JP4130808B2 - Formatting method - Google Patents

Description

  The present invention relates to a recording medium using a nonvolatile memory having an erase block unit, and more particularly to such a recording medium.

  In recent years, semiconductor flash memories that can be electrically recorded and erased and have built-in memory elements in which recorded data is non-volatile even when the power is turned off are becoming widespread. This type of semiconductor flash memory requires fewer mechanical moving parts, and can be configured to be compact and lightweight, so it is increasingly used as an alternative to hard disks in mobile devices and other fields, that is, as secondary storage. I came. In addition, as the capacity of memory devices has increased, semiconductor flash memories have begun to be used for recording video and audio.

Information such as video and audio recorded in a semiconductor flash memory is generally managed as a file by a file system. The file system manages the size of each file, the date and time when it was recorded, the usage status and availability of recording areas such as clusters and sectors, and records such meta information on a recording medium.
For example, in a certain operating system, the minimum unit of file input / output to / from secondary storage is a sector without distinguishing between meta information and data in the user data area. This is because the minimum rewrite unit of a recording medium such as a hard disk or a flexible disk is a sector. Also, in a FAT (File Allocation Table) file system, data in the user data area is managed by a unit called a cluster in which a plurality of continuous sectors are collected.

  On the other hand, in general, a semiconductor flash memory has a characteristic that data of a certain size called an erase block is electrically erased collectively before data recording. Here, when recording data less than the size of the erase block, after reading and holding the data of the block once, erase the data of the block at once, update the data held partially, Will be written back to the block. Such an operation is called “read-modify-write”, which complicates the recording operation and lowers the recording transfer rate. In order to record data at a high transfer rate, in order to avoid partial data update for such erase blocks, the size of the data to be recorded is the same size as the erase block or an integral multiple of that size, and It is necessary to set the recording address to the beginning of the erase block.

  In general, when a semiconductor flash memory is used as secondary storage, the boundary of erase block units in the semiconductor flash memory does not coincide with the boundary of clusters handled by a file system built on the semiconductor flash memory. When a cluster spans multiple erase blocks on a semiconductor flash memory, multiple block erases are performed for one cluster rewrite, which significantly shortens the life of the semiconductor flash memory and As compared with the case where is included in one erase block, the write time is longer.

In order to solve such a problem, for example, the following method has been developed (see Patent Document 1).
FIG. 7 is a block diagram when the semiconductor flash memory is initialized with the FAT16 file system, and particularly shows a volume area in the semiconductor flash memory.

As shown in FIG. 7, the volume area includes a master boot record (MBR), a root partition table sector 701, reserved areas 702 and 703, a system area 708, and a user data area 707.
In the MBR and root partition table sector 701, the start position of the partition is recorded. The partition starts from the system area 708.

The reserved areas 702 and 703 are areas provided between the MBR / root partition table sector 701 and the system area 708. As described above, an area composed of the MBR and root partition table sector 701 and the reserved areas 702 and 703 is referred to as a partition management area.
The system area 708 is further divided into a partition boot sector 704, a duplicated FAT 705, and a root directory entry (RDE) 706.

The user data area 707 is managed in cluster units.
In this way, the data size from the sector including the master boot record and the partition table to the partition boot sector can be determined so that the head of the user data area 707 coincides with the erase block boundary of the semiconductor flash memory. .

Therefore, since the boundary between clusters in the user data area coincides with the boundary of erasable blocks, it is possible to eliminate the waste of erasing two erase blocks when rewriting one cluster.
However, the above method only solves the difference between the file system management unit and the erase block unit in the user data area. Meta information for managing the user data area is recorded in the system area 708 of the semiconductor flash memory. Among these pieces of meta information, there is a piece of information that spans many sectors and is frequently rewritten. Representative examples include FAT recording in the FAT file system and information recording the data allocation status and free space on the recording medium, such as space bitmap in the UDF (Universal Disk Format) file system.

  Since the FAT and the space bitmap are not managed in cluster units, the minimum unit for reading and updating from the operation system is a sector. In many operation systems, sector-by-sector input / output requests are combined as one large input / output request when their addresses are continuous, but there is generally no guarantee that the update location of meta information is continuous. That is, each time a sector at various locations included in one erase block is rewritten, the erase block is read, rewritten, and written back. Since meta information such as FAT and space bitmap has the property of frequently updating relatively small units, this not only shortens the life of the semiconductor flash memory, but also takes a lot of time to update the meta information. As a result, the effective writing rate of the data in the user data area is significantly degraded.

