JPS63247821A - Ram disk controller - Google Patents

Abstract

PURPOSE:To minimize the memory transfer processing by providing a disk image area on a main memory that can receive a direct access from a CPU. CONSTITUTION:A CPU 21 is connected to a main memory 22 and a memory bus 25 and each input/output device 24 is connected to the CPU 21 via a system bus 3. An OS, the shared data, etc., are transferred to the memory 22 from the device 24 via the bus 23. At the same time, a disk image area is included in the memory 22. For the state of the memory 22 set in a task executing state, a task initial value presence/absence data part 9 is provided together with a task stack part 10. A task consists of three elements, i.e., a text part, a data part and a stack part. A header part is provided at the head of the test part 6 of the task stored in a memory area. Furthermore an idle memory is retrieved for security of a memory for the data part and the initial value presence data is transferred from said memory area.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a computer system without a magnetic disk device, and particularly to a RAM disk control device.

[Conventional technology]

The conventional RAM disk device is "Nikkei Byte" (198
6, 4, pages 74 to 81), tasks and shared data in the RAM disk are stored in the CP.
It cannot be accessed directly from U, and the RAM disk and main storage are completely separate in function. When accessing the tasks and shared data in the RAM disk, it is necessary to transfer the tasks and shared data from the RAM disk to the main memory, and a dedicated driver program is required for this purpose.

Also, Unix User's Manual Release 3
, 0 (1980, 6), pages 4-1 (UNIX U
ser'sManual Release 3.0(J
1980) PP4-1) describes a function that allows software to read and write the main memory and the core image. However, this system has a magnetic disk.

[Problem that the invention seeks to solve]

The conventional RAM disk system described above does not take into account memory usage efficiency and system overhead. This is because when executing a task or using shared data, it is necessary to transfer the task or shared data from the disk image area to the main memory, which can be directly accessed from the CPU, which increases system overhead.
Furthermore, tasks and shared data must exist in both the disk image area and main memory, which uses twice as much memory and reduces memory usage efficiency.

An object of the present invention is to provide a RAM disk control device that improves memory usage efficiency and reduces system overhead.

[Means for solving problems]

In the present invention, the O8 is provided with a function that allows the CPU to directly access tasks, shared data, etc. in the disk image area on the main memory.

That is, in the present invention, a readable/writable disk image area is provided on the main memory that can be directly accessed by the CPU using the same means as a conventional magnetic disk, and a table for storing the start address and size of the disk image area and the area are provided. It has a table that stores the positions in the disk image area of registered tasks and shared data.

[Effect]

According to the present invention, tasks, shared data, etc. can be directly accessed by the CPU, and tasks are generated to execute tasks or refer to shared memory in a magnetic diskless system.

The shared memory transfer time was minimized and memory usage efficiency was improved. The magnetic disk image area secured on the main memory can be read and written by the same means as a magnetic disk, and can be directly accessed by the CPU. This improves the memory usage efficiency of the main memory when executing tasks and using shared data, and reduces the transfer time between the disk image area and the main memory during task switching.

〔Example〕

An embodiment of the present invention will be described below.

FIG. 2 shows a system configuration to which the present invention is applied. A central processing unit 1 (CPU) 21 is connected to a main storage device J22 and a memory bus 23, and each input/output device 24 is connected via the system bus 23. O8 (operating system), tasks, shared data, etc. are transferred from the input/output bag [124] to the main storage device i22 through the system bus 23. The disk image area exists within the main storage device 22.

FIG. 1 shows an outline of the RAM disk control method, focusing on the usage status of the main memory.

FIG. 4 shows each related table.

The disk image area configuration table in Figure 4 (a) is
One exists for each disk image area.

The disk image area management table in Figure 4 (b) is
There is one for each task, shared data, etc. in the disk image area. The header section shown in FIG. 4(c) exists at the beginning of the text section within the disk image area, and one header section exists for each task. The PCB in Figure 4 (d) is one for each task.
The fourth one exists and has the address value of the disk image.
The task address management table in Figure (E) has one for each task, and has the start address and size of each element.

