Input/output acceleration in virtualized information handling systems
Methods and systems for I/O acceleration on a virtualized information handling system include loading a storage virtual appliance as a virtual machine on a hypervisor. The hypervisor may execute using a first processor and a second processor. The storage virtual appliance is accessed by the hypervisor using a PCI-E device driver that is mapped to a first PCI-E NTB logical endpoint at the first processor. A second PCI-E device driver may be loaded on the storage virtual appliance that accesses the hypervisor and is mapped to a second PCI-E NTB logical endpoint at the second processor. A data transfer operation may be executed between a first memory space that is mapped to the first PCI-E NTB logical endpoint and a second memory space that is mapped to the second PCI-E NTB logical endpoint. The data transfer operation may be a read or a write operation.
1. A method executed using at least two processors, including a first processor and a second processor, the method comprising:
loading a storage virtual appliance as a virtual machine on a hypervisor executing using the first processor and the second processor, wherein the storage virtual appliance is accessed by the hypervisor using a second Peripheral Component Interconnect Express (PCI-E) device driver that is mapped to a PCI-E non-transparent bridge (NTB) at a second PCI-E NTB logical endpoint at the second processor;
loading a first PCI-E device driver on the storage virtual appliance that accesses the hypervisor and is mapped to the PCI-E NTB at a first PCI-E NTB logical endpoint at the first processor;
loading a first storage driver on the storage virtual appliance that accesses a first central processing unit (CPU) direct memory access (DMA) of the first processor for a first DMA data transfer operation of a plurality of first CPU DMA data transfer operations; and
executing the first DMA data transfer operation between a first memory space that is mapped to the first PCI-E NTB logical endpoint and a second memory space that is mapped to the second PCI-E NTB logical endpoint,
wherein the hypervisor executes in the second memory space,
wherein the storage virtual appliance executes in the first memory space, and
wherein the PCI-E NTB provides address translation between the first memory space and the second memory space.
2. The method of claim 1 , wherein the first DMA data transfer operation is initiated from one of: the first memory space and the second memory space.
3. The method of claim 1 , further comprising:
loading a second storage driver on the hypervisor that accesses a second central processing unit (CPU) direct memory access (DMA) of the second processor for a second DMA data transfer operation of a plurality of second CPU DMA data transfer operations; and
executing the second DMA data transfer operation between the first memory space and the second memory space.
4. The method of claim 3 , wherein the first DMA data transfer operation occurs from the first memory space to the second memory space.
5. The method of claim 3 , wherein the first DMA data transfer operation occurs from the second memory space to the first memory space.
6. The method of claim 1 , wherein the first DMA data transfer operation includes a programmed input/output (PIO).
7. The method of claim 1 , wherein the first processor is associated with a first physical memory bank and the second processor is associated with a second physical memory bank;
wherein loading the storage virtual appliance includes specifying the first processor for the first CPU DMA data transfer operations; and
wherein the first PCI-E NTB logical endpoint is exclusively mapped to the first physical memory bank.
8. An article of manufacture comprising a non-transitory computer-readable medium storing instructions, that, when executed by at least two processors, including a first processor and a second processor, cause the two processors to:
load a storage virtual appliance as a virtual machine on a hypervisor executing using the first processor and the second processor, wherein the storage virtual appliance is accessed by the hypervisor using a second Peripheral Component Interconnect Express (PCI-E) device driver that is mapped to a PCI-E non-transparent bridge (NTB) at a second PCI-E NTB logical endpoint at the second processor;
load a first PCI-E device driver on the storage virtual appliance that accesses the hypervisor and is mapped to the PCI-E NTB at a first PCI-E NTB logical endpoint at the first processor;
load a first storage driver on the storage virtual appliance that accesses a first central processing unit (CPU) direct memory access (DMA) of the first processor for a first DMA data transfer operation of a plurality of first CPU DMA data transfer operations; and
execute the first DMA data transfer operation between a first memory space that is mapped to the first PCI-E NTB logical endpoint and a second memory space that is mapped to the second PCI-E NTB logical endpoint,
wherein the hypervisor executes in the second memory space,
wherein the storage virtual appliance executes in the first memory space, and
wherein the PCI-E NTB provides address translation between the first memory space and the second memory space.
