JPS62236049A - File constitution system - Google Patents

Abstract

PURPOSE:To attain succession of a multi-stage index system and an idle block unitary control system and furthermore to secure an access to a large capacity file at a sufficiently high speed, by performing the input and output of a variable length block. CONSTITUTION:The register processing is carried out by means of a multi-stag index system and an idle block unitary control system in case the file capacity is unknown. However such a file exceeding 10l bytes is decentralized to some blocks and stored there. In other words, a small block is used at first and then the blocks of a section B are used and registered via a point address #P1 if the number of blocks exceeds 10. If a file is not stored in the prescribed number of blocks of a relevant section in case the capacity of the file to be registered is known, the null data are set to the 1st entry of a section A indicating the head address and then execute a section-based block length increasing system. While the file is stored in the 1st ad following entries of the section B if the file is stored in the prescribed number of blocks.

Description

DETAILED DESCRIPTION OF THE INVENTION [Field of Industrial Application] The present invention relates to an operating system in a computer system, and particularly to a file configuration method for storing files in a disk storage medium.

[Conventional technology]

Conventionally, this type of file configuration method has been configured using a multi-stage index method and an empty block unified management method, and files are stored using this method.

For example, in a UNIX file system, one node of the control structure that represents the entity of a file contains information representing the file,
It consists of an area indicating the address of a data block or an indirect block consisting of 128 pointers. This address area indicates which blocks are connected to make up one file. In this case, if the file size is up to 108 blocks (within 512 bytes x 10 = 5 bytes), the
~+b10 is used as the address of the block.

If the size is 5X bytes or more, a method is used in which ÷P1 points to the block that contains the block address to the end block, and then points to individual blocks from that block. Each block address (ie, pointer) is represented by 4 bytes, so a 512-byte block can contain 128 pointers. Therefore, the size of the file capacity that can be indicated by ◆P1 is 128 x 512 bytes = 64 bytes, and a larger file is expressed by referring to P2 in two stages. Therefore, it is expressed by referring to two stages from NaP2. Therefore, the size of the file area that can be pointed to by ÷P2 is 1
28" x 512 bytes = 8192 bytes (8MB). Similarly, larger files can be created using φ
This is shown using a three-stage pointer table using P3. Therefore, 128'X512 pi) = 1024M high)
It can be expressed as 't' and is called a multi-stage index method. In this way, the multi-stage indexing method is a method for representing a group of blocks in which actual data is stored, and sb has a block address that directly points to the block belonging to the first partition, and the end block belonging to the second partition has a block address that directly points to the block belonging to the first partition. It is configured to have the address of the block containing the address group of that block.

Also, all these blocks have the same block input.
Because of the length of the block, empty blocks are managed in a unified manner by an empty block management block containing the address group and number of individual empty blocks, and is called the empty block unified management method.

[Problem that the invention seeks to solve]

The conventional file structure method described above uses a multi-stage index method and a unified empty block management method, so
When accessing a file with a small file size, the number of disk I10 times is small and the access itself is fast, but as the file size increases, the number of times it takes to follow the pointer increases, making file access extremely slow. There is a problem.

[Means for solving problems]

The file configuration method of the present invention is such that when the capacity of the file to be registered is known in advance, it is executed using a block length increase method for each section and an empty block management method for each section.

[Effect]

This invention can also sufficiently speed up access to large files.

〔Example〕

FIG. 1 is a flowchart showing an embodiment of a file configuration method according to the present invention. In the figure, 1 is a start step for accepting and accepting file registration requests;
2 is a file capacity determination step that determines whether the file capacity is known in advance; 3 is a file capacity determination step that determines whether the file will be stored in the number of blocks in the partition; and 4 is a file capacity determination step that determines whether the file capacity is known in advance. Null data registration step 5 sets null data in the first entry pointing to , as shown in FIG. m bytes (where m ) t ), and the length of the block constituting the secondary partition is increased.

a step of executing a so-called partition block length increasing method in which the block length constituting the partition is increased for each partition;
6 is a data registration step in which registration processing is performed using the multi-stage index method described above, and T is a termination step in which registration completion processing is performed.

Next, the operation of the file configuration method shown in the above flowchart will be explained with reference to FIGS. 2, 3, and 4. First, after a process of accepting and accepting a file registration request is performed in a start step 1, a file capacity determining step 2 determines whether or not the file capacity is known in advance.

In this case, if the file capacity is not known, the control moves to data registration step 69, and registration processing is performed using the multi-stage index method and empty block unified management method. When storing a file exceeding the block length (indicates the block length), the file is distributed and stored in each block as shown in FIG. That is, initially small blocks are used, 10 blocks (
10t bytes), the block of section B is used and registered using point address ◆Pi. Then, in the end step 7, registration completion processing is performed and the process ends. Next, if the capacity of the file to be registered is known in advance, the process moves to file capacity determination step 3, in which it is determined whether it will be stored in the number of blocks of the relevant partition, but if it is determined that it will not be stored. If so, proceed to null data registration step 4, and set the null data in the first entry of partition A pointing to the start address of the partition as shown in Figure 3, and then execute the block length increase method for each partition. In step 5, the section a is moved up by one, and the process returns to step 3 for determining the file capacity. Therefore, the file capacity determination step 3 is again performed to determine whether or not the data will be stored in the block number of the partition. If it is determined that the data will be stored, control is transferred to the data registration step 6, As shown in the figure, files can be stored from the first entry of the B partition. In this case, empty blocks are managed for each section. That is, as shown in FIG. 4, a partition-based empty block management method is used in which empty block management blocks are managed for each partition and are chained to the next management block according to the number of empty blocks. In addition, A, B, C,
The block lengths of each block in D are t, m, n, and O, respectively.
Byte (t<m<, n (0)), block addresses ◆b1 to φblo directly indicate blocks in section A, pointer address +P1 indirectly indicate blocks in section B, and in the same way, C, respectively. Refers to the block in section D. One block has a maximum of 12 block addresses.
8 pieces, 1281 pieces can be expressed in the C section, and 1281 pieces can be expressed in the D section.

Then, a registration completion process is performed at end step T, and the process ends. In this way, storage can be performed using the block length increasing method for each section and the empty block management method for each section shown in FIG.

Although the above embodiment has been explained as simply registering data, it is not limited to this, and can be used interactively with large-capacity files such as drawing data and parts data handled by a CAD/CAB system. Of course, it can be used in systems that require a fast response time.

〔Effect of the invention〕

As explained in detail above, 1. According to the file structure method according to the present invention, by performing variable-length block input/output, it is possible to inherit the multi-stage index method and the empty block unified management method, and moreover, it is possible to inherit the multi-stage index method and the empty block unified management method. This also has the effect of allowing access to be achieved sufficiently quickly.

[Brief explanation of drawings]

FIG. 1 is a flowchart showing an embodiment of the file configuration method according to the present invention, and FIGS. 2, 3, and 4 are diagrams for explaining the operation of FIG. 1, respectively. 1...Start step, 2φ...File capacity determination step, 3...File capacity determination step, 4.
-... Null data registration step, No. 5... Section block length increase step, 6... φ data registration step,
7 fists...end step.

Claims (1)

[Claims] The data area of the storage device in a computer system is divided into multiple sections, and when the file capacity is not known, the multi-stage indexing method and the unified empty block management method are used to execute the process, and the capacity of the file to be registered is known in advance. A file configuration method characterized in that, when a file is stored, a block length increase method for each partition and an empty block management method for each partition are used to store the substance of the file.