Method and apparatus for handling transient memory errors

1. A computer-implemented method of managing memory of a data processing system, comprising:

allocating memory objects in response to memory allocation requests, each object having an associated a plurality of addresses;

storing type-identifier codes in association with memory objects, respectively, wherein each type-identifier code is a program counter value from which allocation of memory is requested; and

responsive to a transient memory error at a memory address, identifying the memory object associated with the memory address, obtaining the type-identifier code associated with the memory object, selecting one of a plurality of recovery actions using the type-identifier code as selection criteria, and performing the one of the recovery actions.

2. The method of claim 1 , further comprising storing the type-identifier codes within the memory objects, respectively.

3. The method of claim 2 , wherein an operating system manages resources of the data processing system for use by application programs executing on the data processing system, and a first type-identifier code identifies memory objects used by the operating system and a second type-identifier code identifies memory objects used by the application programs.

4. The method of claim 2 , wherein the one of the recovery actions comprises disregarding the error.

5. The method of claim 2 , wherein the one of the recovery actions comprises signaling an application program if the address of the memory error is associated with a memory object allocated to the application program.

6. The method of claim 2 , wherein the one of the recovery actions comprises halting the operating system.

7. The method of claim 2 , wherein the one of the recovery actions comprises logging information that describes the memory error.

8. The method of claim 2 , wherein an operating system manages resources of the data processing system for use by application programs executing on the data processing system, and a first type-identifier code identifies memory objects of a first type used by the operating system, a second type-identifier code identifies memory objects of a second type used by the operating system, a third type-identifier code identifies memory objects used by the application programs, and for errors in memory objects associated with the second type-identifier code, the one of the recovery actions logs information that describes the memory error.

9. The method of claim 1 , wherein an operating system manages resources of the data processing system for use by application programs executing on the data processing system, and a first type-identifier code identifies memory objects used by the operating system and a second type-identifier code identifies memory objects used by the application programs.

10. The method of claim 1 , wherein the one of the recovery actions comprises disregarding the error.

11. The method of claim 1 , wherein the one of the recovery actions comprises signaling an application program if the address of the memory error is associated with a memory object allocated to the application program.

12. The method of claim 1 , wherein the one of the recovery actions comprises halting the operating system.

13. The method of claim 1 , wherein the one of the recovery actions comprises logging information that describes the memory error.

14. The method of claim 1 , wherein an operating system manages resources of the data processing system for use by application programs executing on the data processing system, and a first type-identifier code identifies memory objects of a first type used by the operating system, a second type-identifier code identifies memory objects of a second type used by the operating system, a third type-identifier code identifies memory objects used by the application programs, and for errors in memory objects associated with the second type-identifier code, the one of the recovery actions logs information that describes the memory error.

15. An apparatus for managing memory of a data processing system, comprising:

means for allocating memory objects in response to memory allocation requests, each object having an associated a plurality of addresses;

means for storing type-identifier codes in association with memory objects, respectively, wherein each type-identifier code is a program counter value from which allocation of memory is requested; and

means, responsive to a transient memory error at a memory address, for identifying the memory object associated with the memory address, obtaining the type-identifier code associated with the memory object, selecting one of a plurality of recovery actions using the type-identifier code as selection criteria, and performing the one of the recovery actions.

16. A processor-based method of managing memory of a data processing system, comprising:

allocating memory objects by an operating system in response to random access memory (RAM) allocation requests, each object having an associated a plurality of RAM addresses, the memory allocation requests received from operating system processes and non-operating system processes;

storing type-identifier codes in association with allocated memory objects, respectively, wherein a first type-identifier code identifies memory objects of a first type used by an operating system process, a second type-identifier code identifies memory objects of a second type used by an operating system process, and a third type-identifier code identifies memory objects used by a non-operating system process; and

responsive to an error detected at a RAM memory address and not corrected, identifying the memory object associated with the RAM memory address, obtaining the type-identifier code associated with the memory object, selecting one of a plurality of recovery actions using the type-identifier code as selection criteria, halting the operating system process in response to an error detected for a memory object having an associated first type-identifier code, and logging information that describes the error in response to an error detected for a memory object having a second type-identifier code.

17. The method of claim 16 further comprising storing the type-identifier codes within the memory objects, respectively.

18. The method of claim 17 , wherein the type-identifier code is a program counter value from which allocation of memory is requested.

19. The method of claim 17 , further comprising responsive to an error detected at a RAM memory address and not corrected, signaling a non-operating system process in response to an error detected for a memory object having a third type-identifier code.

20. An apparatus for managing memory of a data processing system, comprising:

means for allocating memory objects by an operating system in response to random access memory (RAM) allocation requests, each object having an associated a plurality of RAM addresses, the memory allocation requests received from operating system processes and non-operating system processes;

means for storing type-identifier codes in association with allocated memory objects, respectively, wherein a first type-identifier code identifies memory objects of a first type used by an operating system process, a second type-identifier code identifies memory objects of a second type used by an operating system process, and a third type-identifier code identifies memory objects used by a non-operating system process; and

means, responsive to an error detected at a RAM memory address and not corrected, for identifying the memory object associated with the RAM memory address, for obtaining the type-identifier code associated with the memory object, for selecting one of a plurality of recovery actions using the type-identifier code as selection criteria, for halting the operating system process in response to an error detected for a memory object having an associated first type-identifier code, and for logging information that describes the error in response to an error detected for a memory object having a second type-identifier code.

21. A program storage device, comprising:

a processor-readable medium configured with instructions for causing a processor to manage memory of a data processing system by,

allocating memory objects by an operating system in response to random access memory (RAM) allocation requests, each object having an associated a plurality of RAM addresses, the memory allocation requests received from operating system processes and non-operating system processes;

storing type-identifier codes in association with allocated memory objects, respectively, wherein a first type-identifier code identifies memory objects of a first type used by an operating system process, a second type-identifier code identifies memory objects of a second type used by an operating system process, and a third type-identifier code identifies memory objects used by a non-operating system process; and

responsive to an error detected at a RAM memory address and not corrected, identifying the memory object associated with the RAM memory address, obtaining the type-identifier code associated with the memory object, selecting one of a plurality of recovery actions using the type-identifier code as selection criteria, halting the operating system process in response to an error detected for a memory object having an associated first type-identifier code, and logging information that describes the error in response to an error detected for a memory object having a second type-identifier code.