Abstract: A method, system, and computer program product for data transfer to memory over an input/output (I/O) interconnect are provided. The method includes reading a mailbox stored on an I/O adapter in response to a request to initiate an I/O transaction. The mailbox stores a directive that defines a condition under which cache injection for data values in the I/O transaction will not be performed. The method also includes embedding a hint into the I/O transaction when the directive in the mailbox matches data received in the request, and executing the I/O transaction. The execution of the I/O transaction causes a system chipset or I/O hub for a processor receiving the I/O transaction, to directly store the data values from the I/O transaction into system memory and to suppress the cache injection of the data values into a cache memory upon presence of the hint in a header of the I/O transaction.

Abstract: Detection, notification and/or processing of events, such as errors associated with adapters, are facilitated. Hardware detects an event, places one or more adapters in an error state to prevent access to/from the adapters, and notifies the operating system of the event.

Abstract: Notification of hardware actions to be taken responsive to hardware events is facilitated. An operating system coupled, but external to, the hardware notifies firmware of the hardware action to be taken.

Abstract: A source identification facility is provided that enables identification of the one or more types of adapters requesting an interrupt in order to facilitate processing of the interrupt. The adapter types are accessible to the operating system and are used to tailor processing by the operating system of the interrupt.

Abstract: A method for implementing connected input/output (I/O) hub configuration and management includes configuring a first I/O hub in wrap mode with a second I/O hub. The hubs are communicatively coupled via a wrap cable. The method further includes generating data traffic on a computing subsystem that includes the hubs. Generating traffic includes: converting, via the first hub, a request to implement a transaction into an I/O device-readable request packet and transmitting the request packet over the wrap cable; converting, via the second hub, the I/O device-readable (IODR) request packet into a system readable request and transmitting the request over a system bus; converting, via the second hub, the response to an IODR response packet, and transmitting the response packet over the wrap cable; and converting, via the first hub, the IODR response packet into a system readable response packet, and transmitting the response packet over the system bus.

Abstract: The conditions under which adapter interruptions are made pending are controlled. Responsive to an interruption being presented to an operating system, subsequent interruptions are suppressed on all central processing units in the configuration. The operating system processes the interruption, including examining and processing indicators of reported events until the operating system discontinues the suppression. This enables the operating system to control the number of pending interruptions and the number of processors processing those interruptions.

Abstract: Accelerator functions are cascaded, such that a result of one accelerator function is directly forwarded to another accelerator function, bypassing the processor requesting the functions to be performed. The cascading may be provided during compilation of a program specifying the functions to be performed, but can be dynamically reversed during runtime of the program.

Abstract: One or more message signaled interruption requests from one or more input/output (I/O) adapters are converted to I/O adapter event notifications while retaining the message vector indication. An I/O adapter event notification may be routed and presented to a host or to a guest that the host is executing. To present the notification to the correct host or to the correct guest, various data structures in host and/or guest memory are used.

Abstract: One or more message signaled interruption requests from one or more input/output (I/O) adapters are converted to I/O adapter event notifications while retaining the message vector indication. An I/O adapter event notification may be routed and presented to a host or to a guest that the host is executing. To present the notification to the correct host or to the correct guest, various data structures in host and/or guest memory are used.

Abstract: A system to improve a Fibre Channel over Convergence Enhanced Ethernet (FCoCEE) network may include a sender in an FCoCEE network in which data packets having different data link layer structures are transmitted by the sender on a single data link. The system may also include a receiver to receive the data packets at the data link layer and to transmit an ACK and/or NAK in response to a sequence number in the data packets. The system may further include a replay buffer to retransmit the data packets where the replay buffer is sized by the length of the data link, data rate of the data link, the ACK and/or NAK processing time at either the sender and/or the receiver, and/or a threshold time for transmission and/or reception of the data packets.

Abstract: Dropped packets are recovered in a link layer of a Data Center Ethernet (DCE) network. Data packets for transmission are stored in a replay buffer. Each data packet includes a header having a field including data indicating that the data packet is formatted for reliable recovery at the link layer. The data packets are transmitted to a receiver across a link layer in the DCE network. The receiver determines whether a data packet has been dropped. If a data packet has not been dropped, an acknowledgement signal is sent to the transmitter in another data packet across the link layer, indicating that the data packet has been received. If a data packet has been dropped, a non-acknowledgement signal is sent to the transmitter in the other data packet across the link layer, indicating that the data packet has been dropped.

