Linux Driver Verification

Beyer, Dirk; Petrenko, Alexander K.

doi:10.1007/978-3-642-34032-1_1

Cited by 20 publications

(7 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…• RealWorld: RealWorld is the set of C programs obtained from the SoftwareSystems category of SV-COMP '21 [22]. The category contains 3,184 programs from real-world sources, and we exclude programs with no predefined error location [23]. As a result, RealWorld includes 3,042 programs consisting of 2,688 clean programs and 354 buggy programs.…”

Section: Target Programsmentioning

confidence: 99%

Directed Model Checking for Fast Abstract Reachability Analysis

Lee

Kim

et al. 2021

IEEE Access

View full text Add to dashboard Cite

We propose a novel technique (TOUR) to improve both bug detection ability and verification speed of ARMC by detecting a target path quickly. The key idea of TOUR is an error location directed search that utilizes the distance to an error location and function call context at runtime. TOUR applies four different distance metrics and a distance metric selection heuristic using static features of a target program. We have extensively evaluated TOUR on 3,042 real-world C programs in a software verification competition benchmark. The experiment results show that TOUR, due to its error location directed search, finds bugs in 20% more programs in 11% less model checking time than the state-of-the-art ARMC technique (i.e., block-abstraction memoization) for 354 buggy programs. Also, TOUR verifies 15% more programs within 15% less model checking time than the block-abstraction memoization for 652 complex clean programs.

show abstract

Section: Target Programsmentioning

confidence: 99%

Directed Model Checking for Fast Abstract Reachability Analysis

Lee

Kim

et al. 2021

IEEE Access

View full text Add to dashboard Cite

show abstract

“…RELATED WORK [6] explains the various research directories corresponding to Linux Driver Verification. A few of them are: Pointer analysis and dynamic data structures, symbolic verification, explicit-state verification, termination analysis and concurrency.…”

Section: IImentioning

confidence: 99%

Rules based automatic Linux Device Driver Verifier And Code Assistance

Dileep

Raghavendra

Maloji

et al. 2014

International Conference on Recent Advances and Innovations in Engineering (ICRAIE-2014)

View full text Add to dashboard Cite

Development of Linux Device Drivers involves lot of complex data structures and functions. The developer should carefully apply these to the drivers otherwise code will misbehave in the Linux Kernel and degrades the performance of the system. Majorly Linux device driver code requires stability so as to provide correctness and reliability. To achieve this, the need of the hour is a proper and efficient verification tools. Linux driver code verification is a vast application area consisting of different verification methods on proper functional placement, safety and security usages. In this paper we are proposing a tool embedded into Eclipse IDE as a plug-in and this tool works on the principle of "Rules based Linux Device Driver Automatic Verifier and Code Assistance" for the bugs fixation. Majorly the rules are based on Kernel API violations, proper allocations & de allocations, synchronization mechanisms & usage, proper return type usage and interrupt context issues.

show abstract

“…Software verification has been successfully applied to improve the quality and reliability of computer programs [2,3,19,22,28,30,40]. In the last decades, several algorithms and approaches were developed to perform software model checking for various kinds of C programs.…”

Section: Introductionmentioning

confidence: 99%