Clover is a novel hybrid framework that combines static analysis with large language models to automatically detect atomicity violations in interrupt-driven systems.

• Addresses a major security threat in critical systems where interrupts disrupt operations on shared resources
• Overcomes traditional detection challenges including vast program state spaces and complex domain knowledge
• Integrates static analysis techniques with LLM capabilities for more effective violation detection
• Enables improved security and reliability in interrupt-driven programs for critical infrastructure

This research offers significant security advancements for embedded systems, IoT devices, and other critical infrastructure where undetected atomicity violations could lead to catastrophic failures or security breaches.

Automated detection of atomicity violations in large-scale systems