JPH10283230A - File data storage device and machine-readable recording medium with program recorded - Google Patents

Abstract

PROBLEM TO BE SOLVED: To operate a secondary storage device in the state that plural block sizes exist together by dividing the secondary storage device into blocks of plural sizes. SOLUTION: A size-classified idle area reservation means 102 and an idle area management information construction means 103 reserve idle blocks on a secondary storage device 200 in accordance with a construction indication designating plural block sizes and generate idle area management information for idle block management by block sizes and record these management information in the secondary storage device 200. If an idle block is required to write data of a file or the like at the time of operation, a block size determination means 106 determines a required block size in accordance with data to be written, and an idle block extraction means 107 extracts an idle block having this determined block size from idle area management information, and a block assigning means 108 make the extracted block belong to the designated file.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a file data storage device for forming a logical file on a secondary storage device and storing data, and more particularly to a file data storage device capable of operating with a plurality of block sizes and saving storage area waste. The present invention relates to a file data storage device.

[0002]

2. Description of the Related Art There are various file data storage mechanisms such as a UNIX file system as a mechanism for making one secondary storage device appear as a plurality of virtual storage devices (logical files or simply files). In addition, UN
Documents relating to the IX file system include, for example, "Design of UNIX Kernel", Kyoritsu Publishing, and
"4.3 BSD Design and Implementation" Maruzen and "UNIX Int"
ernals: The New Frontiers "
Prince-Hall and the like.

In these file data storage mechanisms, the storage area on the secondary storage device is partitioned into fixed-size sections called blocks, and each block is allocated as a storage area for each file, thereby logically storing a plurality of files. Implemented on one secondary storage device.

If the allocation of each block to a file is fixedly determined, the capacity of each file is fixedly limited. Therefore, most file data storage mechanisms dynamically allocate each block to a file. As a result, there is a degree of freedom in restricting the capacity of each file, and the sum of the capacities of all files is limited to the total capacity.

In order to dynamically allocate each block to a file, in a UNIX file system, blocks not allocated to a file are collectively managed as empty blocks.

When storing data in a certain file, a required number of empty blocks are acquired, and these blocks are assigned as belonging to the file. As a result, only the capacity necessary for storing data is allocated to the file.

Conversely, when the data stored in a certain file is no longer needed, not only the data is erased, but also the blocks allocated as belonging to the file are released, and Return to batch management.

As described above, the capacity of the secondary storage device can be dynamically allocated to each file.

However, in the conventional file data storage mechanism, the size of a block can only be fixed to a fixed value. The problem due to this will be described below.

[0010] To increase the amount of data that can be stored in one file, there are a method of increasing the number of blocks allocated to one file and a method of increasing the size of the block itself.

First, consider a case where the number of blocks allocated to one file is increased. In this case, management data (file management data) for recording which block is allocated to each file and the like increase, and reference to file management information increases when accessing data in the file. For this reason, there is a problem that the time to access the data in the file becomes longer.

When the file management information increases,
Since this file management information is also recorded on the secondary storage device, the capacity of the secondary storage device for recording actual data is reduced accordingly.

Next, consider the case where the block size is increased. When accessing data of a large file, reference to management information is small, and I / O to the secondary storage device is performed.
Since the number of times is also reduced, the access time is reduced.

However, since small data such as small file data and file management information is stored in a block of the same size as compared to the block size, valid data is stored only in a part of the block. There is a problem that the area is wasted.

[0015]

A first problem with the prior art is that when the amount of data that can be stored in a file is increased by increasing the number of blocks, the file management information increases, The access time to the data was long.

Further, in this case, there is a problem that the capacity occupied by the file management information in the secondary storage device increases, and the actual amount of data that can be stored in the secondary storage device decreases accordingly.

A second problem with the conventional technique is that when the amount of data that can be stored in a file is increased by increasing the block size, small data is also stored in the same large size block. Only a part is effectively used for storage, and most of the storage area is wasted.

Therefore, a first object of the present invention is to prepare a plurality of block sizes and use different block sizes according to the data to be written, thereby realizing high-speed access to a huge file and realizing high-speed access to a large file. An object of the present invention is to enable a storage area to be used effectively.

A second object of the present invention is to make it possible to dynamically change the ratio of the storage area occupying the secondary storage device for a plurality of block sizes, so that the entire storage area can be more effectively used. Is to make it possible.

[0020]

According to the present invention, in order to achieve the first object, a file data storage device comprising a plurality of logical files on a secondary storage device is provided. Free space management that secures free blocks for each block size in the secondary storage device and manages information on free blocks for each block size in accordance with a construction instruction that specifies a plurality of block sizes to the input / output control unit A construction unit that creates information and records the information in the secondary storage device, a block size determination unit that determines a required block size according to data to be written in the secondary storage device, and a block size determination unit that determines the block size. A free block extracting means for extracting information on a free block having a block size from the free area management information, and a block extracted by the free block extracting means. Block allocating means for assigning a block to a specified file, block releasing means for removing a block allocated to the file from the file affiliation, and information of the block released by the block releasing means to the block. Empty block registration means for registering in the empty area management information corresponding to the block size of

In the file data storage device configured as described above, when a construction instruction designating a plurality of block sizes is given, the construction unit according to the construction instruction,
An empty block for each block size is secured in the secondary storage device, and free area management information for managing information (for example, a block number) of the empty block for each block size is created and recorded in the secondary storage device. Thereafter, the actual operation is started, and when an empty block is required to record data of a certain file, file management information, and the like, the block size determining means determines the necessary size according to the data to be written on the secondary storage device. The block size is determined, the free block extracting means extracts free block information of the determined block size from the free area management information, and the block allocating means determines the extracted block as belonging to the specified file. I do. When a block belonging to a certain file becomes unnecessary, the block releasing means removes the block assigned to the file from the file belonging to the file, and the free block registration means removes the block of the released block. The information is registered in the free space management information corresponding to the block size of the block.

