Simplified DMA mappings for self-virtualizing input/output device virtual functions

• US 8,561,066 B2
• Filed: 12/08/2010
• Issued: 10/15/2013
• Est. Priority Date: 12/08/2010
• Status: Expired due to Fees

• Current Assignee: International Business Machines Corporation
• Original Assignee: International Business Machines Corporation
• Inventors: Koch, Andrew T., Lucke, Kyle A.
• Primary Examiner(s): To, Jennifer
• Assistant Examiner(s): TEETS, BRADLEY A
• Application Number:
  US12/963,300
• Publication Number:
  US 20120151473A1
• Time in Patent Office: 1,042 Days
• Field of Search: None
• US Class Current: 718/1
• CPC Class Codes: G06F 2009/45579 I/O management, e.g. providing access to device drivers or storage
• CPC Class Codes: G06F 9/45558 Hypervisor-specific management and integration aspects

Abstract

Multiple translation control entries (TCEs) at the same indices in multiple, same size TCE tables are mapped to facilitate data communication between a self-virtualizing input/output (IO) resource and a logical partition. First and second TCE tables used by an adjunct partition that interfaces a self-virtualizing IO resource with a logical partition may be identically sized, so that whenever a direct memory access (DMA) operation between the self-virtualizing IO resource and the logical partition is desired the same TCE entries in the first and second TCE tables may be used to perform a redirected DMA operation, and without the need to perform hashing or other mapping algorithms to map to the respective TCE entries in the respective TCE tables.

Claims

1. A method of accessing a self-virtualizing input/output (IO) resource with multiple logical partitions in a data processing system, the method comprising:

interfacing a logical partition of the multiple logical partitions to an associated adjunct partition through a virtual partition interface, the adjunct partition including a server virtual partition interface device driver configured to interface with a client virtual partition interface device driver in the logical partition and a resource device driver configured to interface with a virtual function or queue pair of the self-virtualizing IO resource, wherein the virtual partition interface interfaces the client virtual partition interface device driver with the server virtual partition interface device driver; and

communicating data between the logical partition and the self-virtualizing IO resource by performing a direct memory access (DMA) operation between a buffer in the logical partition and the self-virtualizing IO resource using first and second translation control entry (TCE) tables, the first TCE table associated with the client virtual partition interface device driver and the second TCE table associated with the resource device driver, wherein the first and second TCE tables each include a same number of TCE entries, and wherein a first TCE entry in the first TCE table that is mapped to the buffer in the logical partition is indexed at a same index in the first TCE table as a second TCE entry in the second TCE table that is used by the resource device driver to initiate the DMA operation between the buffer and the self-virtualizing IO resource.

2. The method of claim 1, wherein performing the DMA operation includes:

with the client virtual partition interface device driver, mapping an address for the buffer into the first TCE entry in the first TCE table and sending a request to the server virtual partition interface driver to transmit data in the buffer;

with the server virtual partition interface device driver, mapping the second TCE entry in the second TCE table at the same index as the first TCE entry in the first TCE table;

with the server virtual partition interface device driver, sending a request to the resource device driver identifying the second TCE entry; and

with the resource device driver, performing the DMA operation using the second TCE entry to copy between the buffer and the self-virtualizing IO resource.

3. The method of claim 2, wherein the second TCE entry is a redirected TCE entry, and wherein the DMA operation is a redirected DMA operation.

4. The method of claim 2, wherein the first and second TCE tables are resident in a hypervisor, and wherein mapping the address with the client virtual partition interface device driver and mapping the second TCE entry with the server virtual partition interface device driver are performed via hypervisor calls.

5. The method of claim 2, further comprising, with the resource device driver, notifying the server virtual partition interface device driver that the DMA operation is complete, and in response thereto, and with the server virtual partition interface device driver, unmapping the second TCE entry from the second TCE table.

6. The method of claim 2, further comprising interfacing the server virtual network partition interface device driver and the resource device driver through a device driver application programming interface (API), wherein sending the request to the resource device driver with the server virtual partition interface device driver is performed by a call to the device driver API.

7. The method of claim 1, wherein the adjunct partition includes a separate dispatchable state and employs a virtual address space donated from the logical partition or a hypervisor of the data processing system.

8. The method of claim 7, wherein interfacing the logical partition to the associated adjunct partition comprises interfacing the adjunct partition to an associated virtual function of the self-virtualizing IO resource.

9. The method of claim 1, wherein the virtual partition interface comprises a vendor-independent virtual network partition interface, the client virtual partition interface device driver comprises a vendor-independent client virtual network partition interface device driver, and the server virtual partition interface device driver comprises a vendor-independent server virtual network partition interface device driver.

10. The method of claim 1, wherein the self-virtualizing IO resource comprises a single root input/output virtualization (SRIOV) Ethernet adapter.

