Thais Moreira Hamasaki

A sequence of processor instructions can lead to unexpected behavior in some Intel processors, potentially allowing an authenticated user to enable escalation of privilege, information disclosure, or denial of service via local access. This issue is related to the interpretation of redundant prefixes and can cause the system to crash or become unresponsive. The vulnerability, termed Reptar, primarily affects multi-tenant virtualized environments where multiple virtual machines are run on the same host.

The issue is related to the implementation of the Indirect Branch Prediction Barrier (IBPB) mechanism in AMD processor microcode, which may lead to a potential information disclosure due to incorrect clearing of the Return Address Stack (RAS). This could allow an attacker to obtain sensitive information using a side channel. The vulnerability is being actively exploited.

A potential issue in AMD processors using frequency scaling may allow an authenticated attacker to execute a timing attack, potentially enabling information disclosure. The vulnerability is related to the implementation of Dynamic Voltage and Frequency Scaling (DVFS) in AMD processor microcode, which is associated with inadequate protection of physical side channels. Exploitation of the issue may allow a remote attacker to gain unauthorized access to protected information using a side channel.

The issue is related to the LFENCE/JMP module in AMD CPUs, specifically concerning the branch prediction mechanism. This allows an attacker to access protected memory from a program without the necessary privileges by exploiting indirect branch prediction.