In order to achieve the second object, the present invention further comprises at least one of the following empty block dividing means and block combining means in the input / output control section.

A free block dividing means for extracting a free block having a block size other than the minimum size among a plurality of block sizes, and dividing the extracted block into free blocks having a block size smaller than the block size. Block combining means for collecting a plurality of empty blocks of a block size other than the maximum size among a plurality of block sizes and combining them into one block having a larger block size.

If the number of blocks for each block size is fixed, the use concentrates on a specific block size depending on the actual operation state, and there is a possibility that empty blocks of that block size are exhausted. In such a case, in a configuration including an empty block dividing unit, an empty block having a block size other than the minimum size is extracted from a plurality of block sizes, and the extracted block is extracted as an empty block having a block size smaller than the block size. Since the blocks are divided into blocks, the surplus large-sized empty blocks can be diverted to the small-sized empty blocks, and the depletion of the small-sized empty blocks can be prevented. On the other hand, in the configuration including the block combining means, a plurality of free blocks having a block size other than the maximum size among a plurality of block sizes are collected and combined into one block having a larger block size. It is possible to divert the used small-sized empty blocks to large-sized empty blocks, thereby preventing the large-sized empty blocks from being depleted.

In another aspect of the present invention, the input / output control unit includes a block size-based occupancy rate managing means for managing an occupation state for each block size, and the empty block dividing means includes a step of dividing an empty block. If the occupation status for each block size does not satisfy the predetermined condition due to
The block combining means does not divide empty blocks, and the block combining means does not combine empty blocks when the occupation status for each block size does not satisfy a predetermined condition due to the combination of empty blocks. Have.
As a result, the occupancy of the storage area for each block size can be maintained within a desired range, and the division of the large-sized blocks is repeated endlessly, so that the large-sized blocks are depleted or vice versa. In addition, since the joining of the small-sized blocks is repeated endlessly, it is possible to prevent the small-sized blocks from being exhausted.

[0026]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Next, embodiments of the present invention will be described in detail with reference to the drawings.

FIG. 1 is a block diagram of an embodiment of the file data storage device of the present invention. The file data storage device of this example includes a secondary storage device 200 constituted by a magnetic disk device or the like, an input / output control unit 100 for controlling the same,
And a recording medium 300.

In this embodiment, the input / output control unit 100
Construction instruction analysis means 101, size-specific free area securing means 102, free area management information construction means 103, file writing / reading means 104, cache 105
, A block size determining unit 106, an empty block extracting unit 107, a block allocating unit 108, a block releasing unit 109, and an empty block registering unit 110.

The input / output control unit 100 can be realized by a data processing device including, for example, a memory and an arithmetic processing unit.

The recording medium 300 is a recording medium such as a magnetic disk, a semiconductor memory, or the like. The program recorded in the recording medium 300 is read by a data processing device constituting the input / output control unit 100 and controls the operation of the data processing device. The data processing device is caused to function as each of the above units.

Hereinafter, the function of each part of the present embodiment will be described through its entire operation.

Generally, a user who manages the operation of the file data storage device is called an administrator, and is distinguished from a general user. The administrator makes the file data storage device available to general users and monitors and manages how much the storage area of the file data storage device is used during operation.

In order to use the secondary storage device 200 for storing file data, the administrator must use the secondary storage device 200.
Build a file data storage mechanism above. At this time, a configuration such as a size of the block is determined. In the present invention, since a plurality of block sizes can be used, the administrator determines a plurality of types of block sizes to be used. Although a plurality of types of block sizes to be used are arbitrary, in the present embodiment, block sizes other than the minimum block size are determined so as to be integral multiples of the minimum size.

When the manager determines a plurality of types of block sizes to be used, the administrator inputs a construction instruction designating those block sizes to the input / output control unit 100 from an input device such as a keyboard (not shown). The construction instruction is, for example, UNI
It is implemented in the form of a command in X. In this case, the designation of the block size can be given by a command parameter.

The input construction instruction is analyzed by the construction instruction analyzing means 101, and the information on the block size is transmitted to the size-specific free area securing means 102. The size-specific free area securing means 102 and the subsequent free area management information constructing means 103 secure free blocks for each block size in the secondary storage device 200 and manage free blocks for each block size. A construction unit that creates management information and records it in the secondary storage device 200 is configured.

First, means for securing free area by size 102
Secures a free area (free capacity) for each block size on the secondary storage device 200. FIG. 2 shows an example of the processing.

The size-specific free area securing means 102 extracts the maximum block size from the plurality of block sizes passed from the construction instruction analyzing means 101 (S1), and uses this maximum block size on the secondary storage device 200. The storage area is partitioned, and a unique number is assigned to each block (S2). Next, the ratio of the free area to be allocated to each block size is determined, and the number of blocks to be allocated for each block size is calculated from the total number of blocks formed in step S2 (S3). Here, in addition to the method of statically determining the ratio of the free area to be allocated to each block size, there is a method of specifying together with the block size as one parameter of the construction instruction.

