Abstract:
Aiming to expose security risks in hardware designs, we describe a novel usage of symbolic simulation that led to discoveries of previously unknown potential local data l...Show MoreMetadata
Abstract:
Aiming to expose security risks in hardware designs, we describe a novel usage of symbolic simulation that led to discoveries of previously unknown potential local data leakages on an Intel Core processor design. Symbolic simulation is an established formal verification method, the main vehicle for verification of arithmetic data-paths in Intel Core processor designs for twenty years. It extends traditional simulation by allowing symbolic variables in the stimulus, covering the circuit behavior for all possible values simultaneously. A special trait of symbolic simulation is that every variable has a name. In the security context, named values allow us to know the exact origin of data and identify data leakages by determining whether values are expected to be read by an operation or present a risk. Leveraging the existing formal verification infrastructure and observing an operation’s data dependencies we could identify local leaks without the need to have a complete functional specification for the operation.
Published in: 2021 Formal Methods in Computer Aided Design (FMCAD)
Date of Conference: 19-22 October 2021
Date Added to IEEE Xplore: 29 November 2021
ISBN Information: