A directed greybox fuzzing tool for continuous integration

Abstract

Changes are occurred frequently during continuous integration. Existing testing methods often suffer from weak specificity or insufficiency when applied to continuous integration. To solve this problem, we implement a fuzzing tool called CIDFuzz for continuous integration. First, difference analysis is performed to locate the change points, and the distances between basic blocks and the change points are calculated. Then, the distances are instrumented into the program under test. During fuzz testing, testing resources are allocated according to the coverage of seeds to test the change points effectively.

Keywords

• Continuous integration
• Taint analysis
• Fuzz testing
• Change analysis