It is widely accepted that the top sides of neural networks convey high-level semantic features and the bottom sides contain low-level details. Therefore, most of recent salient object detection models aim at designing effective fusion strategies for the side-output features of convolutional neural networks (CNNs). Although significant progress has been achieved in this direction, the network architectures become more and more complex, which will make the future improvement difficult and heavily engineered. Moreover, the manually designed fusion strategies would be sub-optimal due to the large search space of possible solutions. To address above problems, we propose an Automatic Top-Down Fusion (ATDF) model, in which the global information at the top sides are flowed into bottom sides to guide the learning of low layers. We design a novel module at each side to control the information flowed into a specific side, called valve module, by which each side is expected to receive the necessary top information. We perform extensive experiments to demonstrate that ATDF is simple yet effective and thus opens a new path for saliency detection. Code is available at https://github.com/yun-liu/ATDF.