Android 谈谈我理解的Surface绘制原理

源码：8.x系统
我们知道ViewRootImpl是处理绘制流程的地方，具体分析如下：

ViewRootImpl.java
private void performTraversals() {
    ······
    performMeasure(childWidthMeasureSpec, childHeightMeasureSpec);
    ······
    performLayout(lp, mWidth, mHeight);
    ······
    performDraw();
    ······
}

测量、布局、绘制，我们重点看绘制performDraw
private void performDraw() {
    ······
    draw(fullRedrawNeeded);
    ······
}

private void draw(boolean fullRedrawNeeded) {
    ······
    if (!dirty.isEmpty() || mIsAnimating || accessibilityFocusDirty) {
        if (mAttachInfo.mThreadedRenderer != null && mAttachInfo.mThreadedRenderer.isEnabled()) {
            ······//硬件加速
        } else {
            ·····
            //drawSoftware为软件加速
            if (!drawSoftware(surface, mAttachInfo, xOffset, yOffset, scalingRequired, dirty)){
                ······
            }
        }
    }
}

private boolean drawSoftware(Surface surface, AttachInfo attachInfo, int xoff, int yoff,
            boolean scalingRequired, Rect dirty) {
     ······
     canvas = mSurface.lockCanvas(dirty);
     ······
     mView.draw(canvas);
     ······
     surface.unlockCanvasAndPost(canvas);
     ······     
}
终于来到了，我们需要重点分析的地方，接下来我们重点分析：
1.mSurface.lockCanvas(dirty)
2.mView.draw(canvas)是否可以用其他的替代?
3.surface.unlockCanvasAndPost(canvas)

重点分析一：mSurface.lockCanvas(dirty)

Surface.java
public Canvas lockCanvas(Rect inOutDirty)
         throws Surface.OutOfResourcesException, IllegalArgumentException {
      ·····
      mLockedObject = nativeLockCanvas(mNativeObject, mCanvas, inOutDirty);
      return mCanvas;       
}
首先，我们重点关注点在返回值mCanvas，而mCanvas初始化的地方为private final Canvas mCanvas = new CompatibleCanvas();
，说明这个mCanvas初始化的时候是空壳。提问：canvas和surface有什么必然的关联？

其次，我们重点看nativeLockCanvas具体做了什么？
android_view_Surface.cpp
static jlong nativeLockCanvas(JNIEnv* env, jclass clazz,
        jlong nativeObject, jobject canvasObj, jobject dirtyRectObj) {
    sp<Surface> surface(reinterpret_cast<Surface *>(nativeObject));
    ······
    //关注点1
    ANativeWindow_Buffer outBuffer;
    status_t err = surface->lock(&outBuffer, dirtyRectPtr);
    ······
    //关注点2
    SkBitmap bitmap;
    ssize_t bpr = outBuffer.stride * bytesPerPixel(outBuffer.format);
    bitmap.setInfo(info, bpr);
    if (outBuffer.width > 0 && outBuffer.height > 0) {
        bitmap.setPixels(outBuffer.bits);
    } else {
        // be safe with an empty bitmap.
        bitmap.setPixels(NULL);
    }

    Canvas* nativeCanvas = GraphicsJNI::getNativeCanvas(env, canvasObj);
    nativeCanvas->setBitmap(bitmap);       
    ·····
}
初步分析:
关注点1：
一个buffer跟surface绑定在一起

关注点2：
bitmap和canvas绑定在一起，而bimap和buffer绑定在一起

小结论：
surface和canvas绑定在一起，用了一个bitmap，进一步说，是一个buffer。

详细分析：
关注点1：
status_t err = surface->lock(&outBuffer, dirtyRectPtr);
Surface.cpp
status_t Surface::lock(
        ANativeWindow_Buffer* outBuffer, ARect* inOutDirtyBounds)
{

      //2.backBuffer来自out，而out由dequeueBuffer函数赋值
      ANativeWindowBuffer* out;
      int fenceFd = -1;
      status_t err = dequeueBuffer(&out, &fenceFd);
      ·····
      sp<GraphicBuffer> backBuffer(GraphicBuffer::getSelf(out));
      
