1. Abstract Interpretation with Alien Expressions and Heap Structures Bor-Yuh Evan ChangK. Rustan M. Leino University of California, BerkeleyMicrosoft Research January 18, 2005 VMCAI 2005 Paris, France

2. 2 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Verifying Object-Oriented Programs OO Program Verifier Inference… Java /C#   Abstract Interpretation

3. 3 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Problem and Motivation Standard abstract interpretation infer properties following a domain specific- schema of relations among (program) variables –e.g., can infer this with Polyhedra [CH78] 0 · x · y z := 2 ¢ y – 2 ¢ x; 0 · z

4. 4 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Problem and Motivation But … this.x 0 · this.x · y this.x z := 2 ¢ y – 2 ¢ this.x; 0 · z? length(x) 0 · length(x) · y length(x) z := 2 ¢ y – 2 ¢ length(x); 0 · z? this.xo  this 0 · this.x · y Æ o  this o.x := o.x := 2 ¢ y this.x z := 2 ¢ y – 2 ¢ this.x; 0 · z? alien expression to Polyhedra

5. 5 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Goal Given a base abstract domain that can represent certain kind of constraints on variables, use it to represent constraints on arbitrary alien expressions (e.g., fields of objects)

6. 6 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Outline Overview Handling Alien Expressions Handling Heap Updates Concluding Remarks

7. 7 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Overview of Contributions To extend base domains to work with alien expressions –use a general abstract domain parameterized by base domains that hide alien expressions as fresh variables (cf. Nelson-Oppen) –congruence-closure abstract domain To deal with heap updates –track successive heaps as a separate base domain –heap succession abstract domain

8. 8 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Fooling the Base Domains Congruence-Closure Abstract Domain Polyhedra Constrain( sel(H,o,f) ¸ 8 ) assume o.f ¸ 8 Constrain(  ¸ 8 ) sel(H,o,f)   Base Domains SymbolicValue

9. 9 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Understands : FunSymbol £ Expr[] ! bool Understandable to the Base Domain + sel Hof ² Abs 2 ¢ x + sel(H,o,f) · Abs(y – z) 2xyz Yes No Understands ·

10. 10 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Understandable to the Base Domain · +  ² Abs 2 ¢ x +  · Abs(y – z) 2xyz Understands : FunSymbol £ Expr[] ! bool No No

11. 11 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Understandable to the Base Domain +  ²  2 ¢ x +  ·  2xyz Understands : FunSymbol £ Expr[] ! bool No Yes   = y - z Also, add this constraint to Polyhedra ·

12. 12 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Congruence-Closure Domain equivalence graph (e-graph)Store mappings in an equivalence graph (e-graph) –give the same symbolic value for equivalent expressions Tracks equalities of uninterpreted functions –an e-graph with abstract domain operations –symbolic values “name” equivalence classes of expressions –implements congruence closure

13. 13 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures E-Graph w = f(x) Æ g(x,y) = f(y) Æ w = h(w) A set of mappings: w   x   f(  )   y   g( ,  )   f(  )   h(  )   Always congruence-closed    w x g h  y f f

14. 14 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Join Roughly, join the e-graphs, then join the base domains G0G0 P0P0 Base Domains G1G1 P1P1 G 0 t G 1 P 0 t P 1 Base Domains

15. 15 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Join of E-Graphs Think of the lattice over conjunctions of equalities (including infinite ones) Let G = Join(G 0,G 1 ) x  G h  0,  0 i if x  G 0  0 and x  G 1  f( h ,  i )  G h  0,  0 i if f(  )  G 0  0 and if f(  )  G 1  0 Rename distinct pairs to fresh symbolic values  x f  x f  f  x f  f h,i à h,i à  h,i à h,i à  Tell base domains about renaming

16. 16 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Join of E-Graphs Complexity: O(n ¢ m) Complete? As precise as possible? –No, e-graphs do not form a lattice! x = y t g(x) = g(y) Æ x = f(x) Æ y = f(y) = Æ i : i ¸ 0 g(f i (x)) = g(f i (y)) –Only relatively complete [Gulwani et al. 2004]

17. 17 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Widen Widen the e-graphs, then widen the base domains Widen of e-graphs is a join of e-graphs that limits the number of new names introduced (see paper)

