首页 > 代码库 > vsync信号产生与分发

vsync信号产生与分发

以下分析基于android 4.4代码

 

vsync信号的产生、分发涉及到以下几个类,先主要了解下他们各自的功能:

HWComposer:产生hardware vsync,post fb

VSyncThread : 如果没有硬件支持,那么通过软件方式模拟hw vsync
DispSync,DispSyncThread: 接受HWComposer的hw vsync信号作为校准,开始模拟产生vsync信号+偏移,并且会不时地进行校准,如postComposition后。
EventControlThread:sf中的一个线程,仅用来控制hw vsync开关
EventThread:负责分发vsync到sf或app

DispSyncSource:EventThread和DispSyncThread的信息传递者, 把vsync信号从DispSyncThread传递到EventThread;同时可以用来设置相位偏移参数。

 

在4.4之后,vsync信号不再完全由硬件产生,hw vsync信号主要用来做时间校准,vsync信号发生者是DispSyncThread, 当DispSyncThread产生的信号得到校准后,hw vsync会被关闭。

几个问题分析:

1.硬件vsync开启-关闭过程
(1)HWComposer在hwc_composer_device_1中注册回调:
mCBContext->procs.vsync = &hook_vsync;
mHwc->registerProcs(mHwc, &mCBContext->procs);
(2)vsync过来:HWComposer::hook_vsync => HWComposer::vsync
(3)HWComposer::vsync=>mEventHandler.onVSyncReceived(disp, timestamp);
mEventHandler即surfaceflinger
(4)SurfaceFlinger::onVSyncReceived:

needsHwVsync = mPrimaryDispSync.addResyncSample(timestamp); //加入样本,校准mPrimaryDispSync(DispSync),返回是否还需要hwsync
if (needsHwVsync) {
enableHardwareVsync(); //未同步,接续接受hw vsync
} else {
disableHardwareVsync(false); //已同步,关闭hw vsync
}

 

开启或关闭硬vsync:
SurfaceFlinger::enableHardwareVsync() => mEventControlThread->setVsyncEnabled(true); //sf专门启动了一个线程EventControlThread来开关硬件vsync

 

2.DispSyncThread 信号的产生:
(1) 在sf类中声明一个成员变量:

class SurfaceFlinger
{
    ...
    DispSync mPrimaryDispSync;
    ...
}

 

(2) 构造函数里启动DispSyncThread线程

DispSync::DispSync() {
    mThread = new DispSyncThread();
    mThread->run("DispSync", PRIORITY_URGENT_DISPLAY +         PRIORITY_MORE_FAVORABLE);

    reset();
    beginResync();
    ...
}

(3) 进入DispSyncThread::threadLoop(): 计算vsync周期,调用所有存在mEventListeners的Listener的callback

fireCallbackInvocations()方法中调用callbacks[i].mCallback->onDispSyncEvent(callbacks[i].mEventTime);
mCallback的类型为DispSync::Callback

(4) mEventListeners的注册通过调用DispSync::addEventListener()

(5) 谁注册成为了vsync事件的监听者? 是DispSyncSource
DispSyncSource继承自DispSync::Callback, 在它setVSysncEnabled(true)中调用:

status_t err = mDispSync->addEventListener(mPhaseOffset,
static_cast<DispSync::Callback*>(this));  //DispSyncSource把自己设为了DispSyncThread vsync信号的监听者

所以vsync消息传到了DispSyncSource中的onDispSyncEvent(nsecs_t when)中:
virtual void onDispSyncEvent(nsecs_t when) {
sp<VSyncSource::Callback> callback;
{
  Mutex::Autolock lock(mMutex);
  callback = mCallback;

  if (mTraceVsync) {
  mValue = (mValue + 1) % 2;
  ATRACE_INT("VSYNC", mValue);
  }
}

if (callback != NULL) {
  callback->onVSyncEvent(when); //又把vsync事件传递到了它自己的mCallback里, 这个mCallback 在EventThread::enableVSyncLocked()中设置, 就是EventThread, 这样vsync传递到了EventThread里
}
}

 

(6)继续看谁又调用了DispSyncSource::setVSyncEnabled()?
是EventThread::enableVSyncLocked():

void EventThread::enableVSyncLocked() {
if (!mUseSoftwareVSync) {
// never enable h/w VSYNC when screen is off
  if (!mVsyncEnabled) {
  mVsyncEnabled = true;
  mVSyncSource->setCallback(static_cast<VSyncSource::Callback*>(this)); //EventThread设为DispSyncSource的callback
  mVSyncSource->setVSyncEnabled(true);
  mPowerHAL.vsyncHint(true);
}
}
mDebugVsyncEnabled = true;
}

