Finding All Potential Run-Time Errors and Data Races in Automotive Software - Inria - Institut national de recherche en sciences et technologies du numérique
Communication Dans Un Congrès Année : 2017

Finding All Potential Run-Time Errors and Data Races in Automotive Software

Daniel Kästner
  • Fonction : Auteur
Antoine Miné
Heinz Hille
  • Fonction : Auteur
Laurent Mauborgne
  • Fonction : Auteur
Stephan Wilhelm
  • Fonction : Auteur
Patrick Cousot
  • Fonction : Auteur
  • PersonId : 951015
Christian Ferdinand
  • Fonction : Auteur

Résumé

Safety-critical embedded software has to satisfy stringent quality requirements. All contemporary safety standards require evidence that no data races and no critical run-time errors occur, such as invalid pointer accesses, buffer overflows, or arithmetic overflows. Such errors can cause software crashes, invalidate separation mechanisms in mixed-criticality software, and are a frequent cause of errors in concurrent and multi-core applications. The static analyzer Astrée has been extended to soundly and automatically analyze concurrent software. This novel extension employs a scalable abstraction which covers all possible thread interleavings, and reports all potential run-time errors, data races, deadlocks, and lock/unlock problems. When the analyzer does not report any alarm, the program is proven free from those classes of errors. Dedicated support for ARINC 653 and OSEK/AUTOSAR enables a fully automatic OS-aware analysis. In this article we give an overview of the key concepts of the concurrency analysis and report on experimental results obtained on concurrent automotive software. The experiments confirm that the novel analysis can be successfully applied to real automotive software projects.
Fichier non déposé

Dates et versions

hal-01674831 , version 1 (03-01-2018)

Identifiants

Citer

Daniel Kästner, Antoine Miné, André Schmidt, Heinz Hille, Laurent Mauborgne, et al.. Finding All Potential Run-Time Errors and Data Races in Automotive Software. WCX™ 2017 - SAE World Congress Experience, Apr 2017, Detroit United States. pp.1-9, ⟨10.4271/2017-01-0054⟩. ⟨hal-01674831⟩
399 Consultations
0 Téléchargements

Altmetric

Partager

More