JPH0973369A - File storage device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To eliminate the generation of an incomplete file by providing a file information management part which stores a file name, a head sector number, sector connection order, and data size, a cache part, and a file control part. SOLUTION: The device consists of a 1st file information management part 101 which manages data in a data storage part 103, a 2nd file information management part 102 which holds data of the same kind with the 1st file information management part 101, and a semiconductor memory which is fast in read and write speed, and is equipped with the cache part 104 which holds the whole contents of the 1st file information management part 101, a file control part 105 which performs processing when an application part 108 on a main device makes a request to read and write a file, a data transfer part 106, and an initial diagnostic part 107 which detects and corrects an error in file contents. Then when the writing of data to a sector is completed, data of the cache part 104 as management information on the file name, file length, etc., is written in the 1st information management part 101.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a file storage device using a magnetic material or a semiconductor memory and capable of writing and reading a plurality of variable length data at any time.

[0002]

2. Description of the Related Art Conventionally, a large-capacity readable / writable external storage device has been used for computer program storage and data storage. In these external storage devices, as a characteristic of the storage medium, even if it is possible to read and write in an arbitrary storage area in units of 1 byte, the storage medium is divided into a small capacity of several hundred bytes to several kilobytes (hereinafter,
Divided into "sectors") and used as a unit for reading and writing, and by connecting and using a plurality of sectors, variable-length data storage is enabled. Further, a configuration is generally used in which a file name is given to variable-length data and a plurality of variable-length files can be stored. As described above, in a large-capacity external storage device, a method of storing a plurality of variable-length data by sector division has been established, but as described above, the storage medium can be read and written in sector units at any time. A magnetic disk has been the mainstream as a large-capacity storage medium that is necessary and meets the specifications. However, with the recent increase in the degree of integration of non-volatile semiconductor memories, the development of memory cards called flash ROMs, which have a storage capacity of several tens of megabytes, has progressed, and they are smaller, lighter, and more resistant to mechanical shock than magnetic disks. Because of its features, it has come to be considered to be used for an external storage device. Taking advantage of the above features, these memory cards are being introduced as external storage devices for portable computers and industrial equipment under severe operating conditions.

A conventional file storage device will be described below with reference to the drawings. FIG. 6 shows a conventional file storage device 60.
It is a block diagram showing 0. In FIG. 6, reference numeral 601 manages the usage status of sectors of a data storage unit 602, which will be described later, and manages the size of each data, the head sector number, and the file name that is an identifier of each data in a batch. As a file information management unit capable of storing a plurality of data, 602 is a data storage unit that stores data by dividing it into small partition sectors, and 603 is an application to be described later when reading and writing to the data storage unit 602, which is a storage medium, is slow. A cache unit 604 that temporarily stores a copy of the contents of the file information management unit 601 and the data storage unit 602 in a semiconductor memory whose read / write response time is short in order to speed up response time to read / write from the unit 606. Is a sector required to create a new file according to an instruction from the application unit 606 described later. Acquisition, connection, file control unit that performs data size, a storage file name, 605 below the application unit 606
After instructing the file control unit 604 to create a new file, the data transfer unit that transfers the data to be actually stored and writes the data in the data storage unit 602, and 606 instructs the file control unit 604 to create a new file. It is an application unit on the main device.

A method of using the file storage device configured as described above will be described.

As a method of using the cache unit 603, in the case of writing, a write-through method of writing in the cache unit 603 and immediately writing in the data storage unit 602, and a write-through method in which only the cache unit 603 is written and the write data is written for a fixed time Data storage unit 60 if not updated
There is a write-back method for writing to 2. In the case of reading, there is a method of reading the contents of the cache unit 603 instead of the contents of the data storage unit 602 when the data of the sector designated to be read is stored in the cache unit 603. It is a common method. The above-mentioned write-through method shortens only the read time, and the write-back method shortens both the read time and the write time. Further, when storing important data in such a file storage device, the file storage device is duplicated to
A method of preventing data loss by storing data in one device at the same time and using the other device even if one device fails is used.

[0006]

However, in the conventional file storage device, the file name and data size are written to the file information management unit 601 on the storage medium during data writing, and the sector state is also changed. When a power failure or a failure of the main device occurs during this time and the writing process cannot be continued, there is a problem that data that has not been written is also created as a file. Moreover, it is impossible to distinguish whether the writing is incomplete or not unless a separate means such as taking a checksum is taken for these data, and the means for taking the checksum complicates the configuration, resulting in high cost and large size. Was there. Further, since the file information management unit 601 is not duplicated in the device alone, if a similar failure occurs during the updating of the file information management unit 601 after the completion of data writing, the sector connection state mismatch occurs. However, there is a problem that not only the file being written but also the management information of other files is destroyed. Furthermore, when the cache unit of the write-back method is used, writing from the cache unit to the file information management unit 601 is also delayed, which poses a problem that the risk of file destruction due to power failure or the like increases. . Furthermore, although the method of duplicating the device itself is effective for the failure of the device itself, it becomes expensive and
It had a problem that it became large.

