Thread-Local Semantics and Its Efficient Sequential Abstractions for Race-Free Programs

Abstract

Data race free (DRF) programs constitute an important class of concurrent programs. In this paper we provide a framework for designing and proving the correctness of data flow analyses that target this class of programs. These analyses are in the same spirit as the "sync-CFG" analysis proposed in earlier literature. To achieve this, we first propose a novel concrete semantics for DRF programs, called L-DRF that is thread-local in nature-each thread operates on its own copy of the data state. We show that abstractions of our semantics allow us to reduce the analysis of DRF programs to a sequential analysis. This aids in rapidly porting existing sequential analyses to sound and scalable analyses for DRF programs. Next, we parameterize L-DRF with a partitioning of the program variables into "regions" which are accessed atomically. Abstractions of the region-parameterized semantics yield more precise analyses for region-race free concurrent programs. We instantiate these abstractions to devise efficient relational analyses for race free programs, which we have implemented in a prototype tool called RATCOP.

Data Races and Static Analysis for Interrupt-Driven Kernels

“…One problem with the value-set analysis is that it may not be able to prove relational invariants (like x ≤ y) for a program. One way to remedy this is to exploit the fact that concurrent programs often ensure racefree access to a region of variables, and to essentially do a region-based value-set analysis, as originally done in [21]. More precisely, let us say we have a partition of the set of variables V of a program P into a set of regions R 1 , .…”

“…Chugh et al [7] use race information to do thread-modular null-dereference analysis, by killing facts at a point whenever a notional read of a variable is found to be racy. De et al [11] propose the sync-CFG and value-set analysis for race-free programs, while Mukherjee et al [21] extend the framework to region and relational analyses. Gotsman et al [12] and Miné et al [18,19] define relational shape/value analyses for concurrent programs that exploit race-freedom and lock invariants respectively.…”

“…This is an important class of programs, as most developers aim to write race-free code, and one could try to exploit this property to give an efficient way of analyzing programs that fall in this class. In recent years there have been many techniques [7,11,12,18,21] that exploit the race-freedom property to perform sound and efficient static analysis. In particular [11,21] create an appealing structure called a "sync-CFG" which is the union of the control flow graphs of the threads augmented with possible "synchronization" edges, and essentially perform sequential analysis on this graph to obtain sound facts about the concurrent program.…”

Relational Thread-Modular Abstract Interpretation Under Relaxed Memory Models

“…This graph has a size which is exponential in the number of threads and linear in each thread size: it does not scale up. Thread-modular analyses have been designed to alleviate this combinatorial explosion [8,10,[15][16][17]. Amongst them, we use the formal system of interferences, that has been proposed by Miné [15] to analyse each thread in isolation, generating the effects it can have on the execution of other threads, and taking into account the effects generated by these other threads.…”

“…Thread-modular and weak memory analyses has been investigated by several authors [1,2,4,7,8,10,13,[15][16][17]], yet few works combine both. Nonetheless, it was shown [3,14] that non-relational analyses that are sound under sequential consistency remain sound under relaxed models.…”