Driver Generation for Configuration-Rich Software

Develop driver-generation strategies that adequately represent the execution modes of configuration-rich software and expose structurally diverse regions for driver-aware fuzzing.

Background

The study manually constructs drivers from primary command-line options and major execution modes rather than exhaustively enumerating all configuration possibilities. The authors acknowledge that this abstraction only partially captures the full configuration space and that alternative driver-construction strategies could expose different structural regions, overlap patterns, effectiveness distributions, and residual-coverage regimes. Consequently, automatically or systematically generating effective drivers remains an unresolved methodological problem for multi-driver fuzzing.