Therefore, it is considered effective to make meta information such as FAT and space bitmap independent of the sector unit input / output mechanism employed in the standard operating system. That is, the meta information is copied onto the main memory in units of erase blocks, and the operation system updates the meta information accompanying file input / output with respect to the copy of the meta information. Then, the operation system checks whether or not the meta information copy is updated at regular intervals, and writes the erase block including the updated meta information back to the semiconductor flash memory. As a result, it is possible to suppress the occurrence of unnecessary read, rewrite, and writeback of the erase block associated with the update of the management information.
JP 2001-188701 A

  However, when the semiconductor flash memory is initialized as described above, there is a possibility that a plurality of meta information may share one erase block, so that input / output management of meta information becomes complicated. FIG. 8 shows the data structure of such a semiconductor flash memory (secondary storage). In FIG. 8, a secondary storage 807 having an erase block unit 808 inside is initialized by the FAT file system. In this initialization, FAT 803 and FAT 804 share the same erase block 810, and FAT 804 and root directory entry 805 share the same erase block 810.

  In order to manage the FAT of the initialized semiconductor flash memory by the operation method, it is necessary to hold the copy 809 of the erase block and the copy 810 of the erase block in the main memory 811. Since the FAT 803 and the FAT 804 are the same data, they are redundant and deteriorate the usage efficiency of the main memory 811. In addition, since the beginning of FAT 803 and the beginning of FAT 804 have offsets in erase blocks 809 and 810, respectively, FAT reference and update accompanying file input / output are complicated, and the FAT is updated on the main memory 811. Therefore, it is necessary to update the corresponding portions of the FATs 803 and 804, respectively.

Further, since the management information other than the FAT (the root directory entry 805 in FIG. 8) is initialized in the erase block in which the FAT is recorded, the update of the erase block is associated with the file input / output. Since it occurs even in cases other than updating, the design of the operation means of the file system becomes complicated, and the management of whether or not the erase block is updated becomes complicated.
The present invention has been made in view of the above-described problems, and achieves high-speed reading / writing of meta information and long life of a semiconductor flash memory, and easy design of file system operation means. It is an object of the present invention to provide a formatting method for initializing so that

  In order to achieve the above object, a semiconductor flash memory according to the present invention is a semiconductor flash memory from which data is erased in units of erase blocks, a user area for recording data as a file, and data for managing the file A system area for recording meta information, and the user area and the system area do not share an erasure block, and when one data is recorded or erased in the user area, it is updated in association therewith. The plurality of meta information to be recorded are recorded in different erase blocks.

In this way, reading and writing of meta information can be speeded up and the life of the semiconductor flash memory can be extended.
The semiconductor flash memory according to the present invention is a semiconductor flash memory initialized in accordance with the FAT file system standard, wherein the system area includes a partition boot sector and a file allocation table, and the partition boot sector Is recorded in the first sector of the partition, and the file allocation table is recorded at a predetermined number of sectors from the partition boot sector, and the last sector of the file allocation table matches the last sector of the erase block. As described above, the size of the file allocation table is determined.

In this way, the effects of the present invention can be obtained for a semiconductor flash memory that is initialized in accordance with the FAT file system standard.
The semiconductor flash memory according to the present invention is a semiconductor flash memory initialized in accordance with the UDF file system standard, and a space bitmap is recorded at the head of the partition, and the space bitmap The size of the space bit map is determined so that the last sector of the first and second sectors coincides with the last sector of the erase block.

In this way, the effects of the present invention can be obtained for a semiconductor flash memory that is initialized in accordance with the UDF file system standard.
With this configuration, management information such as meta information that frequently rewrites can be written back to the semiconductor flash memory independently in units of erase blocks, so that the reading / writing speed of meta information and the length of the semiconductor flash memory can be increased. The design of file system operation means is facilitated while realizing a long life.

  As described above, according to the present invention, the lifetime of the semiconductor flash memory can be extended, and when recording data with high real-time property using the semiconductor flash memory as a secondary medium, the meta information across the erase blocks It is possible to prevent the effective recording speed from being reduced by writing data or rewriting the same erase block multiple times.

Hereinafter, an embodiment of a recording medium according to the present invention will be described with reference to the drawings for an example of a semiconductor flash memory employing a FAT16 file system.
[1] Embodiment FIG. 1 is a diagram showing an entire space starting from physical address 0 of a semiconductor flash memory according to an embodiment of the present invention. As shown in FIG. 1, the data structure of the semiconductor flash memory 1 includes a partition management area, a system area, and a user data area.

