Dwyer, M. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachusetts, Sept. 1995. http:##www.cis.ksu.edu#~dwyer#papers#thesis.ps.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....practical in general, with its help we were able to compute precise information for many examples. For these examples, the information computed by the MHP algorithm was remarkably close to that of the reachability technique. At present, the MHP algorithm is being used as part of the FLAVERS tool [6, 19] for data flow analysis of concurrent programs. In the future, we plan to extend the MHP algorithm to apply to programs containing procedure and function calls without using inlining. Even in its current form, the MHP algorithm can be easily used to support a limited form of interprocedural MHP ....

M. Dwyer. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachussetts, Amherst, 1995.

....After passing through the node marked RecordState, which does not affect the product automaton, the transition on the next event in the execution trace, Pressed=t, leads to the violation state for the product automaton, which signifies that this execution trace corresponds to an infeasible path. [2] proves convergence of this algorithm to the minimal fixed point and reports the analysis complexity for concurrent systems as O(jSjjN j 2 ) In the worst case a task automaton needs to be constructed for each task. Since the number of states in a task automaton is linear in the number of nodes ....

M. Dwyer. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachussetts, Amherst, 1995.

....issue for any conservative analysis is its accuracy : how closely do the behaviors of the model match the behaviors of the program. Data flow analysis can be employed to yield a conservative analysis of a program s properties, from potential cyclic deadlocks [30] to general safety properties [15, 16]. Although these techniques do not really analyze the flow of data in a program, they employ the same algorithms to find the fixed point of a set of flow equations. Integer programming has been used in the analysis of certain kinds of deadlocks [34] and in a conservative analysis for general ....

M. B. Dwyer. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachusetts at Amherst, 1995.

....is very modest, comparing timings for different requirements suggests some interesting tendencies. For example, verifying the passive states requirement on the 3WHS protocol confirms the hypothesis that FLAVERS analysis cost is influenced by the degree to which a requirement is centralized [Dwy95] We call a requirement centralized when it is specified in terms of events from only a small group of distributed processes in a protocol specification. Centralized requirements are usually easier to check than non centralized ones because fewer variables have to be modeled. For example, the ....

Matthew Dwyer. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachussetts, Amherst, 1995.

....1, 2. From there the transition on event RecordState is taken to state 1, 5. Finally, the transition on the next event in the execution trace, Car.CloseDoors, leads to the violation state for the product automaton, which signifies that this execution trace corresponds to an infeasible path. [2] proves convergence of this algorithm to the maximal fixed point and reports the complexity to be O(jSjjN j 2 ) In the worst case a task automaton needs to be constructed for each task. Since the number of states in a task automaton is linear in the number of nodes in the control flow graph ....

M. Dwyer. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachussetts, Amherst, 1995.

....results of the comparison of our algorithm and non concurrency analysis, and Section 5 discusses some conclusions and describes future work. 2. THE MHP ALGORITHM 2.1. Program representation The program representation used in this work is the trace flow graph (TFG) introduced by Dwyer and Clarke [3,4]. This representation is conservative in the sense that it models a superset of all feasible program executions. Informally, TFGs are forests of control flow graphs (CFGs) one for each concurrent process, or task, in the program, with nodes and edges added to represent intertask communications. ....

....in general, with its help we were able to compute perfectly precise information for many examples. For these examples, the information computed by the MHP algorithm was remarkably close to that of the reachability technique. At present, the MHP algorithm is being used as part of the FLAVERS tool [3, 11] for data flow analysis of concurrent programs. In the future, we plan to extend the MHP algorithm to apply to programs containing procedure and function calls without using inlining. Even in its current form, the MHP algorithm can be easily used to support a limited form of interprocedural MHP ....

M. Dwyer. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachussetts, Amherst, 1995.

....results of the comparison of our algorithm and nonconcurrency analysis, and Section 5 discusses some conclusions and describes future work. 2 THE MHP ALGORITHM 2. 1 Program representation The program representation used in this work is the trace flow graph (TFG) introduced by Dwyer and Clarke [3, 4]. This representation is conservative in the sense that it models a superset of all feasible program executions. Informally, TFGs are forests of control flow graphs (CFGs) one for each concurrent process, or task, in the program, with nodes and edges added to represent intertask communications. ....

