Enhance GNNs with Reliable Confidence Estimation via Adversarial Calibration Learning