For example, there are three types of block sizes specified by the construction instruction, A, B, and C, where A is the largest size, B is the next largest, and C is the smallest size, and assigned to each block size. A: B:
When C = 4: 5: 3, the storage area of the secondary storage device 200 is partitioned by the block size A. At this time, for example, if a total of 12 blocks are formed as shown in FIG. ..,..., In order from the block at the start address (0) of the secondary storage device 200.
A block number continuous with 11 is assigned, and the number of blocks shown in FIG. It is to be noted that, by allocating such a continuous number from 0, by multiplying the block number by the block size A, the head address of the block is obtained, and each block can be uniquely identified by the block number.

Next, the free area management information construction means 103
The block of the maximum size formed by the size-specific free area securing means 102 is allocated to each block size according to the calculated number of allocated blocks for each block size, and allocated to block sizes other than the maximum size. The block is divided into blocks each having the size of the block size and a unique block number is allocated. Then, free space management information for managing free blocks is created for each block size and recorded in the secondary storage device 200. FIG. 4 shows an example of the processing.

The free area management information construction means 103 puts all the block numbers of the blocks of the maximum size formed by the size-specific free area securing means 102 into the set U (S1).
1) First, focus on the largest block size (S1).
2). Then, block numbers corresponding to the number of allocations to the maximum block size determined by the size-specific free area securing means 102 are extracted from the set U (S13), and free area management information for the maximum block size is created. The information is recorded in the next storage device 200 (S14).

Next, it is checked whether or not an unprocessed block size remains (S15). If it remains, attention is paid to the next largest block size (S16), and each block included in the set U is checked. The block having the number is divided into N, the numbers of the individual blocks formed by the division are calculated, the set U is once emptied, and all the block numbers calculated this time are stored in the set U (S17). Here, N is a value represented by the following equation. N = block size of interest immediately before / block size of interest now ... (1)

Then, the free area securing means 102 for each size
The number of block numbers corresponding to the number of allocations to the block size determined in (1) is extracted from the set U (S13),
The free space management information for the block size is created and recorded in the secondary storage device 200 (S14).

On the other hand, if it is determined in step S15 that there is no unprocessed block size, the process ends.

The operation of the free area management information construction means 103 will be described below with reference to FIG.

First, block numbers 0 to 11 are put into a set U (S11). Pay attention to the maximum block size A (S1
2) Extract four block numbers from the set U according to the number of blocks "4" to be allocated to the blocks (S1).
3). The block numbers may be taken out sequentially from the beginning or may be taken out at random. In this case, it is assumed that block numbers 0, 1, 2, and 3 have been extracted. Next, free area management information for the maximum block size A is created and recorded in the secondary storage device 200 (S14). Here, the free space management information created for the maximum block size A is stored in any one or a plurality of blocks allocated for the block size A.

Next, attention is paid to the block size B (S
16), N = block size A / block size B = 2
Then, the blocks of block numbers 4 to 11 included in the set U are divided into two equal parts as shown in FIG. 3C, and the individual block numbers are calculated as follows and stored in the set U ( S17). Block number after division = Block number before division × N + n (2) where n = 0,..., (N−1)

Accordingly, the numbers of the respective blocks after division are shown in FIG.
As shown in (c). Here, when the divided block number is multiplied by the block size B, the starting address of the block is obtained, and thus each block number is information for uniquely identifying each block.

Then, the free area securing means 102 for each size
According to the allocated block “5” of the block size B determined in the above, the maximum number of blocks is 5,
The size of the block size B is 5 × 2, that is, 1
Zero block numbers are extracted from the set U (S13).
This extraction may be in order from the beginning or may be random. Here, it is assumed that ten block numbers 8 to 17 have been extracted. Next, free area management information for block size B is created and recorded in the secondary storage device 200 (S14). The free space management information created for the block size B is also stored in one or a plurality of blocks allocated for the block size B.

Next, attention is paid to the block size C (S
16), N = block size B / block size C = 2
Then, each block number 18 to 23 included in the set U
Is divided into two equal parts as shown in FIG. 3D, the individual block numbers are calculated by the above equation (2), and stored in the set U (S17). Accordingly, the numbers of the respective blocks after the division are as shown in FIG. Here, when the divided block number is multiplied by the block size C, the starting address of the block is obtained, and thus each block number is information for uniquely identifying each block.

Then, a free area securing means 102 for each size
According to the allocation block “3” of the block size C determined in the above, the maximum number of blocks is 3,
3 × 2 × 2, that is, 12 block numbers are extracted from the set U as the size of the block size C (S1).
3). In this case, all the block numbers 36 to 47 stored in the set U are extracted. Next, free space management information for the block size C is created, and the secondary storage device 20 is created.
0 is recorded (S14). The free space management information created for the block size C is also stored in one or more blocks allocated for the block size C.

FIG. 5 is a schematic diagram of the operation of a specific example of the above-mentioned size-specific free area securing means 102 and free area management information construction means 103. In FIG. 5, reference numeral 1021 denotes a set of block numbers assigned to the maximum block size by the size-based free area securing unit 102;
Reference numerals 1, 1032 and 1033 denote free area management information construction means 1
03 indicates a set of block numbers of the maximum block size allocated to each of the block sizes A, B, and C. 10
34 is free space management information for block size A;
For example, it is stored in the block of block number 0 on the secondary storage device 200. 1035 is a set of block numbers of individual blocks obtained by dividing the block of the maximum size allocated to the block size B into the block size B, and 1036 is free space management information for the block size B, For example, it is stored in the block of block number 8 on the storage device 200. Reference numeral 1037 denotes a set of block numbers of individual blocks obtained by dividing a block having the maximum size allocated to the block size C into block sizes C. Reference numeral 1038 denotes free space management information for the block size C. For example, the block number 3 on the storage device 200
It is stored across 6, 37 blocks.

