JPH0219938A - File control system - Google Patents

Abstract

PURPOSE:To control files at a high speed by using a continuous block control table to control the block length prescribed at initialization of a file and the number of blocks or preparing each idle block header group. CONSTITUTION:A continuous block control table 10 stores the idle block list headers 1 for each of sizes BL, 2BL... m X BL arranged in order of larger ones. At initialization of a file, a user area 12 is divided every integer multiple of the minimum block size BL and continuously. Thus the list structures are initialized for each size m X BL. When a host machine requests a writing action of the data having its length (x), a file system returns the address of an idle block. In other words, the size of a desired idle block is calculated as (x/BL) + 1. Thus a single access suffices to the files allocated to blocks and therefore the access speed is increased.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to a file management system for an information processing device.

[Conventional technology]

Figure 5 is, for example, Kyoritsu Shuppan, Haruhisa Ishida, rUNIXJ No. 7.
This is an explanatory diagram of the conventional free block file management system shown in the chapter "Internal structure of UNIX file system". In the figure, t indicates the free block that is the first list for managing free block groups in a list structure. List header, 2 is a list for holding addresses of up to N (N=50) free blocks, 3 is an area for storing the number F of addresses held in list 2, 4 is for storing the next list. 5 is an area to store the address of a free block. When a free block is created, there are N-F areas in which the address can be written, and T is the free block indicated by the area.

Next, the operation will be explained. At the time of file initialization, that is, formatting, in a UNIX system, the file system is stored in a management area 11 (super block 11a) as shown in FIG.

An i-list 11b, which is a collection of i-nodes, is provided, and the user area 12 is all empty blocks (fixed size 512 bytes), and the list structure shown in FIG. 5 is generated to manage the empty blocks.

The free block list header 1 shown in FIG. 5 is stored in this super block Ila, and each block in the user area 12 corresponds to a free block 7.

Next, we will explain the operation when the host machine requests block allocation when writing a file. When the host machine requests writing of data with a data length of It is carried out.

That is, the free program F required in step STI is calculated as the number of free blocks managed in the i-th list, and is calculated as the number of free blocks managed in the i-th list, and
is set in the i-th list.

Next, the host machine accesses the i-th list in step ST2, determines whether F-K>0 in step ST3, and when F-K>0, moves to step ST4.

In step ST4, the value of area 3 that stores the number F of addresses held in list 2 is updated, F is changed to F-, and in step ST5, the bottom of area 5 that stores addresses of empty blocks is updated. Notify the host machine of the free block addresses from the direction.

Further, in step ST3, if F-K>0 is not found, the process proceeds to step ST6, and the host machine is notified of F free block addresses from the bottom direction of area 5, and in step ST7, area 4 of the i-th list is The list indicating the pointer is set as the next first list, the value of the required number of blocks is updated in step ST8, K is set to -F, and the process returns to step ST2.

[Problem to be solved by the invention]

Conventional file management systems are configured as described above, so although the required number of blocks is secured, the block addresses are not consecutive, so multiple accesses are required when writing data. However, in order to read data written by this file management system, K addresses are required, which is a problem that makes it unsuitable for high-speed processing.

This invention was made to solve the above-mentioned problems, and enables high-speed processing by returning consecutive free blocks to the host machine so that they can be written or read with a single access. The purpose is to obtain a file management system.

[Means to solve the problem]

The file management system according to the present invention adds a continuous block management table for managing the block length and number of blocks specified by the file system when initializing a file, or provides a group of individual free block headers. .

[Effect]

In this invention, the continuous block management table or free block header group is created by the host machine at the time of file initialization.
It preformats accessed empty blocks while maintaining address continuity, and returns block addresses that satisfy the predetermined block length to the host machine.

〔Example〕

An embodiment of the present invention will be described below with reference to the drawings. 1st
In the figure, 1 is a free block list header indicating the first list for managing a group of free blocks of size (mXBL) in a list structure, and 2 is a list for holding the addresses of a maximum of N free blocks (size mXBL). ,3
is an area that stores the number F of addresses held in list 2, 4 is an area that stores a pointer indicating the next list (if the next list does not exist, φ is stored as an end mark), and 5 is an area that stores the pointer to the next list. Area 6 is used to store the address of a free block of size (mXBL).
L) is an area where the address can be written when a free block is created, I is a free block of size BL, and similarly 8 is a free block of size 2BL.

