Method and apparatus for recovery of partitions in a logical partitioned data processing system

1. A method in a logical partitioned data processing system for recovering terminated partitions, the method comprising:

detecting a termination of a partition in a set of partitions associated with a host bridge in the logical partitioned data processing system;

responsive to detecting the termination, checking a state of other partitions within the set of partitions;

initiating a recovery process if all partitions in the set of partitions have terminated;

resetting input/output slots associated with the host bridge to a normal state if the recovery process is successful; and

rebooting the set of partition after resetting the input/output slots associated with the host bridge without rebooting the logical partitioned data processing system.

2. The method of claim 1 , wherein the termination of the partition is caused by an error in the host bridge.

3. The method of claim 2 , wherein the error comprises an input/output slot in the input/output slots being marked as bad.

4. The method of claim 1 , wherein the recovery process includes:

masking off interrupts;

disabling direct memory access;

resetting the host bridge;

clearing error registers associated with the host bridge; and

enabling direct memory access after clearing the error registers.

5. The method of claim 1 , wherein the recovery process includes reprogramming the host bridge.

6. The method of claim 5 , wherein the host bridge is reprogrammed using previously saved hardware state information.

7. The method of claim 1 , wherein the method is implemented in a partition manager.

8. The method of claim 1 , wherein the recovery process is for the host bridge.

9. A method in a logical partitioned data processing system for handling an error state for a set of partitions, the method comprising:

detecting an error in a shared device assigned to the set of partitions in the logical partitioned data processing system;

checking a state of other partitions within the set of partitions;

initiating a recovery process for the shared device if all the partitions in the set of partitions have terminated;

resetting the shared device if the recovery process is successful; and

rebooting the set of partitions after the share device has been reset.

10. The method of claim 9 , wherein the shared device is a host bridge.

11. A method in a logical partitioned data processing system for handling an error state for a set of partitions caused by an input/output slot being marked bad, the method comprising:

responsive to the set of partitions associated with a host bridge going into the error state in the logical partitioned data processing system, initiating a recovery process to reset the host bridge;

determining whether the recovery process is successful;

restoring hardware states in the host bridge if the recovery process is successful;

changing the input/output slot to a normal state; and

booting the set of partitions.

12. A logical partitioned data processing system for recovering terminated partitions, the logical partitioned data processing system comprising:

detecting means for detecting a termination of a partition in a set of partitions associated with a host bridge in the logical partitioned data processing system;

checking means, responsive to detecting the termination, for checking a state of other partitions within the set of partitions;

initiating means for initiating a recovery process if all partitions in the set of partitions have terminated;

resetting means for resetting input/output slots associated with the host bridge to a normal state if the recovery process is successful; and

rebooting means for rebooting the set of partitions after resetting the input/output slots associated with the host bridge without rebooting the logical partitioned data processing system.

13. The logical partitioned data processing system of claim 12 , wherein the termination of the partition is caused by an error in the host bridge.

14. The logical partitioned data processing system of claim 13 , wherein to error comprises an input/output slot in the input/output slots being marked as bad.

15. The logical partitioned data processing system of claim 12 , wherein the recovery process includes:

masking means for masking off interrupts;

disabling means for disabling direct memory access;

resetting means for resetting the host bridge;

clearing means for clearing error registers associated with the host bridge; and

enabling means for enabling direct memory access alter clearing the error registers.

16. The logical partitioned data processing system of claim 12 , wherein the recovery process includes reprogramming means for reprogramming the host bridge.

17. The logical partitioned data processing system of claim 16 , wherein the host bridge is reprogrammed using previously saved hardware state information.

18. The logical partitioned data processing system of claim 12 , wherein the detecting means, checking means, initiating means, resetting means, and rebooting means are implemented in a partition manager.

19. The logical partitioned data processing system of claim 12 , wherein the recovery process is for the host bridge.

20. A logical partitioned data processing system in a logical partitioned data processing system for handling an error state for a set of partitions, the logical partitioned data processing system comprising:

detecting means for detecting an error in a shared device assigned to the set of partitions in the logical partitioned data processing system;

checking means for checking a state of other partitions within the set of partitions;

initiating means for initiating a recovery process for the shared device if all the partitions in the set of partitions have terminated;

resetting means for resetting the shared device if the recovery process is successful; and

rebooting means for rebooting the set of partitions after the share device has been reset.

