Abstract: A central processing unit (2) and a debug module (10) execute concurrent operations without requiring a data processor (3) to operate in a special debug mode. The use of a bus (25) to communicate data, address, and control information between a core (9) and debug module (10) allows debug module (10) to have access the same internal registers and memory locations as central processing unit (2). While debug module (10) and central processing unit (2) both have the ability to access the same internal registers and memory locations, central processing unit (2) may not modify a value stored in a plurality of breakpoint registers (50) when an Inhibit Processor Writes to Debug Registers (IPW) bit in a CSR (FIG. 8) of a plurality of control registers (40) is set. The IPW bit may only be modified by a command provided by an external development system (7).

Abstract: A central processing unit (2) and a debug module (10) execute concurrent operations without requiring a data processor (3) to operate in a special debug mode. The use of a bus (25) to communicate data, address, and control information between a core (9) and debug module (10) allows debug module (10) to have access the same internal registers and memory locations as central processing unit (2). While debug module (10) and central processing unte (2) both have the ability to access the same internal registers and memory locations, central processing unit (2) may not modify a value stored in a plurality of breakpoint registers (50) when an Inhibit Processor Writes to Debug Registers (IPW) bit in a CSR (FIG. 8) of a plurality of control registers (40) is set. The IPW bit may only be modified by a command provided by an external development system (7).

Abstract: A data processor (3) executes a real time trace function which allows an external development system (7) to dynamically observe internal operations of data processor (3) without assuming a type or availability of an external bus and without significantly impacting the efficiency and speed of the data processor (3). A debug module (10) of data processor (3) provides a parallel output port for providing internal operating information via a DDATA signal and a PST signal. The DDATA signal provides data which reflects operand values and the PST signal provides encoded status information which reflects an execution status of a central processing unit 92). Furthermore, the DDATA signal also provides captured instruction address program flow changes to allow external development system (7) to trace an exact program flow without requiring an externally visible address bus or an externally visible data bus.

Abstract: A computer processor (110) automatically generates block-size operand references during execution of standard instructions. As such a standard instruction is executed, the processor (110) continually examines the number of bytes to be moved (342) and the relative alignment of the operand address (352). At any time during instruction execution, if the operand address is zero modulo the block size, and at least a block sized number of bytes remain to be moved (354), the operand transfer is marked as a block-sized reference.This provides a convenient method for generating block-sized memory references to/from the targeted address space, independent of cache modes such as copyback, write-through, or non-cacheable. This may produce burst accesses, maximizing performance of the data transfer. Additionally, cache memory writes can be optimized to avoid cache line fill reads.

Abstract: A system bus controller (103) within a processor (101) includes programmable logic for different modes of chip enable signals on a per-address-space basis. This allows for a "glueless" interface (107) between the processor (101) and different types of external devices (111, 112, 113), such as memory devices. A chip select register value 604, 608, 612 is preprogrammed with respect to each external device coupled to the processor (101). This preprogrammed register value 604, 608, 612 is used by the system bus controller (103) to uniquely configure a read/write access signal to be sent to each of the external devices (111, 112, 113).

Abstract: A most recently used bit (25) is used to determine if a data transfer should occur from a fill buffer (20) into a data memory (32) in a cache (15) If the data to be displaced in the data memory (32) has been referenced more recently than the data present in the fill buffer (20), then the transfer should not occur. When a cache miss does occur, a control register (50) is used to determine the conditions for loading the fill buffer (20).

Abstract: A data processor (3) implements a single system stack pointer (244) to create records of both supervisor and user stack operations when hardware support for misalignment of such stack operations is optional. The single system stack pointer is implemented as a self-aligning stack pointer which automatically aligns itself to a nearest 0-modulo-4 address below a current address setting such that no alignment system errors occur even when there is no hardware support for misaligned operands. Once the automatic alignment has occurred, the data processor stores a format field in an exception stack frame to indicate information about the alignment of the stack pointer at the time of the error. When the exception has been serviced, the processor uses the four bit format field stored in the exception stack frame to restore the stack pointer to its original value at the time of the exception.

