Formalized Verification of Snapshotable Trees: Separation and Sharing

Mehnert, Hannes; Sieczkowski, Filip; Birkedal, Lars; Sestoft, Peter

doi:10.1007/978-3-642-27705-4_15

Cited by 11 publications

(13 citation statements)

References 15 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In [12], the verification of snapshotable trees is proposed as a challenge. The problem is very similar to the CCoWLs: the clients see a mutable tree and immutable snapshots of previous states of that tree.…”

Section: Discussionmentioning

confidence: 99%

A Discipline for Program Verification Based on Backpointers and Its Use in Observational Disjointness

Kassios

Kritikos

2013

Programming Languages and Systems

View full text Add to dashboard Cite

Abstract. In the verification of programs that manipulate the heap, logics that emphasize localized reasoning, such as separation logic, are being used extensively. In such logics, state conditions may only refer to parts of the heap that are reachable from the stack. However, the correct implementation of some data structures is based on state conditions that depend on unreachable locations. For example, reference counting depends on the invariant that "the number of nodes pointing to a certain node is equal to its reference counter ". Such conditions are cumbersome or even impossible to formalize in existing variants of separation logic. In the first part of this paper, we develop a minimal programming discipline that enables the programmer to soundly express backpointer conditions, i.e., state conditions that involve heap objects that point to the reachable part of the heap, such as the above-mentioned reference counting invariant. In the second part, we demonstrate the expressiveness of our methodology by verifying the implementation of concurrent copy-on-write lists (CCoWL). CCoWL is a data structure with observational disjointness, i.e., its specification pretends that different lists depend on disjoint parts of the heap, so that separation logic reasoning is made easy, while its implementation uses sharing to maximize performance. The CCoWL case study is a very challenging problem, to which we are not aware of any other solutions.

show abstract

Section: Discussionmentioning

confidence: 99%

A Discipline for Program Verification Based on Backpointers and Its Use in Observational Disjointness

Kassios

Kritikos

2013

Programming Languages and Systems

View full text Add to dashboard Cite

show abstract

“…The client code is computationally equivalent to the original challenge [10], we do not introduce an unnecessary boolean variable for the loop condition.…”

Section: The Methods Of the Itree Interface Have The Following Effectsmentioning

confidence: 99%

“…We have verified the A1B1 implementation [10], which does not implement rebalancing and uses path copy persistence: when a snapshot is present, the complete path from the root down to the newly inserted node is copied in a call to add. This ensures that add does not mutate any node that is shared between the snapshot and the tree.…”

Section: Proof Patterns and Verification Of The Implementationmentioning

confidence: 99%

“…Snapshotable trees have been proposed as a verification challenge [10], because they contain abstract separation and internal sharing: the implementation uses sharing, while the user sees each tree and snapshot separately. The complete verified code is available at http://www.itu.dk/people/hame/SnapTree.java.…”

Section: Introductionmentioning

confidence: 99%

“…We will first recapitulate the snapshotable tree verification challenge [10], typestate, and access permissions. Then we will briefly describe Plural and introduce our solution to the challenge.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Verification of Snapshotable Trees Using Access Permissions and Typestate

Mehnert

Aldrich

2012

Objects, Models, Components, Patterns

Self Cite

View full text Add to dashboard Cite

Abstract. We use access permissions and typestate to specify and verify a Java library that implements snapshotable search trees, as well as some client code. We formalize our approach in the Plural tool, a sound modular typestate checking tool. We describe the challenges to verifying snapshotable trees in Plural, give an abstract interface specification against which we verify the client code, provide a concrete specification for an implementation and describe proof patterns we found. We also relate this verification approach to other techniques used to verify this data structure.

show abstract

A Fully Verified Container Library

Polikarpova¹,

Tschannen

Furia

2015

FM 2015: Formal Methods

View full text Add to dashboard Cite

Abstract. The comprehensive functionality and nontrivial design of realistic general-purpose container libraries pose challenges to formal verification that go beyond those of individual benchmark problems mainly targeted by the state of the art. We present our experience verifying the full functional correctness of EiffelBase2: a container library offering all the features customary in modern language frameworks, such as external iterators, and hash tables with generic mutable keys and load balancing. Verification uses the automated deductive verifier AutoProof, which we extended as part of the present work. Our results indicate that verification of a realistic container library (135 public methods, 8400 LOC) is possible with moderate annotation overhead (1.4 lines of specification per LOC) and good performance (0.2 s per method on average).

show abstract

Formalized Verification of Snapshotable Trees: Separation and Sharing

Cited by 11 publications

References 15 publications

A Discipline for Program Verification Based on Backpointers and Its Use in Observational Disjointness

A Discipline for Program Verification Based on Backpointers and Its Use in Observational Disjointness

Verification of Snapshotable Trees Using Access Permissions and Typestate

A Fully Verified Container Library

Contact Info

Product

Resources

About