【Android多媒体（重采样与混频）】从playback流程分析AudioResampler，AudioMixer

文章大致内容介绍

本文主要讲述播放音乐流程，AudioResampler是如何生效的

• 播放流程
• AudioMixer分析
• 从MixerThread分析

播放流程

在Android中间，如果使用硬解码OffloadThread是不会出现混频和重采样的，但是播放需要软解的音频，则会按需要重采样。
为了方便研究，这里修改软件配置文件，让primary compress_offload仅仅支持8000的采样率播放，音乐播放会去跑MixerThread线程，这个线程上面才会出现重采样。
configs/msm8909/audio_policy_configuration.xml修改如下：

@@ -69,7 +69,7 @@
                 
                     -                             samplingRates="8000,11025,12000,16000,22050,24000,32000,44100,48000"
+                             samplingRates="8000"
                              channelMasks="AUDIO_CHANNEL_OUT_STEREO,AUDIO_CHANNEL_OUT_MONO"/>

使用播放流程如下：

AudioMixer分析：

我们从log出发，查看代码调用流程

AudioMixer: add track (0)
AudioMixer::prepareForDownmix(0xb4f7d060) with mask 0x3
AudioMixer::prepareForReformat(0xb4f7d060) with format 0x1
AudioMixer: enable(0) AudioMixer: setParameter(VOLUME, VOLUME0: 0000)
AudioMixer: setParameter(VOLUME, VOLUME1: 0000)
AudioMixer: Creating resampler from track 44100 Hz to device 48000 Hz
AudioResampler: resampler load 0 -> 6 MHz due to delta +6 MHz from quality 6
AudioResampler: Create dynamic Resampler = 6
AudioMixer: setParameter(RESAMPLE, SAMPLE_RATE, 44100) AudioMixer: setParameter(TRACK, MIXER_FORMAT, 0x5) AudioMixer: setParameter(TRACK, MAIN_BUFFER, 0xb4f7a000) AudioMixer: mixer configuration change: 1
AudioMixer:activeTracks (00000001) all16BitsStereoNoResample=0, resampling=1, volumeRamp=0
AudioMixer: track__Resample

这段log的调用流程

1，AudioFlinger::MixerThread::checkForNewParameter_l,
2，AudioFlinger::MixerThread::getTrackName_l
3， AudioMixer::getTrackName(audio_channel_mask_t
channelMask,audio_format_t format, int sessionId)
4，调用prepare处理

prepareForDownmix这个函数，是用来处理多通道，将多通道转化，按照条件选择DownMixer或者Remixer，函数如下：

status_t AudioMixer::track_t::prepareForDownmix()
{
    ALOGV("AudioMixer::prepareForDownmix(%p) with mask 0x%x",
            this, channelMask);

    // discard the previous downmixer if there was one
    unprepareForDownmix();
    // MONO_HACK Only remix (upmix or downmix) if the track and mixer/device channel masks
    // are not the same and not handled internally, as mono -> stereo currently is.
    if (channelMask == mMixerChannelMask
            || (channelMask == AUDIO_CHANNEL_OUT_MONO
                    && mMixerChannelMask == AUDIO_CHANNEL_OUT_STEREO)) {
    //mMixerChannelMask 在getTrackName中间被赋值为双声道，判断是否需要downMixer
        return NO_ERROR;
    }
    // DownmixerBufferProvider is only used for position masks.
    if (audio_channel_mask_get_representation(channelMask)
                == AUDIO_CHANNEL_REPRESENTATION_POSITION
            && DownmixerBufferProvider::isMultichannelCapable()) {
     //条件满足channelMask&0x3为0，即声道数大于2，并且系统支持兼容多声道，则初始化相应的bufferprovider用来改变通道数
        DownmixerBufferProvider* pDbp = new DownmixerBufferProvider(channelMask,
                mMixerChannelMask,
                AUDIO_FORMAT_PCM_16_BIT /* TODO: use mMixerInFormat, now only PCM 16 */,
                sampleRate, sessionId, kCopyBufferFrameCount);

        if (pDbp->isValid()) { // if constructor completed properly
            mDownmixRequiresFormat = AUDIO_FORMAT_PCM_16_BIT; // PCM 16 bit required for downmix
            downmixerBufferProvider = pDbp;
            reconfigureBufferProviders();
            return NO_ERROR;
        }
        delete pDbp;
    }
    //使用remixer的情况

