49 results about "Single-root input/output virtualization" patented technology

A method and an information handling system (IHS) selectively configures a peripheral component interconnect express (PCIe) device with single or multiple root ability to service various compute node requests. A multi-root aware (MRA) module that manages an MRA switch receives requests for powering on PCIe adapters in respective chassis slot locations. If slot locations of respective PCIe adapters are designated as single root and unshared PCIe device slots, the MRA module provides power to the PCIe adapters and configures the PCIe adapter with single root ability. If slot locations of respective PCIe adapters are designated as multi-root and shared PCIe device slots, and corresponding PCIe adapters are single root input/output virtualization (SR-IOV) aware, the MRA module provides power to corresponding PCIe adapters, and configures, using the appropriate device drivers, the PCIe adapters with multi-root ability. The MRA module connects compute nodes via the MRA switch to requested PCIe device functions.

The invention relates to a dynamic scheduling method with a virtual function, belongs to the field of computer virtualization, and solves the problems of insufficient expandability and poor flexibility in using of virtual functions provided by a single root input and output virtualization standard network card so as to improve use ratio of the network card. The dynamic scheduling method with the virtual function comprises the steps of initializing, establishing queues, classifying virtual machines, distributing virtual functions, judging and wakening, adding virtual machines, dynamically adjusting and sleeping. During a period, the dynamic scheduling method with the virtual function performs dynamic scheduling on the virtual functions provided by the single root (input/output) I/O virtualization standard network card through statistics of I/O interruption times and sleeping time of the virtual machines, guarantees that the virtual function is utilized by a virtual machine with the largest I/O priority value, improves, network performance, fully utilizes hardware, and is used for supporting virtual machine platforms with a device direct distribution method.

A method, system and computer program product are provided for implementing distributed debug data collection and analysis for a hardware I/O adapter, such as, a Single Root Input/Output Virtualization (SRIOV) adapter in a virtualized system. A predetermined event triggers a base error collection sequence. Adapter driver data is collected and virtual function (VF) device drivers are locked out responsive to triggering the base error collection sequence. Adapter debug data is collected and the adapter is reinitialized including an indication to VF drivers if VF error data is to be collected. The virtual function (VF) device drivers are unlocked allowing the VF device drivers to commence recovery of a respective virtual function (VF).

A method, system and computer program product are provided for implementing shared adapter configuration updates concurrent with maintenance actions for a Single Root Input/Output Virtualization (SRIOV) adapter in a computer system. A configuration of the adapter is decoupled from the state of the adapter during a recovery period. When a configuration request is received, the configuration request is validated. Responsive to a valid configuration request, the saved configuration state of the adapter is updated. Once the adapter completes recovery, the adapter is restored to the new configuration instead of the configuration prior to failure.

The invention provides a queue scheduling method and a computing system. The method is applied to the computing system based on SR-IOV (Single Root Input Output Virtualization); the computing system comprises a universal processor and a special hardware chip. The universal processor is used for running one or a plurality of virtual machines. The special hardware chip is used for realizing a PF (Program Function) and one or a plurality of VFs (Virtual Functions) based on the SR-IOV protocol. Each virtual machine corresponds to at least one VF. The special hardware chip comprises a first queue idle link table which comprises a plurality of idle first queues connected according to the structure of the link table; the first queues are queues located in the special hardware chip. The method comprises the steps of distributing one or a plurality of first queues to one or a plurality of VFs corresponding to the virtual machines from the first queue idle link table in such a manner that the number of the first queues distributed to each VF is consistent with the number of second queues required by a program in the virtual machine corresponding to each VF, and obtaining data in the second queues in the virtual machines and storing the data in the second queues in the first queues for processing.

A system to allow reallocation of credit among virtual machines associated with separate operating systems includes drivers in each virtual machine to independently track credit usage and a host board adapter configured to report a false maximum to each operating system and track credit usage. The host board adapter allocates credits and reports the allocated credits to virtual functions accessed by the virtual machines. A hypervisor reallocates credits by reporting the new allocation to the host board adapter and consequently to each virtual function and each associated virtual machine. Each operating system maintains resources defined by the false maximum and never knows about the reallocation.

