LAMeD: LLM-generated Annotations for Memory Leak Detection

2025-05-05

AI Summary

Existing static analyzers for memory leak detection suffer from low efficiency and poor scalability due to heavy reliance on manual function labeling as “sources” or “sinks.” To address this, this paper proposes the first LLM-based automated function-level semantic labeling method. Leveraging prompt engineering and fine-grained function behavior understanding, the approach enables large language models to generate source/sink annotations directly consumable by industrial-grade analyzers (e.g., Cooddy), eliminating manual effort and supporting zero-shot generalization to third-party libraries. By incorporating semantically grounded labels, the method mitigates path explosion and enhances data-flow tracking precision. Evaluated on real-world C/C++ projects, it achieves a 37% increase in memory leak detection rate, a 29% reduction in false positives, and reduces annotation time from hours to seconds. This work establishes a new, efficient, scalable, and semantics-driven paradigm for static analysis.

Technology Category

• Machine Learning: Large Multimodal Models (LMMs)
• Natural Language Processing: Code Generation / Program Synthesis from Natural Language
• Cognitive Modeling & Cognitive Systems: Symbolic Representations

Application Category

• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Search and Retrieval-Augmented AI: Search Tool Learning with LLM: Teaching LLMs to invoke search and make use of retrieved information

Abstract

Static analysis tools are widely used to detect software bugs and vulnerabilities but often struggle with scalability and efficiency in complex codebases. Traditional approaches rely on manually crafted annotations -- labeling functions as sources or sinks -- to track data flows, e.g., ensuring that allocated memory is eventually freed, and code analysis tools such as CodeQL, Infer, or Cooddy can use function specifications, but manual annotation is laborious and error-prone, especially for large or third-party libraries. We present LAMeD (LLM-generated Annotations for Memory leak Detection), a novel approach that leverages large language models (LLMs) to automatically generate function-specific annotations. When integrated with analyzers such as Cooddy, LAMeD significantly improves memory leak detection and reduces path explosion. We also suggest directions for extending LAMeD to broader code analysis.

Problem

• Automating memory leak detection annotations using LLMs
• Reducing manual effort in static analysis annotation
• Improving scalability of code analysis tools

Innovation

• Uses LLMs to generate function annotations
• Improves memory leak detection accuracy
• Reduces path explosion in static analysis