Further, the number 10 is added by this invention, and if the minimum block length is BL, then the size BL, 2
The m-th table in BL, 3BL, -.

In addition, 11 in Fig. 2 is the management area, and 11a is the area for management.
Block 11b is one list, and 12 is a user area.

Next, the operation will be explained. In the file initialization operation, the user area 12 shown in FIG.
It is initialized as a list structure for each size (mXBL) by dividing it into continuous areas that are integral multiples of the minimum block size BL.

Here, information on how many blocks are to be preformatted for each size is given as a preformatting parameter, and the sum of the values of area 3 on list 2 of free block list header 1 in Fig. 1 is given as a preformatting parameter. equals the value.

Next, the operation when the host machine requests block allocation when writing a file will be explained. When the host machine requests writing of data with a data length of X bytes, the file system allocates the block by returning the address of one free block (size mXBL) according to the flowchart in FIG.

That is, in step 5TII, the size of the required free block is calculated as (-i+x). Here, X is the data length and BL is the minimum block size.

Next, in step 5T12, list 2 indicated by free block list header 1 of the th block size KXBL in the continuous block management table 10 is set in the i-th list.

Next, in step 5T13, the host machine accesses the i-th list, and in step 5T14, the number of free blocks of size mXBL managed in the i-th list is F-
It is determined whether 1>O or not, and if F-1>O, the process moves to step 5T15.

In step 5T15, the value of the number of free blocks F stored in area 3 is updated, and in step 5T16, the host machine is notified of the address of one free block of size KXBL from the bottom of area 5.

Further, in step 5T14, the number of free blocks of size m x BL managed in the i-th list is F
If -1>O, the process moves to step 5T17, and the list indicated by the pointer stored in area 4 of the i-th list is set as the next first list.

With such block allocation, files can be written or read using the access shown in FIG.

In the above embodiment, the continuous block management table 1 is stored together with the free block list header group for each size.
0 is shown, but such a continuous block management table 10 is not provided, and the list structure is changed to the first node 13 with a free block list header group of size 1, as shown in FIG. The data is arranged in such a way that the second node 14 has a free block list header group of size 2, the third node 15 has a free block list header group of size 3, etc., and the free block list header group for each size is managed. However, the same effects as in the above embodiment can be achieved.

〔Effect of the invention〕

As described above, according to the present invention, when a file is initialized, a free block list header is stored in a continuous block management table for each size that is an integer multiple of the minimum block size, or for each node that is an integer multiple of the minimum block size. Since the configuration is configured to manage a free block list group for each block, a file to which a block has been allocated can be accessed only once, which has the effect of realizing high-speed access such as real-time processing.

[Brief explanation of the drawing]

FIG. 1 is an explanatory diagram of a file management system according to an embodiment of the present invention, FIG. 2 is an internal configuration diagram of a disk for explaining file initialization in the embodiment, and FIG. 3 is an illustration of the file management system in the embodiment. Flowchart showing the flow of block allocation operations when reserving files; FIG. 4 is a list structure in a file management system according to another embodiment of the present invention; FIG. 5 is an explanatory diagram of free block list header groups for each size; is an explanatory diagram of a conventional file management system, and FIG. 6 is an internal configuration diagram of a disk for explaining file initialization in a conventional file management system.
FIG. 7 is a flowchart showing the flow of block allocation operations when writing a file in a conventional file management system. 1 is a free block list header, 2 is a list, 3 to 6 are areas, T, 8 is a free block, 10 is a continuous block management table, 13 is a first node, 14 is a second node, and 15 is a third node. In addition, the same symbols in the figures indicate the same or equivalent parts. 512B Procedural amendment (voluntary)

Claims (1)

[Claims] Free block list header groups for each size that is an integral multiple of the minimum block length stored in the continuous block management table or individual free block list header groups, an area for storing the number of addresses, and a pointer indicating the next list. It has an area for storing and an area for storing addresses of multiple free blocks divided into continuous areas by integral multiples of the minimum block size, and is contiguous with the area for storing the addresses of the free blocks during file initialization. A file management system is provided with a list that holds free blocks and sets a number indicated by the free block list header group in an area for storing the addresses and notifies the host machine of the address when accessed by the host machine.