Deterministic replay of multithreaded applications

• US 6,101,524 A
• Filed: 10/23/1997
• Issued: 08/08/2000
• Est. Priority Date: 10/23/1997
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Choi, Jong-Deok, Srinivasan, Harini
• Primary Examiner(s): Teska, Kevin J.
• Assistant Examiner(s): BOOKER, KELVIN
• Application Number:
  US08/956,717
• Time in Patent Office: 1,020 Days
• Field of Search: 395/704, 395/650, 709/102
• US Class Current: 718/102
• CPC Class Codes: G06F 9/4881 Scheduling strategies for dispatcher, e.g. round robin, multi-level priority queues
• CPC Class Codes: G06F 9/52 Program synchronisation; Mutual exclusion, e.g. by means of semaphores

Abstract

A multithreaded program includes sequences of events wherein each sequence is associated with one of a plurality of execution threads. In a record mode, the software tool of the present invention records a run-time representation of the program by distinguishing critical events from non-critical events of the program and identifying the execution order of such critical events. Groups of critical events are generated wherein, for each group G_i, critical events belonging to the group G_i belong to a common execution thread, critical events belonging to the group G_i are consecutive, and only non-critical events occur between any two consecutive critical events in the group G_i. In addition, the groups are ordered and no two adjacent groups include critical events that belong to a common execution thread. For each execution thread, a logical thread schedule is generated that identifies a sequence of said groups associated with the execution thread. The logical thread schedules are stored in persistent storage for subsequent reuse. In a replay mode, for each execution thread, the logical thread schedule associated with the execution thread is loaded from persistent storage and the critical events identified by the logical thread schedule are executed.

Claims

1. A program storage device, readable by a machine, tangibly embodying instructions to perform method steps for recording a representation of run-time behavior of a program, wherein said run-time behavior of said program includes sequences of events, each sequence associated with one of a plurality of execution threads, said method steps comprising:

identifying execution order of critical events of said program, wherein said program comprises critical events and non-critical events;

generating groups of critical events of said program,

wherein, for each given group, critical events belonging to said given group belong to a common execution thread, critical events belonging to said given group are consecutive, and only non-critical events occur between any two consecutive critical events in said given group; and

wherein said groups are ordered and no two adjacent groups include critical events that belong to a common execution thread;

generating, for each given execution thread, a logical thread schedule that identifies a sequence of said groups associated with said given execution thread, and

storing said logical thread schedule in persistent storage for subsequent reuse.

2. The program storage device of claim 1, wherein each of said critical events represent one of a shared-variable access operation and a synchronization operation.

3. The program storage device of claim 1, wherein critical events belonging to a given group are represented by an ordered pair <

FirstCriticalEvent, LastCriticalEvent>

, wherein FirstCriticalEvent identifies the first critical event in the given group and LastCriticalEvent identifies the last critical event in the given group.

4. The program storage device of claim 1, wherein each critical event is identified by a global clock value that reflects execution order of said critical events.

5. The program storage device of claim 4, wherein critical events belonging to a group are ordered based upon global clock values associated with said critical events, and wherein said groups of critical events are ordered based upon global clock values associated with said critical events.

6. The program storage device of claim 4, wherein said global clock value that identifies a particular critical event is updated within an atomic action.

7. The program storage device of claim 6, wherein, if said particular critical event represents a monitorenter synchronization operation, said particular critical event is executed prior to performing said atomic action that updates said global clock value.

8. The program storage device of claim 6, wherein, if said particular critical event represents a wait synchronization operation, said particular critical event is executed subsequent to performing said atomic action that updates said global clock value.

9. The program storage device of claim 6, wherein, if said particular critical event represents one of a shared-variable access operation, a notify synchronization operation, a notifyAll synchronization operation, a resume synchronization operation, an interrupt synchronization operation, a suspend(self) synchronization operation, and a suspend(other) synchronization operation, said global clock value is updated and said particular critical event is executed within an atomic action.

10. The program storage device of claim 9, further comprising the steps of:

associating a flag with a thread executing a suspend(self) synchronization operation, wherein said flag when set indicates that said thread executing the suspend(self) synchronization operation has the intent to suspend operation;

if said particular critical event represents a suspend(self) synchronization operation, setting said flag associated with said thread;

for threads other the thread executing the suspend(self) synchronization operation that cannot perform atomic actions,

evaluating said flag to determine if said flag is set;

upon determining that said flag is set, determining if the thread executing the suspend(self) synchronization operation is in a self-suspended state; and

if the thread executing the suspend(self) synchronization operation is in a self-suspended state, performing atomic actions associated with the suspend(self) synchronization operation.

