Making Lock-free Data Structures Verifiable with Artificial Transactions

Among all classes of parallel programming abstractions, lock-free data structures are considered one of the most scalable and efficient thanks to their fine-grained style of synchronization. However, they are also challenging for developers and tools to verify because of the huge number of possible interleavings that result from finegrained synchronizations. This paper addresses this fundamental problem between performance and verifiability of lock-free data structure implementations. We present TXIT, a system that greatly reduces the set of possible interleavings by inserting transactions into the implementation of a lock-free data structure. We leverage hardware transactional memory support from Intel Haswell processors to enforce these artificial transactions. Evaluation on six popular lock-free data structure libraries shows that TXIT makes it easy to verify lock-free data structures while incurring acceptable runtime overhead. Further analysis shows that two inefficiencies in Haswell are the largest contributors to this overhead.