automated debugging for arbitrarily long executions
play

Automated Debugging for Arbitrarily Long Executions Cristian - PowerPoint PPT Presentation

Jun 29, 2023 •306 likes •945 views

Automated Debugging for Arbitrarily Long Executions Cristian Zamfir, Baris Kasikci, Johannes Kinder, Edouard Bugnion, George Candea Debugging is Hard Debugging = diagnose + fix the root cause May take days-months to diagnose bugs in the

  1. Automated Debugging for Arbitrarily Long Executions Cristian Zamfir, Baris Kasikci, Johannes Kinder, Edouard Bugnion, George Candea

  2. Debugging is Hard • Debugging = diagnose + fix the root cause • May take days-months to diagnose bugs in the real world 1 1 Concurrency at Microsoft – An Exploratory Survey , CAV workshop 2008

  3. Real World Debugging Debugging during development

  4. Real World Debugging Debugging during development $ gdb ./program

  5. Real World Debugging Debugging during development $ gdb ./program (gdb) record

  6. Real World Debugging Debugging during development $ gdb ./program (gdb) record (gdb) run

  7. Real World Debugging Debugging during development $ gdb ./program (gdb) record (gdb) run Segmentation fault

  8. Real World Debugging Debugging during development $ gdb ./program (gdb) record (gdb) run Segmentation fault (gdb) reverse-step

  9. Real World Debugging Debugging during development $ gdb ./program (gdb) record (gdb) run Segmentation fault (gdb) reverse-step

  10. Real World Debugging Debugging during development Debugging in the real world $ gdb ./program (gdb) record (gdb) run Segmentation fault (gdb) reverse-step

  11. Real World Debugging Debugging during development Debugging in the real world $ ./program $ gdb ./program (gdb) record (gdb) run Segmentation fault (gdb) reverse-step

  12. Real World Debugging Debugging during development Debugging in the real world $ ./program $ gdb ./program Segmentation fault (gdb) record (gdb) run Segmentation fault (gdb) reverse-step

  13. Real World Debugging Debugging during development Debugging in the real world $ ./program $ gdb ./program Segmentation fault (gdb) record (core dumped) (gdb) run Segmentation fault $ gdb ./program core (gdb) reverse-step

  14. Real World Debugging Debugging during development Debugging in the real world $ ./program $ gdb ./program Segmentation fault (gdb) record (core dumped) (gdb) run Segmentation fault $ gdb ./program core (gdb) reverse-step (gdb) reverse-step

  15. Real World Debugging Debugging during development Debugging in the real world $ ./program $ gdb ./program Segmentation fault (gdb) record (core dumped) (gdb) run Segmentation fault $ gdb ./program core (gdb) reverse-step (gdb) reverse-step Target core command unsupported

  16. Debug Without Recording

  17. Debug Without Recording What are the classes of information necessary for debugging?

  18. Debug Without Recording What are the classes of information Coredump Program necessary for debugging? 10101010 11101011 + 10101011 10001001

  19. Debug Without Recording What are the classes of information Coredump Program necessary for debugging? 10101010 11101011 + 10101011 10001001 Synthesize program inputs thread schedule

  20. Debug Without Recording What are the classes of information Coredump Program necessary for debugging? 10101010 11101011 + 10101011 10001001 Synthesize Original Program Binary Debugger program inputs Replay Library thread schedule

  21. Debug Without Recording Synthesize program inputs thread schedule

  22. Debug Without Recording Synthesize program inputs thread schedule

  23. Debug Without Recording Synthesize program inputs thread schedule

  24. Debug Without Recording Synthesize program inputs thread schedule

  25. Debug Without Recording Synthesize program inputs thread schedule

  26. Debug Without Recording Synthesize program inputs thread schedule

  27. Debug Without Recording Synthesize program inputs thread schedule Exact same execution is not necessary 
 ODR, PRES (SOSP’09) and ESD (EuroSys’10)

  28. Debug Without Recording Synthesize program inputs thread schedule Exact same execution is not necessary 
 ODR, PRES (SOSP’09) and ESD (EuroSys’10) Reproduce the root cause and the failure 
 Debug Determinism (HotOS’11)

  29. Debug Without Recording Synthesize program inputs thread schedule Exact same execution is not necessary 
 ODR, PRES (SOSP’09) and ESD (EuroSys’10) Reproduce the root cause and the failure 
 Debug Determinism (HotOS’11)

