Abstract: In a multiprocessor system, a method for dynamically establishing an I/O path between a processor node and an I/O device for routing data there between. The method starts with providing a configuration graph. The graph has objects associated with elements (devices) of the system and links connecting the objects. The node is identified, and links are then followed in the graph from an object associated with the I/O device to an object associated with a node. If multiple I/O paths exist and an optimal path is desired, the method includes providing in the links routing information containing nodes that can be directly reached via the link. Links are then followed, if possible, whose routing information contains the identified node. If no link having such routing information exists at an object along the path, then another link is chosen having routing information containing another node. This other link may be chosen in a round robin manner if there are multiple links to choose from.

Abstract: A method for configuring devices within a computer system provides device configuration information for types of devices that may be included within the computer system. The device configuration information for a device type includes interface types by which devices of that device type connect to other devices. The method includes determining the interface types of a current object associated with a device. Device types that can exist on a determined interface type of the current object are probed for. A child object is created for a device found from the probing and a link between the current object and the child object is also created. A child object is then made the current object, and the above actions are repeated at least until the devices to be configured within the system have associated objects linked together.

Abstract: Computer-based devices, whether initiators or targets, are assured access to a bus having a fixed priority arbitration scheme (such as a SCSI bus) by assigning to each initiator a “fair share” of the bus bandwidth. This share is defined as a number of bytes per a unit of time such as a time period. The shares together total a fraction of the total bus bandwidth, with a margin of bus bandwidth left unassigned. To prevent initiator starvation, each initiator monitors its bus requests to determine if it is being prevented by higher-priority initiators from using its assigned share of the bandwidth. If not, the initiator periodically pings each higher-priority initiator to indicate that it is not being starved. So long as a higher-priority initiator continues to receive pings from all lower-priority initiators, the higher-priority initiator can continue to use as much bandwidth as it needs.