Formal fault analysis of branch predictors: attacking countermeasures of asymmetric key ciphers

Abstract

Implementations of asymmetric key algorithm have been threatened via timing side channels due to the behavior of the underlying branch predictors. However, the effect of faults on such predictors and the consequences thereof on the security of crypto-algorithms have not been studied. Motivated by the fact that unknown branch predictors of standard processors bear a strong correlation with 2-bit dynamic predictors, this paper develops a formal analysis of such a bimodal predictor under the effect of faults. Assuming a popular bit-flip fault model, the analysis shows that differences of branch misses under the effect of such faults can be exploited to attack implementations of RSA-like asymmetric key algorithms, based on square and multiplication operations. Furthermore, these attacks can be also threatening against Montgomery ladder of CRT-RSA (RSA implemented using Chinese Remainder Theorem) and even against fault attack countermeasures which stop or randomize the output in case of a fault. The theoretical claims have been substantiated by detailed fault simulations, where the difference of branch misses has been observed using the “perf” tool in Linux.