Abstract
We present a high-assurance and high-speed implementation of the SHA-3 hash function. Our implementation is written in the Jasmin programming language, and is formally verified for functional correctness, provable security and timing attack resistance in the EasyCrypt proof assistant. Our implementation is the first to achieve simultaneously the four desirable properties (efficiency, correctness, provable security, and side-channel protection) for a non-trivial cryptographic primitive. Concretely, our mechanized proofs show that: 1) the SHA-3 hash function is indifferentiable from a random oracle, and thus is resistant against collision, first and second preimage attacks; 2) the SHA-3 hash function is correctly implemented by a vectorized x86 implementation. Furthermore, the implementation is provably protected against timing attacks in an idealized model of timing leaks. The proofs include new EasyCrypt libraries of independent interest for programmable random oracles and modular indifferentiability proofs.

Abstract
We present a new tool for the generation and verification of high-assurance high-speed machine-level cryptography implementations: a certified C compiler supporting instruction extensions to the x86. We demonstrate the practical applicability of our tool by incorporating it into supercop: a toolkit for measuring the performance of cryptographic software, which includes over 2000 different implementations. We show i. that the coverage of x86 implementations in supercop increases significantly due to the added support of instruction extensions via intrinsics and ii. that the obtained verifiably correct implementations are much closer in performance to unverified ones. We extend our compiler with a specialized type system that acts at pre-assembly level; this is the first constant-time verifier that can deal with extended instruction sets. We confirm that, by using instruction extensions, the performance penalty for verifiably constant-time code can be greatly reduced. © Springer Nature Switzerland AG 2020.