Method, system and product for identifying and executing locked read regions and locked write regions in programming languages that offer synchronization without explicit means to distinguish between such regions

Abstract

A technique for identifying and executing synchronized read regions and synchronized write regions is provided. The technique processes program code with a software tool to convert the code to an internal representation for the software tool and constructs a control flow graph of code blocks which also identifies the entry and exit points for each region of code subject to a lock. For each code block in the control flow graph, visibility analysis is performed. Each determined entry and exit point is moved to enclose the smallest set of code blocks subject to a lock which can be achieved without altering code semantics. A set of trees is created to represent the manner in which the synchronized regions are nested. If the tree contains at least one block previously identified, then the tree is marked a write. Otherwise, the tree is marked a read tree. A write lock is then created for the region in each tree identified as a write tree and a read lock is created for the regions in each tree identified as a read tree.