FIG. 6 shows another example of allocating blocks to the respective block sizes A, B, and C. This example shows an example in which blocks of respective block sizes A, B, and C are scattered and allocated in the secondary storage device 200.

FIG. 7 is a diagram showing an example of the configuration of the free area management information for each block size and an example of its storage in the secondary storage device. The free area management information in this example is a list of block numbers of free blocks. If the number of free blocks is large, the list length becomes proportionally longer. As described above, the free area management information is also recorded in the blocks on the secondary storage device 200, so the free area management information having a long list length is recorded over a plurality of blocks. In this case, the list is divided into a plurality of list parts and recorded in each block. At this time, the next list part is recorded at the head of the list part stored in each block, and the list itself has a chain structure. The number of the block that stores the first list part is recorded and managed, for example, in a specific block that is not to be allocated or released. One such free space management information is prepared for each block size as described above, and is recorded in the secondary storage device 200 using one or a plurality of blocks of each block size.

The above is the operation when the file data storage mechanism is constructed. Hereinafter, the functions and operations of each unit during operation will be described.

When the file data storage mechanism is constructed, all the blocks on the secondary storage device 200 are free areas except for the block for storing the management information. From this state, the user writes data to a certain file.

File writing / reading means 10 of FIG.
Reference numeral 4 denotes a portion that controls writing and reading of a file in accordance with a request from a higher-level device (for example, a central processing unit). When writing data to a certain file, an empty block for storing management information of the file and actual data is required. File writing and reading means 10
When a free block is needed, the block 4 gives the block size determining means 106 a request for a free block by passing information such as a file name using the free block.

The block size determining means 106 determines a required block size according to data to be written on the secondary storage device 200. Several implementations can be considered as criteria for determining the block size.

One example is a method in which the block size is fixedly determined according to the type of data to be written. The data to be written to the secondary storage device 200 includes management information such as file management information in addition to the data requested by the user to write. It is enough. When such a method is performed, the file writing / reading means 10
No. 4 notifies the block size determining means 106 of the type of data to be written to the block. The block size determining means 106 checks the type of the data and determines which block size to use accordingly.

Another example is a method of determining a block size according to the size of data to be written. Since the data size of file management information is almost fixed, the above method is also effective. However, since the write data requested by the user has a large variation in size,
Not valid. Such a problem can be solved by determining the block size according to the size of the data to be written. When this method is performed, when a free block is requested, the file writing / reading unit 104 notifies the block size determining unit 106 of the size of data to be written in the block. The block size determining means 106 selects a block size with no excess or shortage from a plurality of prepared block sizes.

When the block size is determined by the block size determining means 106, the block size is notified to the empty block extracting means 107 together with the file name passed from the file writing / reading means 104.
The empty block extracting unit 107 extracts one empty block of the notified block size, and the subsequent block allocating unit 108 assigns the extracted empty block to the file belonging to the notified file name, and Is returned to the file writing / reading means 104. The file writing / reading unit 104 writes data to an empty block to which the file belongs.

Next, the operation of the free block extracting means 107 and the block allocating means 108 will be described. The input / output control unit 100 of this embodiment has a cache 105 for speeding up the processing. Cache 105
First, the cache 105 will be described because a part of the management information is cached.

FIG. 8 shows an example of information stored in the cache 105. In this example, the secondary storage device 200 is used with two types of block sizes, a large block and a small block. Therefore, there are two types of free area management information, one for a large block and one for a small block. In the cache 105, a part of the respective free area management information (the head part of the list)
Is cached. Similarly, if two files α and β are generated at this time, the management information of each file is recorded in the secondary storage device 200. The part is cached.

Now, a processing example of the empty block extracting means 107 will be described. FIG. 9 shows an example of the processing. When the block size is designated by the block size determining means 106, the empty block extracting means 107, if the head list part of the free area management information of the designated block size is not in the cache 105, deletes the list part. Read from the secondary storage device 200 onto the cache 105 (S2
1, S22), and proceeds to step S23. If the first list part already exists on the cache 105, the process proceeds to step S
Skip to step S22 and proceed to step S23. If there is no free area management information of the designated block size, there is no free block of that block size. To cope with this case, for example, a method of substituting an empty block of another size can be adopted.

Next, whether or not the first list portion is empty, that is, whether or not there is only one entry storing the block number of a block whose list portion includes the next list portion and there is no entry storing the number of an empty block Is determined (S2
3) If not empty, set the block number in the list to 1
Then, the entry is cleared, the extracted block number is notified to the block allocating means 108 as an empty block number (S24), and the process is terminated.

On the other hand, if the list portion is empty, the block number of the block on the secondary storage device 200 in which the list portion has been written is temporarily recorded therein (S2).
5) If there is a next list part, it is stored in the secondary storage device 20.
The block number is read from 0 to the cache 105 (S26), and the recorded block number is notified to the block allocating unit 108 as an empty block number (S27), and the process is terminated. FIG. 10 is a schematic diagram of the operation at this time. As described above, since the free area management information for a certain block size is stored in a block of that block size,
When a block storing the free space management information becomes free, it is used as a free block.

