What is it about?
Input minimization is a crucial stage in coverage-guided kernel fuzzing. However, we observe that this minimization stage accounts for 57.5% of the total fuzzing overhead, consuming more than half of the available resources. This paper aims to reduce the minimization cost by optimizing both call-level and argument-level minimization strategies.
Featured Image
Photo by Andrey Tikhonovskiy on Unsplash
Why is it important?
We are the first to explore and mitigate the significant tension between the benefits and costs of input minimization in kernel fuzzing. We introduce two general and novel optimization strategies, respectively, for call removal and argument simplification in the minimization stage. Our extensive evaluation results demonstrate the effectiveness, general applicability, and benefits of our optimization strategies.
Perspectives
This is a solid research work. I am sincerely grateful to all co-authors for their thoughtful guidance on ideas and experimental design, from which I have greatly benefited. This experience has not only sharpened my research skills but also strengthened my determination to continue advancing kernel security.
Hui Guo
East China Normal University
Read the Original
This page is a summary of: Efficient Input Minimization for Kernel Fuzzing via Relation- and Type-Guided Optimization, ACM Transactions on Computer Systems, August 2026, ACM (Association for Computing Machinery),
DOI: 10.1145/3840388.
You can read the full text:
Resources
Contributors
The following have contributed to this page
