ARMISTICE: Microarchitectural Leakage Modeling for Masked Software Formal Verification

Side-channel attacks are powerful for retrieving secret data by exploiting physical measurements, such as power consumption or electromagnetic emissions. Masking is a popular countermeasure it can be proven secure against an attacker model. In practice, software-masked implementations suffer from security reduction due to mismatch between the considered leakage sources in proof and real ones, which depend on microarchitecture. We propose ARMISTICE, framework formally verifying absence of first-order masked taking into account modeled microarchitectural leakage. As concept, we present modeling Arm Cortex-M3 core its RTL description test vectors, well memory STM32F1 board, exclusively using vectors. show that, with these models, ARMISTICE pinpoints vulnerable instructions real-world helps design software practically secure.