Merge "Change approach for windowPositionLostRT" into nyc-dev (16038a0c) · Commits · e / os / android_frameworks_base

core/java/android/view/SurfaceView.java

+44 −17

Original line number	Diff line number	Diff line
		@@ -124,6 +124,7 @@ public class SurfaceView extends View {
		int mWindowType = WindowManager.LayoutParams.TYPE_APPLICATION_MEDIA;

		boolean mIsCreating = false;
		private volatile boolean mRtHandlingPositionUpdates = false;

		final Handler mHandler = new Handler() {
		@Override
		@@ -649,7 +650,9 @@ public class SurfaceView extends View {
		TAG, "Layout: x=" + mLayout.x + " y=" + mLayout.y +
		" w=" + mLayout.width + " h=" + mLayout.height +
		", frame=" + mSurfaceFrame);
		} else if (!isHardwareAccelerated()) {
		} else {
		// Calculate the window position in case RT loses the window
		// and we need to fallback to a UI-thread driven position update
		getLocationInWindow(mLocation);
		final boolean positionChanged = mWindowSpaceLeft != mLocation[0]
		\|\| mWindowSpaceTop != mLocation[1];
		@@ -670,8 +673,9 @@ public class SurfaceView extends View {
		mTranslator.translateRectInAppWindowToScreen(mWinFrame);
		}

		if (!isHardwareAccelerated() \|\| !mRtHandlingPositionUpdates) {
		try {
		Log.d(TAG, String.format("%d updateWindowPosition UI, " +
		if (DEBUG) Log.d(TAG, String.format("%d updateWindowPosition UI, " +
		"postion = [%d, %d, %d, %d]", System.identityHashCode(this),
		mWinFrame.left, mWinFrame.top,
		mWinFrame.right, mWinFrame.bottom));
		@@ -683,6 +687,7 @@ public class SurfaceView extends View {
		}
		}
		}
		}

		private Rect mRTLastReportedPosition = new Rect();

		@@ -698,6 +703,15 @@ public class SurfaceView extends View {
		// Guess we got detached, that sucks
		return;
		}
		// TODO: This is teensy bit racey in that a brand new SurfaceView moving on
		// its 2nd frame if RenderThread is running slowly could potentially see
		// this as false, enter the branch, get pre-empted, then this comes along
		// and reports a new position, then the UI thread resumes and reports
		// its position. This could therefore be de-sync'd in that interval, but
		// the synchronization would violate the rule that RT must never block
		// on the UI thread which would open up potential deadlocks. The risk of
		// a single-frame desync is therefore preferable for now.
		mRtHandlingPositionUpdates = true;
		if (mRTLastReportedPosition.left == left
		&& mRTLastReportedPosition.top == top
		&& mRTLastReportedPosition.right == right
		@@ -731,13 +745,26 @@ public class SurfaceView extends View {
		Log.d(TAG, String.format("%d windowPositionLostRT RT, frameNr = %d",
		System.identityHashCode(this), frameNumber));
		}
		// TODO: This is a bit of a hack as we don't have an API to report to WM
		// to hide a window with a frameNumber, so just shift the window very far into
		// negative space which will do effectively the same thing.
		// Use the last reported size to avoid influencing the size of the bufferqueue
		int x = -1000 - mRTLastReportedPosition.width();
		int y = -1000 - mRTLastReportedPosition.height();
		updateWindowPositionRT(frameNumber, x, y, -1000, -1000);
		if (mRtHandlingPositionUpdates) {
		mRtHandlingPositionUpdates = false;
		// This callback will happen while the UI thread is blocked, so we can
		// safely access other member variables at this time.
		// So do what the UI thread would have done if RT wasn't handling position
		// updates.
		if (!mWinFrame.isEmpty() && !mWinFrame.equals(mRTLastReportedPosition)) {
		try {
		if (DEBUG) Log.d(TAG, String.format("%d updateWindowPosition, " +
		"postion = [%d, %d, %d, %d]", System.identityHashCode(this),
		mWinFrame.left, mWinFrame.top,
		mWinFrame.right, mWinFrame.bottom));
		mSession.repositionChild(mWindow, mWinFrame.left, mWinFrame.top,
		mWinFrame.right, mWinFrame.bottom, frameNumber, mWinFrame);
		} catch (RemoteException ex) {
		Log.e(TAG, "Exception from relayout", ex);
		}
		}
		mRTLastReportedPosition.setEmpty();
		}
		}

		private SurfaceHolder.Callback[] getSurfaceCallbacks() {