首页> 外文会议>2013 16th International Multi Topic Conference >VG-MIPS: A dynamic binary instrumentation framework for multi-core MIPS processors
【24h】

VG-MIPS: A dynamic binary instrumentation framework for multi-core MIPS processors

机译:VG-MIPS:用于多核MIPS处理器的动态二进制检测框架

获取原文
获取原文并翻译 | 示例

摘要

Valgrind is a Dynamic Binary Analysis tool used for debugging and profiling purposes. It's mostly used to analyze the memory usage of software applications. Currently it supports the ×86, AMD, ARM, PPC and S390X architectures. Recently it has been ported to MIPS/Linux. This paper describes VG-MIPS a port of Valgrind to Cavium Networks®'s Octeon Processor for intelligent networking which hosts a MIPS64 ISA. The Octeon processor has a large number of specialized MIPS instructions used for optimizing the network and security processing features of its Internet applications. This paper discusses how we added support for n32 MIPS dialect and Octeon-MIPS specific instructions. In the process we made modifications to the Valgrind codebase such as support for new system call wrapping and the ability to automatically detect the ABI dialect from the client program ELF at runtime. Our port has been contributed to the open-source Valgrind developer community and has been successfully incorporated in the code base. We have evaluated VG-MIPS on the multi-core MIPS64 processor: Octeon® and verified its profiling accuracy. We hope that our experiences and key design decisions shared in this paper will help future developers in writing DBI frameworks for new architectures.
机译:Valgrind是用于调试和分析的动态二进制分析工具。它主要用于分析软件应用程序的内存使用情况。当前,它支持×86,AMD,ARM,PPC和S390X体系结构。最近,它已被移植到MIPS / Linux。本文介绍了VG-MIPS,这是Valgrind的端口,是CaviumNetworks®Octeon处理器的端口,用于智能网络,该处理器承载MIPS64 ISA。 Octeon处理器具有大量专用的MIPS指令,用于优化其Internet应用程序的网络和安全处理功能。本文讨论了我们如何添加对n32 MIPS方言和Octeon-MIPS特定指令的支持。在此过程中,我们对Valgrind代码库进行了修改,例如对新系统调用包装的支持以及在运行时从客户端程序ELF自动检测ABI方言的功能。我们的港口已为开源Valgrind开发人员社区做出了贡献,并已成功纳入代码库中。我们已经在多核MIPS64处理器Octeon上评估了VG-MIPS,并验证了其分析精度。我们希望本文分享的经验和关键设计决策将有助于未来的开发人员为新架构编写DBI框架。

著录项

相似文献

  • 外文文献
  • 中文文献
  • 专利
获取原文

客服邮箱:kefu@zhangqiaokeyan.com

京公网安备:11010802029741号 ICP备案号:京ICP备15016152号-6 六维联合信息科技 (北京) 有限公司©版权所有
  • 客服微信

  • 服务号