JPS6226553A - Program constituting method - Google Patents

Abstract

PURPOSE:To improve the execution speed of a program by making the whole operating system area resident substantially in a physical memory space. CONSTITUTION:A physical memory capacity being larger than a size of the whole operating system area 1 is provided, physical space can be assigned to plural spaces by a memory managing mechanism 3, and the whole operating system area 1 can be constituted by splitting it into several logical spaces. In each logical space, one main routine is always provided, respectively, and each object program is suppressed to a size which can be contained in one logical space. Also, a mechanism for executing inter-process communication is added to each logical space. In this way, a memory state 2 which has been constituted by splitting the whole operating system 1 can be made a program image 4 on a main storage.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a computer program (system program, user program), and particularly to a program configuration method for implementing a program in a memory space.

Traditionally, software in computer systems has been
The physical memory area is divided appropriately, and a single program is logically constructed between each division and a shared area that serves as a resident area for the operating system, etc., but in the case of large programs, the address space Its limitations make it virtually immovable.

permission The present invention has been made in consideration of the above points.

The present invention provides a program configuration method that makes it possible to implement a large-scale program that cannot be accommodated in one logical address space in a computer system with a narrow logical address space.

In order to achieve the object, the present invention provides a computer system having a memory management function for mapping a plurality of logical address spaces to a physical address space, and a function for making state transitions to different logical address spaces, each of which has a main routine. , and the program size in the executable format is large enough to be accommodated in one logical address space.A single program is divided into multiple parts, and a means for mutual inter-process communication is provided. This allows programs that have sufficient physical memory capacity to implement one program but have insufficient logical address space to reside.

Hereinafter, one embodiment of the present invention will be described in detail with reference to the accompanying drawings.

Currently, the hardware requirements are CPU and memory.

Consider the case where software is constructed in a computer system that is equipped with a memory management unit and has a trap instruction that can transition between a user program area and a system program area as a software requirement.

FIG. 3 shows an example of logical space correspondence in such a multi-program environment.

In such a correspondence example of each logical space LO to Ln, normally the logical space LO is a resident area for a system program such as an operating system, and the other logical spaces L1 to
Ln is set in the user program area. A user using a computer system is assigned an appropriate one of the logical spaces L1-Ln, and processes are executed there. This is becoming common, especially in the smallest microcombinant application systems.

On the other hand, you may want to make your program large-scale.

It may become impossible to accommodate all the data in one logical space.

Let us now assume a case where the operating system has grown so large that it cannot be accommodated in the logical space LO.

In that case, conventionally, as shown in FIG. 4, the operating system area O is smaller than the logical system space LSA.
If the entire 8A is quite large, the same logical system space L
By appropriately overmapping a part of the operating system into SA, the entire operating system is overlaid and implemented.

In this case, if a portion of the entire operating system that is not allocated to the logical system space LSA and is saved in the secondary storage device is accessed, a segmentation fault will occur and the portion will be remapped to the corresponding space. The corresponding part is then accessed and program execution continues. Thus, the entire operating system area SA can be made to behave as if it existed in the memory area.

However, in such conventional methods, when a segmentation fault occurs as described above, it is necessary to remap to the corresponding space, but at that time, the part that should be newly allocated is saved on the secondary storage device. Therefore, it is necessary to load it into memory. Therefore, it takes time to access the secondary storage device.

In contrast, in the present invention, as shown in FIG. 2, the entire operating system area O5A can be made to reside substantially in the physical memory space PMA, thereby making it possible to improve the program execution speed. ing.

In other words, if it is possible to provide a physical memory capacity larger than the entire size of the operating system area ○SA and make the physical space into a logical space using a memory management mechanism, the entire operating system area O3A can be divided into several parts. Configure. Here, logical user space AI, logical system space A2. This shows the case where it is decomposed into a logical user space A3.

Each of the parts A1 to A3 must have one main routine, and each object program is kept to a size that can be accommodated in one logical space (for example, 64 KB). Further, a mechanism is added to allow inter-process communication between the parts A1 to A3.

However, by configuring the program in this way, the image of each object program on the main memory can be made as shown in FIG. That is, part 1 in the figure shows an image of the entire operating system area, part 2 shows a memory state configured by decomposing the entire operating system area into multiple parts, and the memory state divided into multiple parts is divided into multiple parts by the memory management mechanism 3.
It is possible to create a program image in an executable format on the main memory as shown in the figure below.

Here, as a method for allocating each program, the functions of a general program management unit may be used as is.

The program management function is capable of specifying the space of the currently running process and the space of the process running before it by referring to the control register of the CPU.

Therefore, among the entire operating system area, 1. Instructions that can only be executed in system mode (CP
Part 2, which includes commands to change the U priority, commands to stop the CPU, etc., and a part that performs message communication (parameter passing) between multiple processes, are combined into one, and the memory management mechanism is used to handle multiple processes. Add a mechanism to switch between the two and assign it to the logical system space. Then, the remaining parts are appropriately decomposed and allocated to the user space. However, this means that you are running a system program in user mode, not a user program.

A request for inter-process communication generated by each process is accepted by the other process once via the logical system space, and each space can be switched by a trap instruction that performs this.

As described above, according to the program configuration method of the present invention, a large program that cannot be accommodated in one logical space is stored in the main memory, making it possible to execute a large-scale program and This allows for faster execution.

Note that although the above description has been made regarding the case where an operating system is implemented, it goes without saying that the same implementation is possible when configuring a general user program.

Unfortunately, in the program configuration method according to the present invention, in a computer system with a narrow logical address space,
It has the excellent advantage of being able to implement a large-scale program that cannot be accommodated in one logical address space in a manner that allows it to be executed and to execute the program quickly.

[Brief explanation of drawings]

FIG. 1 is a diagram schematically showing a program configuration method according to the present invention, FIG. 2 is a diagram showing an implementation state of an operating system according to the present invention, and FIG. 3 is a diagram showing the correspondence between logical spaces in a multi-program environment. 4 are diagrams showing the implementation state of a conventional operating system. LSA...Logical system space O3A...Operating system area AI, A3...Logical user space A2...Logical system space LO~Ln°゛°
logical space

Claims (1)

[Claims] A computer system has a memory management function that maps multiple logical address spaces to a physical address space, and a function that transitions states to different logical address spaces. A single program is divided into multiple parts each having a size that can be accommodated in one logical address space, and a means for inter-process communication is provided, with sufficient physical space to logically implement one program. A program configuration method that allows a program that has memory capacity but has insufficient logical address space to reside permanently.