Monitoring Translation Lookahead Buffers to Detect Code Injection Attacks.

By identifying memory pages that external I/O operations have modified, a proposed scheme blocks malicious injected code activation, accurately distinguishing an attack from legitimate code injection with negligible performance impact and no changes to the user application.