Crash Path Analysis Method Based on Fused Dynamic Control and Data Flows

To address the limitations of path redundancy and insufficient precision in traditional static analysis for crash path localization, this paper proposes an automated crash path construction method that synergistically integrates dynamic control flow and data flow analysis. Through dynamic instrumentation techniques, we systematically record runtime function call chains and basic block transition sequences, while incorporating taint analysis to trace input-dependent data relationships. This integrated approach enables the construction of a Crash Hybrid Path Graph (CHPG) that comprehensively embeds both control flow semantics and data dependency constraints. Experimental results demonstrate the effectiveness of our method in precisely identifying vulnerability-triggering execution paths, providing substantive support for vulnerability localization.