Abstract: A data processor (3) executes a debug operation by minimally intruding on the real time operation of the data processor and without halting the data processor. The data processor implements a control register (40) which stores trigger response value for determining a function executed by the data processor when a breakpoint signal is asserted. The trigger response values indicates whether a central processing unit (2) of the data processor executes a special debug exception processing routine or suspends operation when the breakpoint signal is asserted.

Abstract: A data processor (3) executes a breakpoint operation before an exception processing routine for a reset operation is initiated. When an External Reset signal is asserted and subsequently negated, a window of time exists in which data processor (3) is quiescent before beginning an actual reset exception processing routine. If an external breakpoint signal, BKPT is asserted during a quiescent time by external development system (7), data processor (3) downloads a target memory value into a memory (6) such that any hardware register configuration may be performed.

Abstract: A processor (10) has a data cache unit (16) wherein the data cache unit includes a memory management unit (MMU) (32). The MMU contains memory locations within transparent translation registers (TTRs), an address translation cache (40), or a table walk controller (42) which store or generate cache mode (CM) bits which indicate whether a memory access (i.e., a write operation) is precise or imprecise. Precise operations require that a first write operation or bus write instruction be executed with no other operationsnstructions executing until the first operation/instruction completes with or without a fault. Imprecise operations are operations/instruction which may be queued, partially performed, or execution simultaneously with other instructions regardless of faults or bus write operations. By allowing the logical address to determine whether the bus write operation is precise or imprecise, a large amount of system flexibility is achieved.

Abstract: A scan chain architecture which has a controller (10), and a multiplexer (24) is used to route test data through functional units (12, 14, 16, 18, 20, and 22). The controller (10) receives as input a serial data stream from an STDI terminal and demultiplexes this data stream to one of the functional units (six functional units are illustrated in FIG. 1). Each of the functional units is considered as one scan chain and therefore FIG. 1 has six scan chains (one for each functional unit). In addition, a seventh scan chain couples all output flip-flops in each of the functional units together between an output of the MUX (24) and the STDO terminal/pin. Therefore, a serial scan of a data stream can be done through one functional unit, the multiplexer (24) and into the output flip-flops of each function unit to make testing easier to set-up. In addition, various new scan chain cells and low power methods are used herein.

Abstract: A branch cache (40) has a plurality of storage levels (120, 122, 140, and/or 142) wherein at least two write registers (114 and 116) are used to perform a parallel write operation to at least two of the storage levels in the plurality of storage levels (120, 122, 140, and/or 142). The two write registers (114 and 116) are provided due to the fact that the branch cache 40 is implemented as a multi-state (typically five state--see FIG. 5) branch prediction unit having instruction folding. Instruction folding, as taught herein, allows a branch instruction which is predicted as being taken to be executed along with an instruction that precedes the branch in execution flow. The instruction which directly precedes the branch in execution flow is usually the instruction which is used to "fold" the branch. Effectively, this instruction folding allows branches, which are predicted as being taken, to be executed in zero clock cycles.

Abstract: A processor (10) has two modes of operation. One mode of operation is a normal mode of operation wherein the processor (10) accesses user address space or supervisor address space to perform a predetermined function. The other mode of operation is referred to as a debug, test, or emulator mode of operation and is entered via an exception/interrupt. The debug mode is an alternate operational mode of the processor (10) which has a unique debug address space which executes instructions from the normal instruction set of the processor (10). Furthermore, the debug mode of operation does not adversely affect the state of the normal mode of operation while executing debug, test, and emulation commands at normal processor speed. The debug mode is totally non-destructive and non-obtrusive to the "suspended" normal mode of operation. While in debug mode, the existing processor pipelines, bus interface, etc. are utilized.

