On the Insecurity of Constraint-sensitive Control Flow Transfer

We still need to defend large projects like Python and Nginx that are built on top of well-knownlegacy languages like C and C++, despite the popularity of more secure languages rising as a result of the introduction of high security capabilities. By establishing a novel control flow integrity policy called Constraint Sensitive Indirect Call (CS-iCall), this thesis focuses on safeguarding indirect calls. We would first introduce the CS-iCall notion before going into the difficulties and complexities of building an LLVM tool to locate all CS-iCalls in the input binary. By contrasting CS-iCalls with current cutting-edge CFI policies, we would also be able to comprehend the relevance of these calls.