The semiconductor flash memory 1 has an erase block for every adjacent N sectors (N = 2 ⁱ : i is a natural number), and data is rewritten by reading to the buffer in units of the erase block, rewriting on the buffer, semiconductor flash memory This is done by writing back to Note that when the data size and position to be rewritten completely match the erase block, reading into the buffer does not occur, so rewriting is executed at high speed.

  In the master boot record written in the physical address 0 and the partition table 101, the start position of the partition, that is, the position of the partition boot sector 103 is written. A sector between these is treated as a reserved area 102 and no data is written. In the present embodiment, the size of the reserved area 102 is set so that the head of the partition boot sector 103 coincides with the head of the erase block 112.

  FIG. 2 is a table showing the configuration of the partition boot sector 103 of FAT16 defined in JIS-X0605. As shown in FIG. 2, the size of the reserved area 104 that immediately follows the partition boot sector 103 can be determined using the 14th to 15th bytes of the partition boot sector 103. In the present embodiment, the number of sectors in the reserved area 104 is defined as (M × N−1), so that the head of the FAT 105 immediately following the reserved area 104 is made coincident with the boundary of the erase block. Here, M is a positive integer indicating the number of erase blocks occupied by the partition boot sector 103 and the reserved area 104. In the present embodiment, M = 1. However, if M is a positive integer, the boundary between the head of the FAT 105 and the erase block always holds, so a value other than 1 may be taken.

In addition, the size of the area per FAT can be set in the 22nd to 23rd bytes of the partition boot sector 103. That is, it means that the size of the FAT can be determined regardless of the effective cluster number T _c . The following equation is an equation for determining the effective cluster number T _c .
T _c = [(T _S −S _sys ) / S _c ]
Here, T _S represents the number of sectors of the entire medium, and S _sys is occupied by management information (master boot record and partition table 101, partition boot sector 103, FAT 105, 106, root directory entry 107 and reserved areas 102, 104). Represents the number of sectors. S _c represents the number of sectors per cluster. The Gaussian symbol [x] represents the maximum integer that does not exceed x. That is, in the above formula, the number of sectors obtained by subtracting management information from the total number of sectors of the semiconductor flash memory 1, that is, the number of sectors in the user data area 109 is divided by the number of sectors per cluster to obtain the effective cluster number _Tc .

In the present embodiment, the size of the FAT is determined as follows. FIG. 3 is a flowchart showing a procedure for determining the number of sectors per FAT. As shown in FIG. 3, when determining the number of sectors per FAT, first, the number of sectors S _R between the partition boot sector 103 and the reserved areas 102 and 104 is subtracted from the total number of sectors T _{S in the} partition. Te, the user data area 109, FAT105,106 and root directory entry 107 to calculate the number of sectors S _o occupied (step S100).

Subsequently, after initializing the variable S representing the number of erase blocks per FAT by substituting the value zero (step S101), the variable S is incremented by 1 (step S102).
And it is judged whether an inequality is materialized.
_{[(S o -S × N ×} F N -RDE S) / S c] ≦ (S × N × S S × 8/16) -2
Here, N represents the number of sectors included in one erase block. F _N represents the number of FATs. In the present embodiment, F _N is set to 2 corresponding to the FATs 105 and 106. Moreover, RDE _S represents the number of sectors in the root directory entry 107, S _c represents the number of sectors per cluster. S _S represents the number of bytes per sector. [X] is a Gaussian symbol as above. Note that F _N may take a positive integer value other than 2 in accordance with JIS-X0605.

  That is, the left side represents the number of clusters secured in the user data area in the partition, and the right side represents the number of clusters that can be managed with the prepared FAT size. Therefore, if the size of the secured FAT does not cover the number of clusters in the user data area, the above inequality does not hold. Conversely, if the size of the FAT can cover the number of clusters in the user data area, the above inequality holds.

If the inequality is not satisfied (step S103: NO), the number of erase blocks S per FAT is increased by 1 (step S102), and it is determined whether the inequality is satisfied.
If it is determined that the above inequality is satisfied (step S103: YES), it is assumed that the necessary number of erase blocks can be allocated to the FAT, and the expression F _S = S × N
Is used to calculate the number of sectors F _S per FAT (step S104), and the process is terminated.