FIG. 5 shows a main memory usage management table. This table manages the use/unuse of a certain size area in the main memory using flags.

6(a) and 6(b) show a flowchart of task/shared data registration processing, and FIG. 7 shows a flowchart of task activation processing.

The disk image defined diagram in FIG. 1(A) shows the state after the disk image area 3 has been secured on the main memory 22. The main memory 22 includes a fixed allocation area 2 where O8 etc. exist, a disk image area 3. Variable allocation area 4-3
divided into. The start address of the disc image area 5 is stored in the disc image area configuration table T□ shown in FIG. 4(a).

FIG. 1(b) shows the state of the main memory after the task/shared data is registered. Tasks and shared data are registered in the free area obtained by searching the disk image area management table T2 in FIG. 4(b) through the disk image area driver program. In the image area 3, 6 indicates the text part of the task, 7 indicates data with an initial value of the task, and 8 indicates shared data. At this point, shared data 8
can be accessed by referring to the address stored in the disk image area management table T2.

FIG. 1(c) shows the state of the main memory when a task is executed. 9 indicates the initial presence/absence data part of the task; 10
indicates the stack part of the task. The components of a task are a text part and a data part.

There are three stack parts. At the beginning of the text section 6 of the task in the disk image area is a header section H shown in FIG. 4(c), which stores the capacity of the constituent elements of the task. Regarding the data section, refer to the header and the PCB table T3 in FIG. 4(d), find the capacity of the data section with initial value and the data section without initial value of the task, and use the memory usage management table T5 in FIG. A free memory corresponding to the size of the data section is searched, memory for the data section is secured, and data with initial values is transferred from the disk image area.

Regarding the stack section, the stack size is determined from the header section, and a free space corresponding to the size is secured from the memory use management table T.

Regarding the text part, the text part in the disk image area is used when executing the task. Furthermore, referring to the tables in FIGS. 4(a) to (d) for each pixel,
The address and size of each risk component are stored in the task address management table T4 shown in Figure (E), and the task is made ready for execution. This process is repeated during task switching.

FIG. 3 explains the conventional method.

The disk image area 31 and the main memory 32 are functionally separated, and task components 33, 34, 36, 37, 3
8. The shared data 35 and 39 exist both in the disk image area and in the main memory, which is a major difference from the embodiment of the present invention.

〔Effect of the invention〕

According to the present invention, by providing a disk image area on the main memory that can be accessed directly from the CPU, memory efficiency is improved when executing tasks or using shared data in a magnetic diskless system, and when executing tasks or using shared data. This has the effect of minimizing memory transfer processing and reducing system overhead.

Furthermore, by making the read/write program for the disk image area similar in logic to the magnetic disk driver program, there is an effect of facilitating the porting of a system with a magnetic disk (O8) to a system without a magnetic disk (O8).

[Brief explanation of drawings]

Figure 1 is a schematic diagram of the RAM disk control method, Figure 2 is a system configuration diagram of the present invention, Figure 3 is an explanatory diagram of the conventional method, Figures 4 and 5 are related table diagrams, and Figure 6 is task/sharing. FIG. 7 is an explanatory diagram of the data registration process, and FIG. 7 is an explanatory diagram of the task activation process. 21... Central processing unit, 22... Main storage device, 23
...System bus, 24...I/O device, 25...
・Memory bus.

Claims (1)

[Claims] 1. In a computer system that has a CPU and a main storage device but does not have a magnetic disk, the program is provided on the main storage device that can be directly accessed from the CPU, and is similar to a magnetic disk driver program. a disk image area that can be read and written using a logic program; a management table that instructs the disk image area configuration of the disk image area; and a management table for managing tasks and shared data in the disk image area; A RAM disk control device having: