Hypervisor Page Fault Processing in a Shared Memory Partition Data Processing System

• US 20090307436A1
• Filed: 03/13/2009
• Published: 12/10/2009
• Est. Priority Date: 06/06/2008
• Status: Active Grant

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Prosser, Edward C., Larson, David A., Vossen, Kenneth C.
• Granted Patent: US 8,166,254 B2
• Time in Patent Office: Days
• US Class Current: 711/148
• CPC Class Codes: G06F 11/0712 in a virtual computing platform, e.g. logically partitioned systems
• CPC Class Codes: G06F 11/0724 in a multiprocessor or a multi-core unit multiprocessors per se G06F15/80
• 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 11/0745 in an input/output transactions management context input/output processing in general G06F13/00
• CPC Class Codes: G06F 11/0793 Remedial or corrective actions recovery from an exception in an instruction pipeline G06F9/3861; by retry G06F11/1402; for recovering from a failure of a protocol instance or entity H04L69/40

Abstract

Hypervisor page fault processing logic is provided for a shared memory partition data processing system. The logic, responsive to an executing virtual processor of the shared memory partition data processing system encountering a hypervisor page fault, allocates an input/output (I/O) paging request to the virtual processor from an I/O paging request pool and increments an outstanding I/O paging request count for the virtual processor. A determination is then made whether the outstanding I/O paging request count for the virtual processor is at a predefined threshold, and if not, the logic places the virtual processor in a wait state with interrupt wake-up reasons enabled based on the virtual processor'"'"'s state, otherwise, it places the virtual processor in a wait state with interrupt wake-up reasons disabled.

Claims

1. A method of addressing a hypervisor page fault in a shared memory partition data processing system, the method comprising:

responsive to an executing virtual processor encountering a hypervisor page fault, allocating an input/output (I/O) paging request to the virtual processor from an I/O paging request pool and incrementing an outstanding I/O paging request count for the virtual processor; and

determining whether the outstanding I/O paging request count for the virtual processor is at a predefined threshold, and if not, placing the virtual processor in a wait state with interrupt wake-up reasons enabled based on the virtual processor'"'"'s state, and if the outstanding I/O paging request count for the virtual processor is at the predefined threshold, then placing the virtual processor in a wait state with interrupt wake-up reasons disabled.

2. The method of claim 1, further comprising pre-assigning N I/O paging requests per virtual processor of a plurality of virtual processors of the shared memory partition data processing system, wherein N is the predefined threshold.

3. The method of claim 2, further comprising, responsive to completing the I/O paging request, freeing the I/O paging request to the I/O paging request pool, decrementing the outstanding I/O paging request count for the virtual processor, and determining whether the virtual processor is still blocked waiting on the completed I/O paging request, and if so, dispatching the virtual processor.

4. The method of claim 1, further comprising, prior to the allocating, determining whether there is a free I/O paging request in the I/O paging request pool, the I/O paging request pool being maintained by a hypervisor of the shared memory partition data processing system, and if so, allocating the free I/O paging request from the I/O paging request pool to the executing virtual processor encountering the hypervisor page fault, otherwise, placing the virtual processor on a list of virtual processors waiting for a free I/O paging request, and placing the virtual processor in a wait state with interrupt wake-up reasons enabled based on the virtual processor'"'"'s state.

5. The method of claim 4, further comprising, responsive to freeing of an I/O paging request to the I/O paging request pool, determining whether the virtual processor is still blocked waiting on the freed I/O paging request, and if so, dispatching the virtual processor, otherwise allowing the virtual processor to continue execution responsive to a prior interrupt while the freed I/O paging request was outstanding.

6. The method of claim 4, further comprising, responsive to freeing of the I/O paging request to the I/O paging request pool, determining whether there is a virtual processor on the list of virtual processors waiting for an I/O paging request, and if so, removing the virtual processor from the list of virtual processors waiting for an I/O paging request and initiating an I/O paging request for that virtual processor'"'"'s pending hypervisor page fault.

7. The method of claim 1, further comprising pre-assigning N I/O paging requests to the virtual processor of the shared memory partition data processing system, and wherein the predefined threshold is a number larger than N, wherein when executing, at any point in time, the virtual processor can employ more than N I/O paging requests, notwithstanding the pre-assigning of only N I/O paging requests to the virtual processor.

8. A shared memory partition data processing system comprising:

a physical memory comprising a shared memory pool for at least one shared memory partition of the data processing system, the at least one shared memory partition comprising a plurality of virtual processors;

a processor supporting the plurality of virtual processors, the plurality of virtual processors sharing at least a portion of the shared memory pool of the physical memory of the data processing system; and

wherein the at least one shared memory partition interfaces with the shared memory pool through a hypervisor memory manager, the hypervisor memory manager facilitating allocating, responsive to an executing virtual processor of the plurality of virtual processors encountering a hypervisor page fault, an input/output (I/O) paging request to the virtual processor from an I/O paging request pool, and determining whether an outstanding I/O paging request count for the virtual processor is at a predefined threshold, and if not, placing the virtual processor in a wait state with interrupt wake-up reasons enabled based on the virtual processor'"'"'s state, otherwise, placing the virtual processor in a wait state with interrupt wake-up reasons disabled.

