1 / 13

sbdb: A Semi-Automated Bidirectional DeBugger

sbdb: A Semi-Automated Bidirectional DeBugger. Prasenjit Mitra Stanford University CS 343: Research Proposal. Introduction. Debugging requires a lot of manual effort. Problem: Overshoot a bug - re-execute. Solution: Backtracking. Related Work. Execution History:

baakir
Télécharger la présentation

sbdb: A Semi-Automated Bidirectional DeBugger

An Image/Link below is provided (as is) to download presentation Download Policy: Content on the Website is provided to you AS IS for your information and personal use and may not be sold / licensed / shared on other websites without getting consent from its author. Content is provided to you AS IS for your information and personal use only. Download presentation by click this link. While downloading, if for some reason you are not able to download a presentation, the publisher may have deleted the file from their server. During download, if you can't get a presentation, the file might be deleted by the publisher.

E N D

Presentation Transcript


  1. sbdb: A Semi-Automated Bidirectional DeBugger Prasenjit Mitra Stanford University CS 343: Research Proposal

  2. Introduction • Debugging requires a lot of manual effort. • Problem: Overshoot a bug - re-execute. • Solution: Backtracking.

  3. Related Work • Execution History: • Balzer, Spring Joint Comp. Conf., 1969. • Lieberman and Fry, Software Viz., 1998. • Agrawal, et al. IEEE Software, 1991. • Netzer and Weaver, PLDI, 1994 . • Re-execution: • Tolmach and Appel, J. Func. Prog. 1993. • Bob Boothe, PLDI, 2000. • Checkpointing

  4. Execution History • Step Counter: • number of steps from start of a program. • Change-Set of a Statement: • registers, memory variables that change value • Store: • Step counter, program counter, values of variables in change-set. • Portion of stack while entering/exiting function.

  5. Coarser Granularity • Jump over loops. • Statement-level to Basic-Block-level. • Bb = {S1;S2} • Cs(Bb) = Cs(S1) U Cs(S2) - DefinedIn( S1) • Store <Var, Val> • where Val = Value of Var before S1

  6. Exponential Thinning of Execution History Log • For distant points, granularity of logs coarser. • To jump within a Bb, rollback to start and then execute forward. Per 8Bb. Per 4Bb. Per Bb. Per 2Bb. Per Stmt.

  7. Checkpoints • Large number of backward steps - inefficient. • Maintain checkpoints of program states at intermediate points. • Fork the program and use the OS’ copy on write policy. • Exponential thinning of checkpoints.

  8. Hybrid Approach • Move backward incrementally using execution history log. • Move backward using coarse-grained log and then re-execute forward. • Jump back to a checkpoint prior to program point and re-execute forward. • Jump back to a checkpoint after program point and then use execution history log.

  9. Operations • bstep n - step n • prev n - next n • bcontinue - continue • before - finish • until <cond.> - buntil <cond.> • undo • prove <condition>

  10. Comparisons • Boothe • History • Checkpoints X X X X X X X X O O O O

  11. Comparisons (contd.) • + Less checkpointing interval. • + Option to move forward or backward. • - History logs

  12. Conclusion • Bidirectional Debugger. • Hybrid approach involving • execution history logs • checkpoints • Potential to analyze programs using stored program state.

  13. And that will be…... • A million dollars please………… • Questions ???

More Related