Abstract: A context-selection mechanism is provided for selecting a best context from a pool of contexts for processing a data packet. The context selection mechanism comprises, an interface for communicating with a multi-streaming processor; circuitry for computing input data into a result value according to logic rule and for selecting a context based on the computed value and a loading mechanism for preloading the packet information into the selected context for subsequent processing. The computation of the input data functions to enable identification and selection of a best context for processing a data packet according to the logic rule at the instant time such that a multitude of subsequent context selections over a period of time acts to balance load pressure on functional units housed within the multi-streaming processor and required for packet processing. In preferred aspects, programmable singular or multiple predictive rules of logic are utilized in the selection process.

Abstract: A software program extension for a dynamic multi-streaming processor is disclosed. The extension comprising an instruction set enabling coordinated interaction between a packet management component and a core processing component of the processor. The software program comprises, a portion thereof for managing packet uploads and downloads into and out of memory, a portion thereof for managing specific memory allocations and de-allocations associated with enqueueing and dequeuing data packets, a portion thereof for managing the use of multiple contexts dedicated to the processing of a single data packet; and a portion thereof for managing selection and utilization of arithmetic and other context memory functions associated with data packet processing. The extension complements standard data packet processing program architecture for specific use for processors having a packet management unit that functions independently from a streaming processor unit.

Abstract: A logic system for resolving port contentions associated with memory-access requests in data packet routing is provided. The logic system comprises a determination logic for assessing and reporting port status of busy or not busy, a command mechanism for issuing commands contingent on determination results and a staged buffer memory for holding pending requests waiting for permission to access the memory. A single request at the head of the buffer memory is considered for port access whereupon if a port is determined to be busy, the command logic issues appropriate commands to units responsible for downloading packets from the memory and for sending new memory-access requests.

Abstract: A hardware/software system is provided for allocating memory in the form of a buffer zone surrounding a data packet to be stored in the memory. The hardware/software system comprises, first and second registers for storing separate values representing in one register, an amount of memory preceding the first line of the data packet to be stored and in the other the amount succeeding the last line of the packet to be stored, a hardware mechanism for allocating the memory according to computational results computed using the register values and the size of a data packet to be stored, and software for processing stored data packet and for writing any new growth data into the designated buffer zones surrounding the data packet.

Abstract: A system is provided for enabling a non-speculative pre-fetch operation for processing instructions to be performed in the background ahead of immediate packet processing by a packet processor. The system comprises a packet-management unit for accepting data packets and en-queuing them for processing, a processor unit for processing the data packets, a processor core memory for holding context registers and functional units for processing, a memory for holding a plurality of instruction threads and a software-configurable hardware table for relating queues to pointers to beginnings of instruction threads. The packet-management unit selects an available context in the processor core for processing of a data packet, consults the table, and communicates the pointer to the processor, enabling the processor to perform the non-speculative pre-fetch for instructions.

Abstract: A logic system in a data packet processor is provided for selecting and releasing one of a plurality of contexts. The selected and released context is dedicated for enabling the processing of interrupt service routines corresponding to interrupts generated in data packet processing and pending for service. The system comprises, a first determination logic for determining control status of all of the contexts, a second determination logic for determining if a context is idle or not, a selection logic for selecting a context and a context release mechanism for releasing the selected context. Determination by the logic system that all contexts are singularly owned by an entity not responsible for packet processing and that at least one of the contexts is idle, triggers immediate selection and release of an idle one of the at least one idle contexts to an entity responsible for packet processing.

Abstract: A system for allocating storage of incoming data packets into a memory of a packet processor has a first facility mapping a first block of memory of a fixed block size in bytes into an ordered plurality of atomic pages comprising each a fixed byte size, a second facility mapping the same block of memory into ordered virtual pages of different sizes, ranging from a smaller virtual page size equal to the atomic page size up to a larger virtual page size equal to the fixed block size, a third facility to allocate virtual pages as unavailable for storage or de-allocate virtual pages as available for storage, a fourth facility to receive a data packet, ascertain packet size for the received packet, and to determine fit by checking allocation state for virtual pages of a smallest size that is equal to or larger than the packet size, then allocation state for next larger virtual pages, and so on, until a de-allocated, available virtual page is found; and a fifth facility to select a virtual page to store the packet,

Abstract: A bypass system for a data cache has two ports to the data cache, registers for multiple data entries, a bus connection for accepting read and write operations to the cache, and address matching and switching logic. The system is characterized in that write operations that hit in the data cache are stored as elements in the bypass structure before the data is written to the data cache, and read operations use the address matching logic to search the elements of the bypass structure to identify and use any one or more of the entries representing data more recent than that stored in the data cache memory array, such that a subsequent write operation may free a memory port for a write stored in the bypass structure to be written to the data cache memory array. In a preferred embodiment there are six entries in the bypass system, and stalls are eliminated.

Abstract: In a data-packet processor, a configurable queueing system for packet accounting during processing has a plurality of queues arranged in one or more clusters, an identification mechanism for creating a packet identifier for arriving packets, insertion logic for inserting packet identifiers into queues and for determining into which queue to insert a packet identifier, and selection logic for selecting packet identifiers from queues to initiate processing of identified packets, downloading of completed packets, or for requeueing of the selected packet identifiers.

Abstract: A pipelined processor in which the decoder can consume a portion of an instruction and hold that portion in sub-field shadow registers while retrieving the remainder of the instruction in a subsequent cycle or cycles. Each byte in a prefetch buffer is individually tagged such that the decoder can clear individual bytes in the prefetch buffer in order to allow additional instruction bytes to be prefetched before the current instruction is completely consumed and decoded by the decode stage. This allows for an optimal buffer size that is less than the maximum possible instruction length but large enough to hold a complete copy of the vast majority of instructions.

Abstract: A microcode based decoder circuit for microprocessors that uses fast access tables to decode instructions. The pointers to the tables are generated directly from the instruction prefetch buffers. Information bits about the instruction are added to the tables at no extra cost and enable the faster decode of the instruction. The present invention includes the decode of an instruction using an entry ROM, which contains information regarding the instruction that can directly be used in generating the decoder outputs. This information is also used in selecting the correct ROM entry, thus enhancing the flexibility of the decoder, and to dynamically generate a generic microcode entry. Thus, microcode space requirements are reduced. A generic microcode instruction is used for commonly used, similar macroinstructions. This avoids duplication of microcode instructions and thus reduces the required microcode space.

Abstract: An apparatus and method for efficiently generating the consecutive addresses needed to access misaligned or doubleword length data stored in the memory of a general purpose microprocessor. The apparatus shares the address generation operations between a small 3 bit adder, typically contained in the bus unit, and the execution unit. Control logic is used to determine whether a data misalignment situation exists based on the length of the data which is to be retrieved and the starting address of the data. When misalignment is indicated, the control unit acts to assign the address calculations to either the 3 bit adder alone or the execution unit together with the 3 bit adder depending upon how much the present address must be incremented to obtain the new addresses.

Abstract: An efficient method for implementing string operations used to process blocks of data within a memory. The registers used to track the memory addresses are updated and committed before the outcome of a read or write address operation is known. In the event an exception occurs, exception handling hardware and microcode is used to restore the state of the registers to the condition they were in prior to the iteration which produced the exception. This reduces the number of microcode instructions required to implement the string operation, producing a faster cycling of the code through multiple iterations. The result is a more optimal code for string operations and a decrease in the time required to carry out the string instruction.

Abstract: A method and apparatus for prefetching instructions in a pipelined processor including first and second prefetch buffers arranged in a two tier system. As instruction bytes are fetched from cache memory or external memory, those instruction bytes from memory for which there is space in the first level buffer are loaded therein, and, simultaneously, those valid instruction bytes in the second tier buffer for which there is room in the first tier buffer are also loaded into the first tier buffer. Those instruction bytes from memory for which there is not currently room in the first tier buffer are loaded into the second tier buffer. The second tier buffer is also used as a buffer for loading the instruction cache memory from the external memory.

Abstract: A microprocessor execution unit includes an arithmetic unit and an addressing unit. The arithmetic unit performs arithmetic and logical operations on operands. The addressing unit operates in conjunction with the arithmetic unit to calculate offsets, limits, and linear addresses in a single cycle.

Abstract: A circuit designates the values of an M bit first pointer and of an N+M bit second pointer. A first register circuit network holds the M bit first pointer, and a second register circuit network into holds an M bit portion of the N+M bit second pointer. A third register circuit network holds the remaining N bit portion of the N+M bit second pointer. A combiner circuit network, connected to receive the M bit first pointer from the first register circuit network, combines the received M bit first pointer with an externally provided data element length value to generate a new M bit first pointer. The combiner circuit network selectively generates a carry signal. The new M bit first pointer is selectively provided for loading into the first register circuit network and for loading into the second register circuit network as the M bit portion of the N+M bit portion of the second pointer.

Abstract: A system comprises a service processor and a plurality of operating units dependent on the service processor. The service processor responds to service requests from the operating units and services the operating units one at a time. A scheduler is responsible for assigning priority to the operating units and determining the order in which the service requests are handled. A register contains a value indicative of the operating unit currently being serviced and is under control of the scheduler. According to one aspect of the present invention the register is also under control of the service processor itself. Another register, under control of the service processor, is coupled to the scheduler to generate service requests thereto independent of the operating units. A memory addressable by the service processor stores data. The service processor is capable of generating addresses for the memory derived from the contents of the register indicative of the operating unit currently being serviced.

Abstract: A timer subsystem which provides a data processor servicing the timer subsystem with the ability to inhibit the match recognition logic of the timer subsystem while the processor is servicing the subsystem. The disclosed embodiment comprises a sixteen-channel timer subsystem with a dedicated service processor. The service processor, under control of the micro-coded programs executing thereon, is capable of disabling a match recognition latch in the timer channel currently being serviced. This feature provides the ability to prevent unwanted matches which occur while the service processor is updating the match register, for instance. Another feature of the timer subsystem is the inhibition of multiple matches to a single match register value by disabling the match recognition latch upon the occurrence of a match and re-enabling it only when the match register is written by the data processor.

Abstract: A timer system comprises a plurality of timer channels serviced by a single service processor. Each of the timer channels is capable of both input (capture) and output (match) functions. The microprogrammed service processor is responsible for configuring each of the channels for their intended uses and for responding to service requests generated by the channels in response to the occurrence of timer events. Features of the timer channels include the ability to continuously execute capture functions without generating service requests, the ability to execute a single capture function and service request and protect the captured value from being overwritten until the service request has been responded to and the ability to combine match and capture functions in such a way as to place a time-out window on the capture event.