  30. Reverse Execution Synthesis Synthesize program inputs thread schedule

  31. Reverse Execution Synthesis Synthesize program inputs thread schedule The root cause is close to the failure 
 85% of the time (Conseq, ASPLOS’11)

  32. Reverse Execution Synthesis Synthesize program inputs thread schedule The root cause is close to the failure 
 85% of the time (Conseq, ASPLOS’11)

  33. Reverse Execution Synthesis Synthesize program inputs thread schedule The root cause is close to the failure 
 85% of the time (Conseq, ASPLOS’11)

  34. Reverse Execution Synthesis Synthesize program inputs thread schedule The root cause is close to the failure 
 85% of the time (Conseq, ASPLOS’11)

  35. Reverse Execution Synthesis

  36. Reverse Execution Synthesis x = 1; if (f(x) == y) { goto next; } ... x = 2; if (g(x) == y) { goto next; } exit(); next: buffer[y] = 1

  37. Reverse Execution Synthesis x = 1; if (f(x) == y) { goto next; } ... x = 2; if (g(x) == y) { goto next; } exit(); next: buffer[y] = 1 Coredump: x 1 y 10

  38. Reverse Execution Synthesis x = 1; if (f(x) == y) { goto next; } ... x = 2; if (g(x) == y) { goto next; } exit(); next: buffer[y] = 1 buffer[y] = 1; Coredump: (buffer overflow) x 1 y 10

  39. Reverse Execution Synthesis x = 1; if (f(x) == y) { goto next; } ... x = 2; if (g(x) == y) { goto next; } exit(); x = 1; next: if (f(x) == y) buffer[y] = 1 True buffer[y] = 1; Coredump: (buffer overflow) x 1 y 10

  40. Reverse Execution Synthesis x = 1; if (f(x) == y) { goto next; } ... x = 2; if (g(x) == y) { goto next; } exit(); x = 2; x = 1; next: if (g(x) == y) if (f(x) == y) buffer[y] = 1 True True buffer[y] = 1; Coredump: (buffer overflow) x 1 y 10

  41. Reverse Execution Synthesis x = 1; if (f(x) == y) { goto next; } ... x = 2; if (g(x) == y) { goto next; } exit(); x = 2; x = 1; next: if (g(x) == y) if (f(x) == y) buffer[y] = 1 False True True False buffer[y] = 1; Coredump: (buffer overflow) x 1 y 10

  42. Reverse Execution Synthesis x = 1; if (f(x) == y) { goto next; } ... x = 2; x x if (g(x) == y) { goto next; y y } exit(); x = 2; x = 1; next: if (g(x) == y) if (f(x) == y) buffer[y] = 1 False True True False buffer[y] = 1; Coredump: (buffer overflow) x 1 y 10

  43. Reverse Execution Synthesis x = 1; if (f(x) == y) { goto next; } ... x = 2; x x x if (g(x) == y) { goto next; y y y } exit(); x = 2; x = 1; next: if (g(x) == y) if (f(x) == y) buffer[y] = 1 False True True False buffer[y] = 1; Coredump: (buffer overflow) x 1 y 10

  44. Reverse Execution Synthesis x = 1; if (f(x) == y) { goto next; } ... x = 2; x x if (g(x) == y) { goto next; y y } x exit(); x = 2; x = 1; y next: if (g(x) == y) if (f(x) == y) buffer[y] = 1 False True True False buffer[y] = 1; Coredump: (buffer overflow) x 1 y 10

  45. Reverse Execution Synthesis x = 1; if (f(x) == y) { goto next; } ... x = 2; x x if (g(x) == y) { goto next; y y } x 2 exit(); x = 2; x = 1; y next: if (g(x) == y) if (f(x) == y) buffer[y] = 1 False True True False buffer[y] = 1; Coredump: (buffer overflow) x 1 y 10

  46. Reverse Execution Synthesis x = 1; if (f(x) == y) { goto next; } ... x = 2; x x if (g(x) == y) { goto next; y y } exit(); x = 2; x = 1; x 2 next: if (g(x) == y) if (f(x) == y) buffer[y] = 1 y 10 False True True False buffer[y] = 1; Coredump: (buffer overflow) x 1 y 10