Next, the operation of the block allocating means 108 will be described. FIG. 11 shows an example of the processing. When the block number and the file to which the block belongs are designated by the free block extracting means 107, the block allocating means 108, if the management information of the designated file is not in the cache 105, the management information of the file is secondarily stored. Read from the storage device 200 onto the cache 105 (S3
1, S32), and proceeds to step S33. If the management information of the file already exists on the cache 105, the process skips step S32 and proceeds to step S33. In step S33, the number of the block to be assigned is recorded in the management information of the file on the cache in such a manner that the relationship with the block size is understood (S33). Here, as an example of a method of recording in a form in which the relationship with the block size is changed, a method of recording the block size together with the block number, and a case where the block number is divided for each block size in the file management information. And a method of recording a block number in a corresponding place.

As described above, the file writing / reading means 104 obtains the block number of an empty block and calculates the secondary storage device 2 from the block number.
Data is written to an empty block existing at the address above 00. The block to which data has been written belongs to the management of the file as described above, and the management information of the file indicates which data of the file is written in which block of which block size. Since the data is managed, the data can be read.

Next, the operation when the file data is no longer necessary for the user will be described. In this case, in addition to erasing the data in the corresponding block, the empty block in which the data becomes empty
00 must be managed as an empty area. Therefore, the corresponding block is removed from the file management by the block release unit 109, and the empty block is returned to the management as an empty area on the secondary storage device 200 by the empty block registration unit 110. As a result, the block from which data has been erased is managed as an empty block, and can be reused when another data is written. Hereinafter, block release and free block registration will be described.

When erasing the file data, the file writing / reading means 104 specifies the block number of the block containing the corresponding data, the name of the file to which the data belongs, and the block size to the block releasing means 109, and requests its release. The block releasing unit 109 performs a process of releasing the block in response to this.

FIG. 12 shows an example of the processing of the block release means 109. If the management information of the file specified by the file writing / reading means 104 is not in the cache 105, the block release means 109 reads the management information of the file from the secondary storage device 200 into the cache 105 (S41, S42), Proceed to step S43. If the management information of the file already exists on the cache 105, the process skips step S42 and proceeds to step S43. In step S43, the cache 105
The block number of the block to be released is cleared from the management information of the above file, and the free block registration unit 110 is notified of the block number and the block size.

The blocks released by the block releasing means 109 are output by the empty block registering means 108.
It is returned under collective management as an empty block. FIG. 13 shows a processing example of the empty block registration means 110. The free block registration unit 110 stores the list portion of the free area management information for the block size of the released block in the secondary storage device 2 if the first list portion does not exist in the cache 105.
00 to the cache 105 (S51, S5
2), proceed to step S53. If the first list part already exists on the cache 105, the process skips step S52 and proceeds to step S53.

Next, it is determined whether or not the first list portion is full, that is, whether or not all the entries of the list portion have been used and there is no entry for registering a new block number in the block storing the list portion. If it is not full (S53), the block number of the released block is registered in the entry in the list (S54).

On the other hand, when the list part on the cache is full, the list part is written to the secondary storage device 200 at the position of the released block (S55), the list part on the cache is cleared, and the written part is written. Generates an empty list part on the cache that sets the block number of the block storing the list part at the beginning of the list part,
The block number of the released block is registered in the entry in the list (S56). FIG. 14 is a schematic diagram of the operation at this time. As described above, since each list portion of the free area management information for a certain block size is stored in a block of that block size, when the list portion on the cache becomes full, it is used as an empty block for storing it. The block that was released this time is used.

When there is no free area management information of the designated block size, it means that no free block of the block size currently exists. At this time, the current release block is used as a block for storing the head list portion of the free area management information.

FIG. 15 is a block diagram of another embodiment of the file data storage device of the present invention. The difference between the file data storage device of this embodiment and the file data storage device of the embodiment of FIG. 1 is that the input / output control unit 100 is provided with a free block dividing unit 111 and a block size occupancy management unit 112. It is in. As in the previous embodiment, the input / output control unit 100 can be realized by a data processing device including, for example, a memory and an arithmetic processing unit.
0, which is read by the data processing device constituting 0, controls the operation of the data processing device, and causes the data processing device to function as each unit.

In the file data storage device of the embodiment shown in FIG. 1, the ratio of the area occupied by each block size in the secondary storage device 200 is fixedly determined at the time of constructing the data structure. If there are not enough free blocks of the small block size, no matter how many free blocks of the larger block size remain, it is impossible to write more data to the secondary storage device 200.
An empty block with a large block size must be used instead.

To avoid this, it is necessary to carefully plan what percentage of area should be reserved for each block size when constructing the data structure, based on the expected usage of the secondary storage device 200. There is a need to. For this reason, it is not possible to cope with various use methods in which the assumed use state cannot be determined.

Therefore, in the present embodiment, when a small block runs short, a large block is divided and used, and the ratio for each block size is flexibly changed. Hereinafter, the operation of this embodiment will be described focusing on the differences from the embodiment of FIG.

Occupancy rate management means 112 for each block size
Manages the occupation status for each block size. This is managed, for example, by holding the total number of blocks for each block size. The initial value of the total number of blocks of each block size is the total number of free blocks for each block size at the time of constructing the data structure. Thereafter, the block size-based occupancy management unit 112 increases / decreases the total number of blocks for each block size according to the notification from the free block dividing unit 111, and always manages the latest occupation status.

In the case of the present embodiment, the empty block dividing means 111 refers to the empty area management information periodically using the idle time of the input / output control unit 100, etc. Whether the number of empty blocks of the block size is equal to or less than a predetermined number,
That is, it is checked whether or not there is a shortage. If there is a shortage, an empty block having a larger block size is divided, and an empty block having the insufficient block size is newly generated.