11. A program storage device, readable by a machine, tangibly embodying instructions to perform method steps for replaying run-time behavior of a program, wherein said run-time behavior of said program includes sequences of events, each sequence associated with one of a plurality of execution threads, said method steps comprising:

for each given execution thread, loading from persistent storage a logical thread schedule associated with said given execution thread, wherein said logical thread schedule comprises a sequence of groups of critical events of said program,

wherein said program comprises critical events and non-critical events,

wherein, for each given group, critical events belonging to said given group belong to a common execution thread, critical events belonging to said given group are consecutive, and only non-critical events occur between any two consecutive critical events in said given group, and

wherein said groups are ordered and no two adjacent groups include critical events that belong to a common execution thread;

for each given execution thread, executing critical events identified by said logical thread schedule associated with said given execution thread.

12. The program storage device of claim 11, wherein each of said critical events represent one of a shared-variable access operation and a synchronization operation.

13. The program storage device of claim 11, wherein critical events belonging to a given group are represented by an ordered pair <

FirstCriticalEvent, LastCriticalEvent>

, wherein FirstCriticalEvent identifies the first critical event in the given group and LastCriticalEvent identifies the last critical event in the given group.

14. The program storage device of claim 11, wherein each critical event is identified by a global clock value that reflects execution order of said critical events.

15. The program storage device of claim 14, wherein critical events belonging to a group are ordered based upon global clock values associated with said critical events, and wherein said groups of critical events are ordered based upon global clock values associated with said critical events.

16. The program storage device of claim 14, wherein said global clock value that identifies a particular critical event is updated within an atomic action.

17. The program storage device of claim 16, wherein, if said particular critical event represents a monitorenter synchronization operation, said global clock value is updated and said particular critical event is executed within an atomic action.

18. The program storage device of claim 16, wherein, if said particular critical event represents a wait synchronization operation, said particular critical event is executed subsequent to performing said atomic action that updates said global clock value.

19. The program storage device of claim 16, wherein, if said particular critical event represents one of a shared-variable access operation, a notify synchronization operation, a notifyall synchronization operation, a resume synchronization operation, an interrupt synchronization operation, a suspend(self) synchronization operation, and a suspend(other) synchronization operation, said global clock value is updated and said particular critical event is executed within an atomic action.

20. The program storage device of claim 19, further comprising the steps of:

associating a flag with a thread executing a suspend(self) synchronization operation, wherein said flag when set indicates that said thread executing the suspend(self) synchronization operation has the intent to suspend operation;

if said particular critical event represents a suspend(self) synchronization operation, setting said flag associated with said thread;

for threads other the thread executing the suspend(self) synchronization operation that cannot perform atomic actions,

evaluating said flag to determine if said flag is set;

upon determining that said flag is set, determining if the thread executing the suspend(self) synchronization operation is in a self-suspended state; and

if the thread executing the suspend(self) synchronization operation is in a self-suspended state, performing atomic actions associated with the suspend(self) synchronization operation.

21. A program storage device, readable by a machine, tangibly embodying instructions to perform method steps for recording a representation of run-time behavior of a program, wherein said run-time behavior of said program includes sequences of events, each sequence associated with one of a plurality of execution threads, said method steps comprising:

identifying a suspend(self) operation within said sequences of events;

in an atomic action, assigning an identifier to said suspend(self) operation and executing said suspend(self) operation, wherein said identifier identifies execution order of said suspend(self) operation within said sequences of events;

controlling a second thread to complete said atomic action; and

storing said identifier in persistent storage for subsequent reuse.

22. The program storage device of claim 21, wherein said identifier is a global clock value that reflects execution order of said critical events.

23. A program storage devices, readable by a machine, tangibly embodying instructions to perform method steps for replaying run-time behavior of a program, wherein said run-time behavior of said program includes sequences of events, each sequence associated with one of a plurality of execution threads, said method steps comprising:

for each given execution thread, loading from persistent storage a thread schedule associated with said given execution thread, wherein said thread schedule identifies sequences of events of said program;

identifying a suspend(self) operation within said sequences of events; and

in an atomic action, executing said suspend(self) operations; and

controlling a second thread to complete said atomic action in accordance with the thread schedule.

24. The program storage device of claim 23, wherein said suspend(self) operation is identified by a global clock value that reflects execution order of said critical events.