ComputingComputer ScienceCybersecurityComputer SecurityMalware AnalysisVulnerability AnalysisBehavioral AnalysisCode Obfuscation AnalysisControl-flow AnalysisData-flow AnalysisMalware Reverse EngineeringStatic Vulnerability AnalysisAutomatic DeobfuscationCode DeobfuscationDeobfuscationMalware DeobfuscationTaint DetectionTainted Data Analysis
Taint analysis has a wide variety of applications in software analysis, making the precision of taint analysis an important consideration. Current taint analysis algorithms, including previous work on bit-precise taint analyses, suffer from shortcomings that can lead to significant loss of precision (under/over tainting) in some situations. This paper discusses these limitations of existing taint analysis algorithms, shows how they can lead to imprecise taint propagation, and proposes a generalization of current bit-level taint analysis techniques to address these problems and improve their precision. Experiments using a deobfuscation tool indicate that our enhanced taint analysis algorithm leads to significant improvements in the quality of deobfuscation.
17
Heng Yin, Dawn Song, Manuel Egele et al. · 2007 · 688 citations
Artem Dinaburg, Monirul Sharif, Wenke Lee · 2008 · 661 citations
James Clause, Wanchun Li, Alessandro Orso · 2007 · 539 citations