    // Effect downmixer does not accept the channel conversion.  Let's use our remixer.
    RemixBufferProvider* pRbp = new RemixBufferProvider(channelMask,
            mMixerChannelMask, mMixerInFormat, kCopyBufferFrameCount);
    // Remix always finds a conversion whereas Downmixer effect above may fail.
    downmixerBufferProvider = pRbp;
    reconfigureBufferProviders();
    return NO_ERROR;
}

prepareForReformat()判断是否需要格式转换：

status_t AudioMixer::track_t::prepareForReformat()
{
    ALOGV("AudioMixer::prepareForReformat(%p) with format %#x", this, mFormat);
    // discard previous reformatters
    unprepareForReformat();
    // only configure reformatters as needed
    const audio_format_t targetFormat = mDownmixRequiresFormat != AUDIO_FORMAT_INVALID
            ? mDownmixRequiresFormat : mMixerInFormat;
    bool requiresReconfigure = false;
    if (mFormat != targetFormat) {
        mReformatBufferProvider = new ReformatBufferProvider(
                audio_channel_count_from_out_mask(channelMask),
                mFormat,
                targetFormat,
                kCopyBufferFrameCount);
        requiresReconfigure = true;
    }
    if (targetFormat != mMixerInFormat) {
        mPostDownmixReformatBufferProvider = new ReformatBufferProvider(
                audio_channel_count_from_out_mask(mMixerChannelMask),
                targetFormat,
                mMixerInFormat,
                kCopyBufferFrameCount);
        requiresReconfigure = true;
    }
    if (requiresReconfigure) {
        reconfigureBufferProviders();
    }
    return NO_ERROR;
}

从AudioMixer到AudioResampler

调用的时序图如下：

voidAudioMixer::process__genericNoResampling(state_t* state, int64_t pts)
它是把各个track的声音数据相加。所谓声音数据，可以认为是一个个的采样点，Android默认支持的采样精度是16bit的，格式为signedPCM，所以每个采样点用有符号的16位数int16_t表示。如果直接加16bit的数据，肯定会造成16bit的值溢出，Android的做法是强转成int32_t，相加，并把和赋值给了32bit的数。注意，相加前乘上了音量，而表达音量的数据类型也是int32_t。这样，就能保证在这个过程中是不会溢出的。

void ditherAndClamp(int32_t* out, constint32_t *sums, size_t c) 将32bit的输出转成16bit，再把右声道放高16bit，左声道放低16bit输出

AudioMixer->AudioResampler传递设置参数处理主要的函数：prepareTracks_l()

这个函数中设置重采样参数等操作
// prepareTracks_l() must be called with ThreadBase::mLock held
AudioFlinger::PlaybackThread::mixer_state AudioFlinger::MixerThread::prepareTracks_l(
        Vector< sp > *tracksToRemove)
{

    mixer_state mixerStatus = MIXER_IDLE;
    // find out which tracks need to be processed
    size_t count = mActiveTracks.size();
    size_t mixedTracks = 0;
    size_t tracksWithEffect = 0;
    // counts only _active_ fast tracks
    size_t fastTracks = 0;
    uint32_t resetMask = 0; // bit mask of fast tracks that need to be reset

    float masterVolume = mMasterVolume;
    bool masterMute = mMasterMute;
#ifdef DOLBY_ENABLE // DOLBY_DAP_PREGAIN
    // The maximum volume of left channel and right channel for pregain calculation.
    uint32_t max_vol = 0;
    float dvlf = 0.0f;
    float dvrf = 0.0f;
#endif // DOLBY_END

    if (masterMute) {
        masterVolume = 0;
    }
    // Delegate master volume control to effect in output mix effect chain if needed
    sp chain = getEffectChain_l(AUDIO_SESSION_OUTPUT_MIX);
    if (chain != 0) {
        uint32_t v = (uint32_t)(masterVolume * (1 << 24));
        chain->setVolume_l(&v, &v);
        masterVolume = (float)((v + (1 << 23)) >> 24);
        chain.clear();
    }

    // prepare a new state to push
    FastMixerStateQueue *sq = NULL;
    FastMixerState *state = NULL;
    bool didModify = false;
    FastMixerStateQueue::block_t block = FastMixerStateQueue::BLOCK_UNTIL_PUSHED;
    if (mFastMixer != 0) {
        sq = mFastMixer->sq();
        state = sq->begin();
    }

    mMixerBufferValid = false;  // mMixerBuffer has no valid data until appropriate tracks found.
    mEffectBufferValid = false; // mEffectBuffer has no valid data until tracks found.

