Abstract: To solve pattern “double lock & unlock” and pattern “misuse of volatile”, a detection method based on coding rules is proposed. Firstly, the failure mechanism is analyzed, and then 3 coding rules are proposed to avoid corresponding defects in development stage. These rules and corresponding checking methods are researched and implemented by extending an existing static analysis tool named SpaceCCH. The evaluation result based on real world onboard software show that, the extended SpaceCCH can find rules violations efficiently with low false positive rate and low false negative rate, by which the data race bugs of pattern “double lock & unlock” and pattern “misuse of volatile” can be avoided effectively. The main contribution of this paper is that the detection of a really complex bug is transformed to corresponding coding rules and their automatic detection.

Key words: data race, coding rule, static analysis, aerospace embedded software