Towards effective and efficient search-based deterministic replay

Abstract

Deterministic replay tools are a useful asset when it comes to pinpoint hard-to-reproduce bugs. However, no sweet spot has yet been found with respect to the trade-off between recording overhead and bug reproducibility, especially in the context of search-based deterministic replay techniques, which rely on inference mechanisms.In this paper, we argue that tracing the locking order, along with the local control-flow path affected by shared variables, allows to dramatically reduce the inference time to find a fault-inducing trace, while imposing only a slight increase in the overhead during production runs. Preliminary evaluation with a micro-benchmark and third-party benchmarks provides initial evidence that supports our claim.

Concurrency Debugging with Differential Schedule Projections

“…Giving Symbiosis minutes or hours to help solve hard bugs (like pbzip2 (large)) is reasonable. Additionally, Symbiosis could divide the SMT constraint system into different instances and solve them in parallel, like CLAP, or incorporate lock ordering information, like [Bravo et al 2013], to decrease solver time.…”

Concurrency debugging with differential schedule projections

“…Giving Symbiosis minutes or hours to help solve hard bugs (like pbzip2 (large)) is reasonable. Additionally, Symbiosis could use parallel SMT solving, like CLAP or incorporate lock ordering information, like [5], to decrease solver time.…”