    for (size_t i=0 ; i t = mActiveTracks[i].promote();
        if (t == 0) {
            continue;
        }

        // this const just means the local variable doesn't change
        Track* const track = t.get();

        // process fast tracks
        if (track->isFastTrack()) {

            // It's theoretically possible (though unlikely) for a fast track to be created
            // and then removed within the same normal mix cycle.  This is not a problem, as
            // the track never becomes active so it's fast mixer slot is never touched.
            // The converse, of removing an (active) track and then creating a new track
            // at the identical fast mixer slot within the same normal mix cycle,
            // is impossible because the slot isn't marked available until the end of each cycle.
            int j = track->mFastIndex;
            ALOG_ASSERT(0 < j && j < (int)FastMixerState::sMaxFastTracks);
            ALOG_ASSERT(!(mFastTrackAvailMask & (1 << j)));
            FastTrack *fastTrack = &state->mFastTracks[j];

            // Determine whether the track is currently in underrun condition,
            // and whether it had a recent underrun.
            FastTrackDump *ftDump = &mFastMixerDumpState.mTracks[j];
            FastTrackUnderruns underruns = ftDump->mUnderruns;
            uint32_t recentFull = (underruns.mBitFields.mFull -
                    track->mObservedUnderruns.mBitFields.mFull) & UNDERRUN_MASK;
            uint32_t recentPartial = (underruns.mBitFields.mPartial -
                    track->mObservedUnderruns.mBitFields.mPartial) & UNDERRUN_MASK;
            uint32_t recentEmpty = (underruns.mBitFields.mEmpty -
                    track->mObservedUnderruns.mBitFields.mEmpty) & UNDERRUN_MASK;
            uint32_t recentUnderruns = recentPartial + recentEmpty;
            track->mObservedUnderruns = underruns;
            // don't count underruns that occur while stopping or pausing
            // or stopped which can occur when flush() is called while active
            if (!(track->isStopping() || track->isPausing() || track->isStopped()) &&
                    recentUnderruns > 0) {
                // FIXME fast mixer will pull & mix partial buffers, but we count as a full underrun
                track->mAudioTrackServerProxy->tallyUnderrunFrames(recentUnderruns * mFrameCount);
            } else {
                track->mAudioTrackServerProxy->tallyUnderrunFrames(0);
            }

            // This is similar to the state machine for normal tracks,
            // with a few modifications for fast tracks.
            bool isActive = true;
            switch (track->mState) {
            case TrackBase::STOPPING_1:
                // track stays active in STOPPING_1 state until first underrun
                if (recentUnderruns > 0 || track->isTerminated()) {
                    track->mState = TrackBase::STOPPING_2;
                }
                break;
            case TrackBase::PAUSING:
                // ramp down is not yet implemented
                track->setPaused();
                break;
            case TrackBase::RESUMING:
                // ramp up is not yet implemented
                track->mState = TrackBase::ACTIVE;
                break;
            case TrackBase::ACTIVE:
                if (recentFull > 0 || recentPartial > 0) {
                    // track has provided at least some frames recently: reset retry count
                    track->mRetryCount = kMaxTrackRetries;
                }
                if (recentUnderruns == 0) {
                    // no recent underruns: stay active
                    break;
                }
                // there has recently been an underrun of some kind
                if (track->sharedBuffer() == 0) {
                    // were any of the recent underruns "empty" (no frames available)?
                    if (recentEmpty == 0) {
                        // no, then ignore the partial underruns as they are allowed indefinitely
                        break;
                    }
                    // there has recently been an "empty" underrun: decrement the retry counter
                    if (--(track->mRetryCount) > 0) {
                        break;
                    }
                    // indicate to client process that the track was disabled because of underrun;
                    // it will then automatically call start() when data is available
                    track->disable();
                    // remove from active list, but state remains ACTIVE [confusing but true]
                    isActive = false;
                    break;
                }
                // fall through
            case TrackBase::STOPPING_2:
            case TrackBase::PAUSED:
            case TrackBase::STOPPED:
            case TrackBase::FLUSHED:   // flush() while active
                // Check for presentation complete if track is inactive
                // We have consumed all the buffers of this track.
                // This would be incomplete if we auto-paused on underrun
                {
                    size_t audioHALFrames =
                            (mOutput->stream->get_latency(mOutput->stream)*mSampleRate) / 1000;
                    int64_t framesWritten = mBytesWritten / mFrameSize;
                    if (!(mStandby || track->presentationComplete(framesWritten, audioHALFrames))) {
                        // track stays in active list until presentation is complete
                        break;
                    }
                }
                if (track->isStopping_2()) {
                    track->mState = TrackBase::STOPPED;
                }
                if (track->isStopped()) {
                    // Can't reset directly, as fast mixer is still polling this track
                    //   track->reset();
                    // So instead mark this track as needing to be reset after push with ack
                    resetMask |= 1 << i;
                }
                isActive = false;
                break;
            case TrackBase::IDLE:
            default:
                LOG_ALWAYS_FATAL("unexpected track state %d", track->mState);
            }