....help we were able to compute perfectly precise information for small enough examples. For these examples, the information computed by the MHP algorithm was remarkably close to that of the reachability technique. At present, the MHP algorithm is being used implemented as part of the FLAVERS tool [3, 11] for data flow analysis of concurrent programs. In the future, we plan to extend the MHP algorithm to programs containing conventional procedure and function calls. Even in its current form, the MHP algorithm can be easily used to support a limited form of interprocedural MHP analysis, with the ....

M. Dwyer. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachussetts, Amherst, 1995.

No context found.

M.B. Dwyer. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachusetts at Amherst, September 1995.

No context found.

M.B. Dwyer. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachusetts at Amherst, September 1995.

....as a collection of Ada generic packages that provide the interfaces defined by the architecture. This implementation has been used to construct a variety of data flow analyzers that are used in a toolset for analyzing explicitly stated correctness properties of distributed systems [DC94, Dwy95] These include analyzers for traditional data flow problems, e.g. dominators, live variables, constant propagation, and for non traditional problems, e.g. for complex reachability problems. These analyzers have been formulated over a variety of program representations, e.g. sequential and ....

....flow predecessors and inter process synchronization predecessors may be treated differently. Recently data flow frameworks and solution algorithms have been extended to support analysis of concurrent programs by allowing different combining operators to be defined for different classes of nodes[Dwy95] We incorporate this extension in our architecture. For decades data flow analyzers have been an integral part of optimizing compilers. Well engineered compilers often provide standard interfaces to analyzer components. These interfaces can insulate the majority of the compiler from changes to ....

[Article contains additional citation context not shown here]

Matthew B. Dwyer. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachusetts, September 1995.

....of sub system and system level analyses. Since a primary goal of our work is to evaluate the practicality of using formal specifications and verification techniques, we have spent significant effort in developing automated tools that support analysis approaches, for example the FLAVERS [Dwy95] and TPN [DC96a] toolsets. By applying theses tools to the analysis of formal specifications derived from requirements of realistic software systems we can begin to understand the pragmatic issues of scalability and performance that are crucial to wide spread acceptance of automated formal ....

M.B. Dwyer. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachusetts at Amherst, September 1995. 9

....results the precision of the flow analysis can be increased. This flexibility allows users to begin with a low cost analysis, and then incorporate as much information as is necessary to obtain precise analysis results for the specification in question. FLAVERS is described in detail in [5]. Here we give a brief overview of FLAVERS major components. 2.1 Trace Flow Graph The trace flow graph (TFG) is a model of the set of feasible executions of a concurrent program. It is conservative in the sense that all feasible executions are included in the model; some infeasible executions ....

....indicates an infeasible path; in this case, no output tuple is produced for the node. 2.4 The FLAVERS Ada Toolset Two generations of toolsets have been developed that implement the FLAVERS technique for reasoning about Ada tasking programs. The first generation toolset was implemented in Ada [5] and was applied to the analysis of a number of concurrent Ada applications [2, 17] These studies demonstrated that FLAVERS is capable of effectively reasoning about correctness properties of realistic systems. A second generation toolset has re implemented the state propagation analysis in Java. ....

M. Dwyer. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachusetts at Amherst, Sept. 1995.

....to reduce the cost of analysis. Modular analyses can use proven properties of a sub system to reduce the cost of analyzing the remainder of a program. The work describe in this paper is able to leverage off all of these benefits of modular analyses. This paper extends the FLAVERS [DC94, Dwy95] analysis approach. To use FLAVERS, developers define a set of program events that they want to reason about and specify properties of concurrent programs as patterns of those program events. They then formulate an analysis to verify that program execution satisfes the specified property. ....

