Abstract: A set of memory access operations is obtained. The set of memory access operations includes a plurality of memory access operations to be chained, in which the plurality of memory access operations are to be processed as an atomic unit. The plurality of memory access operations are executed in a particular order, and one or more results are provided.

Abstract: A set of memory access operations is obtained. The set of memory access operations includes a plurality of memory access operations to be chained, in which the plurality of memory access operations are to be processed as an atomic unit. The plurality of memory access operations are executed in a particular order, and one or more results are provided.

Abstract: A set of memory access operations is obtained. The set of memory access operations includes a plurality of memory access operations to be chained, in which the plurality of memory access operations are to be processed as an atomic unit. The plurality of memory access operations are executed in a particular order, and one or more results are provided.

Abstract: A set of memory access operations is obtained. The set of memory access operations includes a plurality of memory access operations to be chained, in which the plurality of memory access operations are to be processed as an atomic unit. The plurality of memory access operations are executed in a particular order, and one or more results are provided.

Abstract: A synchronous input/output (I/O) computing system includes a processor and a memory unit that stores program instructions. The system is configured to purge a device table cache (DTC) in response to the processor executing the program instructions. An operating system runs on the synchronous I/O computing system and issues a synchronous I/O command indicating a request to perform a device table entry transaction that has a total data length to be transferred. A device table entry is selected from a device table, loaded into the DTC, and data packets corresponding to the device table entry transaction are transferred using the selected device table entry. A host bridge processor monitors the data packets transferred using the selected table entry, and automatically purges the selected device table entry from the DTC in response to determining the transferred data packets match the total data length.

Abstract: A synchronous input/output (I/O) computing system includes a processor and a memory unit that stores program instructions. The system purges one or more address translation entries in response to the processor executing the program instructions to issue, via an operating system running on the synchronous I/O computing system, a synchronous I/O command indicating a request to perform a transaction. The program instructions further command the operating system to select a device table entry from a device table, load the entry into the DTC, request required address translation entries, install the required address translation entries in the address translation cache, and transfer data packets corresponding to the transaction. The program instructions further command the operating system to automatically purge the address translation cache entries associated with a transaction in response to detect that the transaction is completed.

Abstract: In response to a determination to allocate additional storage, within a real address space employed by a system memory of a data processing system, for translation control entries (TCEs) that translate addresses from an input/output (I/O) address space to the real address space, a determination is made whether or not a first real address range contiguous with an existing TCE data structure is available for allocation. In response to determining that the first real address range is available for allocation, the first real address range is allocated for storage of TCEs, and a number of levels in the TCE data structure is retained. In response to determining that the first real address range is not available for allocation, a second real address range discontiguous with the existing TCE data structure is allocated for storage of the TCEs, and a number of levels in the TCE data structure is increased.

Abstract: In response to a determination to allocate additional storage, within a real address space employed by a system memory of a data processing system, for translation control entries (TCEs) that translate addresses from an input/output (I/O) address space to the real address space, a determination is made whether or not a first real address range contiguous with an existing TCE data structure is available for allocation. In response to determining that the first real address range is available for allocation, the first real address range is allocated for storage of TCEs, and a number of levels in the TCE data structure is retained. In response to determining that the first real address range is not available for allocation, a second real address range discontiguous with the existing TCE data structure is allocated for storage of the TCEs, and a number of levels in the TCE data structure is increased.

Abstract: Embodiments of the present disclosure use vendor defined messages (VDMs) to send high priority information (e.g., cache writebacks) on a designated channel that is separate from a channel used for other commands (e.g., normal memory write commands). By using VDMs and a designated channel to send cache writebacks, the cache writebacks will not be blocked by normal memory write commands. For example, an endpoint device may encode cache writebacks as VDMs to be sent to a root complex. The root complex may store the VDMs in a dedicated VDM buffer and send the VDMs on a dedicated VDM channel.

Abstract: A computer-implemented method for computer-implemented method for communicating completion of synchronous input/output (I/O) commands between a processor executing an operating system and a recipient control unit is described. The method may include issuing, by a processor, a Synchronous I/O command to the recipient control unit; receiving, with the processor, a DMA read request from the recipient control unit; converting, with the processor, the DMA read response to write a data record into memory of the recipient control unit; issuing the DMA read request to the recipient control unit, wherein the DMA read request comprises an echo read portion comprising at least one byte of information at the end of the data record written; receiving, by the processor, a DMA write confirmation comprising the echo read portion of the record; and writing the echo read portion to a status area.

Abstract: A computing system includes a processor and a memory unit that stores program instructions. The system purges an entry from an address translation cache in response to the processor executing the program instructions to perform issuing, via an operating system running on the computing system, a command indicating a request to perform an I/O transaction requiring a translation entry. A host bridge monitors a total data length of the address translation entry to be transferred during the I/O transaction. An address translation entry is selected from an address translation table, loaded into the address translation cache, and data corresponding to the I/O transaction is transferred using the selected address translation entry. The host bridge automatically purges the selected address translation entry from the address translation cache in response to determining the transferred amount of data matches the total data length for the address translation entry.

