抽象层 (HAL) 可将 android.media 中特定于音频的较高级别的框架 API 连接到底层音频驱动程序和硬件.本部分介绍了有关提升性能的实现说明和提示. ...
1.7 深入理解Audio系统
第7章深入理解Audio系统 7.1 概述 Audio系统是Android平台的重要组成部分,它主要包括三方面的内容: AudioRcorder和AudioTrack:这两个类属于Audio系统对外 ...
Android N的Audio系统（五）
AudioFlinger 回放录制线程 AndioFlinger 作为 Android 的音频系统引擎,重任之一是负责输入输出流设备的管理及音频流数据的处理传输,这是由回放线程(PlaybackThr ...
Android ALSA音频系统架构分析（1）----从Loopback了解Audio
/*********************************** * Author:刘江明 * Environment:MTK Android 6.0 * Date:2017年05月25日 * ...
mtk+android+之mt6577驱动笔记,MTK6577+Android之音频(audio)移植
MTK6577+Android之音频(audio)移植备注:audio PA音频功放(power amplifier) 先借用<Y1MT6577 design notice V0.1>关 ...

Android 9 Audio系统笔记：AudioFlinger音频流处理流程