무기체계 소프트웨어의 자료경합을 탐지하기 위한 프레임워크

A Framework for Detecting Data Races in Weapon Software
  • 오진우
  • 최으뜸
  • 전용기

초록

Software has been used to develop many functions of the modern weapon systems which has a high mission criticality. Weapon system software must consider multi-threaded processing to satisfy growing performance requirement. However, developing multi-threaded programs are difficult because of concurrency faults, such as unintended data races. Especially, it is important to prepare analysis for debugging the data races, because the weapon system software may cause personal injury. In this paper, we present an efficient framework of analysis, called ConDeWS, which is designed to determine the scope of dynamic analysis through using the result of static analysis and fault analysis. As a result of applying the implemented framework to the target software, we have detected unintended data races that were not detected in the static analysis.

키워드

Multi-threadWeapon system softwareConcurrency faultsData racesDynamic analysis
제목
무기체계 소프트웨어의 자료경합을 탐지하기 위한 프레임워크
제목 (타언어)
A Framework for Detecting Data Races in Weapon Software
저자
오진우최으뜸전용기
DOI
10.14372/IEMEK.2018.13.6.305
발행일
2018
저널명
대한임베디드공학회논문지
13
6
페이지
305 ~ 312