Triggering Data Races in Multi-Threaded Programs Using Enhanced S2E

Most critical software used and developed today is built with the use of multiple threads, which is quite complicated to get right.This leads to a prevalence of bugs related to concurrency, such as data races.In addition to that, current technological trends include robotics and unmanned vehicle software, which is not only multi-threaded but also made more complicated since it interacts with hardware, requires a specific OS environment, utilizes IPC, and must be rigorously tested.In this paper, we build on top of a multi-core version of S2E and provide an algorithm designed to improve the capabilities of detecting data race bugs in programs running in complicated environments.To achieve this, we utilize S2E's full-system emulation, its plugin system, and the CPU core scheduler from our fork.