Static Code Analysis For Embedded Systems

Abstract

Much software for embedded systems is written in languages such as C. This is known to be error prone, because of manual memory management and similar insecurities. A countermeasure against such problems is static code analysis. This thesis presents an evaluation of techniques for static code analysis, focusing on methods of fault detection. A number of different analysis tools have been tested, at Ascom Wireless Solutions, a developer of embedded system for wireless communication, on production code. The tools were able to detect real faults, but with significant manual interaction required.