JPS61204740A - Overlay controlling system for program - Google Patents

Abstract

PURPOSE:To curtail the memory capacity required for executing a program by using an overlay controlling system of a dynamic link type, and loading only a subroutine required for the execution, onto a memory at each time point. CONSTITUTION:When a program is being executed by a base address shown by a base register 1, an execution of a subroutine X is started by an instruction 'CALLX'. Usually a control is shifted to the subroutine X, but in this case,the control is shifted to a dummy routine 2 of the same name. The dummy routine 2 is prepared at every subroutine, and jumped to a program link control routine 4 by setting consecutive numbers of the subroutine. In the program link control routine 4, such processings as loading of the routine concerned, a change of the base register, and a stack adjustment, etc. are executed by the subroutine consecutive numbers. In an address management table 10, a load address on a memory of each subroutine is stored, and if it is unloaded, a special value such as 'O' is set in advance.

Description

[Detailed Description of the Invention] [Field of Application of the Invention] The present invention is a system that uses a dynamic link type to overlay control routines such as subroutines necessary when executing a program, and is capable of executing a large-scale program with a limited memory capacity. This invention relates to a program overlay control method suitable for

[Background of the invention]

Most conventional overlay control performed by the program itself is a planned overlay method in which the load address on memory and the content of the subroutine to be loaded are fixed in advance. This method requires exclusive execution between the subroutines on each overlay file and the subroutines on other overlay files, but as the program size increases, the number of subroutines increases and the interrelationships become more complex. The overlay file settings are
It was ill. In this case, it is necessary to store commonly used subroutines in duplicate in each overlay file or move them to the resident section.
In either case, the required memory capacity has increased. Furthermore, even if exclusive execution between subroutine groups in each overlay file can be guaranteed, adding, deleting, or changing the configuration of subroutines requires understanding the mutual calling relationships of all subroutines, which is difficult to implement when the scale of the program is large. It was a rotation.

To solve the problems of the above planning overlay method,
A dynamic link function is required that loads the subroutine requested to be executed into a free area in memory from the overlay file at the time of request, and determines the calling address at the time of execution.

By using this function, only the subroutines that are actually executed are loaded onto the memory, so the amount of memory required is reduced, and there is no need to consider mutually exclusive relationships between subroutines. However, dynamic link functionality is generally provided in O8.

In addition to the limited environments in which it can be used, in order to use this function, it is necessary to modify the program according to the O8 system call procedure.

Based on the above circumstances, it has become practical to realize a dynamic link type overlay control method in a microprocessor equipped with a stack mechanism and a pace register mechanism, without assuming O8 functionality and without requiring modification of existing programs. above is valid.

Note that Japanese Patent Application Laid-Open No. 58-217052 is related to this type of overlay control method.

Examples include JP-A-59-53944 and JP-A-59-24344.

[Purpose of the invention]

An object of the present invention is to provide a dynamic link type overlay control function that loads and executes a subroutine requested for execution in a program from an overlay file into a free area of memory at the time of request, and performs the subroutines required for program execution. An object of the present invention is to provide a program overlay control method that has the effect of reducing memory capacity.

[Summary of the invention]

Conventional dynamic link control uses O
Since it is realized by using O8, the scope of applicability is limited and requires program description specific to O8.

The present invention realizes the dynamic link function by the program itself without using the O8 function, and thus has a wide range of application and high versatility regardless of the type of O8 used.

Furthermore, in applying the present invention, since there is no need to modify existing programs, compatibility with conventional systems can be maintained. The present invention mainly relies on the functions of a "program link control subroutine" and a "dummy routine." The former is to load the subroutine requested for execution onto the memory and control the execution, and the point that the execution start address of each subroutine changes sequentially is realized by changing and restoring the pace register. The latter performs execution preparation procedures for subroutines called in the program and sets information required by the "program link control routine." Also, with the addition of a "dummy routine",
This eliminates the need to change subroutine call instructions in existing programs.

[Embodiments of the invention]

An embodiment of the present invention will be described below with reference to the overall configuration diagram in FIG. 1 and the processing flowchart in FIG. 2.

First, the basic system of the present invention will be described with reference to FIG.

While the program is being executed at the pace address indicated by pace register 1 in the figure, execution of subroutine X is started by instruction rCALL XJ. Normally, control is transferred to subroutine X 6, but in this method control is transferred to subroutine 2, which is a dummy routine with the same name. 2 is prepared for each subroutine, and a series of subroutines Nα is set to jump to the program link control routine 4. 4, a series of subroutines Nα performs processing such as loading the routine, changing the pace register, and adjusting the stack, the details of which will be explained with reference to FIG. Referring to FIG. 2, it is first checked whether the routine specified by the subroutine series selection is loaded onto the memory. Specifically, the address management table 10 in FIG. 1 is referred to. 10 stores the load address on the memory of each subroutine, and if it has not been loaded, a special value such as O is set.

When it is determined that the routine has not been loaded, the beginning of the empty area on the memory is set as the load address, and the corresponding routine is loaded onto the memory from the program library 3 in FIG. At the same time, register the load address as 10. Next, change the pace register value to the load address value of subroutine X and set it to point to 5 in FIG. However, if this continues, the first
Since the return address on the system stack shown in FIG. 7 is based on the old pace register value (1 in FIG. 1), it will not be the correct return address after execution of subroutine X is completed. Therefore, immediately before changing the pace register value, the return address stored in 7 is saved onto the user stack 8 in FIG. Then, the return address value above 7 is changed to return to the beginning of the dummy return routine 9 in FIG. 1, which is based on the pace register value 5.

After the above processing is performed in the program link control routine 4, subroutine X 6 in FIG. 1 is executed. By the return command (RETVRN) after execution is completed,
Referring to the return address above 7, control is transferred to the dummy repeat routine 9. At step 9, the pace register value is returned to the position 1, the true return address saved and placed on the user stack at step 8 is restored to 7, and control of the true return address is transferred. Therefore, the next instruction word of rCALL XJ is executed.

〔Effect of the invention〕

According to the present invention, there are first-order effects.

(1) Because the dynamic link type overlay control method loads only the subroutines necessary for execution into memory at each point in time, it is expected that the memory capacity required for program execution will be reduced.

(2) Compared to the planned overlay method, there is no need to consider the mutual exclusivity of subroutines, and module management of overlay files becomes easier.

(3) Since the overlay control function is added to the program itself, no special functions are required on the OS side, and there is no need to modify existing program parts, so the range of application is wide.

[Brief explanation of drawings]

FIG. 1 is a diagram showing the overall configuration of the present invention, and FIG. 2 is a diagram showing the flow of dynamic link control of the present invention. 1... Pace register value (normal state), 2... Dummy routine, 3... Program library. 4...Program link control device, 5...Pace register value (during dynamic link execution), 6...Subroutine X, 7...System stack, 8...User stack, 9...Dummy return routine. 10...Address management table. ¥/Country 2nd item (QALLX/I's life) m-"--Tamil-Chimpuro 2-24'Non 2 System 4 Cape Routineter S-Return Luken m-"-11

Claims (1)

[Claims] A program overlay control method is characterized in that dynamic linking is performed by programs using a program link control routine, a dummy routine, and a dummy return routine.