International Business Machines Corporation

Steven J Martin

Piyush Chaudhary

Jason E Goscinski

John Scott Houston

Leonard W Helmer, Jr

Robert Sidney Blackmore

Fu Chung Chang

Peter H Hochschild

Rama Krishna Govindaraju

Donald G Grice

Patricia E Heywood

Chulho Kim

Jang-Soo Lee

Michael J Miele

Kevin J Gildea

Paul D DiNicola

Gregory Salyer

Deryck X Hong

Carol L Soto

Yuqing Zhu

Jason S Houston

In a multinode data processing system in which the nodes communicate with one another via communication adapters over a network or switch, the adapters are provided with a dual register mechanism for tracking microcode task status. Upon the issuance of a disruptive command that requires attention from one of the nodes, the task status maintained in one register is copied to the snapshot register. As tasks within the adapter are completed, both registers are updated, thus providing a mechanism for the nodes to determine that all tasks active at the time of the disruptive command have completed. This means that the nodes now have a mechanism for determining, as soon as possible, that all tasks that are active when a disruptive command occurs have completed, thus allowing the data processing node to perform such operations as releasing system memory that is associated with the disruptive command, thus eliminating temporal overhead that can affect performance.

Snapshot interface operations

In remote direct memory access transfers in a multinode data processing system in which the nodes communicate with one another through communication adapters coupled to a switch or network, failures in the nodes or in the communication adapters can produce the phenomenon known as trickle traffic, which is data that has been received from the switch or from the network that is stale but which may have all the signatures of a valid packet data. The present invention addresses the trickle traffic problem in two situations: node failure and adapter failure. In the node failure situation randomly generated keys are used to reestablish connections to the adapter while providing a mechanism for the recognition of stale packets. In the adapter failure situation, a round robin context allocation approach is used with adapter state contexts being provided with state information which helps to identify stale packets.

Failover mechanisms in RDMA operations

In a multinode data processing system in which nodes exchange information over a network or through a switch, the mechanism which enables out-of-order data transfer via Remote Direct Memory Access (RDMA) also provides a corresponding ability to carry out broadcast operations, multicast operations, third party operations and conditional RDMA operations. In a broadcast operation a source node transfers data packets in RDMA fashion to a plurality of destination nodes. Multicast operation works similarly except that distribution is selective. In third party operations a single central node in a cluster or network manages the transfer of data in RDMA fashion between other nodes or creates a mechanism for allowing a directed distribution of data between nodes. In conditional operation mode the transfer of data is conditioned upon one or more events occurring in either the source node or in the destination node.

Third party, broadcast, multicast and conditional RDMA operations

In a multinode data processing system in which nodes exchange information over a network or through a switch, a structure and mechanism is provided within the realm of Remote Direct Memory Access (RDMA) operations in which DMA operations are present on one side of the transfer but not the other. On the side in which the transfer is not carried out in DMA fashion, transfer processing is carried out under program control; this is in contrast to the transfer on the DMA side which is characteristically carried out in hardware. Usage of these combination processes is useful in programming situations where RDMA is carried out to or from contiguous locations in memory on one side and where memory locations on the other side is noncontiguous. This split mode of transfer is provided both for read and for write operations.

Half RDMA and half FIFO operations

A system and method that utilizes a dedicated transmission queue to enable expedited transmission of data messages to adaptive 'nearest neighbor' nodes within a cluster. Packet descriptors are pre-fetched by the communications adapter hardware during the transmission of the preceding data element and setup for the next transmission is performed in parallel with the transmission of the preceding data element. Data elements of a fixed length that is equal to the cache line size of the communication hardware can optionally be used to provide optimized transfer between computer memory and communications hardware. The data receiving processing can also be optimized to recognize and handle cache line size data elements.

Speculative method and system for rapid data communications

A method is provided for operating a communications adapter employed in a multinode data processing system in a fashion which enhances the performance of remote direct memory access data transfers. The system is provided with pointers and a table which are employed to determine whether or not an address which has been supplied for the transfer has already been mapped to a real address at the source or destination node. The table is also preferably provided with counters which can be incremented or decremented to enable the use of least recently used mechanisms at the upper level protocol layers to more efficiently control the setting and resetting of table entries.

Lazy deregistration of user virtual machine to adapter protocol virtual offsets

Method for third party, broadcast, multicast and conditional RDMA operations

In remote direct memory access (RDMA) transfers in a multinode data processing system in which the nodes communicate with one another through communication adapters coupled to a switch or network, there is a need for the system to ensure efficient memory protection mechanisms across jobs. A method is thus desired for addressing virtual memory on local and remote servers that is independent of the process ID on the local and/or remote node. The use of global Translation Control Entry (TCE) tables that are accessed/owned by RDMA jobs and are managed by a device driver in conjunction with a Protocol Virtual Offset (PVO) address format solves this problem.

RDMA server (OSI) global TCE tables

A data correlation technique is provided for a computing environment having multiple independent operating environments. The technique includes associating by a first operating environment a first key to a set of data to be obtained responsive to an identified event within the computing environment; obtaining by a second operating environment the set of data with the associated first key, and associating a second key therewith; and using the set of data with the associated first key and second key to match to the set of data at least one other set of data obtained by the first operating environment or the second operating environment responsive to the event. The at least one other set of data has only the first key or the second key associated therewith. In one implementation, the event is a fatal event at a network interface adapter.

Growing community of inventors

Poughkeepsie, NY, United States of America

Steven J Martin