9. The article of manufacture of claim 8 , wherein the first DMA data transfer operation is initiated from one of: the first memory space and the second memory space.
10. The article of manufacture of claim 8 , further comprising:
load a second storage driver on the hypervisor that accesses a second central processing unit (CPU) direct memory access (DMA) of the second processor for a second DMA data transfer operation of a plurality of second CPU DMA data transfer operations; and
execute the second DMA data transfer operation between the first memory space and the second memory space.
11. The article of manufacture of claim 10 , wherein the first DMA data transfer operation occurs from the first memory space to the second memory space.
12. The article of manufacture of claim 10 , wherein the first DMA data transfer operation occurs from the second memory space to the first memory space.
13. The article of manufacture of claim 8 , wherein the first DMA data transfer operation includes a programmed input/output (PIO).
14. The article of manufacture of claim 8 , wherein the first processor is associated with a first physical memory bank and the second processor is associated with a second physical memory bank;
wherein loading the storage virtual appliance includes specifying the first processor for the first CPU DMA data transfer operations; and
wherein the first PCI-E NTB logical endpoint is exclusively mapped to the first physical memory bank.
15. An information handling system comprising:
a processor subsystem having access to a memory subsystem, the processor subsystem including a first processor and a second processor, wherein the memory subsystem stores instructions executable by the processor subsystem, that, when executed by the processor subsystem, cause the processor subsystem to:
load a storage virtual appliance as a virtual machine on a hypervisor executing using the first processor and the second processor, wherein the storage virtual appliance is accessed by the hypervisor using a second Peripheral Component Interconnect Express (PCI-E) device driver that is mapped to a PCI-E non-transparent bridge (NTB) at a second PCI-E NTB logical endpoint at the first second processor;
load a first PCI-E device driver on the storage virtual appliance that accesses the hypervisor and is mapped to the PCI-E NTB at a first PCI-E NTB logical endpoint at the first processor;
load a first storage driver on the storage virtual appliance that accesses a first central processing unit (CPU) direct memory access (DMA) of the first processor for a first DMA data transfer operation of a plurality of first CPU DMA data transfer operations; and
execute the first DMA data transfer operation between a first memory space that is mapped to the first PCI-E NTB logical endpoint and a second memory space that is mapped to the second PCI-E NTB logical endpoint,
wherein the hypervisor executes in the second memory space,
wherein the storage virtual appliance executes in the first memory space, and
wherein the PCI-E NTB provides address translation between the first memory space and the second memory space.
16. The information handling system of claim 15 , wherein the first DMA data transfer operation is initiated from one of: the first memory space and the second memory space.
17. The information handling system of claim 15 , further comprising:
load a second storage driver on the hypervisor that accesses a second central processing unit (CPU) direct memory access (DMA) of the second processor for a second DMA data transfer operation of a plurality of second CPU DMA data transfer operations; and
execute the second DMA data transfer operation between the first memory space and the second memory space.
18. The information handling system of claim 17 , wherein the first DMA data transfer operation occurs from the first memory space to the second memory space.
19. The information handling system of claim 17 , wherein the first DMA data transfer operation occurs from the second memory space to the first memory space.
20. The information handling system of claim 15 , wherein the first DMA data transfer operation includes a programmed input/output (PIO).
21. The information handling system of claim 15 , wherein the first processor is associated with a first physical memory bank and the second processor is associated with a second physical memory bank;
wherein loading the storage virtual appliance includes specifying the first processor for the first CPU DMA data transfer operations; and
wherein the first PCI-E NTB logical endpoint is exclusively mapped to the first physical memory bank.