FIG. 16 shows a processing example of the empty block dividing means 111. First, the missing block size is 1
Block size is selected (S6
1). Next, referring to the free area management information for the selected block size, it is checked whether or not the number of free blocks is equal to or greater than a predetermined number, that is, whether or not a sufficient number remains (S62). . If a sufficient number is not left, if it is used, then the number of free blocks of the block size will be insufficient, and the process is abandoned.

When there is a sufficient number of free blocks of one large block size, the current occupation status held by the block size-based occupancy management means 112 is referred to, and the free blocks are divided to separate each block size. It is determined whether or not the occupation condition does not satisfy a predetermined condition (S63). If not, the process is abandoned.

If there is a sufficient number of empty blocks of one large block size and the occupation status of each block size satisfies a predetermined condition even if the empty blocks are divided, the empty block dividing means 111 The empty block extraction unit 107 is requested to extract an empty block by designating the block size (S64). At this time, the operation of the empty block extracting unit 107 is such that the request source is the empty block dividing unit 11 instead of the block size determining unit 106.
1; instead of notifying the extracted block number to the block allocating means 108, the empty block dividing means 11
1 is substantially the same as that of FIG.

Next, based on the block number notified from the free block extracting means 107, the free block dividing means 111 calculates the block number of each block when the block of the block number is divided by the insufficient block size, for example, It is calculated by an equation (S65). Block number after division = block number before division × N + n (4) where N = (size of block before division) / (size of block after division n = 0,..., (N−1) (4) is the same as the expression (2) used by the free area management information construction means 103 when constructing the data structure.

Next, the free block dividing means 111 stores the calculated N block numbers in the free block registering means 1.
10 to request registration of the missing block size in the free area management information (S66). The operation of the free block registration means 110 at this time is substantially the same as that of FIG. 13 except that the request source is the free block division means 111 instead of the block release means 109. Then, the free block dividing means 111 instructs the block size-based occupancy rate managing means 112 to reduce the number of blocks of the divided block size by one and increase the number of blocks of the divided block size by N. (S67).

FIG. 17 schematically shows the operation of the empty block dividing means 111 as described above.

In this embodiment, the following various modifications are possible.

If the number of empty blocks one size higher than the insufficient block size is not sufficient, or if the occupation status for each block size does not satisfy the predetermined condition, the processing is not abandoned and an empty space of a size higher than the shortest block size is not given. The number of blocks and the occupation status by block size when the blocks are divided are examined. If the blocks can be divided, the blocks are divided and applied.

When there are a plurality of sizes larger than the insufficient block size, a block size having a larger total number of blocks is preferentially set as a division target.

When the empty block extracting means 107 attempts to extract an empty block of the corresponding block size in response to a request from the block size determining means 106, if there is no empty block of the block size, the empty block The empty block dividing unit 111 is started by designating the block size from the extracting unit 107, and the empty block extracting unit 107 extracts the empty block generated by the empty block dividing unit 111.

FIG. 18 is a block diagram of still another embodiment of the file data storage device of the present invention. The difference between the file data storage device of this embodiment and the file data storage device of the embodiment of FIG. 15 is that a block combining means 113 and a continuous block detection means 114 are added to the input / output control unit 100. As in the previous embodiment, the input / output control unit 100 can be realized by a data processing device including, for example, a memory and an arithmetic processing device.
The program recorded in “0” is read by the data processing device that constitutes the input / output control unit 100, controls the operation of the data processing device, and causes the data processing device to function as each unit.

In the file data storage device of the embodiment shown in FIG. 15, the depletion of small blocks could be prevented by dividing large blocks and using them as a plurality of small blocks. However, the opposite, depletion of large blocks cannot be prevented.

Therefore, in this embodiment, before empty blocks of a large size run out, empty blocks of a small size are combined to generate an empty block of a large size.
At this time, simply extracting a plurality of small-sized empty blocks at random cannot be combined into one large-sized block. This is because, in order to combine small empty blocks as large empty blocks, there is a necessary number of consecutive small empty blocks, and these small empty blocks must be accommodated in one large block. Therefore, the continuous block detection means 114 detects whether or not the free blocks are continuous for each block size other than the maximum size block size so as to be combined with the larger size block. Then, the block connecting means 113 connects them. Hereinafter, the operation of this embodiment will be described focusing on the differences from the embodiment of FIG.

In the case of the present embodiment, the block combining means 113 refers to the free area management information periodically using the idle time of the input / output control section 100 and the like, and selects a block size other than the minimum size among a plurality of block sizes. Check whether the number of free blocks of the block size is equal to or less than a predetermined number, that is, whether or not the number of free blocks is insufficient. Merge to generate a new free block of the missing block size.

FIG. 19 shows a processing example of the block combining means 113. First, a block size one size smaller than the missing block size is selected (S71). Next, it is determined whether or not the number of free blocks is equal to or greater than a predetermined number, that is, whether or not a sufficient number of remaining free blocks are left by referring to the free area management information for the selected block size (S72). If you don't have enough left,
If this is used, the number of free blocks of the block size will be insufficient this time, so the process is abandoned.

If there is a sufficient number of free blocks of one small block size, the current occupation status held by the occupancy management means 112 for each block size is referred to, and by combining the free blocks, the occupancy of each block size is determined. It is determined whether or not the occupation condition does not satisfy a predetermined condition (S73). If not, the process is abandoned.

If there is a sufficient number of free blocks of one small block size and the occupation status of each block size satisfies a predetermined condition even when the free blocks are combined, the block combining means 113 Insufficient block size) / (the corresponding block size) = N
Detection is performed using the continuous block detection unit 114 (S7).
4).