            if (isActive) {
                // was it previously inactive?
                if (!(state->mTrackMask & (1 << j))) {
                    ExtendedAudioBufferProvider *eabp = track;
                    VolumeProvider *vp = track;
                    fastTrack->mBufferProvider = eabp;
                    fastTrack->mVolumeProvider = vp;
                    fastTrack->mChannelMask = track->mChannelMask;
                    fastTrack->mFormat = track->mFormat;
                    fastTrack->mGeneration++;
                    state->mTrackMask |= 1 << j;
                    didModify = true;
                    // no acknowledgement required for newly active tracks
                }
                // cache the combined master volume and stream type volume for fast mixer; this
                // lacks any synchronization or barrier so VolumeProvider may read a stale value
                track->mCachedVolume = masterVolume * mStreamTypes[track->streamType()].volume;
                ++fastTracks;
            } else {
                // was it previously active?
                if (state->mTrackMask & (1 << j)) {
                    fastTrack->mBufferProvider = NULL;
                    fastTrack->mGeneration++;
                    state->mTrackMask &= ~(1 << j);
                    didModify = true;
                    // If any fast tracks were removed, we must wait for acknowledgement
                    // because we're about to decrement the last sp<> on those tracks.
                    /* MODIFIED-BEGIN by baowen.zhang, 2016-12-22,BUG-3773166*/
                    // If the fast track was added when output was already suspended
                    // and removed before the output resumed, which means fastmixer never
                    // runs, so no need to wait for acknowledgement for this case.
                    if (state->mCommand & FastMixerState::MIX_WRITE)
                        block = FastMixerStateQueue::BLOCK_UNTIL_ACKED;
                        /* MODIFIED-END by baowen.zhang,BUG-3773166*/
                } else {
                    LOG_ALWAYS_FATAL("fast track %d should have been active; "
                            "mState=%d, mTrackMask=%#x, recentUnderruns=%u, isShared=%d",
                            j, track->mState, state->mTrackMask, recentUnderruns,
                            track->sharedBuffer() != 0);
                }
                tracksToRemove->add(track);
                // Avoids a misleading display in dumpsys
                track->mObservedUnderruns.mBitFields.mMostRecent = UNDERRUN_FULL;
            }
            continue;
        }

        {   // local variable scope to avoid goto warning

        audio_track_cblk_t* cblk = track->cblk();

        // The first time a track is added we wait
        // for all its buffers to be filled before processing it
        int name = track->name();
        // make sure that we have enough frames to mix one full buffer.
        // enforce this condition only once to enable draining the buffer in case the client
        // app does not call stop() and relies on underrun to stop:
        // hence the test on (mMixerStatus == MIXER_TRACKS_READY) meaning the track was mixed
        // during last round
        size_t desiredFrames;
        const uint32_t sampleRate = track->mAudioTrackServerProxy->getSampleRate();
        AudioPlaybackRate playbackRate = track->mAudioTrackServerProxy->getPlaybackRate();

        desiredFrames = sourceFramesNeededWithTimestretch(
                sampleRate, mNormalFrameCount, mSampleRate, playbackRate.mSpeed);
        // TODO: ONLY USED FOR LEGACY RESAMPLERS, remove when they are removed.
        // add frames already consumed but not yet released by the resampler
        // because mAudioTrackServerProxy->framesReady() will include these frames
        desiredFrames += mAudioMixer->getUnreleasedFrames(track->name());

