Pragmatic primitives for non-blocking data structures. Manuscript available from http://www.cs.utoronto.ca/∼tabrown

We define a new set of primitive operations that greatly sim-plify the implementation of non-blocking data structures in asynchronous shared-memory systems. The new operations operate on a set of Data-records, each of which contains multiple fields. The operations are generalizations of the well-known load-link (LL) and store-conditional (SC) oper-ations, which we call LLX and SCX. The LLX operation takes a snapshot of one Data-record. An SCX operation by a process p succeeds only if no Data-record in a specified set has been changed since p last performed an LLX on it. If successful, the SCX atomically updates one specific field of a Data-record in the set and prevents any future changes to some specified subset of those Data-records. We provide an implementation of these new primitives from single-word compare-and-swap and prove it correct. As a simple exam-ple, we show how LLX and SCX can be used to implement a non-blocking multiset data structure in a straightforward way. 1