SHF: Medium: RacePro: Automatically Detecting API Races in Deployed Systems

点击翻译按钮获取中文摘要

While races in multithreaded programs have drawn huge attention from theresearch community, little has been done for API races, a classof errors as dangerous and as difficult to debug as traditional threadraces. An API race occurs when multiple activities, whether they bethreads or processes, access a shared resource via an applicationprogramming interface (API) without proper synchronization. DetectingAPI races is an important and difficult problem as existing racedetectors are unlikely to work well with API races. Software reliability increasingly affects everyone, whether or notthey personally use computers. This research studies andautomatically detects for the first time an important class of racesthat has a significant impact on software reliability. The studyquantitatively demonstrates how API races are numerous, difficult todebug, and a real threat to software reliability. To address thisproblem, this research is developing RacePro, a new system toautomatically detect API races in deployed systems. RacePro checksdeployed systems in-vivo by recording live executions thendeterministically replay and check them later. This approachincreases checking coverage beyond the configurations or executionscovered by software vendors or beta testing sites. RacePro recordsmultiple processes and threads, detects races in the recording among API methods that may concurrently access shared objects, then exploresdifferent execution orderings of such API methods to determine which racesare harmful and result in failures. Technologies developed will helpapplication developers detect insidious software defects, enabling more robust, reliable, and secure software infrastructure.