Debugging system for multi-threaded computer programs

• US 10,445,215 B2
• Filed: 07/27/2017
• Issued: 10/15/2019
• Est. Priority Date: 07/27/2016
• Status: Active Grant

• Current Assignee: Undo Ltd.
• Original Assignee: Undo Ltd.
• Inventors: Bull, Nicholas Peter, Smith, Julian Philip, Law, Gregory Edward Warwick
• Primary Examiner(s): Kang, Insun
• Application Number:
  US15/662,178
• Publication Number:
  US 20180032423A1
• Time in Patent Office: 810 Days
• Field of Search: None
• CPC Class Codes: G06F 11/3636 by tracing the execution of the program
• CPC Class Codes: G06F 11/3664 Environments for testing or debugging software

Abstract

A method of generating program analysis data for analyzing the operation of a computer program. The method comprises, executing an instrumented process of the computer program to define a reference execution of the program, intercepting a call to a library function by the instrumented process, executing the library function in an uninstrumented process, for the uninstrumented process, capturing in a log, only data generated by or modified through the execution of the library function required by the instrumented process to continue execution of the program, and wherein the captured log is arranged to enable deterministically reproducing the effect of the library function call on the instrumented process upon re-running of the reference execution based upon the captured log to generate the program analysis data.

Claims

1. A method of executing a system call in a multi-threaded computer program for generating program analysis data, the method comprising:

running an instrumented version of machine code representing the program wherein the instrumented version of machine code initializes a plurality of threads in a first instrumented process;

generating a second instrumented process;

executing the machine code of a first thread of the first instrumented process in the second instrumented process;

intercepting a system call during the execution of the machine code of the first thread in the second instrumented process; and

executing the system call in the first instrumented process.

2. A method according to claim 1,

wherein generating the second instrumented process comprises generating a plurality of second instrumented processes, each one of the plurality of second instrumented processes being associated with a respective central processing unit (CPU) core;

wherein the first thread of the first instrumented process is associated with a CPU core;

the method further comprising;

selecting the second instrumented process of the plurality of second instrumented processes based upon a corresponding CPU core association between the selected second instrumented process and the first thread of the first instrumented process; and

wherein executing the machine code of the first instrumented process in the second instrumented process comprises executing the machine code in the selected second instrumented process.

3. A method according to claim 2, wherein a second instrumented process is associated with a CPU core by binding the second instrumented process to the CPU core.

4. A method according to claim 2, wherein the association between the first thread of the first instrumented process and a CPU core is based upon the CPU core that the first thread is currently running on or was most recently run on.

5. A method according to claim 2, further comprising:

executing a second thread of the first instrumented process in the second instrumented process whilst executing the system call in the first instrumented process;

wherein the second thread is associated with the CPU core that the second instrumented process is associated with.

6. A method according to claim 2,

wherein the number of second instrumented processes generated is based upon a CPU core count.

7. A method according to claim 1, wherein the second instrumented process is generated based upon the first thread of the plurality of threads in the first instrumented process.

8. A method according to claim 1, further comprising:

providing data resulting from the execution of the system call in the first instrumented process from the first instrumented process to the second instrumented process that was executing the machine code of the first thread.

9. A method according to claim 1, wherein executing the system call in the first instrumented process is caused by a remote procedure call.

10. A method according to claim 1, wherein intercepting the system call comprises:

marshalling input data and input buffers into the area of memory shared between the first instrumented process and the second instrumented process; and

notifying the first instrumented process to execute the system call.

11. A non-transitory computer readable medium comprising processor readable instructions configured to cause one or more processors to carry out a method comprising:

running an instrumented version of machine code representing the program wherein the instrumented version of machine code initializes a plurality of threads in a first instrumented process;

generating a second instrumented process;

executing the machine code of a first thread of the first instrumented process in the second instrumented process;

intercepting a system call during the execution of the machine code of the first thread in the second instrumented process; and

executing the system call in the first instrumented process.