Vulnerability-oriented directed fuzzing for binary programs

Directed greybox fuzzing (DGF) is an effective method to detect vulnerabilities of the specified target code. Nevertheless, there are three main issues in the existing DGFs. First, the target vulnerable code of the DGFs needs to be manually selected, which is tedious. Second, DGFs mainly leverage distance information as feedback, which neglects the unequal roles of different code snippets in reaching the targets. Third, most of the existing DGFs need the source code of the test programs, which is not available for binary programs. In this paper, we propose a vulnerability-oriented directed binary fuzzing framework named VDFuzz, which automatically identifies the targets and leverages dynamic information to guide the fuzzing. In specific, VDFuzz consists of two components, a target identifier and a directed fuzzer. The target identifier is designed based on a neural-network, which can automatically locate the target code areas that are similar to the known vulnerabilities. Considering the inequality of code snippets in reaching the given target, the directed fuzzer assigns different weights to basic blocks and takes the weights as feedback to generate test cases to reach the target code. Experimental results demonstrate that VDFuzz outperformed the state-of-the-art fuzzers and was effective in vulnerability detection of real-world programs.

Abstract licence: CC BY-NC-ND

39 more references, whose wording the paper’s licence keeps private.

A person counts once a day per paper. We keep no cookie and no address: a reader is a hash that changes every day. Any month under five reads shows no number.