        uint32_t minFrames = 1;
        if ((track->sharedBuffer() == 0) && !track->isStopped() && !track->isPausing() &&
                (mMixerStatusIgnoringFastTracks == MIXER_TRACKS_READY)) {
            minFrames = desiredFrames;
        }

        size_t framesReady = track->framesReady();
        if (ATRACE_ENABLED()) {
            // I wish we had formatted trace names
            char traceName[16];
            strcpy(traceName, "nRdy");
            int name = track->name();
            if (AudioMixer::TRACK0 <= name &&
                    name < (int) (AudioMixer::TRACK0 + AudioMixer::MAX_NUM_TRACKS)) {
                name -= AudioMixer::TRACK0;
                traceName[4] = (name / 10) + '0';
                traceName[5] = (name % 10) + '0';
            } else {
                traceName[4] = '?';
                traceName[5] = '?';
            }
            traceName[6] = '\0';
            ATRACE_INT(traceName, framesReady);
        }
        if ((framesReady >= minFrames) && track->isReady() &&
                !track->isPaused() && !track->isTerminated())
        {
            ALOGVV("track %d s=%08x [OK] on thread %p", name, cblk->mServer, this);

            mixedTracks++;

            // track->mainBuffer() != mSinkBuffer or mMixerBuffer means
            // there is an effect chain connected to the track
            chain.clear();
            if (track->mainBuffer() != mSinkBuffer &&
                    track->mainBuffer() != mMixerBuffer) {
                if (mEffectBufferEnabled) {
                    mEffectBufferValid = true; // Later can set directly.
                }
                chain = getEffectChain_l(track->sessionId());
                // Delegate volume control to effect in track effect chain if needed
                if (chain != 0) {
                    tracksWithEffect++;
                } else {
                    ALOGW("prepareTracks_l(): track %d attached to effect but no chain found on "
                            "session %d",
                            name, track->sessionId());
                }
            }


            int param = AudioMixer::VOLUME;
            if (track->mFillingUpStatus == Track::FS_FILLED) {
                // no ramp for the first volume setting
                track->mFillingUpStatus = Track::FS_ACTIVE;
                if (track->mState == TrackBase::RESUMING) {
                    track->mState = TrackBase::ACTIVE;
                    param = AudioMixer::RAMP_VOLUME;
                }
                mAudioMixer->setParameter(name, AudioMixer::RESAMPLE, AudioMixer::RESET, NULL);
            // FIXME should not make a decision based on mServer
            } else if (cblk->mServer != 0) {
                // If the track is stopped before the first frame was mixed,
                // do not apply ramp
                param = AudioMixer::RAMP_VOLUME;
            }

            // compute volume for this track
            uint32_t vl, vr;       // in U8.24 integer format
            float vlf, vrf, vaf;   // in [0.0, 1.0] float format
            if (track->isPausing() || mStreamTypes[track->streamType()].mute) {
                vl = vr = 0;
                vlf = vrf = vaf = 0.;
                if (track->isPausing()) {
                    track->setPaused();
                }
            } else {

                // read original volumes with volume control
                float typeVolume = mStreamTypes[track->streamType()].volume;
                float v = masterVolume * typeVolume;
                AudioTrackServerProxy *proxy = track->mAudioTrackServerProxy;
                gain_minifloat_packed_t vlr = proxy->getVolumeLR();
                vlf = float_from_gain(gain_minifloat_unpack_left(vlr));
                vrf = float_from_gain(gain_minifloat_unpack_right(vlr));
                // track volumes come from shared memory, so can't be trusted and must be clamped
                if (vlf > GAIN_FLOAT_UNITY) {
                    ALOGV("Track left volume out of range: %.3g", vlf);
                    vlf = GAIN_FLOAT_UNITY;
                }
                if (vrf > GAIN_FLOAT_UNITY) {
                    ALOGV("Track right volume out of range: %.3g", vrf);
                    vrf = GAIN_FLOAT_UNITY;
                }
                // now apply the master volume and stream type volume
                vlf *= v;
                vrf *= v;
                // assuming master volume and stream type volume each go up to 1.0,
                // then derive vl and vr as U8.24 versions for the effect chain
                const float scaleto8_24 = MAX_GAIN_INT * MAX_GAIN_INT;
                vl = (uint32_t) (scaleto8_24 * vlf);
                vr = (uint32_t) (scaleto8_24 * vrf);
                // vl and vr are now in U8.24 format
                uint16_t sendLevel = proxy->getSendLevel_U4_12();
                // send level comes from shared memory and so may be corrupt
                if (sendLevel > MAX_GAIN_INT) {
                    ALOGV("Track send level out of range: %04X", sendLevel);
                    sendLevel = MAX_GAIN_INT;
                }
                // vaf is represented as [0.0, 1.0] float by rescaling sendLevel
                vaf = v * sendLevel * (1. / MAX_GAIN_INT);
            }

