Researchers Find New Way to Sneak Past Intel and AMD's Spectre Chip Defenses

Security researchers from MIT CSAIL have identified a new attack method, dubbed Interrupt Injection, that can bypass Spectre v2 defenses built into Intel and AMD processors. The technique works by carefully timing a hardware interrupt to land in a brief gap between when a processor clears its branch predictor and when the operating system kernel actually uses it. This timing trick allows an unprivileged program to "re-poison" the predictor right after the security check has already run, effectively undoing the protection.

The researchers, Daniël Trujillo and Mengjia Yan, demonstrated the attack on an AMD Zen 2 machine running Linux 6.14 with all default Spectre v2 mitigations turned on. Spectre-style attacks exploit how modern CPUs try to predict what instructions to run next, and defenses like this one are designed to stop attackers from tricking the processor into leaking sensitive data. This new finding shows that even fully patched, up-to-date systems using standard protections may still be vulnerable to a determined attacker with local code execution.

While this type of attack requires an attacker to already run code on the target system, it highlights a broader truth: chip-level security defenses are complex and can have subtle gaps. For small businesses, this is a reminder that hardware vulnerabilities are an ongoing part of the threat landscape, even if they are less common than everyday risks like phishing or weak passwords.