  47. Reverse Execution Synthesis x = 1; if (f(x) == y) { goto next; } ... x = 2; x x if (g(x) == y) { goto next; y y } exit(); x = 2; x = 1; next: if (g(x) == y) if (f(x) == y) buffer[y] = 1 False True True False buffer[y] = 1; Coredump: (buffer overflow) x 1 x 2 y 10 y 10 g(2) != 10

  48. Reverse Execution Synthesis x = 1; if (f(x) == y) { goto next; } ... x = 2; x x x ? if (g(x) == y) { goto next; y y y 10 } exit(); x = 2; x = 1; next: if (g(x) == y) if (f(x) == y) buffer[y] = 1 False True True False buffer[y] = 1; Coredump: (buffer overflow) x 1 x 2 y 10 y 10 g(2) != 10

  49. Reverse Execution Synthesis x = 1; if (f(x) == y) { goto next; } ... x = 2; x x x ? if (g(x) == y) { goto next; y y y 10 } exit(); x = 2; x = 1; next: if (g(x) == y) if (f(x) == y) buffer[y] = 1 False True True False buffer[y] = 1; Coredump: (buffer overflow) x 1 x 2 y 10 y 10 g(2) != 10 no match

  50. Reverse Execution Synthesis x = 1; if (f(x) == y) { goto next; } ... x = 2; x x x ? if (g(x) == y) { goto next; y y y 10 } exit(); x = 2; x = 1; next: if (g(x) == y) if (f(x) == y) buffer[y] = 1 False True True False buffer[y] = 1; Coredump: (buffer overflow) x 1 x 2 y 10 y 10 g(2) != 10

  51. Reverse Execution Synthesis x = 1; if (f(x) == y) { goto next; } ... x = 2; x x x x ? if (g(x) == y) { goto next; y y y y 10 } exit(); x = 2; x = 1; next: if (g(x) == y) if (f(x) == y) buffer[y] = 1 False True True False buffer[y] = 1; Coredump: (buffer overflow) x 1 x 2 y 10 y 10 g(2) != 10

Recommend

Where the executions are Frank R. Baumgartner Introduction Data compiled on all US executions

Where the executions are Frank R. Baumgartner Introduction Data compiled on all US executions

Where the executions are Frank R. Baumgartner Introduction Data compiled on all US executions since 1976 as of April 11 2011, 1245 executions in total. Maps by county do not include 3 executions by the federal government not assigned to

371 views • 21 slides
Automated Debugging In Data Intensive Scalable Computing Systems Muhammad Ali Gulzar 1 Siman Wang

Automated Debugging In Data Intensive Scalable Computing Systems Muhammad Ali Gulzar 1 Siman Wang

Automated Debugging In Data Intensive Scalable Computing Systems Muhammad Ali Gulzar 1 Siman Wang 1,2 Miryung Kim 1 1 University of California, Los Angeles 2 Hunan University 1 Big Data Debugging in the Dark Develop locally Hope it works Run

392 views • 22 slides
Comparative Causality: Explaining the Differences Between Executions William N. Sumner Xiangyu

Comparative Causality: Explaining the Differences Between Executions William N. Sumner Xiangyu

Comparative Causality: Explaining the Differences Between Executions William N. Sumner Xiangyu Zhang {wsumner,xyzhang} @ cs.purdue.edu ICSE 2013 22 May 2013 Background Debugging requires understanding how a program behaves. Background

1.41k views • 122 slides
Failure Sketching: A Technique for Automated Root Cause Diagnosis of In-Production Failures

Failure Sketching: A Technique for Automated Root Cause Diagnosis of In-Production Failures

Failure Sketching: A Technique for Automated Root Cause Diagnosis of In-Production Failures Baris Kasikci, Benjamin Schubert, Cristiano Pereira, Gilles Pokam, George Candea Debugging In-Production Software Failures Today 2 Debugging

937 views • 93 slides
Enhancing End-to-End Tracing Systems for Automated Performance Debugging in Distributed Systems

Enhancing End-to-End Tracing Systems for Automated Performance Debugging in Distributed Systems

Enhancing End-to-End Tracing Systems for Automated Performance Debugging in Distributed Systems Jethro S. Sun January 23, 2018 MassOpenCloud Research Group 1 Introduction A Sad Story ... 2 A Sad Story ... A distributed system is one in

