Cargo-call-stack Static Call-stack Analysis for Rust

Lindgren, Per; Fitinghoff, Nils; Aparicio, Jorge Martínez

doi:10.1109/indin41052.2019.8972088

Cited by 2 publications

(1 citation statement)

References 4 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This lint simply compares the equality of caller def id and callee def id. Cargo-call-stack [21] develops a call graph on LLVM IR to analysis stack usage of each function, but it evades recursive functions.…”

Section: B Related Workmentioning

confidence: 99%

Detect Stack Overflow Bugs in Rust via Improved Fuzzing Technique

Ren,

2023

International Conferences on Software Engineering and Knowledge Engineering

View full text Add to dashboard Cite

Stack overflow has been a common memory vulnerability for a long time due to limited stack memory. Deep or infinite recursion serves as the main cause to exhaust the stack memory and crash the program. As a relatively new system programming language, Rust suffers from stack overflow problem inevitably. However, there is no relevant tool to detect those stack overflow bugs in Rust programming language. In this paper, we propose a novel approach using fuzz technique to trigger stack overflow bugs in Rust projects. We first build a call graph on Rust MIR (Middle Intermediate Representation). In the call graph, recursions appear as cycles lying in the SCCs (strong connect components). To find the entry APIs of those SCCs, we leverage Tarjan's algorithm to locate the SCCs and then reversely BFS (Breadth First Search) to search for the APIs. After that, we modify the underlying logic of AFL (American Fuzzing Loop) to trigger stack overflow bugs through fuzzing those dangerous APIs. Specifically, we add a function call time counter to accelerate the fuzzing process. We conduct our experiments on several existing Rust CVEs (Common Vulnerabilities and Exposures) related to stack overflow. Experiments show that our approach can trigger stack overflow bugs in a short time.

show abstract

Section: B Related Workmentioning

confidence: 99%