The vulnerability of the way predictor implementation in AMD processor CPUs’ L1D data cache allows attackers to restore the encryption key in vulnerable AES implementations, reduce the effectiveness of ASLR protection, or extract data from the processor’s core.

The vulnerability of the way predictor implementation in AMD processors’ L1D cache involves collisions with the hash function used to calculate tags, which are used to locate a channel in the L1D cache. Exploiting this vulnerability can allow an attacker to recover encryption keys in vulnerable A...