摘要:
Finding bugs in lock-free concurrent programs is hard. This is due in part to the difficulty of reasoning about the correctness of concurrent algorithms and the timing-sensitive nature of concurrent programs. One of the most widely used tools for reasoning about the correctness of concurrent algorithms is the linearization property. This thesis presents a tool for automatic dynamic checking of concurrent programs under the Total-Store-Order (TSO) memory model and a methodology for finding linearization violations automatically with the tool.