      //1.逆向思维分析，outBuffer是被赋值的？
      //是被backBuffer，backBuffer是怎么来的？
      mLockedBuffer = backBuffer;
      outBuffer->width  = backBuffer->width;
      outBuffer->height = backBuffer->height;
      outBuffer->stride = backBuffer->stride;
      outBuffer->format = backBuffer->format;
      outBuffer->bits   = vaddr;
}

int Surface::dequeueBuffer(android_native_buffer_t** buffer, int* fenceFd) {
    ······
    //gbuf由mGraphicBufferProducer(GBP)产生
    result = mGraphicBufferProducer->requestBuffer(buf, &gbuf);
    ······
    //逆向思维分析：buffer来自gbuf
    *buffer = gbuf.get();
}

关注点2：
java层的canvas是一个空壳，通过native层赋予了值，如下：
nativeCanvas->setBitmap(bitmap);  
所以，canvas的关键在于对这个bitmap赋予生机。

进一步小结：
surface和canvas绑定在一起，用了一个bitmap，进一步说，是一个buffer。
这个buffer是由Surface的mGraphicBufferProducer生产。

重点分析二： mView.draw(canvas)

顾明思意，canvas会把view中元素draw到自身身上。那是否可以通过其他途径来实现此意图呢？
mCanvas.drawBitmap
这种方法，就是直接赋值一张图片，简单粗暴。

例如，现在车载产品上流行的UsbCamera，在绘制usbcamera成像时，就会采用这种方法。
具体操作可以这样做：
1.SurfaceView负责展示成像
2.我们把SurfaceView中的Surface拿出来
3.之后通过Surface中的canvas.drawBitmap进行设置成像图案
因为surface可以进行跨进程传递，所以，负责展示和图像处理，可以分为两个进程执行。
这样就减小了耦合。

重点分析三：surface.unlockCanvasAndPost(canvas)

Surface.java
public void unlockCanvasAndPost(Canvas canvas) {
      if (mHwuiContext != null) {//硬件绘制
            mHwuiContext.unlockAndPost(canvas);
      } else {//软件绘制
            unlockSwCanvasAndPost(canvas);
      }    
}
我们关注软件绘制
private void unlockSwCanvasAndPost(Canvas canvas) {
    ······
    try {
         nativeUnlockCanvasAndPost(mLockedObject, canvas);
    } finally {
         ······
    }    
    ······
}

android_view_Surface.cpp
static void nativeUnlockCanvasAndPost(JNIEnv* env, jclass clazz,
        jlong nativeObject, jobject canvasObj) {
    ······

    // detach the canvas from the surface
    Canvas* nativeCanvas = GraphicsJNI::getNativeCanvas(env, canvasObj);
    nativeCanvas->setBitmap(SkBitmap());

    // unlock surface
    status_t err = surface->unlockAndPost();
    if (err < 0) {
        doThrowIAE(env);
    }
}
这里重点处理了两块：
a.把canvas的bitmap设置为了空壳，也就是此时surface与canvas已经失去深入的关联
b.我们具体分析surface->unlockAndPost()
Surface.cpp
status_t Surface::unlockAndPost()
{
    ······
    int fd = -1;
    status_t err = mLockedBuffer->unlockAsync(&fd);
    ALOGE_IF(err, "failed unlocking buffer (%p)", mLockedBuffer->handle);

    err = queueBuffer(mLockedBuffer.get(), fd);
    ALOGE_IF(err, "queueBuffer (handle=%p) failed (%s)",
            mLockedBuffer->handle, strerror(-err));

    mPostedBuffer = mLockedBuffer;
    mLockedBuffer = 0;
    return err;
}
这里，我们可以看到native层的Surface有两个buffer，一个是mPostedBuffer(frontBuffer)，一个是mLockedBuffer(backBuffer),这两个buffer在相互切换。
也就是说，当unlock的时候，mLockedBuffer就切换到了前台，而后台的buffer就变成了0.

int Surface::queueBuffer(android_native_buffer_t* buffer, int fenceFd) {
    ····
    status_t err = mGraphicBufferProducer->queueBuffer(i, input, &output);
    ····
}

小结：
Surface的buffer绘制完后，通过mGraphicBufferProducer提交。

参考学习

https://zhuanlan.zhihu.com/p/25477828
https://blog.csdn.net/huagjie/article/details/78572040