The embodiment of the invention discloses a virtual machine thermal migration method, which can realize SR-IOV network card state thermal migration. The method is applied to a source server running afirst virtual machine, the method is used for migrating the first virtual machine to a destination server, the source server comprises a single input/output virtualization SR-IOV network card, and thefirst virtual machine, a first virtual network card used by the first virtual machine, a virtual machine manager and a thermal migration module run in the source server. The first virtual network card is virtualized based on the SR-IOV network card, and the method comprises the following steps that the virtual machine manager sends a heating migration command to the thermal migration module; thethermal migration module obtains information needing to be migrated of the first virtual machine according to the thermal migration command, the information needing to be migrated of the first virtualmachine comprises state information of the first virtual network card and memory data of the first virtual machine, and the information needing to be migrated of the first virtual machine is reportedto the virtual machine manager; and the virtual machine manager sends the state information and the memory data of the first virtual machine to the destination server.

The invention relates to a virtual network resource allocation method and system, and an electronic device. The method comprises the following steps: step a: generating a task map according to the network information of all virtual machines on a physical machine, and dividing the task map by using a multilevel map partitioning algorithm to obtain k mutually disjoint divided sub-sets; step b: calculating a critical path with the longest completion time in the divided sub-sets to obtain the virtual machine located in a critical path node; and step c: allocating a single input and output virtualized virtual device for the virtual machine located in the critical path node. According to the virtual network resource allocation method and system provided by the invention, the network is optimizedin a manner of combining the para-virtualization technology and the single single input and output virtualization technology, it is defaulted that each virtual machine is allocated with a para-virtualized virtual network card and the single input and output virtualized virtual device is only allocated with the virtual machine located in the critical path node after the multilevel map division, sothat the purpose of improving the overall network bandwidth utilization rate and shortening the overall execution time of the parallel applications is achieved.

The embodiment of the invention relates to enabling Virtual Functions on storage media. The present disclosure describes apparatuses and methods for automatically mapping virtual functions to storagemedia to enable single root input output virtualization. A storage media switch manages access to virtual functions that execute behind a storage media interface managed by the switch. The switch includes a host interface through which the switch receives host commands. The switch determines virtual function identifiers associated with the host commands and automatically selects the virtual functions of the storage media based on the virtual function identifiers. The switch executes the host commands over the storage media interface using the virtual functions, and after execution, responds via the host interface to each of the host commands. By automatically mapping virtual functions in this way, the switch automatically enables single root input output virtualization of storage media, including storage media that is without native support for input output virtualization.

The invention discloses a flow control method of a virtual machine and a physical machine. The method comprises the following steps of obtaining the number of interrupt signals which are sent to the virtual machine within the preset time bucket when a first interrupt signal needs to be sent to the virtual machine; judging whether the number of the interrupt signals which are sent to the virtual machine is smaller than a preset threshold; if the number of the interrupt signals which are sent to the virtual machine is smaller than the preset threshold, sending the first interrupt signal to the virtual machine in order to make the virtual machine respond to the first interrupt signal to receive first data; and if the number of the interrupt signals which are sent to the virtual machine is equal to the preset threshold, storing the first interrupt signal into a waiting queue. The method solves the technical problem that the flow control cannot be carried out on the data received by the virtual machine with an SRIOV (Single Root Input Output Virtualization) physical network card when the data directly enters the virtual machine and is not subjected to virtualization network processing of the physical machine in the prior art.

Managing a pool of virtual functions including generating a virtual function pool comprising a plurality of virtual functions for at least one single root input/output virtualization (SR-IOV) adapter; creating a control path from a client virtual network interface controller (VNIC) driver in a first client partition to a target network using an active virtual function; receiving a failure alert indicating that the control path from the client VNIC driver in the first client partition to the target network using the active virtual function has failed; selecting, from the virtual function pool, a backup virtual function for the first client partition based on the failure alert; and recreating the control path from the client VNIC driver in the first client partition to the target network using the backup virtual function.