International audienceInformation Flow Control at Operating System (OS) level features interesting properties and have been an active topic of research for years. However, no implementation can work reliably if there does not exist a way to correctly and precisely track all information flows occurring in the system. The existing implementations for Linux are based on the Linux Security Modules (LSM) framework which implements hooks at speciic points in code where any security mechanism may interpose a security decision in the execution. However, previous works on the verification of LSM only addressed access control and no work has raised the question of the reliability of information flow control systems built on LSM. In this work, we present a compiler-assisted and reproducible static analysis on the Linux kernel to verify that the LSM hooks are correctly placed with respect to operations generating information flows so that LSM-based information flow monitors can properly track all information flows. Our results highlight flaws in LSM that we propose to solve, thus improving the suitability of this framework for the implementation of information flow monitors

Abstract. Information flow control can be used at the Operating System level to enforce restrictions on the diffusion of security-sensitive data. In Linux, information flow trackers are often implemented as Linux Security Modules. They can fail to monitor some indirect flows when flows occur concurrently and affect the same containers of information. Furthermore, they are not able to monitor the flows due to file mappings in memory and shared memory between processes. We first present two attacks to evade state-of-the-art LSM-based trackers. We then describe an approach, formally proved with Coq [12] to perform information flow tracking able to cope with concurrency and in-memory flows. We demonstrate its implementability and usefulness in Rfblare, a race conditionfree version of the flow tracking done by KBlare [4].