Presentation is loading. Please wait.

Presentation is loading. Please wait.

Bebop: A Symbolic Model Checker for Boolean Programs Thomas Ball Sriram K. Rajamani

Similar presentations


Presentation on theme: "Bebop: A Symbolic Model Checker for Boolean Programs Thomas Ball Sriram K. Rajamani"— Presentation transcript:

1 Bebop: A Symbolic Model Checker for Boolean Programs Thomas Ball Sriram K. Rajamani http://research.microsoft.com/slam/

2 Outline  Boolean Programs and Bebop What? Why? Results Demo  Semantics of Boolean Programs  Technical details of algorithm  Evaluation  Related Work

3 Boolean Programs: What  Model for representing abstractions of imperative programs in C, C#, Java, etc.  Features:  Boolean variables  Control-flow: sequencing, conditionals, looping, GOTOs  Procedures  Call-by-value parameter passing  recursion  Control non-determinism

4 Boolean programsBoolean programs: Why bool x,y; [1] while (true) { [2] if(x == y) { [3]y = !x; } else{ [4]x = !x; [5]y = !y; } [6]if (?) break; } [7] if(x == y) [8]assert (false); Representation of program abstractions, a la Cousots Each boolean variable represents a predicate: (i < j) (*p==i) && ( (int) p == j) (p  T), where T is recursive data type [Graf-Saidi]

5 Bebop - Results  Reachability in boolean programs reduced to context-free language reachability  Symbolic interprocedural dataflow analysis  Adaptation of [Reps-Horwitz-Sagiv, POPL’95] algorithm  Complexity of algorithm is O(E  2 n )  E = size of interprocedural control flow graph  n = max. number of variables in the scope of any label

6 Bebop - Results  Admits control flow + variables Existing pushdown model checkers don’t use variables (encode variable values explicitly in state) [Esparaza, et al.] Analyzes procedures separately  exploits procedural abstraction + locality of variable scopes  Uses hybrid representation  Explicit representation of control flow graph, as in a compiler  Implicit representation of reachable states via BDDs  Generates hierarchical trace

7 Bebop Demo!

8 Outline  Boolean Programs and Bebop  Semantics of Boolean Programs “stackless” semantics using context-free grammar  Technical details of algorithm  Evaluation  Related Work

9 Stackless Semantics  State  = p = program counter  = valuation to variables in scope at p No stack!   (B): finite alphabet over boolean program B Call (with return to p),  a valuation to Locals(p) Return to p,  a valuation to Locals(p)

10 State transition -  ->  =  (x) =  (x), x in Locals(c)  =  ’(x) =  (x), x in Locals(c)  ’(g) =  (g), g a global

11 Trace Semantics  Context-free grammar L(B) constrains allowable traces M -> M M -> M M M ->    0 -  1 ->  1 -  2 -> …  m-1 -  m ->  m is a trajectory of B iff  i -  i+1 ->  i+1 is a state transition, for all i  1  2 …  m  L(B)

12 Outline  Boolean Programs and Bebop  Semantics of Boolean Programs  Technical details of reachability algorithm Binary Decision Diagrams (BDDs) Path edges Summary edges Example  Preliminary Evaluation  SLAM Project

13 Binary Decision Diagrams  Acyclic graph data structure for representing a boolean function (equivalently, a set of bit vectors)  F(x,y,z) = (x=y) x y z 11 z 00 y z 00 z 11

14 Hash Consing + Variable Elimination x yy z 0 z 1 x yy 0 1 x y z 11 z 00 y z 00 z 11

15 Path Edges   PE(p), iff Exists initialized trajectory ending in, where e = entry(Proc(p)) Exists trajectory from to  PE(p) is a set of pairs of valuations to boolean variables in scope in Proc(p) Can be represented with a BDD!

16 Representing Path Edges with BDDs  Example PE(p) for boolean variables x,y and z: PE(p) = F(x,y,z,x’,y’,z’) = (x’=x)^(y’=y)^(z’=x^y)  BDDs also used to represent transfer functions for statements Transfer(z := x^y) = F(x,y,z,x’,y’,z’) = (x’=x)^(y’=y)^(z’=x^y)

17 decl g; void main() begin decl h; h := !g; A(g,h); skip; A(g,h); skip; if (g) then R: skip; fi end void A(a1,a2) begin if (a1) then A(a2,a1); skip; else g := a2; fi end 1 g'=0^h'=1 |g'=1^h'=0 g=g’=0^a1=a1’=0^a2=a2’=1 | g=g’=1^a1=a1’=1^a2=a2’=0 g=g’=0^a1=a1’=0^a2=a2’=1g=0^g’=1^a1=a1’=0^a2=a2’=1 Join(S,T) = { |  S,  T }

18 Summary Edges = Lift(, Pr)   1 (x) =  2 (x), x in Locals(c)  Locals don’t change   1 (g) =  d (g) and  r (g) =  2 (g), g global  Propagation of global state c: Pr() d: Proc Pr() e r <d,r><d,r> <1,2><1,2>

