JPH01263728A - Microprocessor equipped with relative instructable register group - Google Patents

Abstract

PURPOSE:To secure a register only by the number needed with a process routine by moving a reference pointer, securing the necessary register only and defining and using the use at the relative position from the reference pointer when a new processing is started by a programmer. CONSTITUTION:In a system having N+1 registers (0 to N), when a newly used register is secured, the position of a reference pointer is moved to a register 0 side and when the register during the use is abandoned, the position of the reference pointer is moved to a register N side. A program executes a certain routine R1, at this time, the reference pointer of the register points n1, and when the routine R1 calls a routine R2, the routine R2 moves the reference pointer to n2, secures newly j2+1 registers and defines and uses registers n2 to (n2+j1); (n2+j1+1) to (n2+j2+j2); n1 to (n1+i1). When the routine of R2 is completed, the reference pointer is returned to n1 and n2 to (n2+j2) are abandoned.

Description

[Detailed Description of the Invention] (Field of Industrial Application) This invention can secure as many operation registers and pointers as required by each processing routine, and can use any number of passing registers to receive parameters between routines. The present invention relates to a microprocessor equipped with a register group that can be secured and relatively designated.

(Prior Art) Conventionally, register selection is an absolute designation, so programmers must always manage the storage state of registers when performing sub-chip calls or interrupt processing. Therefore,
As a method to solve this problem, a method is mainly used in which a plurality of register groups are provided and the register group to be used is changed over at appropriate times. At this time, a part of the register group may be duplicated for passing parameters between routines.

(Figure 2: BLKO-BLKM indicates M+1 register groups, and REGI to REGN indicate the absolute position of each register when each register group is selected.) (Problem to be solved by the invention) However, In the above means, l) Since the number of registers in one register group is fixed, there are excess registers depending on the processing, resulting in poor usage efficiency, and conversely, there may be a shortage of registers in complex processing.

2) Since the number of registers for passing parameters between routines is fixed, this is often a constraint when using registers.

There are drawbacks such as.

(Means for Solving the Problem) A microprocessor equipped with a register group that can be relatively designated according to the present invention has a means for relatively designating register selection from a reference pointer, and a means for specifying register selection relative to a reference pointer, and an arbitrary reference pointer for each processing routine. The present invention is characterized in that it is equipped with a means for moving the registers to a certain position, so that as many registers as required by the processing routine can be secured.

(Function) When starting a new process, the programmer moves the reference pointer to secure as many registers as necessary, and defines and uses the registers based on the relative position from the reference pointer. When processing is completed, the reference pointer is returned to its original position and the reserved register is discarded.

(Embodiment) FIG. 1 is a diagram showing the register usage state of a microprocessor according to the present invention. Note that p is the reference pointer, D
-N is the absolute position of the register, rl.

, r2 are the register usage ranges in routines R1 and R2, respectively, n1+n2 is the position indicated by the reference pointer, i,
~i3, j, and +j2 indicate the number of registers used.

In the figure, in a system with °N+1 registers (0 to N), when securing a new register to be used, the position of the reference pointer is placed on the register O side, and when discarding a register in use, the position of the reference pointer is placed on the register side. Let's move it to the N side. The program is currently executing a certain routine R1, and at this time, the reference pointer of the register points to nl, and this routine R is set to n1 to n, +i, : register n for storing parameters to be passed to the routine to be called from now on. + + + r ++-n, ++ 2: This routine uses registers for work as n, + i 2+l~n, +i3: Used as registers to store parameters received from the routine that called this routine. .

When routine R, calls routine R2 in this state, routine R2 moves the reference pointer to 02 and newly
Secure 2+1 registers, n2 to n2+j,
: Register for storing parameters passed to the routine to be called from now on (from R2) n2+ji+1-n2+j2: Registers for work in this routine 01 to nl+1. : Define and use a register that stores the parameters received from the routine that called this routine (R+ in this example). Here, n1 to n 1 +1
1 stores input parameters from routine R to routine R2.

When routine R2 completes all predetermined processing, it returns the reference pointer to nl and abandons registers n2 to n2+j2. At this time, output parameters from routine R2 to routine R are stored in n1 to n and i.

During this time, all registers within each routine are specified by relative positions to the reference pointer. For example, routine R
, the instructions MV REG and REG4 "transfer the contents of register n1+4 to register n1".
Transfer to +2. ” means.

With the architecture and operating procedure described above, programmers can always secure and define new registers for subroutine calls and interrupt processing, so there is no need to manage the storage status of registers, and Since the number of registers to be secured and defined is arbitrary, the degree of freedom in register definition is high and the burden on the programmer is lightened.

(Effects of the Invention) Since the programmer can always secure and define new registers during subroutine calls and interrupt processing, there is no need to manage the storage state of registers. Further, since the number of registers to be secured and defined for each processing routine is arbitrary, the degree of freedom in register definition is higher than in a method of switching between a plurality of register groups, and the burden on the programmer is reduced.

[Brief explanation of the drawing]

FIG. 2 is a diagram showing a conventional system having a plurality of register groups. P is the reference pointer, 0~N is the absolute position of the register, J+
12 is the position indicated by the reference pointer, 11 to i3 and j
, + j2 is the number of registers used.

Claims (1)

[Claims] 1. Equipped with a means for specifying register selection relative to a reference pointer and a means for moving the reference pointer to an arbitrary position for each processing routine, making it possible to secure as many registers as required by the processing routine. A microprocessor equipped with a register group that allows relative instructions.