Download presentation
Presentation is loading. Please wait.
Published byKelly Brook Shields Modified over 9 years ago
1
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University Merging Equivalent Contexts for Scalable Heap-cloning-based Points-to Analysis Guoqing Xu and Atanas Rountev Ohio State University Supported by NSF under CAREER grant CCF-0546040
2
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University 2 Precise and Scalable Points-to Analysis Analysis precision - Context sensitivity – e.g. chain of call sites - Heap cloning [Nystrom-PASTE’04, Lhotak-CC’06] - The most precise analysis: refinement-based analysis [Sridharan-PLDI’06] Analysis scalability - Millions of distinct call chains in a moderate-size Java program - Sacrifice precision: k-length chain - Merging equivalent relationships using BDDs - BDDs incurs running time overhead, and may not scale for heap-cloning-based analysis
3
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University Merge Equivalent Contexts Equivalence classes exists in the representation of calling contexts [Lhotak-CC’06] - Merging such contexts will not affect precision Can we find and merge equivalent calling contexts? - We would be able to scale the points-to analysis without relying on the merging inside the BDD “black box” A unique replacement context (URC) can replace all contexts from the same equivalence class 3
4
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University Outline A model of equivalent contexts - Abstraction functions for pointer variables and targets - Proposed for pointer analysis, but can be applied to other context-sensitive analysis algorithms A whole-program points-to analysis for Java - Implements the model - Context-sensitive for both pointer variables and targets - Does not limit the length of context strings (not k-CFA) - Bottom-up, summary-based Experimental evaluation - Much more precise and efficient than state-of-the-art 1-object-sensitive analysis with BDDs [Lhotak-CC’06] - More efficient than the refinement-based analysis 4
5
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University Motivating Example void main(String[] args){ A a1 = new A(); A a2 = new A(); foo(a1); //call site 1 foo(a2); //call site 2 } void foo(A a){ t = new B(); bar(t); //call site 3 } void bar(B b){ p = b; } 5 t (1) new B (1) t (2) new B (2) p (1,3) new B (1) p (2,3) new B (2) Observation: 1. t points to new B, under all calling contexts 2. p points to new B, under calling contexts (*,3)
6
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University A Better Representation? Can we represent the points-to relationships like this? - t new B - p 3 new B - 1 copy of t, p, and new B Key insights - Context-sensitivity corresponds to inlining; full context- sensitivity is achieved if all reachable methods are inlined in main - If a points-to relationship can be determined at method m during inlining, it will not be affected by m’s callers This is the conceptual source of context equivalence and merging 6
7
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University Hypothetical Inlining-based Analysis (v, c v, o, c o ) represents a fully context-sensitive points-to relationship; v is local var; o is alloc site Conceptual inlining - At call graph edge e, statement p := q from the callee is cloned as p (e) := q (e) in the caller - One caller up, the clone is p (e2, e) := q (e2, e) and so on … Bottom-up analysis - After all call sites in a method m are inlined, an intraprocedural analysis is performed for m - Suppose (v, c v, o, c o ) is produced for m - For a call edge e from n to m: (v, (e) c v, o, (e) c o ) will definitely be produced later for n 7
8
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University Calling Context Reduction Consider a tuple (v, c v, o, c o ) computed for main Its lifetime consists of - A single creation event in some method m This method is the flowing point for the tuple - A sequence of inlining steps that increase both c v and c o in synch URCs computed by abstraction functions for calling context; m is the flowing point - c v is mapped to a suffix (e 0, e 1, …, e i ) where e 0.src = m - c o is mapped to a suffix (f 0, f 1, …, f j ) where f 0.src = m - Keep only the relevant suffix of the call chains 8
9
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University Using the Reduced Contexts A URC is used to represent a set of calling contexts in the points-to relationships A query (v, c) can be answered as follows: - Find all (v, urc v, o, urc o ) such that suffix(urc v, c) holds - Return all (o, c o ) such that suffix(urc o, c o ) holds Generalization for recursion – see the paper A BDD may not be as effective for an analysis with heap cloning - Without heap cloning, an equivalence class is defined by a single string urc v - For an analysis with heap cloning, an equivalence class is defined by a pair (urc v, urc o ) - More classes = fewer opportunities for merging 9
10