19 decl g; void main() begin decl h; h := !g; A(g,h); skip; A(g,h); skip; if (g) then R: skip; fi end void A(a1,a2) begin if (a1) then A(a2,a1); skip; else g := a2; fi end g=0^g’=1^a1=a1’=0^a2=a2’=1 g=0^g’=1^a1=a1’=1^a2=a2’=0g=0^g’=1^h=h’=1

20 decl g; void main() begin decl h; h := !g; A(g,h); skip; A(g,h); skip; if (g) then R: skip; fi end void A(a1,a2) begin if (a1) then A(a2,a1); skip; else g := a2; fi end 1 g'=0^h'=1 |g'=1^h'=0 g=0^g’=1^h=h’=1 g’=h’=1 g=g’=a1=a1’=a2=a2’=1

21 decl g; void main() begin decl h; h := !g; A(g,h); skip; A(g,h); skip; if (g) then R: skip; fi end void A(a1,a2) begin if (a1) then A(a2,a1); skip; else g := a2; fi end 1 g'=0^h'=1 |g'=1^h'=0 g=g’=0^a1=a1’=0^a2=a2’=1 | g=g’=1^a1=a1’=1^a2=a2’=0 g=g’=1^a1=a1’=1^a2=a2’=0 g=g’=0^a1=a1’=0^a2=a2’=1 | g=g’=1^a1=a1’=1^a2=a2’=0 |a1=a1’=0^a2=a2’=1 g=g’=1^a1=a1’=0^a2=a2’=1

22 decl g; void main() begin decl h; h := !g; A(g,h); skip; A(g,h); skip; if (g) then R: skip; fi end void A(a1,a2) begin if (a1) then A(a2,a1); skip; else g := a2; fi end g'=0^h'=1 |g'=1^h'=0 g=g’=1^a1=a1’=0^a2=a2’=1g=g’=1^a1=a1’=1^a2=a2’=0g=g’=1^h=h’=0 g'=1^h'=0 g=g’=1^a1=a1’=1^a2=a2’=0

23 Worklist Algorithm  while PE(v) has changed, for some v Determine if any new path edges can be generated New path edge comes from  Existing path edge + transfer function  Existing path edge + summary edge (transfer function for procedure calls) New summary edges generated from path edges that reach exit vertex

24 Generating Error Traces  Partition reachable states into “rings” A ring R at stmt S is numbered N iff there is a shortest trace of length N to S ending in a state in R  Hierarchical generation of error trace Skip over or descend into called procedures

25 Outline  Boolean Programs and Bebop  Semantics of Boolean Programs  Technical details of algorithm  Preliminary Evaluation Linear behavior if # vars in scope remains constant Self application of Bebop  Related Work

26 decl g; void main() begin level1(); if(!g) then reach: skip; else skip; fi end void level () begin decl a,b,c; if (g) then while(!a|!b|!c) do if (!a) then a := 1; elsif (!b) then a,b := 0,1; elsif (!c) then a,b,c := 0,0,1; else skip; fi od else ; fi g := !g; end

27

28 Application: Analysis Validation  Live variable analysis (LVA) A variable x is live at s if there is a path from s to a use of x (with no intervening def of x) Used to optimize bebop Quantify out variables as soon as they become dead  How to check correctness of LVA? Analysis validation Create a boolean program to check results of LVA Model check boolean program (w/out LVA)

29 Analysis Validation  Output of LVA: { (s,x) | x is dead at s }  Boolean program Two variables per original program var x: x_dead (initially 0) x_defined (initially 0) For each fact (s,x): x_dead, x_defined := 1, 0; For each def of x: x_defined := 1; For each use of x if (x_dead && !x_defined) LVAError();  Query: is LVAError reachable?

30 Results  Found subtle error in implementation of LVA  Was able to show colleague that there was another error, in his code  Analysis validation now part of regression test suite

31 Related Work  Pushdown Automata (PDA) decidability results  [Hopcroft-Ullman]  Model checking PDAs  [Bouajjani-Esparza-Maler] [Esparza-Hansel-Rossmanith-Schwoon]  Model checking Hierarchical State Machines  [Alur, Grosu]  Interprocedural dataflow analysis  [Sharir-Pnueli] [Steffen] [Knoop-Steffen] [Reps-Horwitz-Sagiv]

32 Related Work  Reps-Horwitz-Sagiv (RHS) algorithm Handles IFDS problems Interprocedural Finite domain D Distributive dataflow functions (MOP=MFP) Subsets of D Dataflow as CFL reachability over “exploded graph”  Our results RHS algorithm can be reformulated as a traditional dataflow algorithm over original control-flow graph with same time/space complexity Reformulated algorithm is easily lifted to powersets of D using BDDs Arbitrary dataflow functions Path-sensitive

33 Summary  Bebop: a model checker for boolean programs Based on interprocedural dataflow analysis using BDDs Exploits procedural abstraction Admits many traditional compiler optimizations Hierarchical trace generation + DHTML user interface Release at end of year  SLAM project Iteratively refine boolean program models of C programs Use path simulation to discover relevant predicates (simcl) Automated predicate abstraction (c2bp)

34 Software Productivity Tools Microsoft Research http://research.microsoft.com/slam/


Download ppt "Bebop: A Symbolic Model Checker for Boolean Programs Thomas Ball Sriram K. Rajamani"

Similar presentations


Ads by Google