Fuzzing with data dependency information

Mantovani, Alessandro; Fioraldi, Andrea; Balzarotti, Davide
EuroS&P 2022, 7th IEEE European Symposium on Security and Privacy, 6-10 June 2022, Genoa, Italy

Recent advances in fuzz testing have introduced several forms of feedback mechanisms, motivated by the fact that for a large range of programs and libraries, edgecoverage alone is insufficient to reveal complicated bugs. Inspired by this line of research, we examined existing program representations looking for a match between expressiveness of the structure and adaptability to the context of fuzz testing. In particular, we believe that data dependency graphs (DDGs) represent a good candidate for this task, as the set of information embedded by this data structure is potentially useful to find vulnerable constructs by stressing combinations of def-use pairs that would be difficult for a traditional fuzzer to trigger. Since some portions of the dependency graph overlap with the control flow of the program, it is possible to reduce the additional instrumentation to cover only “interesting” data-flow dependencies, those that help the fuzzer to visit the code in a distinct way compared to
standard methodologies. To test these observations, in this paper we propose DDFuzz, a new approach that rewards the fuzzer not only with code coverage information, but also when new edges in the data dependency graph are hit. Our results show that the adoption of data dependency instrumentation in coverage-guided fuzzing is a promising solution that can help to discover bugs that would otherwise remain unexplored by
standard coverage approaches. This is demonstrated by the 72 different vulnerabilities that our data-dependency driven approach can identify when executed on 38 target programs from three different datasets.

DOI
Type:
Conference
City:
Genoa
Date:
2022-06-06
Department:
Digital Security
Eurecom Ref:
6831
Copyright:
© 2022 IEEE. Personal use of this material is permitted. However, permission to reprint/republish this material for advertising or promotional purposes or for creating new collective works for resale or redistribution to servers or lists, or to reuse any copyrighted component of this work in other works must be obtained from the IEEE.

PERMALINK : https://www.eurecom.fr/publication/6831