Citation Context

...ety for the concurrent program. However, we emphasize that developing a scalable, precise shape analysis for concurrent programs is an open problem; only very recently, some interesting ideas such as =-=[12, 18, 5]-=- have been proposed, which give promising new lines of attack, but on which further, especially experimental, work is needed. 3 Abstract States and Setting In this section we describe the abstract sta...

Citation Context

...had automated the derivation of such proofs. Amit et al. [6] used shape analysis to verify linearizability for a fixed (small) number of threads. More recently, Manevich et al. [7] and Berdine et al. =-=[8]-=- extended this analysis, so that it works for a larger (fixed) number of threads and for an unbounded number of threads respectively. These works require a specialized abstract domain, do not handle m...

Citation Context

...ization points to be annotated, are good at quickly finding bugs, and return concrete counterexample traces for failed verifications. Second, there are static analyses (shape analyses, in particular) =-=[2, 3, 19]-=-. With the exception of [2], these analyses work for an unbounded number of threads and result in a proof of linearizability. Unfortunately, all of these analyses require the programmer to specify the...

Citation Context

...fect instantaneously at some point between its invocation and its response. Such points are commonly referred to as linearization points. Automatic verification and checking of linearizability (e.g., =-=[7, 8, 1, 23, 21, 2]-=-) and of related correctness conditions (e.g., [5, 4]) is an active area of research. Most of these methods rely on the user to specify linearization points, which typically requires an insight on how...

Citation Context

...ction that tracks the difference between two heaps. This approach works well if the concrete heap and the abstract heap have almost identical shapes. Recently, Manevich et al. [17] and Berdine et al. =-=[5]-=- extended it to handle larger number and unbounded number of threads, respectively. Vafeiadis [26] further improved this solution to allow linearization points in different threads. The main limitatio...

Citation Context

...ine-grained concurrency, but also required us to introduce concepts such as stabilization. In contrast, Gotsman et al. can handle only coarse-grained concurrency. Manevich et al. [12], Berdine et al. =-=[3]-=-, and Segalov et al. [16] have developed a series of related shape analyses that suitably restrict the correlations between the states of different threads that are tracked. This gives them a very str...

Citation Context

...rams, since predicate abstraction does not make any considerations for threads (a very recent work [11] takes a first step towards closing this gap by making predicate abstraction symmetry-aware). In =-=[4]-=-, an abstract domain construction was presented which can represent invariants of the form ∀t.ϕ(t), where t is a variable representing a thread ϕ(t) is an abstract value in some base domain. Since suc...

Citation Context

...ot tracking the state of the heap at each program point, such analyses can be directly used in a concurrent setting, while our approach requires adaptations of the TVLA approach to handle concurrency =-=[4,22]-=-. 8 Conclusion We presented a shape analysis for overlapping data-structures, which are ubiquitous in low-level systems code. Using our prototype implementation we were able to establish subtle reacha...