  By the above method, the FATs 105 and 106 having a size necessary for managing the user data area can be secured in the size of the erase block unit. Further, as described above, by adjusting the size of the reserved area 104, the beginning of the FAT is made to coincide with the boundary of the erase block, so that the end of the FAT secured by the above method also coincides with the boundary of the erase block. . That is, the FAT is arranged in units of erase blocks.

In the 17th to 18th bytes of the partition boot sector 103, the size of the root directory entry 107 is set to N × L × S _S / 32
Then, the beginning of the user data area and the boundary of the erase block can be matched. Here, L represents the number of erase blocks occupied by the root directory entry 107, and an arbitrary positive integer value can be given. S _S represents the number of bytes per sector. The denominator 32 is the number of bytes consumed per directory entry as defined in JIS-X0605.

By initializing the semiconductor flash memory 1 as described above, it is possible to avoid two FATs coexisting in one erase block. In addition, the head of the user data area can be made coincident with the erase block boundary.
Therefore, when the duplicated FAT is updated, the situation that the same erase block is rewritten twice is avoided, and the lifetime of the semiconductor flash memory 1 that can reduce the number of rewrites for each erase block is extended. be able to.

  In the present embodiment, the FATs 105 and 106 are arranged in the same manner in the corresponding erase blocks. In other words, each FAT is placed so that the head is placed at the head of the erase block and occupies the area for one erase block. In the present invention, it goes without saying that the number of erase blocks occupied by the FAT is not limited to one, and two or more erase blocks may be occupied.

  Therefore, when an access device that accesses the semiconductor flash memory reads the FAT into the built-in memory in units of erase blocks in order to access the FAT, it is not necessary to read the number of FATs into the memory, and only one FAT is stored in the built-in memory. Just read. That is, after the copy of the FAT in the built-in memory is updated, the plurality of FATs on the semiconductor flash memory can be updated by overwriting the contents on the plurality of FATs on the semiconductor flash memory.

[2] Modifications Although the present invention has been described based on the embodiments, it is needless to say that the present invention is not limited to the above-described embodiments, and the following modifications can be implemented. .
(1) In the above embodiment, the case where the head of the partition boot sector 103 is made to coincide with the boundary of the erase block has been described, but the present invention is not limited to this, and the partition boot sector is included in the same erase block. Unless 103 and FAT 105 are mixed, it is not necessary to make the head of the partition boot sector 103 coincide with the erase block boundary. For example, if the head of the partition boot sector 103 is in the Kth sector from the erase block boundary, the size of the reserved area 104 between the FAT head and the sector (M × N−K) may be set. .

(2) Although the case where the FAT is duplexed has been described in the above embodiment, it goes without saying that the present invention is not limited to this, and the number of FATs can be arbitrarily set by specifying the 16th byte of the partition boot sector 103. It may be changed to
Further, if there is only one FAT and therefore no FAT 106, it is not necessary to make the head of the FAT 105 coincide with the head of the erase block. This is because the erase block rewrite frequency does not increase.

In this case, in order to prevent the FAT 105 and the root directory entry 107 from coexisting in the same erase block, it is preferable to match the end of the FAT 105 with the boundary of the erase block. For this purpose, each parameter may be set so as to satisfy the following equation.
S _R + F _S = N × I
Here, S _R represents the number of sectors occupied by management information other than the FAT 105 and the root directory entry 107. F _S represents the number of sectors of the FAT 105. N represents the number of sectors per erase block, and I is a positive integer.

The left side of the above formula is the sum of all management information (master boot record and partition table 101, partition boot sector 103 and reserved areas 102, 104) before FAT 105 and the number of sectors consumed by FAT 105, and the right side is The left side indicates that the erase block is secured.
(3) In the above embodiment of the present invention, the present invention has been described by taking the case of using the FAT16 file system as an example. However, it goes without saying that the present invention is not limited to this, and instead of the FAT16 file system, a FAT12 file system, A FAT32 file system may be used.

For the FAT16 file system, the inequality is used when determining the FAT size:
_{[(S o -S × N ×} F N -RDE S) / S c] ≦ (S × N × S S × 8/16) -2
In the case of the FAT32 file system, the inequality is:
_{[(S o -S × N ×} F N -RDE S) / S c] ≦ (S × N × S S × 8/32) -2
It is good to use. Also, in the case of the FAT12 file system, the inequality:
_{[(S o -S × N ×} F N -RDE S) / S c] ≦ (S × N × S S × 8/12) -2
It is good to use.