Abstract: A method and apparatus for placing a data processor (12) into a low-power mode of operation using a system (10). The system (10) has a processor (12). The processor (12) has access to a bus (18). The bus (18) is coupled to a bus controller (14). The processor (12) sends a broadcast cycle out through the bus (18) when the processor (12) desires to enter a low-power mode of operation. The bus controller (14) determines that the broadcast cycle has been sent on the bus (18). The bus controller (14) waits a predetermined amount of time to process the low-power request and grants permission to the processor (12) to enter the low-power mode via the communication of a transmission termination signal. The processor (12) conditionally drives either logic ones or a tri-state value onto the bus (18) depending upon whether or not the processor (12) has been granted ownership of the bus (18).

Abstract: A system and technique for providing early decoding of complex instructions in a pipelined processor uses a programmed logic array to decode instruction segments and loads both the instruction bits and the associated predecoded bits into a FIFO buffer to accumulate a plurality of such entries. Meanwhile, an operand execute pipeline retrieves such entries from the FIFO buffer as needed, using the predecoded instruction bits to rapidly decode and execute the instructions at rates determined by the instructions themselves. Delays due to cache misses are substantially or entirely masked, as the instructions and associated predecoded bits are loaded into the FIFO buffer more rapidly than they are retrieved from it, except during cache misses. A method is described for increasing the effective speed of executing a three operand construct.

Abstract: A system and technique for providing early decoding of complex instructions in a pipelined processor uses a programmed logic array to decode instruction segments and loads both the instruction bits and the associated predecoded bits into a FIFO buffer to accumulate a plurality of such entries. Meanwhile, an operand execute pipeline retrieves such entries from the FIFO buffer as needed, using the predecoded instruction bits to rapidly decode and execute the instructions at rates determined by the instructions themselves. Delays due to cache misses are substantially or entirely masked, as the instructions and associated predecoded bits are loaded into the FIFO buffer more rapidly than they are retrieved from it, except during cache misses. A method is described for increasing the effective speed of executing a three operand construct.

Abstract: A multiprocessing system includes a cache coherency technique that ensures that every access to a line of data is the most up-to-date copy of that line without storing cache coherency status bits in a global memory and any reference thereto. An operand cache includes a first directory which directly, on a one-to-one basis, maps a range of physical address bits into a first section of the operand cache storage. An associative directory multiply maps physical addresses outside of the range into a second section of the operand cache storage section. All stack frames of user programs to be executed on a time-shared basis are stored in the first section, so cache misses due to stack operations are avoided. An instruction cache having various categories of instructions stores a group of status bits identifying the instruction category with each instruction.

Abstract: A multiprocessing system includes a cache coherency technique that ensures that every access to a line of data is the most up-to-date copy of that line without storing cache coherency status bits in a global memory and any reference thereto. An operand cache includes a first directory which directly, on a one-to-one basis maps a range of physical address bits into a first section of the operand cache storage. An associative directory multiply maps physical addresses outside of the range into a second section of the operand cache storage section. All stack frames of user programs to be executed on the time-shared basis are stored in the first section, so cache misses due to stack operations are avoided. An instruction cache haivng various categories of instructions stores a group of status bits identifying the instruction category with each instruction.

Abstract: An associative memory used to translate a virtual page number (VPN) of a virtual word address to a physical page number (PPN) of a physical word address of a random access memory of a digital computer system is provided with a pair of independently addressable validity bit arrays, each of which arrays can store a validity bit in each of the addressable locations of each array. A pointer enables only one of the validity bit arrays to receive address signals corresponding to the lower virtual page number (LVPN) of a VPN. The validity bit read out of the memory location corresponding to the LVPN of the enabled array is used in determining if the PPN read out of the corresponding memory location of the associative memory is valid. The bits of the disabled array, immediately after it is disabled, are all reset, or cleared.

Abstract: In a data processing system including a central processing unit capable of operation with a plurality of operating systems, a VMSM unit is described for producing a composite decor descriptor from a plurality of possible decor descriptor formats. The VMSM unit includes an input buffer unit and an output buffer unit, a control unit to analyze an incoming DATA and provide appropriate control signals, a reconfiguration unit for reformatting the plurality of descriptor formats into a composite format, a descriptor fetch unit for retrieving a descriptor when the signals applied to the VMSM unit contain a descriptor address, and a descriptor master copy unit which contains a copy of the descriptors stored in the addressing apparatus.