....model and flow analysis algorithms can be increased. This flexibility allows the analyst to begin with a low cost analysis then incorporate only as much information as is necessary to obtain precise analysis results for the specification in question. FLAVERS is described in detail in [DC94, Dwy95] Our goal here is to provide sufficient background to enable the reader to understand how we have modified the analysis approach for modular analysis. FLAVERS consists of three major components. 3.1 Trace Flow Graph The trace flow graph (TFG) is a model of the set of feasible program ....

[Article contains additional citation context not shown here]

M.B. Dwyer. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachusetts at Amherst, September 1995.

....local nodes. These local nodes are re labeled with , since their joint execution is now captured by the label on the communication node. The third, and final, phase creates MIP edges between all TFG nodes in different tasks. The full details of this construction algorithm are given in [Dwy95] 4.5 Conservativeness of the TFG Model Conservative state propagation analysis requires that the TFG be conservative with respect to all executable sequences of program events. To satisfy this requirement, we could produce a program model that includes a path for each interleaving of the ....

....eliminates MIP edges from the TFG that connect nodes in communicating tasks that, because of the pattern of task synchronization used, cannot immediately precede one another. In Section 8, we present data on how CI refinement increases the precision of analysis results, but we refer the reader to [Dwy95] for the details. We note that recent work [NA98] that exploits may happen in parallel analysis subsumes CI refinement. Alphabet Refinement The resulting TFG can be further transformed to eliminate edges and nodes that do not add to the set of non event sequences. These transformations ....

[Article contains additional citation context not shown here]

M.B. Dwyer. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachusetts at Amherst, September 1995.

....A wide variety of approaches have been developed to address this problem for whole programs. It is well known that early life cycle validation of system components offers the potential to reduce overall development cost. In this paper, we present an extension to the FLAVERS analysis approach [8, 10] that can be applied to component implementations at the unit level, such that, the analysis results can be leveraged effectively for integration and system level analyses. FLAVERS applied to program components is refered to as modular FLAVERS. This work was supported in part by NSF and DARPA ....

....of both the program model and flow analysis algorithms can be increased. This flexibility allows users to begin with a low cost analysis then incorporate only as much information as is necessary to obtain precise analysis results for the specification in question. FLAVERS is described in detail in [8, 10]. Our goal here is to provide sufficient background so that the reader can understand how we have modified the analysis approach for modular analysis. FLAVERS consists of three major components. 3.1. Trace Flow Graph The trace flow graph (TFG) is a model of the set of feasible program ....

[Article contains additional citation context not shown here]

M. Dwyer. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachusetts at Amherst, Sept. 1995.

....generators as a collection of Ada generic packages that provide the interfaces defined by the architecture. This implementation has been used to construct a variety of data flow analyzers that are used in a toolset for analyzing explicitly stated correctness properties of distributed systems [8, 9]. These include analyzers for traditional data flow problems, such as dominators, live variables, and constant propagation, and for non traditional problems, such as complex reachability problems. These analyzers have been formulated over a variety of program representations, including sequential ....

....and inter process synchronization predecessors may be treated differently. Recently, data flow frameworks and solution algorithms have been extended to support analysis of concurrent and distributed programs by allowing different combining operators to be defined for different classes of nodes[8]. We incorporate this extension in our architecture. For decades data flow analyzers have been an integral part of optimizing compilers. Well engineered compilers often provide standard interfaces to analyzer components. These interfaces can insulate the majority of the compiler from changes to a ....

[Article contains additional citation context not shown here]

Matthew B. Dwyer. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachusetts, September 1995.

No context found.

Dwyer, M. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachusetts, Sept. 1995. http:##www.cis.ksu.edu#~dwyer#papers#thesis.ps.

No context found.

Dwyer, M. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachusetts, Sept. 1995. http:##www.cis.ksu.edu#~dwyer#papers#thesis.ps.

No context found.

M. Dwyer. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachussetts, Amherst, 1995.

No context found.

Dwyer, M. Data Flow Analysis for Verifying Correctness Properties of Concurrent Programs. PhD thesis, University of Massachusetts, Sept. 1995. #http:##www.cis.ksu.edu#~dwyer#papers#thesis.ps#.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC