Method and apparatus for structured memory analysis of data processing systems and applications

• US 6,158,024 A
• Filed: 03/31/1998
• Issued: 12/05/2000
• Est. Priority Date: 03/31/1998
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Mandal, Manas
• Primary Examiner(s): Sheikh, Ayaz R.
• Assistant Examiner(s): BACKER, FIRMIN
• Application Number:
  US09/052,331
• Time in Patent Office: 980 Days
• Field of Search: 714/37, 714/41, 714/47, 714/25, 714/38, 714/45, 395/704, 395/708, 702/182, 702/183, 702/186
• US Class Current: 714/37
• CPC Class Codes: G06F 11/073 in a memory management context, e.g. virtual memory or cache management memory management G06F12/00; testing of static memory units G11C29/00
• CPC Class Codes: G06F 11/0751 Error or fault detection not based on redundancy power supply failures G06F1/30; network fault management H04L41/06
• CPC Class Codes: G06F 11/079 Root cause analysis, i.e. error or fault diagnosis in a hardware test environment G06F11/22; in a software test environment G06F11/36
• CPC Class Codes: G06F 11/3466 Performance evaluation by tracing or monitoring
• CPC Class Codes: G06F 2201/86 Event-based monitoring

Abstract

A method for monitoring memory performance of a program. A frequently occurring event is detected and a stack associated with the program is identified in response to a detection of the periodically occurring event, which may be a page fault. The stack is examined to identify each routine (and specific invocation point, or offset) that is currently executing in association with the program. Each routine (and offset) is represented as a node in a tree structure. Classification of the page fault is performed. Page faults classified as data faults can be further sub-classified to provide additional information.

Claims

1. A method in a data processing system for analyzing memory use in the data processing system, the method comprising:

initiating a page fault in response to an occurrence of a selected type of memory access;

monitoring for the page fault in the data processing system;

identifying a call stack associated with the page fault in response to detecting the page fault; and

examining the call stack to identify each routine executing when the page fault was detected; and

representing each routine as a node in a tree structure.

2. The method of claim 1 further comprising:

identifying a page fault type in response to detecting the page fault.

3. The method of claim 2, wherein the page fault includes an address used to identify the page fault type and further comprising:

identifying a type of data in response to an identification of the page fault being a data fault.

4. The method of claim 1 further comprising:

associating information relating to each routine with a node used to represent a routine.

6. The method of claim 3, wherein the data fault is for a call stack.

7. The method of claim 3, wherein the data fault is for a heap.

8. The method of claim 3, wherein the data fault is for a module.

9. The method of claim 3 further comprising:

identifying a variable associated with the data address.

5. A method in a data processing system for analyzing memory use in the data processing system, the method comprising:

initiating a page fault each time a selected type of memory access occurs;

monitoring for a page fault in the data processing system;

identifying a call stack associated with a page fault in response to detecting the page fault; and

examining the call stack to identify each routine executing when the page fault was detected;

reading the call stack, wherein the call stack includes a list of routines being executed when the page fault occurs;

representing each routine as a node in the tree structure;

determining whether a selected routine is found within the tree structure; and

adding a new node to the tree structure in response to a determination that the selected routine is absent from the tree structure, wherein the new node is a child node descending from a parent node.

10. A method in a data processing system for analyzing memory use in the data processing system, the method comprising:

detecting an occurrence of a selected type of memory access in the data processing system;

identifying a call stack associated with the memory access;

examining the call stack to identify each routine executing when the memory access occurred; and

representing each routine as a node in a tree structure.

11. The method of claim 10, wherein the occurrence of the memory access is detected by detecting a page fault.

12. The method of claim 11, wherein a page fault interrupt is initiated each time a memory access occurs.

13. The method of claim 11 further comprising:

determining whether the page fault is a data fault.

14. The method of claim 13 further comprising:

identifying a type of page fault.

15. A method in a data processing system for analyzing memory use in the data processing system, the method comprising:

detecting an occurrence of a selected type of memory access in the data processing system wherein the occurrence of the memory access is detected by detecting a page fault;

determining whether the page fault is a data fault;

identifying a type of page fault;

obtaining a page fault address and a page fault instruction address for the page fault;

comparing the page fault address to the page fault instruction address;

identifying the type of page fault as an instruction fault in response to a determination that the page fault address is equal to the page fault instruction address;

identifying the type of page fault as an instruction fault in response to a determination that the page fault address not equal to the page fault instruction address;

identifying a call stack associated with the memory access;

examining the call stack to identify each routine executing when the memory access occurred; and

representing each routine as a node in a tree structure.

16. A data processing system for analyzing memory use in the data processing system comprising;

initiation means for selectively initiating a page fault in response to an occurrence of a memory access;

monitoring means for monitoring the page fault in the data processing system;

identification means for identifying a call stack associated with the page fault in response to detecting the page fault;

examination means for examining the call stack to identify each routine executing when the page fault was detected; and

representation means for representing each routine as a node in a tree structure.

17. The data processing system of claim 16 further comprising:

second identification means for identifying a page fault type in response to detecting the page fault.

18. The data processing system of claim 17, wherein the page fault includes an address used to identify the page fault type and further comprising:

identification means for identifying a type of data in response to an identification of the page fault being a data address.

19. The data processing system of claim 16 further comprising:

association means for associating information relating to each routine with a node used to represent a routine.

20. A data processing system for analyzing memory use in the data processing system comprising:

initiation means for selectively initiating a page fault each time a memory access occurs;

monitoring means for monitoring a page fault in the data processing system;

identification means for identifying a call stack associated with the page fault in response to detecting the page fault;

examination means for examining the call stack to identify each routine executing when the page fault was detected;

reading means for reading the call stack, wherein the call stack includes a list of routines being executed when the page fault occurs;

representation means for representing each routine as a node in a tree structure;

determination means for determining whether a selected routine is found within the tree structure; and

adding means for adding a new node to the tree structure in response to a determination that the selected routine is absent from the tree structure, wherein the new node is a child node descending from a parent node.

21. A data processing system for analyzing memory used in the data processing system, the data processing system comprising:

detection means for detecting an occurrence of a selected type of memory access in the data processing system;

identification means for identifying a call stack associated with the memory access;

examination means for examining the call stack to identify each routine executing when the memory access occurred; and

representation means for representing each routine as a node in a tree structure.

22. The data processing system of claim 21, wherein the occurrence of the memory access is detected by detecting a page fault interrupt.

23. The data processing system of claim 22, wherein a page fault interrupt is initiated each time a memory access occurs.

24. The data processing system of claim 22 further comprising:

second determination means for determining whether the page fault is a data fault.

25. The data processing system of claim 24 further comprising:

identification means for identifying a type of page fault.

26. The data processing system of claim 25, wherein the identification means includes:

first means for obtaining a page fault address and a page fault instruction address for the page fault;

second means for comparing the page fault address to the page fault instruction address;

third means for identifying the type of page fault as an instruction fault in response to a determination that the page fault address is equal to the page fault instruction address; and

fourth means for identifying the type of page fault as an instruction fault in response to a determination that the page fault address is not equal to the page fault instruction address.

27. A computer program product for use with a data processing system for structured memory analysis, the computer program product comprising:

first instructions for initiating a page fault in response to an occurrence of a selected type of memory access;

second instructions for monitoring for the page fault in the data processing system;

third instructions for identifying a call stack associated with the page fault in response to detecting the page fault; and

fourth instructions for examining the call stack to identify each routine executing when the page fault was detected; and

fifth instructions for representing each routine as a node in a tree structure.

28. The computer program product of claim 27 further comprising:

sixth instructions for identifying a page fault type in response to detecting the page fault.

29. The computer program product of claim 28, wherein the page fault includes an address used to identify the page fault type and further comprising:

seventh instructions for identifying a type of data in response to an identification of the page fault being a data fault.

30. The computer program product of claim 27 further comprising:

fifth instructions for associating information relating to each routine with a node used to represent a routine.

31. A computer program product for use with a data processing system for structured memory analysis, the computer program product comprising:

first instructions for initiating a page fault each time a selected type of memory access occurs;

second instructions for monitoring for a page fault in the data processing system;

third instructions for identifying a call stack associated with the page fault in response to detecting the page fault;

fourth instructions for examining the call stack to identify each routine executing when the page fault was detected;

fifth instructions for reading the call stack, wherein the call stack includes a list of routines being executed when the page fault occurs;

sixth instructions for representing each routine as a node in a tree structure;

seventh instructions for determining whether a selected routine is found within the tree structure; and

eighth instructions for adding a new node to the tree structure in response to a determination that the selected routine is absent from the tree structure, wherein the new node is a child node descending from a parent node.