 

(7)EventThread::enableVSyncLocked()又被EventThread::waitForEvent()调用

Vector< sp<EventThread::Connection> > EventThread::waitForEvent(
        DisplayEventReceiver::Event* event)
{
 
    ...
    // Here we figure out if we need to enable or disable vsyncs
        if (timestamp && !waitForVSync) {
            // we received a VSYNC but we have no clients
            // don‘t report it, and disable VSYNC events
            disableVSyncLocked();
        } else if (!timestamp && waitForVSync) {
            // we have at least one client, so we want vsync enabled
            // (TODO: this function is called right after we finish
            // notifying clients of a vsync, so this call will be made
            // at the vsync rate, e.g. 60fps.  If we can accurately
            // track the current state we could avoid making this call
            // so often.)
            enableVSyncLocked();
        }
    ...      
}

waitForEvent()的逻辑是:

接到vsync信号,但是当前EventThread中没有请求 vsync 的connection, EventThread向下不再监听vsync

EventThread中有请求 vsync 的connection, EventThread继续监听vsync

 

总结: vsync传递路径 DispSyncThread => DispSyncSource => EventThread
DispSyncSource: 这个类比较简单,其实就是vsync传递者, 同时负责传递相位偏移phase offset到dispsyncThread。
EventThread: 负责接收vsync, 分发给sf或者app


所以在Sf init中的代码就比较好理解了

// start the EventThread
sp<VSyncSource> vsyncSrc = http://www.mamicode.com/new DispSyncSource(&mPrimaryDispSync,
vsyncPhaseOffsetNs, true);
mEventThread = new EventThread(vsyncSrc);
sp<VSyncSource> sfVsyncSrc = http://www.mamicode.com/new DispSyncSource(&mPrimaryDispSync,
sfVsyncPhaseOffsetNs, false);
mSFEventThread = new EventThread(sfVsyncSrc);
//上面代码注册了两个vsync信号的监听者, 当vsync产生时,分发给EventThread,这里有两个,一个处理sf,一个处理app

mEventQueue.setEventThread(mSFEventThread); //

mEventControlThread = new EventControlThread(this);
mEventControlThread->run("EventControl", PRIORITY_URGENT_DISPLAY);

 

3. 两个EventThread 线程是怎样把vsync分发给sf和app的?

(1)app与EventThread
EventThread里有一个 class Connection : public BnDisplayEventConnection,

BnDisplayEventConnection 继承自 IDisplayEventConnection

这个接口实现了:
getDataChannel()
setVsyncRate()
requestNextVsync()

App端请求vsync:

app端通过DisplayEventReceiver 中的sp<IDisplayEventConnection> mEventConnection 来requestNextVsync
mEventConnection实际上是BpDisplayEventConnection, 所以requestNextVsync是通过Binder通信从app端传递到EventThread并由它来处理。
同时通过mEventConnection->getDataChannel(); 保存一个BitTube类型的mDataChannel。

在EventThread端:
app调用createEventConnection创建DisplayEventConnection时,EventThread将这个connection注册到它的mDisplayEventConnections集合里
当vsync过来时,通知所有connection

App端接受vsync是通过建立连接时保存的那个BitTube:

//file:android_view_DisplayEventReceiver.cpp
status_t NativeDisplayEventReceiver::initialize() {
  status_t result = mReceiver.initCheck();
  f (result) {
    ALOGW("Failed to initialize display event receiver, status=%d", result);
    return result;
  }

  int rc = mMessageQueue->getLooper()->addFd(mReceiver.getFd(), 0, ALOOPER_EVENT_INPUT, // 监听BitTube的fd,有数据时,调用handler
  this, NULL);
  if (rc < 0) {
    return UNKNOWN_ERROR;
  }
  return OK;
}

 

所以app接受vsync信号是通过监听BitTube的fd, BitTube底层是通过socket实现。

 

(2) sf与EventThread
SurfaceFlinger::init()中调用mEventQueue.setEventThread(mSFEventThread);

void MessageQueue::setEventThread(const sp<EventThread>& eventThread)
{
  mEventThread = eventThread;
  mEvents = eventThread->createEventConnection();
  mEventTube = mEvents->getDataChannel();
  mLooper->addFd(mEventTube->getFd(), 0, ALOOPER_EVENT_INPUT,
  MessageQueue::cb_eventReceiver, this);
}

在sf进程中直接监听BitTube的fd,当有vsync过来时, 直接由SF的MessageQueue处理

vsync信号产生与分发