#ifdef DOLBY_ENABLE // DOLBY_DAP_PREGAIN
            dvlf = vlf;
            dvrf = vrf;
#endif // DOLBY_END
            // Delegate volume control to effect in track effect chain if needed
            if (chain != 0 && chain->setVolume_l(&vl, &vr)) {
                // Do not ramp volume if volume is controlled by effect
                param = AudioMixer::VOLUME;
                // Update remaining floating point volume levels
                vlf = (float)vl / (1 << 24);
                vrf = (float)vr / (1 << 24);
                track->mHasVolumeController = true;
            } else {
                // force no volume ramp when volume controller was just disabled or removed
                // from effect chain to avoid volume spike
                if (track->mHasVolumeController) {
                    param = AudioMixer::VOLUME;
                }
                track->mHasVolumeController = false;
            }
#ifdef DOLBY_ENABLE // DOLBY_DAP_PREGAIN
            // Select the maximum volume by scanning all the active audio tracks but not the output one.
            if (!track->isOutputTrack() && !EffectDapController::instance()->bypassTrack(track)) {
                if (track->mHasVolumeController) {
                    const float scaleto8_24 = MAX_GAIN_INT * MAX_GAIN_INT;
                    max_vol = max(max_vol, max(((uint32_t) (scaleto8_24 * vlf * dvlf)), ((uint32_t) (scaleto8_24 * vrf * dvrf))));
                } else {
                    max_vol = max(max_vol, max(vl, vr));
                }
            }
#endif // DOLBY_END

            // XXX: these things DON'T need to be done each time
            mAudioMixer->setBufferProvider(name, track);
            mAudioMixer->enable(name);

            mAudioMixer->setParameter(name, param, AudioMixer::VOLUME0, &vlf);
            mAudioMixer->setParameter(name, param, AudioMixer::VOLUME1, &vrf);
            mAudioMixer->setParameter(name, param, AudioMixer::AUXLEVEL, &vaf);
            mAudioMixer->setParameter(
                name,
                AudioMixer::TRACK,
                AudioMixer::FORMAT, (void *)track->format());
            mAudioMixer->setParameter(
                name,
                AudioMixer::TRACK,
                AudioMixer::CHANNEL_MASK, (void *)(uintptr_t)track->channelMask());
            mAudioMixer->setParameter(
                name,
                AudioMixer::TRACK,
                AudioMixer::MIXER_CHANNEL_MASK, (void *)(uintptr_t)mChannelMask);
            // limit track sample rate to 2 x output sample rate, which changes at re-configuration
            uint32_t maxSampleRate = mSampleRate * AUDIO_RESAMPLER_DOWN_RATIO_MAX;
            uint32_t reqSampleRate = track->mAudioTrackServerProxy->getSampleRate();
            if (reqSampleRate == 0) {
                reqSampleRate = mSampleRate;
            } else if (reqSampleRate > maxSampleRate) {
                reqSampleRate = maxSampleRate;
            }
            mAudioMixer->setParameter(
                name,
                AudioMixer::RESAMPLE,
                AudioMixer::SAMPLE_RATE,
                (void *)(uintptr_t)reqSampleRate);

            AudioPlaybackRate playbackRate = track->mAudioTrackServerProxy->getPlaybackRate();
            mAudioMixer->setParameter(
                name,
                AudioMixer::TIMESTRETCH,
                AudioMixer::PLAYBACK_RATE,
                &playbackRate);

