When designing concurrent algorithms, Load-Link/Store-Conditional (LL/SC) is a very useful primitive since it avoids ABA problems. The full semantics of LL/SC are not supported in hardware by any modern architecture, so there has been a significant amount of work on simulations of LL/SC using CAS. However, all previous algorithms that are constant time either use unbounded sequence numbers (and thus base objects of unbounded size), or require ??(MP) space to implement M LL/SC objects for P processes. We present the first constant time implementation of LL/SC from bounded-sized CAS objects using only constant space overhead per LL/SC variable. In particular, our implementation uses ?~(M kP?2) space, where k is the number of outstanding LL operations per process, and only requires pointer-width CAS operations. In most algorithms that use LL/SC, k is a small constant which reduces our additive space overhead to ?~(P?2). Our algorithm can also be extended to implement L word LL/SC objects in ?~(L) time for LL and SC, O(1) time for VL, and ?~((M kP?2)L) space. To achieve these bounds, our main technical contribution is implementing a new primitive called Single-Writer Copy which takes a pointer to a word sized memory location and atomically copies its contents into another object. The restriction is that only one process is allowed to write/copy into the destination object at a time. The ability to read from one memory location and write to another atomically, and in constant-time, is very powerful and we believe this primitive will be useful in designing other algorithms.
展开▼