21. The logical partitioned data processing system of claim 20 , wherein the shared device is a host bridge.

22. A logical partitioned data processing system in a logical partitioned data processing system for handling an error state for a set of partitions caused by an input/output slot being marked bad, the logical partitioned data processing system comprising:

initiating means, responsive to the set of partitions associated with a host bridge going into the error state in the logical partitioned data processing system, for initiating a recovery process to reset the host bridge;

determining means for determining whether the recovery process is successful;

restoring means for restoring hardware states in the host bridge if the recovery process is successful;

changing means for changing the input/output slot to a normal state; and

booting means for booting the set of partitions.

23. A logical partitioned data processing system for recovering terminated partitions, the logical partitioned data processing system comprising:

a bus system;

a memory connected to the bus system, wherein the memory includes a set of instructions;

a processing unit having a plurality of processors and being connected to the bus system, wherein the processing unit executes the set of instructions to detect a termination of a partition in a set of partitions associated with a host bridge in the logical partitioned data processing system; check a state of other partitions within the set of partitions responsive to detecting the termination; initiate a recovery process if all partitions in the set of partitions have terminated; reset input/output slots associated with the host bridge to a normal state if the recovery process is successful; and reboot the set of partitions after resetting the input/output slots associated with the host bridge without rebooting the logical partitioned data processing system.

24. A logical partitioned data processing system for handling an error state for a set of partitions, the logical partitioned data processing system comprising:

a bus system;

a memory connected to the bus system, wherein the memory includes a set of instructions; and

a processing unit having a plurality of processors and being connected to the bus system, wherein the processing unit executes the set of instructions to detect an error in a shared device assigned to the set of partitions in the logical partitioned data processing system; check a state of other partitions within the set of partitions; initiate a recovery process for the shared device if all the partitions in the set of partitions have terminated; reset the shared device if the recovery process is successful; and reboot the set of partitions after the share device has been reset.

25. A logical partitioned data processing system for handling an error state for a set of partitions caused by an input/output slot being marked bad, the logical partitioned data processing system comprising:

a bus system;

a memory connected to the bus system, wherein the memory includes a set of instructions; and

a processing unit having a plurality of processors and being connected to the bus system, wherein the processing unit executes the set of instructions to initiate a recovery process to reset the host bridge, responsive to the set of partitions associated with a host bridge going into the error state in the logical partitioned data processing system; determine whether the recovery process is successful; restore hardware states in the host bridge if the recovery process is successful; change the input/output slot to a normal state; and boot the set of partitions.

26. A computer program product in a computer recordable medium for recovering terminated partitions, in a logical partition in a data processing system the computer program product comprising:

first instructions for detecting a termination of a partition in a set of partitions associated with a host bridge in the logical partitioned data processing system;

second instructions, responsive to detecting to termination, for checking a state of other partitions within the set of partitions;

third instructions for initiating a recovery process if all partitions in the set of partitions have terminated;

fourth instructions for resetting input/output slots associated with the host bridge to a normal state if the recovery process is successful; and

fifth instructions for rebooting the set of partitions after resetting the input/output slots associated with the host bridge without rebooting the logical partitioned data processing system.

27. A computer program product in a computer recordable medium for handling an error state for a set of partitions in a logical partitioned data processing system, the computer program product comprising:

first instructions for detecting an error in a shared device assigned to the set of partitions in the logical partitioned data processing system;

second instructions for checking a state of other partitions within the set of partitions;

third instructions for initiating a recovery process for the shared device if all the partitions in the set of partitions have terminated;

fourth instructions for resetting the shared device if the recovery process is successful; and

fifth instructions for rebooting the set of partitions after the share device has been reset.

28. A computer program product in a logical partitioned data processing system for handling an error state for a set of partitions caused by an input/output slot being marked bad, the computer program product comprising:

first instructions, responsive to the set of partitions associated with a host bridge going into the error state in the logical partitioned data processing system, for initiating a recovery process to reset the host bridge;

second instructions for determining whether the recovery process is successful;

third instructions for restoring hardware states in the host bridge if the recovery process is successful;

fourth instructions for changing the input/output slot to a normal state; and

fifth instructions for booting the set of partitions.