            /*
             * Select the appropriate output buffer for the track.
             *
             * Tracks with effects go into their own effects chain buffer
             * and from there into either mEffectBuffer or mSinkBuffer.
             *
             * Other tracks can use mMixerBuffer for higher precision
             * channel accumulation.  If this buffer is enabled
             * (mMixerBufferEnabled true), then selected tracks will accumulate
             * into it.
             *
             */
            if (mMixerBufferEnabled
                    && (track->mainBuffer() == mSinkBuffer
                            || track->mainBuffer() == mMixerBuffer)) {
                mAudioMixer->setParameter(
                        name,
                        AudioMixer::TRACK,
                        AudioMixer::MIXER_FORMAT, (void *)mMixerBufferFormat);
                mAudioMixer->setParameter(
                        name,
                        AudioMixer::TRACK,
                        AudioMixer::MAIN_BUFFER, (void *)mMixerBuffer);
                // TODO: override track->mainBuffer()?
                mMixerBufferValid = true;
            } else {
                mAudioMixer->setParameter(
                        name,
                        AudioMixer::TRACK,
                        AudioMixer::MIXER_FORMAT, (void *)AUDIO_FORMAT_PCM_16_BIT);
                mAudioMixer->setParameter(
                        name,
                        AudioMixer::TRACK,
                        AudioMixer::MAIN_BUFFER, (void *)track->mainBuffer());
            }
            mAudioMixer->setParameter(
                name,
                AudioMixer::TRACK,
                AudioMixer::AUX_BUFFER, (void *)track->auxBuffer());

            // reset retry count
            track->mRetryCount = kMaxTrackRetries;

            // If one track is ready, set the mixer ready if:
            //  - the mixer was not ready during previous round OR
            //  - no other track is not ready
            if (mMixerStatusIgnoringFastTracks != MIXER_TRACKS_READY ||
                    mixerStatus != MIXER_TRACKS_ENABLED) {
                mixerStatus = MIXER_TRACKS_READY;
            }
        } else {
            if (framesReady < desiredFrames && !track->isStopped() && !track->isPaused()) {
                ALOGV("track(%p) underrun,  framesReady(%zu) < framesDesired(%zd)",
                        track, framesReady, desiredFrames);
                track->mAudioTrackServerProxy->tallyUnderrunFrames(desiredFrames);
            } else {
                track->mAudioTrackServerProxy->tallyUnderrunFrames(0);
            }

            // clear effect chain input buffer if an active track underruns to avoid sending
            // previous audio buffer again to effects
            chain = getEffectChain_l(track->sessionId());
            if (chain != 0) {
                chain->clearInputBuffer();
            }

            ALOGVV("track %d s=%08x [NOT READY] on thread %p", name, cblk->mServer, this);
            if ((track->sharedBuffer() != 0) || track->isTerminated() ||
                    track->isStopped() || track->isPaused()) {
                // We have consumed all the buffers of this track.
                // Remove it from the list of active tracks.
                // TODO: use actual buffer filling status instead of latency when available from
                // audio HAL
                size_t audioHALFrames = (latency_l() * mSampleRate) / 1000;
                int64_t framesWritten = mBytesWritten / mFrameSize;
                if (mStandby || track->presentationComplete(framesWritten, audioHALFrames)) {
                    if (track->isStopped()) {
                        track->reset();
                    }
#ifdef DOLBY_ENABLE // DOLBY_UDC_VIRTUALIZE_AUDIO
                    // Since this track is being removed from active tracks, check if we
                    // should re-enable content processing in DAP.
                    EffectDapController::instance()->trackStateChanged(track->mId, track->mState);
#endif // DOLBY_END
                    tracksToRemove->add(track);
                }
            } else {
                // No buffers for this track. Give it a few chances to
                // fill a buffer, then remove it from active list.
                if (--(track->mRetryCount) <= 0) {
                    ALOGI("BUFFER TIMEOUT: remove(%d) from active list on thread %p", name, this);
                    tracksToRemove->add(track);
                    // indicate to client process that the track was disabled because of underrun;
                    // it will then automatically call start() when data is available
                    track->disable();
                // If one track is not ready, mark the mixer also not ready if:
                //  - the mixer was ready during previous round OR
                //  - no other track is ready
                } else if (mMixerStatusIgnoringFastTracks == MIXER_TRACKS_READY ||
                                mixerStatus != MIXER_TRACKS_READY) {
                    mixerStatus = MIXER_TRACKS_ENABLED;
                }
            }
            mAudioMixer->disable(name);
        }