9. The shared memory partition data processing system of claim 8, wherein the hypervisor memory manager facilitates pre-assigning of N I/O paging requests per virtual processor of the plurality of virtual processors of the shared memory partition data processing system, wherein N is the predefined threshold.

10. The shared memory partition data processing system of claim 9, wherein the hypervisor memory manager, responsive to completing of the I/O paging request, frees the I/O paging request to the I/O paging pool, decrements the outstanding I/O paging request count for the virtual processor, and determines whether the virtual processor is still blocked waiting on the completed I/O paging request, and if so, dispatches the virtual processor.

11. The shared memory partition data processing system of claim 8, wherein the hypervisor memory manager, prior to the allocating, determines whether there is a free I/O paging request in the I/O paging request pool, the I/O paging request pool being maintained by a hypervisor of the shared memory partition data processing system, and if so, allocates the free I/O paging request from the I/O paging request pool to the executing virtual processor encountering the hypervisor page fault, otherwise, places the virtual processor on a list of virtual processors waiting for a free I/O paging request, and places the virtual processor in a wait state with interrupt wake-up reasons enabled based on the virtual processor'"'"'s state.

12. The shared memory partition data processing system of claim 11, wherein the hypervisor memory manager, responsive to freeing of an I/O paging request to the I/O paging request pool, determines whether the virtual processor is still blocked waiting on the freed I/O paging request, and if so, dispatches the virtual processor, otherwise, allows the virtual processor to continue execution responsive to a prior interrupt received while the freed I/O paging request was outstanding.

13. The shared memory partition data processing system of claim 11, wherein the hypervisor memory manager, responsive to freeing of the I/O paging request to the I/O paging request pool, determines whether there is a virtual processor on the list of virtual processors waiting for an I/O paging request, and if so, removes the virtual processor from the list of virtual processors waiting for an I/O paging request and initiates an I/O paging request for that virtual processor'"'"'s pending hypervisor page fault.

14. The shared memory partition data processing system of claim 8, wherein N I/O paging requests are pre-assigned to each virtual processor of the shared memory partition data processing system, and wherein the predefined threshold is a number larger than N, wherein the executing virtual processor encountering the hypervisor page fault can employ more than N I/O paging requests at a time, notwithstanding the pre-assigning of only N I/O paging requests to the virtual processor.

15. An article of manufacture comprising:

at least one computer-readable medium having computer-readable program code logic to address a hypervisor page fault in a shared memory partition data processing system, the computer-readable program code logic when executing on a processor performing;

responsive to an executing virtual processor encountering a hypervisor page fault, allocating an input/output (I/O) paging request to the virtual processor from an I/O paging request pool and incrementing an outstanding I/O paging request count for the virtual processor; and

determining whether the outstanding I/O paging request count for the virtual processor is at a predefined threshold, and if not, placing the virtual processor in a wait state with interrupt wake-up reasons enabled based on the virtual processor'"'"'s state, and if the outstanding I/O paging request count for the virtual processor is at the predefined threshold, then placing the virtual processor in a wait state with interrupt wake-up reasons disabled.

16. The article of manufacture of claim 15, further comprising pre-assigning N I/O paging requests per virtual processor of a plurality of virtual processors of the shared memory partition data processing system, wherein N is the predefined threshold.

17. The article of manufacture of claim 16, further comprising, responsive to completing the I/O paging request, freeing the I/O paging request to the I/O paging request pool, decrementing the outstanding I/O paging request count for the virtual processor, and determining whether the virtual processor is still blocked waiting on the completed I/O paging request, and if so, dispatching the virtual processor.

18. The article of manufacture of claim 15, further comprising, prior to the allocating, determining whether there is a free I/O paging request in the I/O paging request pool, the I/O paging request pool being maintained by a hypervisor of the shared memory partition data processing system, and if so, allocating the free I/O paging request from the I/O paging request pool to the executing virtual processor encountering the hypervisor page fault, otherwise, placing the virtual processor on a list of virtual processors waiting for a free I/O paging request, and placing the virtual processor in a wait state with interrupt wake-up reasons enabled based on the virtual processor'"'"'s state.

19. The article of manufacture of claim 18, further comprising, responsive to freeing of an I/O paging request to the I/O paging request pool, determining whether the virtual processor is still blocked waiting on the freed I/O paging request, and if so, dispatching the virtual processor, otherwise allowing the virtual processor to continue execution responsive to a prior interrupt while the freed I/O paging request was outstanding.

20. The article of manufacture of claim 18, further comprising, responsive to freeing of the I/O paging request to the I/O paging request pool, determining whether there is a virtual processor on the list of virtual processors waiting for an I/O paging request, and if so, removing the virtual processor from the list of virtual processors waiting for an I/O paging request and initiating an I/O paging request for that virtual processor'"'"'s pending hypervisor page fault.