SHF: Small: Dynamic Detection of Heap-Based Bugs

SHF：小：基于堆的错误的动态检测

• 批准号: 1018038
• 负责人: Samuel Guyer
• 金额: $ 48.49万
• 依托单位: Tufts University
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2010
• 资助国家: 美国
• 起止时间: 2010-09-01 至 2014-08-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1018038&HistoricalAwards=false
• 关键词: SHF; Small; Dynamic; Detection; Heap

Identifying bugs in software continues to be a challenging, but essential problem to solve. One particularly difficult task is ensuring the integrity of large-scale data structures stored in memory. Existing bug-finding techniques, such as static analysis of the code, have not been effective on this problem, especially for complex and highly dynamic software, such as web applications.This project explores a new technique for checking data structures dynamically as the program executes. Dynamic checking is effective and precise, but must be efficient in order to avoid significantly slowing program execution. The key idea in this work is to piggyback checking on the garbage collector, which already periodically visits all data structures in the program. An efficient and precise tool for detecting data structure errors could be widely deployed to improve the reliability of critical software infrastructure.The project consists of three specific avenues of research. The first involves developing a declarative language for expressing dynamic data structure properties, building on existing techniques from static analysis and verification. The second investigates the class of properties that can be checked during a single pass of the garbage collector. The third builds on the machinery of concurrent garbage collection, allowing heap checks to proceed concurrently with the application on available extra CPU cores.

识别软件中的错误仍然是一个具有挑战性但必须解决的问题。一个特别困难的任务是确保存储在内存中的大规模数据结构的完整性。现有的错误发现技术，如代码的静态分析，在这个问题上并不有效，特别是对于复杂和高度动态的软件，如Web应用程序。本项目探索了一种新的技术，用于在程序执行时动态检查数据结构。动态检查是有效和精确的，但必须是高效的，以避免显着减缓程序执行。这项工作的关键思想是在垃圾收集器上进行检查，垃圾收集器已经定期访问程序中的所有数据结构。一个有效和精确的工具来检测数据结构错误可以广泛部署，以提高关键软件基础设施的可靠性。第一个涉及到开发一种声明性语言，用于表达动态数据结构属性，建立在现有的静态分析和验证技术的基础上。第二种方法研究在垃圾收集器的单次传递过程中可以检查的属性类。第三种方法建立在并发垃圾收集机制的基础上，允许堆检查在可用的额外CPU内核上与应用程序并发进行。