Automatic fine-grained locking generation for shared data structures

Autor: Tingting Hu, Zongyan Qiu, Haiyang Liu
Rok vydání: 2017
Předmět:
Zdroj: TASE
DOI: 10.1109/tase.2017.8285633
Popis: Correct mutual-exclusion is one of the key challenges in concurrent programming. Although the fine-grained locking schema can be more efficient compared with the coarse-grained techniques, it is tough to use, as well as error-prone. Here we present a static approach, based on program analysis, to automatically add fine-grained locking primitives to data structures implemented as classes. For tree-like structures, the modified class definitions are guaranteed to be thread-safe. Experiments show that the approach can successfully deal with programs which are challenging to be handled manually, and it works efficiently.
Databáze: OpenAIRE