Abstract: Accelerators of a computing environment are managed in order to optimize energy consumption of the accelerators. To facilitate the management, virtual queues are assigned to the accelerators, and a management technique is used to enqueue specific tasks on the queues for execution by the corresponding accelerators. The management technique considers various factors in determining which tasks to be placed on which virtual queues in order to manage energy consumption of the accelerators.

Abstract: An address controller includes a bit selector that receives a first portion of a requester id and selects a bit from a vector that identifies whether a requesting function is an SR-IOV device or a standard PCIe device. The controller also includes a selector coupled to the bit selector that forms an output comprised of either a second portion of the RID or a first portion of the address portion based on an input received from the selector and an address control unit that receives the first portion of the RID and the output and determines the LPAR that owns the requesting function based thereon, the address control unit providing the corrected memory request to the memory.

Abstract: Systems, methods and computer program products for automatically triggering operations on a queue pair (QP). Methods include receiving a command at a remote direct memory access (RDMA) capable adapter. A trigger event element associated with the command is determined. The trigger event element is posted on a triggered QP. A triggeror element on a triggeror QP is posted, where the triggeror QP includes a reference to the triggered QP. A notification that the triggeror element has completed is received. The trigger event element is automatically initiated in response to receiving the notification.

Abstract: Detecting and recovering from soft errors in a network comprising a first device. A first device receives a first data packet. Responsive to receiving a second data packet, the first device determines whether the two data packets are identical. Responsive to the determination that the two data packets are not identical, the first device discards the two data packets, and requests retransmission of the two data packets.

Abstract: Disclosed are a PCI switch assembly, having automatic link failover, and a computer system including that switch assembly. The switch assembly comprises first and second interconnected, peripheral component switches. Each of the these switches has first and second primary ports and a plurality of secondary ports. The switch assembly has a normal mode and a failover mode. In the normal mode, each switch routes data through the switch to the secondary ports of the switch. In the failover mode, a failover path is defined and data are routed from the first switch to the second switch and then to one of the secondary ports of the second switch. The second switch detects a predefined fail condition, and changes the switch assembly from the normal mode to the failover mode in response to detecting the predefined fail condition.

Abstract: An energy management system for a data center network may include a central computer to establish an energy use policy for the computer data center network. The system may also include computer nodes in the computer data center network to receive a packet which is a query that obtains energy information from the nodes and/or a transmission reservation that provides instructions for the energy use policy to be implemented at the computer nodes. The computer nodes may reallocate data traffic on the computer data center network based on the energy use policy to improve energy consumption of the computer data center network.

Abstract: A method for determining data packet transmission times in an Ethernet protocol including, receiving a first data packet having a first data packet send time, wherein the first data packet send time is a time the data packet is sent by a source transmitter, subtracting the first data packet send time from a first data packet receive time to yield a first data packet transmission time, wherein the first data packet receive time is the time the data packet is received by a destination receiver, comparing the first data packet transmission time to a third time, determining whether a difference between the first data packet transmission time and the third time exceeds a threshold value, and sending a notification of a transmission delay responsive to determining that the difference between the first data packet transmission time and the third time exceeds the threshold value.

Abstract: A receive queue provided in a computer system holds work completion information and message data together. An InfiniBand hardware adapter sends a single CQE+message data to the computer system that includes the completion Information and data. This information is sufficient for the computer system to receive and process the data message, thereby providing a highly scalable low latency receiving mechanism.

Abstract: One or more message signaled interruption requests from one or more input/output (I/O) adapters are converted to I/O adapter event notifications. Each I/O adapter event notification includes the setting of one or more specific indicators in system memory and an interruption request, the first of which results in a pending I/O adapter interruption request. While a request for an I/O adapter interruption is pending, subsequent message signaled interruption requests are converted to I/O adapter event notifications, but do not result in additional requests for I/O adapter interruptions.