Abstract
This disclosure describes techniques based on static analysis and dynamic profiling to accelerate the identification and resolution of root causes of bugs in software binaries. The techniques narrow down the search space by determining code paths that are involved in the causation of the issue, changes in such code paths, changes to the methods in such code paths, and by attributing the changes to change lists. To perform issue resolution, various components are provided, including a service that returns a list of methods changed within a given period; an on-device profiling service with on-demand start-and-stop; a static analysis component accepts a change list (CL) as input and produces as output the methods changed by the change list; and a culprit identifier that accepts trace data and a list of changed methods as input and produces as output the CLs associated with the changed methods and in the trace data.
Creative Commons License
This work is licensed under a Creative Commons Attribution 4.0 License.
Recommended Citation
Wang, Fei; Hu, Yue; Turner, Joseph; Owens, David; Huang, Andrew; Al-Muhtasib, Leila; Andrews, Bradley; Ding, Oscar; Krasichkov, Eugene; Li, Zening; Kodur, Kamakshi; Burns, Christopher Charles; and Cheng, Shu-Wei, "Reducing Time to Identify Bugs in Software Binaries Using Static Analysis and Dynamic Profiling", Technical Disclosure Commons, (March 14, 2024)
https://www.tdcommons.org/dpubs_series/6797