841 views • 58 slides
Automated Debugging In Data Intensive Scalable Computing Systems Muhammad Ali Gulzar 1 , Matteo

Automated Debugging In Data Intensive Scalable Computing Systems Muhammad Ali Gulzar 1 , Matteo

Automated Debugging In Data Intensive Scalable Computing Systems Muhammad Ali Gulzar 1 , Matteo Interlandi 3 , Xueyuan Han 2 , Mingda Li 1 , Tyson Condie 1 , and Miryung Kim 1 1 University of California, Los Angeles 2 Harvard University 3

397 views • 25 slides
Debugging Debugging with High Level Languages Same goals as low-level debugging Examine and

Debugging Debugging with High Level Languages Same goals as low-level debugging Examine and

Chapter 15 Debugging Debugging with High Level Languages Same goals as low-level debugging Examine and set values in memory Execute portions of program Stop execution when (and where) desired Want debugging tools to operate on

274 views • 9 slides
Debugging Debugging Tools Module Overview Introduction to Debugging Problems in Production

Debugging Debugging Tools Module Overview Introduction to Debugging Problems in Production

Welcome to Advanced .NET Debugging Debugging Tools Module Overview Introduction to Debugging Problems in Production Challenges in debugging Production issues Production Environment Debugging Timeline Tools for .NET Debugging Review 3

622 views • 47 slides
Automorphism Groups of Projective Planes with Arbitrarily Many Point and Line Orbits G. Eric

Automorphism Groups of Projective Planes with Arbitrarily Many Point and Line Orbits G. Eric

Dembowski-Hughes-Parker Theorem Desarguesian examples Examples with arbitrarily many orbits Automorphism Groups of Projective Planes with Arbitrarily Many Point and Line Orbits G. Eric Moorhouse Department of Mathematics University of Wyoming

525 views • 40 slides
Capacity Region of the Gaussian Arbitrarily-Varying Broadcast Channel Fatemeh Hosseinigoki and

Capacity Region of the Gaussian Arbitrarily-Varying Broadcast Channel Fatemeh Hosseinigoki and

Capacity Region of the Gaussian Arbitrarily-Varying Broadcast Channel Fatemeh Hosseinigoki and Oliver Kosut A RIZONA S TATE U NIVERSITY ISIT 2020 1 / 12 Gaussian Arbitrarily-Varying Broadcast Channel Transmitter power constraint X 2

557 views • 44 slides
Visual Debugging Software What is Debugging Visualization Visualizing

Visual Debugging Software What is Debugging Visualization Visualizing

Visual Debugging Software What is Debugging Visualization Visualizing Program State Incremental interactive unfolding (DDD) Visual Memory Abstract representations (traversal-based) Debugging Focusing: memory

433 views • 5 slides
Debugging Techniques for C Programs Debugging Basics Will focus on the gcc/gdb combination.

Debugging Techniques for C Programs Debugging Basics Will focus on the gcc/gdb combination.

Debugging Techniques for C Programs Debugging Basics Will focus on the gcc/gdb combination. Will also talk about the ddd gui for gdb (lots of value added to gdb). First, debugging in the abstract: Program state is a snapshot

475 views • 20 slides
1 Crashes, interrupts, and backtrace Watchpoints GDB will automatically stop if the program

1 Crashes, interrupts, and backtrace Watchpoints GDB will automatically stop if the program

Debugging and debuggers Debugging in the development cycle You have probably already had the experience of making Edit a mistake in a program Speaking roughly, debugging is the process: After you know that your code is wrong

911 views • 3 slides
Coroutines Update Seva Tolstopyatov @qwwdfsad October 13, 2020 Coroutines debugging Coroutines

Coroutines Update Seva Tolstopyatov @qwwdfsad October 13, 2020 Coroutines debugging Coroutines

Kotlin 1.4 Online Event Coroutines Update Seva Tolstopyatov @qwwdfsad October 13, 2020 Coroutines debugging Coroutines debugging Coroutines debugging Coroutines debugging Coroutines debugging IDEA 2020.1 Coroutines debugging Debugging

737 views • 41 slides
Scalable Post-Mortem Debugging Abel Mathew CEO - Backtrace amathew@backtrace.io @nullisnt0

Scalable Post-Mortem Debugging Abel Mathew CEO - Backtrace amathew@backtrace.io @nullisnt0

Scalable Post-Mortem Debugging Abel Mathew CEO - Backtrace amathew@backtrace.io @nullisnt0 Debugging or Sleeping? Debugging Debugging: examining (program state, design, code, output) to identify and remove errors from software.

