Researchers have uncovered a new Spectre variant that exploits the way processors remember the locations of dynamically generated code, demonstrating that sensitive Linux memory can be exposed even when existing speculative-execution protections are enabled.
