Balancing the LoadLeveraging a Semantics Stack for Systems Verification

We have developed a stack of semantics for a high-level C-like language and low-level assembly code, which has been carefully crafted to support the pervasive verification of system software. It can handle mixed-language implementations and concurrently operating devices, and permits the transferral...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:Journal of automated reasoning 2009-04, Vol.42 (2-4), p.389-454
Hauptverfasser: Alkassar, Eyad, Hillebrand, Mark A, Leinenbach, Dirk C, Schirmer, Norbert W, Starostin, Artem, Tsyban, Alexandra
Format: Artikel
Sprache:eng
Schlagworte:
Online-Zugang:Volltext
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
Beschreibung
Zusammenfassung:We have developed a stack of semantics for a high-level C-like language and low-level assembly code, which has been carefully crafted to support the pervasive verification of system software. It can handle mixed-language implementations and concurrently operating devices, and permits the transferral of properties to the target architecture while obeying its resource restrictions. We demonstrate the applicability of our framework by proving the correct virtualization of user memory in our microkernel, which implements demand paging. This verification target is of particular interest because it has a relatively simple top-level specification and it exercises all parts of our semantics stack. At the bottom level a disk driver written in assembly implements page transfers via a swap disk. A page-fault handler written in C uses the driver to implement the paging algorithm. It guarantees that a step of the currently executing user can be simulated at the architecture level. Besides the mere theoretical and technical difficulties the project also bore the social challenge to manage the large verification effort, spread over many sites and people, concurrently contributing to and maintaining a common theory corpus. We share our experiences and elaborate on lessons learned.
ISSN:0168-7433
DOI:10.1007/s10817-009-9123-z