11. An apparatus, comprising:

at least one processor; and

program code configured upon execution by the at least one processor to interface a self-virtualizing input/output (IO) resource with multiple logical partitions in a data processing system, wherein the program code is further configured to;

interface a logical partition of the multiple logical partitions to an associated adjunct partition through a virtual partition interface, the adjunct partition including a server virtual partition interface device driver configured to interface with a client virtual partition interface device driver in the logical partition and a resource device driver configured to interface with a virtual function or queue pair of the self-virtualizing IO resource, wherein the virtual partition interface interfaces the client virtual partition interface device driver with the server virtual partition interface device driver; and

communicate data between the logical partition and the self-virtualizing IO resource by performing a direct memory access (DMA) operation between a buffer in the logical partition and the self-virtualizing IO resource using first and second translation control entry (TCE) tables, the first TCE table associated with the client virtual partition interface device driver and the second TCE table associated with the resource device driver, wherein the first and second TCE tables each include a same number of TCE entries, and wherein a first TCE entry in the first TCE table that is mapped to the buffer in the logical partition is indexed at a same index in the first TCE table as a second TCE entry in the second TCE table that is used by the resource device driver to initiate the DMA operation between the buffer and the self-virtualizing IO resource.

12. The apparatus of claim 11, wherein the program code is configured to perform the DMA operation by:

with the client virtual partition interface device driver, mapping an address for the buffer into the first TCE entry in the first TCE table and sending a request to the server virtual partition interface driver to transmit data in the buffer;

with the server virtual partition interface device driver, mapping the second TCE entry in the second TCE table at the same index as the first TCE entry in the first TCE table;

with the server virtual partition interface device driver, sending a request to the resource device driver identifying the second TCE entry; and

with the resource device driver, performing the DMA operation using the second TCE entry to copy between the buffer and the self-virtualizing IO resource.

13. The apparatus of claim 12, wherein the second TCE entry is a redirected TCE entry, and wherein the DMA operation is a redirected DMA operation.

14. The apparatus of claim 12, wherein the first and second TCE tables are resident in a hypervisor, and wherein the client virtual partition interface device driver is configured to map the address the server virtual partition interface device driver is configured to map the second TCE entry via hypervisor calls.

15. The apparatus of claim 12, wherein the resource device driver is further configured to notify the server virtual partition interface device driver that the DMA operation is complete, and in response thereto, the server virtual partition interface device driver, is configured to unmap the second TCE entry from the second TCE table.

16. The apparatus of claim 12, wherein the program code is further configured to interface the server virtual network partition interface device driver and the resource device driver through a device driver application programming interface (API), and wherein the server virtual partition interface device driver is configured to send the request to the resource device driver by calling the device driver API.

17. The apparatus of claim 11, wherein the adjunct partition includes a separate dispatchable state and employs a virtual address space donated from the logical partition or a hypervisor of the data processing system.

18. The apparatus of claim 17, wherein the program code is configured to interface the logical partition to the associated adjunct partition by interfacing the adjunct partition to an associated virtual function of the self-virtualizing IO resource.

19. The apparatus of claim 11, wherein the virtual partition interface comprises a vendor-independent virtual network partition interface, the client virtual partition interface device driver comprises a vendor-independent client virtual network partition interface device driver, and the server virtual partition interface device driver comprises a vendor-independent server virtual network partition interface device driver.

20. The apparatus of claim 11, wherein the self-virtualizing IO resource comprises a single root input/output virtualization (SRIOV) Ethernet adapter.

21. A program product, comprising:

a non-transitory computer readable medium; and

program code stored on the computer readable medium and configured upon execution to interface a self-virtualizing input/output (IO) resource with multiple logical partitions in a data processing system, wherein the program code is further configured to;

interface a logical partition of the multiple logical partitions to an associated adjunct partition through a virtual partition interface, the adjunct partition including a server virtual partition interface device driver configured to interface with a client virtual partition interface device driver in the logical partition and a resource device driver configured to interface with a virtual function or queue pair of the self-virtualizing IO resource, wherein the virtual partition interface interfaces the client virtual partition interface device driver with the server virtual partition interface device driver; and

communicate data between the logical partition and the self-virtualizing IO resource by performing a direct memory access (DMA) operation between a buffer in the logical partition and the self-virtualizing IO resource using first and second translation control entry (TCE) tables, the first TCE table associated with the client virtual partition interface device driver and the second TCE table associated with the resource device driver, wherein the first and second TCE tables each include a same number of TCE entries, and wherein a first TCE entry in the first TCE table that is mapped to the buffer in the logical partition is indexed at a same index in the first TCE table as a second TCE entry in the second TCE table that is used by the resource device driver to initiate the DMA operation between the buffer and the self-virtualizing IO resource.