A Traced-based Automated System Diagnosis and Software Debugging Methodology for Embedded Multi-core Systems
Li Lin · mediaTUM – the media and publications repository of the Technical University Munich (Technical University Munich) · 2019
This dissertation proposes an automated system diagnosis and debugging methodology tackling debugging challenges such as shared resource contentions and multicore synchronization issues. The methodology named “embedded health” is inspired by medical blood tests, which can rapidly and precisely detect several diseases. It is based on non-intrusive hardware tracing and generally applicable on various embedded system issues. Results reveal high effectiveness of the methodology. For example, the methodology is approx. 90% accurate in detecting program flash contention that was previously not directly measurable.