464 views • 31 slides
Advanced Production Debugging About Me Co-founder Takipi, JVM Production Debugging. Director,

Advanced Production Debugging About Me Co-founder Takipi, JVM Production Debugging. Director,

Advanced Production Debugging About Me Co-founder Takipi, JVM Production Debugging. Director, AutoCAD Web & Mobile. Software Architect at IAI Aerospace. Coding for the past 16 years - C++, Delphi, .NET, Java. Focus on real-time,

460 views • 35 slides
Inefficient weights from pairwise comparison matrices with arbitrarily small inconsistency

Inefficient weights from pairwise comparison matrices with arbitrarily small inconsistency

Inefficient weights from pairwise comparison matrices with arbitrarily small inconsistency Sndor Bozki Institute for Computer Science and Control Hungarian Academy of Sciences (MTA SZTAKI), Corvinus University of Budapest 15 September

152 views • 14 slides
Controlling Arbitrarily Intelligent Systems Tom Everitt tomeveritt.se Australian National

Controlling Arbitrarily Intelligent Systems Tom Everitt tomeveritt.se Australian National

Controlling Arbitrarily Intelligent Systems Tom Everitt tomeveritt.se Australian National University Supervisors: Marcus Hutter, Laurent Orseau, Stephen Gould July 19, 2016 Selfmodification of Policy and Utility Function in Rational Agents.

593 views • 21 slides
Nonconvex Distributed Optimization: Novel Algorithmic Design and Arbitrarily Precise Solutions

Nonconvex Distributed Optimization: Novel Algorithmic Design and Arbitrarily Precise Solutions

Nonconvex Distributed Optimization: Novel Algorithmic Design and Arbitrarily Precise Solutions Presenter: Zhiyu He Coauthors: Jianping He*, Cailian Chen and Xinping Guan Department of Automation Shanghai Jiao Tong University July 2020

716 views • 39 slides
Embedded Software TI2726-B 8. Debugging techniques Koen Langendoen Embedded Software Group

Embedded Software TI2726-B 8. Debugging techniques Koen Langendoen Embedded Software Group

Embedded Software TI2726-B 8. Debugging techniques Koen Langendoen Embedded Software Group Grace Hopper 1947 2 Overview Debugging techniques Debugging a distributed system Preparation for the exam Debugging Write good code!

733 views • 45 slides
Kernel Debugging and Virtualization John Baldwin January 15, 2015 What is Kernel Debugging

Kernel Debugging and Virtualization John Baldwin January 15, 2015 What is Kernel Debugging

Kernel Debugging and Virtualization John Baldwin January 15, 2015 What is Kernel Debugging Step in Kernel Development Post-Mortem Crash Analysis Live Inspection Performance Analysis Userland Debugging Wait, Userland

216 views • 10 slides
Multithreading in Qt Doing it wrong, debugging it, doing it right David Faure

Multithreading in Qt Doing it wrong, debugging it, doing it right David Faure

Multithreading in Qt Doing it wrong, debugging it, doing it right David Faure <david.faure@kdab.com> Outline QThread Debugging race conditions Debugging deadlocks Unit-testing for thread-safety How to use QThread A busy

167 views • 16 slides
Automated Reasoning: Some Successes and New Challenges Predrag Jani ci c

Automated Reasoning: Some Successes and New Challenges Predrag Jani ci c

What is this talk about? What is automated reasoning? Automated reasoning in propositional logic Automated reasoning in first-order logic Automated reasoning in higher-order logic Automated reasoning in geometry Conclusions Automated

806 views • 52 slides
Accelerating Real Applications Best Practices for Profiling and Debugging Complex Code Beau

Accelerating Real Applications Best Practices for Profiling and Debugging Complex Code Beau

Accelerating Real Applications Best Practices for Profiling and Debugging Complex Code Beau Paisley Senior Solutions Architect US Allinea: The industry standard tools for HPC (and hundreds more) We have enjoyed a long and productive

354 views • 32 slides

More recommend