{"title":"los:用于mpsoc安全性的低开销评测器驱动的设计流程","authors":"K. Patel, S. Parameswaran","doi":"10.1145/1450135.1450154","DOIUrl":null,"url":null,"abstract":"Security is a growing concern in processor based systems and hence requires immediate attention. New paradigms in the design of MPSoCs must be found, with security as one of the primary objectives. Software attacks like Code Injection Attacks exploit vulnerabilities in \"trusted\" code. Previous countermeasures addressing code injection attacks in MPSoCs have significant performance overheads and do not check every single line of code. The work described in this paper has reduced performance overhead and ensures that all the lines in the program code are checked.\n We propose an MPSoC system where one processor (which we call a MONITOR processor) is responsible for supervising all other application processors. Our design flow, LOCS, instruments and profiles the execution of basic blocks in the program. LOCS subsequently uses the profiler output to re-instrument the source files to minimize runtime overheads. LOCS also aids in the design of hardware customizations required by the MONITOR. At runtime, the MONITOR checks the validity of the control flow transitions and the execution time of basic blocks.\n We implemented our system on a commercial extensible processor, Xtensa LX2, and tested it on three multimedia benchmarks. The experiments show that our system has the worst-case performance degradation of about 24% and an area overhead of approximately 40%. LOCS has smaller performance, area and code size overheads than all previous code injection countermeasures for MPSoCs.","PeriodicalId":300268,"journal":{"name":"International Conference on Hardware/Software Codesign and System Synthesis","volume":"9 1","pages":"0"},"PeriodicalIF":0.0000,"publicationDate":"2008-10-19","publicationTypes":"Journal Article","fieldsOfStudy":null,"isOpenAccess":false,"openAccessPdf":"","citationCount":"2","resultStr":"{\"title\":\"LOCS: a low overhead profiler-driven design flow for security of MPSoCs\",\"authors\":\"K. Patel, S. Parameswaran\",\"doi\":\"10.1145/1450135.1450154\",\"DOIUrl\":null,\"url\":null,\"abstract\":\"Security is a growing concern in processor based systems and hence requires immediate attention. New paradigms in the design of MPSoCs must be found, with security as one of the primary objectives. Software attacks like Code Injection Attacks exploit vulnerabilities in \\\"trusted\\\" code. Previous countermeasures addressing code injection attacks in MPSoCs have significant performance overheads and do not check every single line of code. The work described in this paper has reduced performance overhead and ensures that all the lines in the program code are checked.\\n We propose an MPSoC system where one processor (which we call a MONITOR processor) is responsible for supervising all other application processors. Our design flow, LOCS, instruments and profiles the execution of basic blocks in the program. LOCS subsequently uses the profiler output to re-instrument the source files to minimize runtime overheads. LOCS also aids in the design of hardware customizations required by the MONITOR. At runtime, the MONITOR checks the validity of the control flow transitions and the execution time of basic blocks.\\n We implemented our system on a commercial extensible processor, Xtensa LX2, and tested it on three multimedia benchmarks. The experiments show that our system has the worst-case performance degradation of about 24% and an area overhead of approximately 40%. LOCS has smaller performance, area and code size overheads than all previous code injection countermeasures for MPSoCs.\",\"PeriodicalId\":300268,\"journal\":{\"name\":\"International Conference on Hardware/Software Codesign and System Synthesis\",\"volume\":\"9 1\",\"pages\":\"0\"},\"PeriodicalIF\":0.0000,\"publicationDate\":\"2008-10-19\",\"publicationTypes\":\"Journal Article\",\"fieldsOfStudy\":null,\"isOpenAccess\":false,\"openAccessPdf\":\"\",\"citationCount\":\"2\",\"resultStr\":null,\"platform\":\"Semanticscholar\",\"paperid\":null,\"PeriodicalName\":\"International Conference on Hardware/Software Codesign and System Synthesis\",\"FirstCategoryId\":\"1085\",\"ListUrlMain\":\"https://doi.org/10.1145/1450135.1450154\",\"RegionNum\":0,\"RegionCategory\":null,\"ArticlePicture\":[],\"TitleCN\":null,\"AbstractTextCN\":null,\"PMCID\":null,\"EPubDate\":\"\",\"PubModel\":\"\",\"JCR\":\"\",\"JCRName\":\"\",\"Score\":null,\"Total\":0}","platform":"Semanticscholar","paperid":null,"PeriodicalName":"International Conference on Hardware/Software Codesign and System Synthesis","FirstCategoryId":"1085","ListUrlMain":"https://doi.org/10.1145/1450135.1450154","RegionNum":0,"RegionCategory":null,"ArticlePicture":[],"TitleCN":null,"AbstractTextCN":null,"PMCID":null,"EPubDate":"","PubModel":"","JCR":"","JCRName":"","Score":null,"Total":0}
LOCS: a low overhead profiler-driven design flow for security of MPSoCs
Security is a growing concern in processor based systems and hence requires immediate attention. New paradigms in the design of MPSoCs must be found, with security as one of the primary objectives. Software attacks like Code Injection Attacks exploit vulnerabilities in "trusted" code. Previous countermeasures addressing code injection attacks in MPSoCs have significant performance overheads and do not check every single line of code. The work described in this paper has reduced performance overhead and ensures that all the lines in the program code are checked.
We propose an MPSoC system where one processor (which we call a MONITOR processor) is responsible for supervising all other application processors. Our design flow, LOCS, instruments and profiles the execution of basic blocks in the program. LOCS subsequently uses the profiler output to re-instrument the source files to minimize runtime overheads. LOCS also aids in the design of hardware customizations required by the MONITOR. At runtime, the MONITOR checks the validity of the control flow transitions and the execution time of basic blocks.
We implemented our system on a commercial extensible processor, Xtensa LX2, and tested it on three multimedia benchmarks. The experiments show that our system has the worst-case performance degradation of about 24% and an area overhead of approximately 40%. LOCS has smaller performance, area and code size overheads than all previous code injection countermeasures for MPSoCs.