Library-Aware Doubles and Iterative Repair for Large Language Model-Generated Unit Tests in OpenSIL Firmware

The study presents an automated unit test (UT) authoring workflow for the Open-Source Silicon Initialization Library (openSIL) firmware, utilizing a multi-agent pipeline guided by a large language model (LLM). Key technical features include library-aware generation of test scaffolds and an iterative compile-dispatch repair loop, which significantly improved compilation success rates to 96% across 76 functions and achieved mean line coverage of 98.8% under optimal conditions. This approach is crucial for practitioners as it enhances the efficiency and reliability of UT generation in low-level firmware development, addressing common challenges related to build constraints and manual debugging.