The present invention solves the above-mentioned conventional problems by providing a small-sized and low-cost file storage device which does not create an incomplete file even if a power failure or a failure of the main unit occurs during writing. The purpose is to provide.

[0008]

To achieve this object, a file storage device according to claim 1 of the present invention comprises a file storage device having a readable and writable storage medium composed of sectors each having a plurality of bytes as a unit. A file information management unit that stores a file name, a start sector number, a sector connection order, and a data size, a cache unit that holds data of the same type as the file information management unit, and a file control that controls the cache unit And a part.

A file storage device according to a second aspect of the present invention is the file storage device according to the first aspect, wherein the file control unit writes the data in the cache unit to the file information management unit upon completion of writing the data to the sector. ing.

A file storage device according to a third aspect of the present invention is a file storage device having a readable / writable storage medium composed of sectors each having a plurality of bytes as a unit, and comprises a file name, a head sector number, a sector connection order and data. File contents of the first and second file information management units, which stores the size and the first and second file information management units, and holds the same type of data as the first file information management unit The file in which the error is detected in the data of the file information management unit that has no error when the file content error is detected in either the cache unit that sequentially updates the file information or the first and second file information management units at the time of startup. It has a configuration including an initial diagnosis unit for copying and correcting the information in the information management unit.

[0011]

With this configuration, the file control unit can write the data in the cache unit to the file information management unit when the writing of the data is completed.
Until the data writing is completed, no changes are made to the file information management unit, and even if a power failure or a failure of the main unit occurs in the middle, an incomplete file will not be created. Also,
The device itself is not duplicated, and miniaturization and cost reduction are achieved.

When an error in the file content is detected in either one of the first and second file information management units, the data in the file information management unit without error is copied to the file information management unit in which the error is detected. Since the correction is made by correcting the error, the file contents in which an error is detected can be corrected by using the complete file before or after the file is created, and a halfway file is not created. Further, even if one of the file information management units is destroyed, the same correction can be made by copying the contents of the file which is not destroyed. Further, the device itself is not duplicated, and downsizing and cost reduction are achieved.

[0013]

【Example】

(Embodiment 1) A file storage device according to an embodiment of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing a file storage device 100 according to an embodiment of the present invention.

In FIG. 1, 101 is a data storage unit 10.
A first file information management unit that manages the data on 3
Reference numeral 02 is a second file information management unit that holds the same type of data as the first file information management unit 101, and reference numeral 104 is a semiconductor memory that has a high read / write speed, and stores the entire contents of the first file information management unit 101. A cache unit for holding 105, a file control unit for performing processing when a file read / write is requested from the application unit 108 on the main device, and 106 is a data storage unit 103 for reading / writing data designated by the application unit 108. And a data transfer unit 107 between the file storage device 100.
It is an initial diagnosis unit that detects and corrects errors in the file contents at startup. The first file information management unit 101,
Second file information management unit 102 and data storage unit 1
03 is arranged on the storage medium.

A method of storing a plurality of variable length data in the file storage device configured as described above will be described. FIG. 2 is a data state diagram showing the first file information management unit 101, the second file information management unit 102, and the data storage unit 103. As in the conventional example, the data storage unit 103 is obtained by dividing the storage medium into sectors of small partitions and allocating the sectors. In this embodiment, it is assumed that the data storage unit 103 is composed of 12 sectors each having 512 bytes, and file names of file 1b and file 2b are given to two data having data sizes of 200 bytes and 600 bytes, respectively. The case where the data is stored in the data storage unit 103 will be described. As shown in FIG. 2, the first file information management unit 101 has a directory 1 and a file allocation table (FAT) 1a, and the directory 1 includes files 1b, 2b in areas 1101, 201,
Data size 200, 600 of areas 1102, 202,
The area 1103 is composed of head sector numbers 211 and 212. The second file information management unit 102 uses the directory 2 and file allocation table (FAT).
2a. The second file information management unit 102 will be described in the second embodiment. File 1b has a data size of 200 bytes, and 1 sector (512 bytes)
Since it falls within the range, sector number 211 on FAT1a
However, since the file 2b has a data size of 600 bytes, two 512-byte sectors are required, and the next sector number 213 is written to the sector 212 on the FAT1a, and the last sector 213 has the mark. FF is written as.

The file storage device 100 having such a configuration
In FIG. 3, the operation of newly writing the file 2b of FIG. 2 from the application unit of the main device to the file storage device will be described with reference to FIG. FIG. 3 shows a file storage device 1.
10 is an operation procedure diagram for explaining the operation of No. 00. FIG. 3, 101 is a first file information management unit, 103 is a data storage unit, 104 is a cache unit, 105 is a file control unit, 106 is a data transfer unit, and 108 is an application unit. These are the same as in FIG. Therefore, the same reference numerals are given and the description thereof will be omitted. First, the file control unit 105 that has received the open request from the application unit 108 for instructing the creation of a new file
It is searched whether or not there is an area for storing a new file in the directory area on 04 (S1, S2), and if there is an empty area, the file name designated by the application unit 108 is used as shown in FIG.
In the area 201 as shown in FIG. 3, “0” is registered as the initial value of the data size in the area 202, and an open confirmation, which is a positive response, is returned to the application unit 108 (S
3). After receiving the open confirmation, the application unit 108 issues a 200-byte write request to the data transfer unit 106 (S4, S5). The data transfer unit 106 searches for an empty sector 212 (see FIG. 2) necessary for storage, writes it in the first sector number area 203 of the directory 1 (S6), adds the file length 200 to the data size area 202, and is requested to write. Holds 200 bytes of data (S
7).

Next, the application unit 108 again sets 2
A write request of 00 bytes is made (S8). This makes a total of 400 bytes of write requests (accumulation size),
Since this storage size does not exceed 512 bytes for one sector, it is stored in the data transfer unit 106 (S9, S1).
0). Next, when the third 200-byte write request is received, 512 bytes for one sector are written to the data storage unit 103 on the storage medium (S11). Also, 200
The remaining 88 bytes of the bytes are the data transfer unit 106.
(S12), a new empty sector 213 (see FIG. 2) is searched to hold this 88 bytes, and the sector number 213 is written to the empty sector 212 (S1).
3). Finally, when the file control unit 105 receives from the application unit 108 a close notification notifying the completion of writing, the 88-byte data remaining in the data transfer unit 106 is written in the data storage unit 103 (S15), and at the same time the first file is written. The data of the content held in the cache unit 104 is written in the information management unit 101 (S16).

As described above, according to this embodiment, the data of the cache unit 104, which is management information such as the file name and the file length, is written to the first file information management unit 101 when the writing of the data to the sector is completed. Therefore, no change is made to the contents of the first file information management unit 101 until the data writing is completed, and even if a power failure or a failure of the main device occurs in the middle, the first file information management unit 101 does not Has no effect, and therefore an incomplete file is not created in the first file information management unit 101. If a power failure or a failure of the main unit occurs during the process, file creation will have to be redone.

(Embodiment 2) A file storage device according to a second embodiment of the present invention will be described below. The file storage device according to this embodiment has the same configuration as that in FIG. First
Although the second file information management unit 102 is not used in this embodiment (see FIG. 3), the second file information management unit 102 is used in this embodiment. That is, the file information management unit is duplicated in the first file information management unit 101 and the second file information management unit 102.

First, the data destruction that occurs when a power failure occurs during the file write completion processing in steps 15 and 16 of FIG. 3 will be described with reference to FIG. FIG. 4A is a data state diagram of the first file information management unit 101, and FIG.
(B) is a data state diagram of the second file information management unit 102. In FIGS. 4A and 4B, 400 and 40
1 is a directory, 402 to 404 are file allocation tables, and 405 to 414 are sectors.

In FIG. 4A, sectors 405, 40
When the data writing of the file using 6, 407, 408 is completed, the first file information management unit 101 receives the cache flush notification of the close notification in step 16 of FIG.
If a power failure occurs at the point when the writing of the directories 400 and 401 to the
The sector 407 pointed to by 06 will remain in the previous empty state,
This file remains broken at sector concatenation 406. In this state, when the power failure is restored after this and a new file is created separately and the sector 407 is used, the file indicated by the sector 406 and the newly created file simultaneously use the sector 407 and the link crossing state. Occurs. Further, as shown in FIG. 4B, the order of connecting the sectors of the newly created file is the sectors 411 and 412.
After that, they are arranged like sectors 413 and 414,
If a power outage occurs with the sector 413 updated,
Since the sector 410 points to an empty sector, the above-mentioned link crossing condition may occur, and the two sectors 413 and 414 are not used for any files, and all the files registered in the directory are deleted. However, since it remains in use, there arises a problem that the size of the usable data storage unit 103 becomes small.
This embodiment solves these problems.

