Alternation for Termination

1. Alternation for Termination 1 2 2 2 William Harris, AkashLal, AdityaNori SriramRajamani 1 2

2. Termination bugs are a real problem in systems and application code.

3. A Quick Search “bug code hangs”: “Gecko mediaplayer hangs the browser” “Eclipse hangs after 5 minutes or so of working” “BUG: Silverlight makes browser hang after BeginSaveChanges on some machines” “BUG: VB Hangs While Automating Excel Using OLE Control” …

4. Key challenge to proving termination:Analyzing the context of a loop

5. An Example with Non-Trivial Context f(int d, z) { int x, y; while (x > 0 && y > 0) { if (*) { x := x – d; y := *; z := z – 1; } else { y := y – d; } } } main() { int k; int z = 1; while (z < k) { z := 2 * z; } f(1, z); f(2, z); }

6. Local Termination Provers For a fixed over-approximation of a loop, find a proof of termination

7. Local ProversSucceeding while (x > 0 && y > 0) { • assume(d > 0); if (*) { x := x – d; y := *; z := z – 1; } else { y := y – d; } } x y

8. Local Provers Failing f(intd) { intx, y; while (x > 0 && y > 0) { assume(d > 0); if (*) { x := x – d; y := *; z := z – 1; } else { y := y – d; } } } main() { f(1); f(2); } ??

9. Transition Invariants From stem and cycle of a loop, guess and check a proof of termination

10. Advantage of Transition Invariants A stem to a loop can include information about the loop’s context.

11. Transition Invariants Succeeding f(int d) { while (x > 0 && y > 0) { if (*) { x := x – d; y := *; } else { y := y – d; } } } while (x > 0 && y > 0) { x := x – d; y := *; } x • main() { • f(1); • f(2); }

12. Transition Invariants Succeeding f(int d) { while (x > 0 && y > 0) { if (*) { x := x – d; y := *; } else { y := y – d; } } } while (x > 0 && y > 0) { y := y - d; } y • main() { • f(1); • f(2); }

13. Disadvantage of Transition Invariants Stem and cycle can lead to incorrect guesses for proof of termination.

14. Transition Invariants Failing f(intd) { f(int d, int z) { int x, y; while (x > 0 && y > 0) { if (*) { x := x – d; y := *; z := z – 1; } else { y := y – d; } } } main() { int k; int z = 1; while (z < k) { z := 2 * z; } f(1); f(1, z); f(2); f(2, z); }

15. Key Insight of TREX From cycles through a loop, inferinvariants for proving termination.

16. Context Analysis via TREX f(int d, z) { int x, y; while (x > 0 && y > 0) { assume(d > 0); if (*) { x := x – d; y := *; z := z – 1; } else { y := y – d; } } } main() { int k; int z = 1; while (z < k) { z := 2 * z; } f(1, z); f(2, z); }

17. Payoff of TREX’s Approach TREX can apply local proversto find a proof of termination quickly

18. Analysis via TREX f(int d, z) { int x, y; while (x > 0 && y > 0) { assume(d > 0); if (*) { x := x – d; y := *; z := z – 1; } else { y := y – d; } } } main() { int k; int z = 1; while (z < k) { z := 2 * z; } f(1, z); f(2, z); } x, y

19. TREX in More Detail • TREX by example • Experiments

20. TREX iterativelyfinds a proof of termination,or finds a counterexample to termination, or refinesstronger program invariants The TREX Algorithm

21. TREX IterationStep 1 Find a proof of termination by applying a local termination prover

22. TREX IterationStep 1 f(int d, z) { int x, y; while (x > 0 && y > 0) { if (*) { x := x – d; y := *; z := z – 1; } else { y := y – d; } } } main() { int k; int z = 1; while (z < k) { z := 2 * z; } f(1, z); f(2, z); } ??

23. TREX Iteration Step 2 If local prover fails, then find a counterexample cycle

24. TREX Iteration Step 2 f(int d, z) { int x, y; while (x > 0 && y > 0) { if (*) { x := x – d; y := *; z := z – 1; } else { y := y – d; } } } main() { int k; int z = 1; while (z < k) { z := 2 * z; } f(1, z); f(2, z); } while (x > 0 && y > 0) { y := y – d; }

25. TREX Iteration Step 3 From the counterexample cycle, find a sufficient condition for non-terminationby applying a non-termination prover(TNT)

26. Applying a Non-Termination Prover while (x > 0 && y > 0) { y := y – d; } Non-termination if: y > 0 && d <= 0

27. TREX IterationStep 4 Check if the sufficient condition is reachable

28. TREXIterationStep 4 Non-termination if: y > 0 && d <= 0 f(int d, z) { int x, y; while (x > 0 && y > 0) { assert(d > 0); if (*) { x := x – d; y := *; z := z – 1; } else { y := y – d; } } } main() { int k; int z = 1; while (z < k) { z := 2 * z; } f(1, z); f(2, z); } while (x > 0 && y > 0) { y := y – d; }

29. TREX Iteration Step 5 If the sufficient condition is unreachable, thenassume this as an invariant.

30. TREX Iteration Step 5 f(int d, z) { int x, y; while (x > 0 && y > 0) { assert(d > 0); if (*) { x := x – d; y := *; z := z – 1; } else { y := y – d; } } } main() { int k; int z = 1; while (z < k) { z := 2 * z; } f(1, z); f(2, z); } assume(d > 0); x, y

31. Experiments Windows Vista driver snippets

32. Vista Driver Snippets

33. Conclusion TREX proves termination by using cycles through a loop to infer useful program invariants

34. Extra slides

35. Transition Invariants Succeeding f(int d) { while (x > 0 && y > 0) { if (*) { x := x – d; y := *; } else { y := y – d; } } } • main() { • f(1); • f(2); } x, y

36. Transition Invariants Failing f(intd) { f(int d, int z) { int x, y; while (x > 0 && y > 0) { if (*) { x := x – d; y := *; z := z – 1; } else { y := y – d; } } } main() { int k; int z = 1; while (z < k) { z := 2 * z; } f(1); f(1, z); f(2); f(2, z); } z = 1; f(1, z); while (x > 0 && y > 0) { assume(d = 1 && z = 1); if (*) { x := x – d; y := *; z := z – 1; } } z - 1

37. Transition Invariants Failing f(intd) { f(int d, int z) { int x, y; while (x > 0 && y > 0) { if (*) { x := x – d; y := *; z := z – 1; } else { y := y – d; } } } main() { int k; int z = 1; while (z < k) { z := 2 * z; } f(1); f(1, z); f(2); f(2, z); } z = 1; z := 2 * z; f(1, z); while (x > 0 && y > 0) { assume(d = 1 && z = 2); if (*) { x := x – d; y := *; z := z – 1; } } z - 2