18. 18 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures So Far We Have … Reasoning for uninterpreted functions Base domains that work with alien expressions transparently What we need for field reads – sel is alien to all base domains

19. 19 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Outline Overview Handling Alien Expressions Handling Heap Updates Concluding Remarks

20. 20 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Heap Updates Java/C#if (p.g == 8) { o.f = x; } Guardedassume H[p,g] == 8; CommandsH := H 0 where sel(H 0,o,f) = x and H 0 ´ o,f H

21. 21 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Heap Updates Guardedassume H[p,g] == 8; CommandsH := H 0 where sel(H 0,o,f) = x and H 0 ´ o,f H AbstractConstrain( sel(H,p,g) = 8 ) InterpreterConstrain( sel(H 0,o,f) = x ) Constrain( H 0 ´ o,f H ) Eliminate( H ) Rename( H 0, H ) Tracked by a new base domain: Heap Succession

22. 22 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Heap Update Example Heap Succession H 0 ´ o,f H E-Graph sel(H,p,g)   8   sel(H 0,o,f)   x   H  Hp  p H 0  H 0 g  g o  of  f Constrain( sel(H,p,g) = 8 ) Constrain( sel(H 0,o,f) = x ) Constrain( H 0 ´ o,f H ) Eliminate( H ) Rename( H 0, H ) ToPredicate()

23. 23 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Constrain( sel(H,p,g) = 8 ) Constrain( sel(H 0,o,f) = x ) Constrain( H 0 ´ o,f H ) Eliminate( H ) Rename( H 0, H ) ToPredicate() Heap Update Example Heap Succession H 0 ´ o,f H E-Graph sel(H,p,g)   8   sel(H 0,o,f)   x   H  Hp  p H 0  H 0 g  g o  of  f Only removes mapping “Lazy quantifier elimination” “Garbage values” remain

24. 24 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Constrain( sel(H,p,g) = 8 ) Constrain( sel(H 0,o,f) = x ) Constrain( H 0 ´ o,f H ) Eliminate( H ) Rename( H 0, H ) ToPredicate() Heap Update Example Heap Succession H 0 ´ o,f H E-Graph sel(H,p,g)   8   sel(H 0,o,f)   x   H  Hp  p H  H 0 o  of  f g  gg  g

25. 25 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Constrain( sel(H,p,g) = 8 ) Constrain( sel(H 0,o,f) = x ) Constrain( H 0 ´ o,f H ) Eliminate( H ) Rename( H 0, H ) ToPredicate() Heap Update Example Heap Succession H 0 ´ o,f H E-Graph sel(H,p,g)   8   sel(H 0,o,f)   x   H  Hp  p H  H 0 g  g o  of  f 1.Do Eliminate (H) EquivalentExpr : Queryable £ Expr £ Var ! Expr option Can you give me an equivalent expression without H?

26. 26 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Constrain( sel(H,p,g) = 8 ) Constrain( sel(H 0,o,f) = x ) Constrain( H 0 ´ o,f H ) Eliminate( H ) Rename( H 0, H ) ToPredicate() Heap Update Example Heap Succession H 0 ´ o,f H E-Graph sel(H 0,p,g)   8   sel(H 0,o,f)   x   H  Hp  p H  H 0 g  g o  of  f 1.Do Eliminate (H) EquivalentExpr : Queryable £ Expr £ Var ! Expr option Eliminate(H) on Base 2.ToPredicate() on Base and Convert Expr for Client 3.Conjoin Equalities Yes, use H 0 H 0 To query other abstract domains (e.g., o  p?)

27. 27 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Related Work Join for Uninterpreted Functions [Gulwani, Tiwari, Necula 2004] –same as our join for e-graphs Shape Analysis [many] and TVLA [Sagiv, Reps, Wilhelm, …] –they abstract heap nodes into summary nodes –they use special “instrumentation predicates” whereas we use “off-the-shelf” abstract domains –could use shape analysis as base domain?

28. 28 1/18/2005Chang and Leino: Abstract Interpretation with Alien Expressions and Heap Structures Conclusion and Future Work Extended the power of abstract domains to work with alien expressions using the congruence- closure domain Added reasoning about heap updates with the heap succession domain Close to having “cooperating abstract interpreters”? –missing propagating back equalities inferred by base domains Implementation and experiments in progress

29. Thank you! Questions? Comments?