FIG. 20 shows a processing example of the continuous block detecting means 114. First, by referring to the free area management information of the block size present in the cache 105 and the secondary storage device 200, it is checked whether or not there are N consecutive free blocks starting with a block having a block number that is just divisible by the value of N. (S81). Here, N is the above (the block size one level above the block size) /
(The block size). If such continuous empty blocks exist (S82), those block numbers are removed from the empty area management information (S83). As a result, at the block size, the block with that block number is no longer an empty block. Then, the block numbers are notified to the block combining means 113 (S
84). On the other hand, if there is no continuous free block (S82), the detection failure is notified to the block combining means 113 (S85).

If the detection of N consecutive free blocks fails, the block combining means 113 gives up the processing (S75). On the other hand, if the detection is successful, the number of the combined block is calculated from the detected N block numbers by, for example, the following equation (S76). Block number after combining = [any block number before combining / N] (5) where N is (the block size one higher than the block size) / (the block size) [] is the division result Means taking the integer part of. This equation (5) is an equation for finding a relationship opposite to the above-mentioned equation (2) used by the free area management information construction means 103 when constructing the data structure.

The block combining means 113 designates the calculated block number and designates the empty block registration means 1
10 is requested to register (S77). The operation of the empty block registration unit 110 at this time is substantially the same as that of FIG. 13 except that the request source is the block combining unit 113 instead of the block releasing unit 109. Thereby, a plurality of small empty blocks can be used as one large empty block. Then, the block combining unit 113 instructs the block size-based occupancy management unit 112 to reduce the number of blocks of the combining source block size by N and increase the number of blocks of the block size created by combining by one. (S78).

Next, a description will be given of an example of the configuration of the free area management information so that the continuous block detecting means 114 can easily detect the continuity.

Even if the free area management information is a simple list of free block numbers as described with reference to FIG. 7, it is possible to detect continuous free blocks that can be combined into a larger size, but the processing takes time. . Therefore,
In this example, the free area management information other than for the maximum block size has a matrix structure as shown in FIG.

In each column of the matrix, block numbers having the same value of [(block number) / N] are arranged. However,
N = (upper block size) / (the relevant block size). In each row of the matrix, block numbers having the same value of (block number) mod N are arranged. Here, mod is an operator for obtaining a remainder obtained by dividing a left term by a right term.

In the case where the free space management information has the above-described matrix structure, if one column of the matrix is satisfied, a block having a plurality of block numbers included in the column corresponds to one block size higher by one. The continuous block detecting means 114 may detect such a column.

When the free area management information has the above-described matrix structure, the free block registration means 11
In the case of 0, when registering a certain block number, the value of [(block number) / N] is obtained, a column having an equal value is searched, and a column to be inserted is determined. At this time, if there is no column having the same value, an empty column is inserted. Further, the value of (block number) mod N is obtained, and the column to be inserted is determined.

In the present embodiment, the following various changes are possible.

In the block combining means 113, the number of empty blocks one size below the insufficient block size is not sufficient, the occupation status by block size does not satisfy a predetermined condition due to the block combination, or the continuous free space required for the combination is obtained. If there is no block, the process is not immediately abandoned. For the lower size, whether the number of free blocks is sufficient, whether the occupation status by block size satisfies a predetermined condition by block combination, or continuous free blocks required for combination. Is checked, and if both are satisfied, it is divided and applied.

When there are a plurality of sizes smaller than the shortage block size, the block size having the larger total number of blocks is preferentially set as the combining target.

Since both the free block dividing means 111 and the block combining means 113 are provided, there is a possibility that both means may operate when a free block of an intermediate block size is insufficient. In order to prevent this, the free block dividing means 111 and the block combining means 113 compare the preconditions such as the total number of free blocks that can be used and the occupancy, and operate the means on the more optimal side. .

The empty block dividing means 111 is omitted, and the number of blocks is dynamically changed only by the block combining means 113.

When the free block extraction means 107 attempts to extract a free block of the corresponding block size in response to a request from the block size determination means 106, if no free block of the block size exists, the free block extraction means 107 The extracting unit 107 specifies the block size and activates the block combining unit 113 so that the empty block extracting unit 107 extracts the empty block generated by the block combining unit 111. At this time, similarly to the above, when the intermediate block size is insufficient, there are two methods, a method using the empty block dividing means 111 and a method using the block combining means 113, so that the adjustment is performed.

Although the present invention has been described with reference to several embodiments, the present invention is not limited to the above-described embodiments.
Various additions and changes are possible within the scope of the gist. For example, the calculation example of the block number in the embodiment of FIG.
5 and the example applicable to the embodiment of FIG.
In the embodiment of FIG. 1, since the block number generated at the time of data construction is not changed during operation, the present invention is not limited to the above-described example, and an arbitrary block number can be assigned if it is unique for each block size. it can. Further, as information for identifying each block, any information other than the block number can be used.

[0113]

As described above, according to the present invention, the following effects can be obtained.

The operation in which a plurality of block sizes are mixed can be performed, and the block size can be selectively used according to the data to be written. Therefore, the storage area on the secondary storage device can be effectively used while realizing high-speed access to a huge file. Can be used.

In the configuration provided with the empty block dividing means, even if the empty blocks of the small block size become insufficient during the operation, the empty blocks of the large block size are dynamically divided and distributed, thereby depleting the empty blocks. Can be prevented, and the storage area of the secondary storage device can be more effectively used.

In the configuration provided with the block combining means, even if empty blocks of a large block size run short during operation, the empty blocks of a small block size can be dynamically combined to prevent depletion. Thus, the storage area of the secondary storage device can be more effectively used.