Next, the operation of the file storage device according to this embodiment will be described with reference to FIG. 5 (a), (b),
(C) shows the first and second file information management units 101 and 10.
It is a data state diagram of No. 2. In the present embodiment, the file information management unit is duplicated such as the first file information management unit 101 and the second file information management unit 102, and when the cache close notification of file close of step 16 in FIG. 3 is processed. Then, the directory 500 of the first file information management unit 101, the FAT 501, the directory 502 of the second file information management unit 102, and the FAT 503 are sequentially written. Therefore, even if a power failure occurs during writing, the location of failure is FAT50 in FIG.
It will be limited to one place like 4. in this case,
After the power failure is recovered, the initial diagnosis unit 107 checks the sector connection state of all files in the first file information management unit 101 and the second file information management unit 102, and confirms that the unused sectors of the file are empty. The second file information management unit 102 is inspected to find an empty sector 504.
By copying the contents of the above to the entire surface of the file of the first file information management unit 101, the state before the file creation can be restored. On the contrary, as shown in FIG. 5C, when the failure location is found in the FAT 506 of the second file information management unit 102, the contents of the first file information management unit 101 (the content of which has been updated) ( By copying (corresponding contents of FAT 506) to the second file information management unit 102, the newly created file can be reproduced in a correctly connected state.

As described above, according to this embodiment, the first file information management unit 101 and the second file information management unit 1
When an error in the file content is detected in either of the two, the error-free file information management unit data is copied to the file information management unit that has detected the error and corrected. You can use the complete file of
Never create a half-finished file. Further, even if one of the file information management units is destroyed, the same correction can be made by copying the contents of the file which is not destroyed.

[0025]

As described above, according to the present invention, the file information management unit that stores the file name, the start sector number, the sector connection order, and the data size, and the cache unit that holds the same type of data as the file information management unit And the file control unit for controlling the cache unit, it is possible to write the data in the cache unit to the file information management unit when the writing of the data is completed. No change is made to the contents of the file, and even if a power failure or a failure of the main unit occurs in the middle, the file information management section is not affected, so an incomplete file is created in the file information management section. It is possible to realize a file storage device that does not have a problem. Also, since the device itself is not duplicated,
It is possible to realize a file storage device that can be reduced in size and cost.

Further, the file control unit writes the data in the cache unit to the file information management unit when the writing of the data to the sector is completed, so that even if a power failure or a failure of the main unit occurs before the data writing is completed,
It is possible to prevent the file information management unit from being affected, and it is possible to realize a file storage device in which an incomplete file is not created in the file information management unit.

Further, the first and second file information management sections for storing the file name, the head sector number, the sector connection order and the data size, and the same type of data as the first file information management section are held. When the data writing is completed,
A cache unit that sequentially updates the contents of the second file information management unit, and file information management that is error-free if an error in the file contents is detected in one of the first and second file information management units at startup. By including the initial diagnosis section that copies the data of the section to the file information management section that has detected an error and corrects it, it is possible to use a complete file before file creation, that is, before file update or after file creation, that is, after file update. Since the contents of the file in which an error has been detected can be corrected, a file storage device that does not create a half-finished file can be realized. Further, since the device itself is not duplicated, it is possible to realize a file storage device that can be downsized and reduced in cost.

[Brief description of drawings]

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

FIG. 2 is a data state diagram showing a first file information management unit, a second file information management unit, and a data storage unit.

FIG. 3 is an operation procedure diagram for explaining the operation of the file storage device.

FIG. 4A is a data state diagram of the first file information management unit, and FIG. 4B is a data state diagram of the second file information management unit.

5A, 5B, and 5C are data state diagrams of the first and second file information management units.

FIG. 6 is a block diagram showing a conventional file storage device.

[Description of Reference Signs] 100 file storage device 101 first file information management unit 102 second file information management unit 103 data storage unit 104 cache unit 105 file control unit 106 data transfer unit 107 initial diagnosis unit 108 application unit

Claims (3)

[Claims] 1. A file storage device having a readable / writable storage medium composed of sectors each having a plurality of bytes as one unit, and file information management for storing a file name, a start sector number, a sector connection order and a data size. A file storage device comprising: a unit, a cache unit that holds the same type of data as the file information management unit, and a file control unit that controls the cache unit. 2. The file storage device according to claim 1, wherein the file control unit writes the data of the cache unit to the file information management unit when the writing of the data to the sector is completed. 3. A file storage device having a readable / writable storage medium composed of sectors, each of which has a plurality of bytes as a unit, and which stores a file name, a start sector number, a sector connection order, and a data size. A second file information management unit, a cache unit that holds the same type of data as the first file information management unit, and that sequentially updates the file contents of the first and second file information management units when data writing is completed, When an error in the file content is detected in either the first or second file information management unit at startup, the data in the file information management unit without error is copied to the file information management unit that detected the error and corrected. A file control device comprising: