Do you have 5 min? Improving Call Graph Analysis with Runtime Information

Constructing precise and sound call graphs is fundamental for effective static analysis, yet it remains a significant challenge in today's software. Traditionally, researchers have developed sophisticated algorithms to address this issue, often resulting in increased computational costs. But what if we could provide a simple, cost-effective way to improve call graphs drastically?
This paper introduces a simple method to largely enhance static call graphs almost for free, i.e., with 5 min of dynamic analysis and low overhead. Our approach improves the soundness of call graphs, thereby benefiting any downstream static analyses based on call graphs, such as data flow analysis. We demonstrate the efficacy of our method on Android apps by integrating it with FlowDroid, the leading static analysis tool for Android apps. Additionally, we outline future directions for achieving even more accurate and sound call graphs in static analysis.