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University Points-to Analysis A specific algorithm that implements this model - Using URCs to represent calling contexts - The use of URCs could be applicable to other categories of points-to analysis Resembles bottom-up inlining Heap-cloning-based - Context-sensitively treat both pointer variables and targets Partial unification (bi-directional flow of values) 10
11
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University Intraprocedural Analysis Symbolic points-to graph (SPG) - A symbolic object node is introduced for each (1) formal parameter, (2) base variable v of load a = v.f, and (3) lhs v of a call site v = a.b(…) - Standard points-to analysis algorithm [Lhotak-CC’03] is used for SPG construction - SPG contains much fewer nodes and edges than the original program Example void add (Integer t) { this.names = t; } 11
12
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University Escape Analysis An allocation node new C or symbolic node SO directly escapes a method, if - It is pointed to by a formal parameter - It is pointed to by a returned variable - It is pointed to by a static field A node indirectly escapes a method, if - It is reachable from nodes that directly escape Compute a set of allocation/symbolic nodes that escape the method where they are defined 12
13
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University Interprocedural Analysis Summary-based - Bottom-up traverse the call graph SCC-DAG Summary function definition: set of [O f, G f ] - O f : … - G f : the subgraph of all escaping objects (reachable from O f ) and their points-to edges Clone a summary function for each incoming call graph edge e - If o 1 c1 o 2 c2 where o 1 and o 2 escape: in the caller, create o 1 (e) c1 o 2 (e) c2 - If v c1 s c2 where s is an escaping symbolic node: create v (e) c1 s (e) c2 13 f f
14
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University Interprocedural Analysis Composition of summary functions - For each [O actual, G actual ] - Find [O formal, G formal ] - Merge G actual and G formal Subgraph merging - Simultaneously traverse G actual and G formal from O actual and O formal - Merge b and c, if where d and e have been merged 14
15
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University Merging Two Nodes 15
16
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University Interprocedural Analysis The points-to solution is built on the fly - Once v c1 o c2 is formed, (v, c 1, o, c 2 ) is added to the points-to solution - The edge is removed from the SPG: we have found the flowing point and no further propagation is necessary - (c 1, c 2 ) defines an equivalence class for contexts Node merging is essentially a dynamic transitive closure computation for identifying memory aliases 16
17
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University K-Last-Substring-Merging A scalability-precision tradeoff - When composing summary functions, do not clone node o c1 in the callee if there already exists a node o c2 in the caller such that suffix(c 1, k) == suffix(c 2, k) Does not limit the context length Example - k =3 - O (d,e,f,g) and O (h,e,f,g) are distinct nodes if d.src != h.src - O (d,e,f,g) and O (h,e,f,g) are merged if d.src == h.src 17
18
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University Experiments Benchmark set contains 19 Java programs, from SPECJVM, Ashes, and DaCapo Experimentally compared our analysis with - Refinement: the refinement-based analysis from [Sridharan-PLDI’06] with its default budget for refinement the most precise publicly-available analysis for computing an on-demand solution queried the points-to sets for all possible (v, c) - 1H: 1-object-sensitive analysis with heap cloning, using BDDs [Lhotak-CC’06]; the most precise publicly-available analysis for computing a whole-program solution Our analysis: computes a whole-program solution - 1Equiv: 1-last-substring-merging - 2Equiv: 2-last-substring-merging 18
19
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University Precision: #downcasts proven to be safe 19 2Equiv proves 59% more safe downcasts than 1H But less than Refinement
20
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University Cost: time to get a whole-program solution 20 2Equiv is 13 times faster than 1H 5.3 times faster than Refinement
21
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University Conclusions Context equivalence class identification - Only URCs need to be explicitly represented in the data structures of the analysis A points-to analysis for Java - Heap cloning, summary-based, bottom-up, with partial unification Experimental evaluation - Precision approaches that of the refinement-based analysis, and is much higher than that of 1H - Significantly faster than both refinement-based and 1H
22
PRESTO: Program Analyses and Software Tools Research Group, Ohio State University Approximation in the Presence of Recursion Two-phase approximation: (1) map an infinite call chain to a finite one, collapsing cycles while going backwards; (2) then, use the abstraction function shown earlier 22 Precision loss may result from phase 1: e.g., p points to o under context eabcdabf, but does not under context eabf
Similar presentations
© 2024 SlidePlayer.com. Inc.
All rights reserved.