Logic List Mailing Archive

PhD student position in compositional verification, Rennes (Frances)

Compositional verification of system program modules in Rust

Keywords: theorem proving, programming languages, operating systems, 
formal verification, dependent types, theory of contracts

Project RIOT-fp [b] is an Inria Challenge with the objective of developing 
future-proof operating system libraries [1,2,4] for application to IoT: 
RIOT [a].  Our PhD project is interested in one of the futures of RIOT: 
RIOT-rs, implemented in Rust [c]. This computing base provides access to a 
vast ecosystem of analysis, code generation, verification and proof tools 
[d,e,f]. It offers us to rethink a system software validation process that 
would suit both system programming and verification requirements (as one 
may expect from using, e.g., a theorem prover).

The notion of contract [3] is one ideal such interface between the 
development and verification of system programs in Rust.  A contract 
allows, on one hand, to formally document the hypothesis and guarantees of 
system modules, functions, artifacts, with respect to global safety ad 
security requirements.  Contracts can be sufficiently abstract and 
comprehensible for system programmers, and adequately refined to meet the 
strongest requirements of mechanized verification.

Our project will focus on the development of such a modular validation 
flow by case-studying the core of RIOT's implementation in Rust 
[riot-rs-core].  We define and exercise this workflow to characterize and 
validate global requirements ranging from race-condition, deadlock 
avoidance, priority management and schedulability, and/or memory 
isolation, faul isolation, information flow control.

The project will be implemented with teams Tea and Celtique at Inria, 
Rennes, in close collaboration with teams Tribe and Prosecco at Inria, 
Paris.  It requires a Master degree with solid background in proof theory 
and mathematical logic, programming languages and type theory, as well as 
motivation and interest in both the implementation and verification of 
operating systems.  Prior knowledge and experiences with both Rust, F*, 
Coq, Lean will stand out. The project will require weekly multi-center 
meetings and hence excellent communication and team-working skills in both 
french and english.

https://www.irisa.fr/phd-subject/2022-04/compositional-verification-system-program-modules-rust

BIBLIOGRAPHY
[a] RIOT: http://www.riot-os.org
[b] RIOT-fp: https://future-proof-iot.github.io/RIOT-fp
[c] riot-rs-core: https://github.com/future-proof-iot/RIOT-rs/tree/main/src/riot-rs-core/src
[d] F*: https://www.fstar-lang.org
[e] Lean: https://leanprover.github.io
[f] Electrolysis: https://kha.github.io/electrolysis

REFERENCES
[1] "Verified Functional Programming of an Abstract Interpreter". Static Analysis Symposium. ACM, 2021.
[2] "Verified Functional Programming of an IoT operating system’s boot-loader". International Conference on Formal Methods and Models for System Design. ACM, 2021.
[3] "A Mechanically Verified Theory of Contracts". International Colloquium on Theoretical Aspects of Computing. Springer, 2021.
[4] "End-to-end Mechanized Proof of an eBPF Virtual Machine for Microcontrollers". International Conference on Computer Aided Verification, 2022.
--
[LOGIC] mailing list
http://www.dvmlg.de/mailingliste.html
Archive: http://www.illc.uva.nl/LogicList/

provided by a collaboration of the DVMLG, the Maths Departments in Bonn and Hamburg, and the ILLC at the Universiteit van Amsterdam