        }   // local variable scope to avoid goto warning

    }

    // Push the new FastMixer state if necessary
    bool pauseAudioWatchdog = false;
    if (didModify) {
        state->mFastTracksGen++;
        // if the fast mixer was active, but now there are no fast tracks, then put it in cold idle
        if (kUseFastMixer == FastMixer_Dynamic &&
                /* MODIFIED-BEGIN by baowen.zhang, 2016-12-22,BUG-3773166*/
                //state->mCommand == FastMixerState::MIX_WRITE && state->mTrackMask <= 1) {
                (state->mCommand & FastMixerState::MIX_WRITE) && state->mTrackMask <= 1) {
                /* MODIFIED-END by baowen.zhang,BUG-3773166*/
            state->mCommand = FastMixerState::COLD_IDLE;
            state->mColdFutexAddr = &mFastMixerFutex;
            state->mColdGen++;
            mFastMixerFutex = 0;
            if (kUseFastMixer == FastMixer_Dynamic) {
                mNormalSink = mOutputSink;
            }
            // If we go into cold idle, need to wait for acknowledgement
            // so that fast mixer stops doing I/O.
            block = FastMixerStateQueue::BLOCK_UNTIL_ACKED;
            pauseAudioWatchdog = true;
        }
    }
    if (sq != NULL) {
        sq->end(didModify);
        sq->push(block);
    }
#ifdef AUDIO_WATCHDOG
    if (pauseAudioWatchdog && mAudioWatchdog != 0) {
        mAudioWatchdog->pause();
    }
#endif

    // Now perform the deferred reset on fast tracks that have stopped
    while (resetMask != 0) {
        size_t i = __builtin_ctz(resetMask);
        ALOG_ASSERT(i < count);
        resetMask &= ~(1 << i);
        sp t = mActiveTracks[i].promote();
        if (t == 0) {
            continue;
        }
        Track* track = t.get();
        ALOG_ASSERT(track->isFastTrack() && track->isStopped());
        track->reset();
    }

    // remove all the tracks that need to be...
    removeTracks_l(*tracksToRemove);

    if (getEffectChain_l(AUDIO_SESSION_OUTPUT_MIX) != 0) {
        mEffectBufferValid = true;
    }

    if (mEffectBufferValid) {
        // as long as there are effects we should clear the effects buffer, to avoid
        // passing a non-clean buffer to the effect chain
        memset(mEffectBuffer, 0, mEffectBufferSize);
    }
    // sink or mix buffer must be cleared if all tracks are connected to an
    // effect chain as in this case the mixer will not write to the sink or mix buffer
    // and track effects will accumulate into it
    if ((mBytesRemaining == 0) && ((mixedTracks != 0 && mixedTracks == tracksWithEffect) ||
            (mixedTracks == 0 && fastTracks > 0))) {
        // FIXME as a performance optimization, should remember previous zero status
        if (mMixerBufferValid) {
            memset(mMixerBuffer, 0, mMixerBufferSize);
            // TODO: In testing, mSinkBuffer below need not be cleared because
            // the PlaybackThread::threadLoop() copies mMixerBuffer into mSinkBuffer
            // after mixing.
            //
            // To enforce this guarantee:
            // ((mixedTracks != 0 && mixedTracks == tracksWithEffect) ||
            // (mixedTracks == 0 && fastTracks > 0))
            // must imply MIXER_TRACKS_READY.
            // Later, we may clear buffers regardless, and skip much of this logic.
        }
        // FIXME as a performance optimization, should remember previous zero status
        memset(mSinkBuffer, 0, mNormalFrameCount * mFrameSize);
    }

    // if any fast tracks, then status is ready
    mMixerStatusIgnoringFastTracks = mixerStatus;
    if (fastTracks > 0) {
        mixerStatus = MIXER_TRACKS_READY;
    }
#ifdef DOLBY_ENABLE
    // DOLBY_DAP_BYPASS_SOUND_TYPES
    EffectDapController::instance()->checkForBypass(mActiveTracks, mOutput->flags, mId);
    // DOLBY_DAP_PREGAIN
    // Skip the DS pregain setting if there're no active tracks, or all the active tracks are pausing ones,
    // so that the last pregain will be adopted and zero volume level will not be sent in the 2 cases above.
    if (mMixerStatusIgnoringFastTracks == MIXER_TRACKS_READY) {
        EffectDapController::instance()->updatePregain(mType, mId, mOutput->flags, max_vol);
    }
#endif // DOLBY_END
    return mixerStatus;
}

AudioTrack.cpp
createTrack_l()：
// mTransfer, mCbf, mUserData, mFormat, mFrameSize, mFlags
// mFlags is also assigned by createTrack_l(), but not the bit we care about.

set():
initializing a shared buffer AudioTrack via set()