In the configuration having the occupancy management means for each block size, block division and block combination can be performed within a range in which the occupation state of the storage area for each block size satisfies a predetermined condition.

[Brief description of the drawings]

FIG. 1 is a block diagram of an embodiment of a file data storage device according to the present invention.

FIG. 2 is a flowchart illustrating an example of processing of a free area securing unit for each size.

FIG. 3 is a diagram illustrating the operation of a free area securing means for each size and a free area management information construction means.

FIG. 4 is a flowchart illustrating an example of processing of a free area management information construction unit.

FIG. 5 is a schematic diagram of an operation of a specific example of a size-specific free area securing unit and a free area management information constructing unit;

FIG. 6 is a diagram illustrating an example in which blocks are randomly allocated to each block size.

FIG. 7 is a diagram showing an example of the configuration of free area management information for each block size and an example of storing the information in a secondary storage device.

FIG. 8 is an explanatory diagram of information stored in a cache.

FIG. 9 is a flowchart illustrating a processing example of an empty block extraction unit.

FIG. 10 is an explanatory diagram of the operation of an empty block extraction unit.

FIG. 11 is a flowchart illustrating a processing example of a block allocating unit.

FIG. 12 is a flowchart illustrating an example of a process of a block release unit.

FIG. 13 is a flowchart illustrating a processing example of an empty block registration unit;

FIG. 14 is an explanatory diagram of the operation of the empty block registration means.

FIG. 15 is a block diagram of another embodiment of the file data storage device of the present invention.

FIG. 16 is a flowchart illustrating a processing example of an empty block dividing unit;

FIG. 17 is an explanatory diagram of the operation of the empty block dividing means.

FIG. 18 is a block diagram of still another embodiment of the file data storage device of the present invention.

FIG. 19 is a flowchart illustrating a processing example of a block combining unit.

FIG. 20 is a flowchart illustrating a processing example of a continuous block detection unit;

FIG. 21 is a diagram illustrating an example of free area management information having a matrix structure.

[Explanation of symbols]

 REFERENCE SIGNS LIST 100 input / output control unit 101 construction instruction analysis means 102 free area securing means by size 103 free area management information construction means 104 file writing / reading means 105 cache 106 block size determination means 107 free block extraction means 108 ... block allocating means 109 ... block releasing means 110 ... empty block registering means 111 ... empty block dividing means 112 ... occupancy management means by block size 113 ... block combining means 114 ... continuous block detecting means 200 ... secondary storage device

Claims (7)

[Claims] 1. A file data storage device comprising a plurality of logical files on a secondary storage device, comprising: an input / output control unit for controlling the secondary storage device; A construction unit that secures free blocks for each block size on the secondary storage device, creates free space management information for managing information on free blocks for each block size, and records the information in the secondary storage device; Block size determining means for determining a required block size in accordance with data to be written on the device; and empty block extraction for extracting information on empty blocks of the block size determined by the block size determining means from the free area management information. Means for assigning a block extracted by the empty block extracting means to a designated file. Means, block release means for removing information on blocks allocated to the file from the file affiliation, and free space for registering the blocks released by the block release means in free area management information corresponding to the block size of the block. A file data storage device comprising: a block registration unit. 2. An empty block for extracting an empty block having a block size other than the minimum size from among a plurality of block sizes and dividing the extracted block into empty blocks having a block size smaller than the block size. 2. The apparatus according to claim 1, further comprising a block dividing unit.
File data storage device as described. 3. The input / output control unit includes a block size occupancy management unit that manages an occupation state for each block size, and the empty block dividing unit occupies each block size by dividing an empty block. 3. The file data storage device according to claim 2, wherein, when the condition no longer satisfies the predetermined condition, the empty block is not divided. 4. The input / output control unit includes a block combining unit that collects a plurality of empty blocks having a block size other than the maximum size among a plurality of block sizes and combines the empty blocks into one block having a larger block size. The file data storage device according to claim 1, further comprising: 5. An occupancy management unit for each block size, wherein the input / output control unit includes an occupancy management unit for each block size, which manages an occupancy status for each block size. 5. The file data storage device according to claim 4, wherein when no predetermined condition is satisfied, empty blocks are not combined. 6. An empty block for extracting an empty block having a block size other than the minimum size from among a plurality of block sizes, and dividing the extracted block into empty blocks having a block size smaller than the block size. A block dividing unit; and a block combining unit that collects a plurality of empty blocks having a block size other than the maximum size among the plurality of block sizes and combines the empty blocks into one block having a larger block size. The file data storage device according to claim 1. 7. A data processing device constituting an input / output control unit for controlling a secondary storage device, according to a construction instruction specifying a plurality of block sizes, to secure an empty block for each block size on the secondary storage device. A construction unit that creates free area management information for managing information on free blocks for each block size and records the information in a secondary storage device, and determines a required block size according to data to be written on the secondary storage device A block size determining unit for extracting, from the free area management information, an empty block extracting unit for extracting information of an empty block having the block size determined by the block size determining unit; and a block extracted by the empty block extracting unit. Means for allocating the file to which the file belongs, and information on the blocks allocated to the file A block releasing unit for removing the block released by the block releasing unit from the affiliation of the block in the free area management information corresponding to the block size of the block; a block size other than the minimum size among the plurality of block sizes Empty block dividing means for extracting an empty block of the same size and dividing the extracted block into empty blocks of a block size smaller than the block size, and collecting a plurality of empty blocks of a block size other than the maximum size among a plurality of block sizes. A machine-readable recording medium on which a program for functioning as block combining means for combining one block having a larger block size is recorded.