Detecting Code Reuse Attacks with Branch Prediction

Code reuse attacks (CRAs) allow attackers to produce malicious results by using legitimate code binaries in memory. The authors propose incorporating control-flow validation into the processors instruction execution pipeline, along with a mis-prediction validation unit and a branch prediction unit, to help identify attacks.