Abstract: A system can translate an input/output (I/O) direct memory access (DMA) address to a physical system memory address in a data processing system. In response to receiving a DMA packet containing a requester identity (RID) associated with a partitionable endpoint (PE) number and an I/O DMA address, the system can retrieve an entry associated with the RID from a first translation validation table (TVT). Using that entry, the system can validate the number of TVT entries and extract from the I/O DMA address an offset. This offset can be validated and used to retrieve an entry in a second TVT. Data from this entry can be validated and the system can use this to access another table to retrieve the translation to the physical system memory address.

Abstract: Embodiments include a technique for management of data transactions, where the technique includes receiving, at a link interface, a packet from an I/O device, wherein the packet includes address information, and performing, by a host bridge, an address translation for the address information included in the packet. The technique also includes responsive to performing the address translation, determining a target page associated with a translated address of the packet is for at least one of a payload target page or a signaling target page, and appending a flag to a command based at least in part on the target page being associated with the translated address of the packet. The technique includes transmitting the command to an ordering controller for ordering the packet.

Abstract: Embodiments include a technique for management of data transactions, where the technique includes receiving, at a link interface, a packet from an I/O device, wherein the packet includes address information, and performing, by a host bridge, an address translation for the address information included in the packet. The technique also includes responsive to performing the address translation, determining a target page associated with a translated address of the packet is for at least one of a payload target page or a signaling target page, and appending a flag to a command based at least in part on the target page being associated with the translated address of the packet. The technique includes transmitting the command to an ordering controller for ordering the packet.

Abstract: A system includes a cut-through buffer operable to be asynchronously read while being written at different clock frequencies. The system also includes a controller operatively connected to the cut-through buffer. The controller is operable to write one or more values into the cut-through buffer in a first clock domain and compare a number of values written into the cut-through buffer to a notification threshold. A notification indicator is passed from the first clock domain to a second clock domain based on determining that the number of values written into the cut-through buffer meets the notification threshold. Based on receiving the notification indicator, the cut-through buffer is read from the second clock domain continuously without pausing until the one or more values are retrieved and any additional values written to the cut-through buffer during the reading of the one or more values are retrieved.

Abstract: Embodiments include methods, systems, and computer program products for performing synchronous data I/O. Aspects include a processor of computer system sending a store block to request data from a device through a PCIe connection, requested data having a predetermined number of data blocks, and the processor executing a data transaction loop to retrieve requested data. Executing the data transaction loop may include writing to a table prefetch trigger register on host bridge to queue up speculative prefetches in ETU for each data block. The host bridge may perform a first speculative prefetch to install a device table entry in a device table cache. The processor may further perform a second speculative prefetch to install an address translation in an address translation cache. The host bridge processes the data block received through direct memory access over the PCIe connection using the prefetched device table entry and address translation.

Abstract: A system includes an input/output adapter operable to receive a plurality of packets in a single clock cycle. The system includes a controller operatively connected to the input/output adapter. The controller is operable to receive a first packet at a data link layer and determine a state of a first output indicator to maintain packet ordering. Based on determining that a first receiver formatting interface is selected by the first output indicator, the controller performs an alignment adjustment and output of the first packet by the first receiver formatting interface. Based on determining that a second receiver formatting interface is selected by the first output indicator, the controller performs the alignment adjustment and output of the first packet by the second receiver formatting interface.

Abstract: A computing system includes a processor and a memory unit that stores program instructions. The system purges an entry from an address translation cache in response to the processor executing the program instructions to perform issuing, via an operating system running on the computing system, a command indicating a request to perform an I/O transaction requiring a translation entry. A host bridge monitors a total data length of the address translation entry to be transferred during the I/O transaction. An address translation entry is selected from an address translation table, loaded into the address translation cache, and data corresponding to the I/O transaction is transferred using the selected address translation entry. The host bridge automatically purges the selected address translation entry from the address translation cache in response to determining the transferred amount of data matches the total data length for the address translation entry.

Abstract: Embodiments include a technique for management of data transactions, where the technique includes receiving, at a link interface, a packet from an I/O device, wherein the packet includes address information, and performing, by a host bridge, an address translation for the address information included in the packet. The technique also includes responsive to performing the address translation, determining a target page associated with a translated address of the packet is for at least one of a payload target page or a signaling target page, and appending a flag to a command based at least in part on the target page being associated with the translated address of the packet. The technique includes transmitting the command to an ordering controller for ordering the packet.

Abstract: Embodiments include a technique for management of data transactions, where the technique includes receiving, at a link interface, a packet from an I/O device, wherein the packet includes address information, and performing, by a host bridge, an address translation for the address information included in the packet. The technique also includes responsive to performing the address translation, determining a target page associated with a translated address of the packet is for at least one of a payload target page or a signaling target page, and appending a flag to a command based at least in part on the target page being associated with the translated address of the packet. The technique includes transmitting the command to an ordering controller for ordering the packet.