[088]首次直播网友的问题澄清

王小二

发布于 2023-11-12 08:41:48

1620

发布于 2023-11-12 08:41:48

文章被收录于专栏：王小二的Android站王小二的Android站

背景

很开心今天完成了首次直播，很感谢那么多网友的捧场，希望真的帮助到了大家。针对直播中网友临时提的几个问题，给出我的澄清。一、问题1

如果同一时间多个view快速的调用invalidate是否会触发多次request next vsync

回答：不会

scheduleTraversals触发后会将mTraversalScheduled改成false，直到next vsync来了之之前，都不会重复触发mChoreographer.postCallback(Choreographer.CALLBACK_TRAVERSAL...)，这个函数里面会request next vsync

frameworks/base/core/java/android/view/ViewRootImpl.java

    void scheduleTraversals() {
        if (!mTraversalScheduled) {
            mTraversalScheduled = true;//这里会设置成true，所以在
            mTraversalBarrier = mHandler.getLooper().getQueue().postSyncBarrier();
            mChoreographer.postCallback(
                    Choreographer.CALLBACK_TRAVERSAL, mTraversalRunnable, null);//这里会request next vsync
            notifyRendererOfFramePending();
            pokeDrawLockIfNeeded();
        }
    }

// next vsync来了出发doTraversal
    void doTraversal() {
        if (mTraversalScheduled) {
            mTraversalScheduled = false;//重置mTraversalScheduled
            mHandler.getLooper().getQueue().removeSyncBarrier(mTraversalBarrier);

            if (mProfile) {
                Debug.startMethodTracing("ViewAncestor");
            }

            performTraversals();

            if (mProfile) {
                Debug.stopMethodTracing();
                mProfile = false;
            }
        }
    }

二、问题2

硬件vsync的校准是否会影响界面的卡顿

回答:大概率不会

整个校准有两个路径分别是onComposerHalVsync和postComposition

路径1

onComposerHalVsync是硬件vsync的回调，运行在Hwbinder线程上，不会影响到sf的主线程的合成任务。

void SurfaceFlinger::onComposerHalVsync(hal::HWDisplayId hwcDisplayId, int64_t timestamp,
                                        std::optional<hal::VsyncPeriodNanos> vsyncPeriod) {
    mScheduler->addResyncSample(timestamp, vsyncPeriod, &periodFlushed);
}

void Scheduler::addResyncSample(nsecs_t timestamp, std::optional<nsecs_t> hwcVsyncPeriod,
                                bool* periodFlushed) {
    bool needsHwVsync = false;
    *periodFlushed = false;
    { // Scope for the lock
        std::lock_guard<std::mutex> lock(mHWVsyncLock);
        if (mPrimaryHWVsyncEnabled) {
            needsHwVsync = mVsyncSchedule.controller->addHwVsyncTimestamp(timestamp, hwcVsyncPeriod,
                                                                          periodFlushed);
        }
    }
}

bool VSyncReactor::addHwVsyncTimestamp(nsecs_t timestamp, std::optional<nsecs_t> hwcVsyncPeriod,
                                       bool* periodFlushed) {
    } else {
        ATRACE_NAME("VSR: adding sample");
        *periodFlushed = false;
        mTracker.addVsyncTimestamp(timestamp);
        mMoreSamplesNeeded = mTracker.needsMoreSamples();
    }
}

bool VSyncPredictor::addVsyncTimestamp(nsecs_t timestamp) {
//进行校准更新模型参数
}

路径2

postComposition是每次sf主线程合成接触之后调用的，将上一帧的presentfence的signal时间作为hw vsync采样数据。

void SurfaceFlinger::postComposition() {
    if (display && display->isPrimary() && display->getPowerMode() == hal::PowerMode::ON &&
        mPreviousPresentFences[0].fenceTime->isValid()) {
        mScheduler->addPresentFence(mPreviousPresentFences[0].fenceTime);
    }
}

void Scheduler::addPresentFence(const std::shared_ptr<FenceTime>& fenceTime) {
    if (mVsyncSchedule.controller->addPresentFence(fenceTime)) {
        enableHardwareVsync();
    } else {
        disableHardwareVsync(false);
    }
}

bool VSyncReactor::addPresentFence(const std::shared_ptr<android::FenceTime>& fence) {
    timestampAccepted &= mTracker.addVsyncTimestamp(signalTime);
}

bool VSyncPredictor::addVsyncTimestamp(nsecs_t timestamp) {
//进行校准更新模型参数
}

虽然postComposition运行在sf的主线程，总体来说，addVsyncTimestamp的计算不会太耗时，也不会造成sf合成的卡顿，