Citation Context

...programmers. 1.3 State of the Art Although the problem of atomicity violations has been known for years, few good solutions exist to address this challenging problem. Most state-of-the-art techniques =-=[11, 12]-=- for detecting these bugs require programmers’ annotations on atomic regions, which is too time-consuming for programmers. Additionally, in many cases programmers are not consciously aware of their at...

Citation Context

... colt). The last six columns show the number of warnings produced by each checker using the fine grain analysis. All eight warnings from FASTTRACK reflect real race conditions. As reported previously =-=[16, 28, 38, 39]-=-, some of these are benign (as in tsp, mtrt, and jbb) but others can impact program behavior (as on the checksum field in raytracer and races on several fields related to a thread pool in hedc). ERASE...

Citation Context

...y without interleaved steps from other threads. A number of tools have been developed for checking atomicity violations, including type systems [21, 22, 18, 32]; dynamic analysis such as the Atomizer =-=[20]-=- which combines the theory of reduction [29] and ideas from dynamic race detectors; and model checking techniques [25, 17]. These approaches require atomicity annotations in addition to synchronized b...

Citation Context

... needs to know is a set of statements whose simultaneous execution could lead to a concurrency problem. Such sets of problematic statements could be provided by a static or dynamic analysis technique =-=[23, 22, 2]-=-. We have implemented RACEFUZZER in a prototype tool for Java. The tool has been applied to a number of large benchmarks having a total of 600K lines of code. The results of these experiments demonstr...

Citation Context

...ophic disasters in real world (e.g., the Northeastern Electricity Blackout Incident [21]). Among different types of concurrency bugs, atomicity violation bugs are one of the most common and important =-=[13, 12, 14, 25, 5]-=-. Atomicity violation bugs (a realworld example is shown in Figure 1) widely exist because many programmers are used to sequential thinking and frequently assume code regions to be atomic without appr...

Citation Context

...detected. In this vein, our motivation is similar to recent proposals that have argued in favor of higher-level abstractions that enforce desirable properties on concurrent programs such as atomicity =-=[19, 18, 22]-=- or transaction-based isolation [23, 40, 45] without requiring the low-level (and thus potentially error-prone) operational reasoning demanded by locks. Other approaches include lock-free data structu...

Citation Context

... the user is desirable, but not always available in practice. The reduction-based algorithm is an extension of our original reduction-based algorithm [31] and Flanagan and Freund’s Atomizer algorithm =-=[10]-=-. It first determines whether there is data race on each variable and then uses this information to classify events. If the sequence of events in each transaction matches a given pattern, then the tra...

Citation Context

...Autolocker). Subsequently, Flanagan, Freund, Qadeer have combined guarded by annotations with explicitly declared atomicity constraints (similar to atomic statements). They use static [11] or dynamic =-=[7]-=- analysis to ensure that the locking policy chosen by the programmer actually enforces the declared atomicity constraints. They have also developed analyses to infer guarded by annotations [8], atomic...

Citation Context

...tion tools have been developed [36, 33, 47], including complete race detectors that never produce false alarms [34, 37, 7]. For atomicity, a variety of dynamic analyses have also been developed (e.g. =-=[11, 44, 1, 43, 46]-=-), but these tools are all incomplete in that they report false alarms on some correctly-synchronized programs. For example, the Atomizer [11] uses the Lockset algorithm [36] to reason about standard ...