As described above, when the FAT32 file system is used, since it is not necessary to have the root directory entry 107 in the system area 108, the processing related to the root directory entry 107 can be omitted in the initialization procedure shown in FIG. .
(4) In the above embodiment, the case where one erase block is assigned to the partition boot sector 103 has been described. However, when there is only one FAT and therefore no FAT 106, the partition boot sector 103 is assigned to the FAT 105. May be arranged in the same erase block.

For example, in FIG. 1, the partition boot sector 103 may be arranged at the head sector of the erase block 3, and the subsequent (N−1) sectors may be used as the FAT 105.
Even if the partition boot sector 103 is arranged in any erase block, the partition boot sector 103 is not rewritten after being written when the semiconductor flash memory is formatted. Does not affect the effects of the present invention.

(5) In the above embodiment, the FAT16 file system has been described as an example. However, it goes without saying that the present invention is not limited to this, and a UDF file system may be used instead of the FAT series file system. The UDF file system is standardized as follows.
ISO / IEC 13346-1: 1995, Information technology-Volume and file structure of write-once and rewritable media using nonsequential recording for information interchange-Part 1: General.
ISO / IEC 13346-2: 1995, Information technology-Volume and file structure of write-once and rewritable media using nonsequential recording for information interchange-Part 2: Volume and boot block recognition.
ISO / IEC 13346-3: 1995, Information technology-Volume and file structure of write-once and rewritable media using nonsequential recording for information interchange-Part 3: Volume structure.
ISO / IEC 13346-4: 1995, Information technology-Volume and file structure of write-once and rewritable media using nonsequential recording for information interchange-Part 4: File structure.
ISO / IEC 13346-5: 1995, Information technology-Volume and file structure of write-once and rewritable media using nonsequential recording for information interchange-Part 5: Record structure.
ISO / IEC 13490-1: 1995, Information technology-Volume and file structure of read-only and write-once compact disk media for information interchange-Part 1: General.
ISO / IEC 13490-2: 1995, Information technology-Volume and file structure of read-only and write-once compact disk media for information interchange-Part 2: Volume and file structure.
Now, the data structure of the semiconductor flash memory when the UDF file system is used will be described. FIG. 4 is a diagram showing a space starting from logical sector number 0 in the semiconductor flash memory when the UDF file system is used.

  In the UDF file system, meta information frequently updated during file recording is a space bitmap 403. The space bitmap 403 is recorded at the head of the partition as shown in FIG. The head position of the partition is specified by a partition descriptor 402 included in the main volume descriptor 401. By matching the head of the partition specified by the partition descriptor with the boundary of the erase block, the head of the space bitmap 403 can be made to match the boundary of the erase block.

FIG. 5 is a table showing the configuration of the space bitmap 403. As shown in FIG. 5, the size of the space bitmap 403 is described in the 20th to 23rd bytes of the space bitmap 403. That is, the size of the space bitmap 403 can be defined independently of the number of blocks actually included in the partition.
Accordingly, in the same manner as in the FAT16 file system, the space bitmap 403 is made large enough to cover the entire partition, and the end of the space bitmap 403 coincides with the boundary of the erase block. Can be determined. Thus, the space bitmap 403 and other meta information and user data can be arranged so as not to be mixed in the same erase block.

  In the above, the top sector of the space bitmap 403 and the top sector of the erase block are matched so that the updated meta information does not share the same erase block. If rewriting does not occur, it is not necessary to make the head sector of the space bitmap coincide with the head sector of the erase block. This is because a plurality of meta information existing in the same erase block cannot be updated simultaneously.

Next, a file system operation method for efficiently recording a file in a semiconductor flash memory will be described taking the case of a FAT file system as an example. A similar operation method is also used when using the UDF file system. FIG. 6 is a diagram for explaining a procedure for recording a file in the semiconductor flash memory.
In FIG. 6, an application 601 creates data to be stored in a semiconductor flash memory 605 that is a secondary storage. Further, the application 601 requests the file system configuration unit 602 to store the data in the semiconductor flash memory 605. The file system configuration unit 602 includes a meta information processing unit 606 and a data processing unit 607 inside. The semiconductor flash memory 605 has a data structure as described above.

  When the semiconductor flash memory 605 is mounted, the file system configuration unit 602 performs meta information (for example, from logical sector N to logical sector M) from the system area 613 of the partition on the semiconductor flash memory 605, which frequently rewrites. .) Is cut out in units of erase blocks composed of continuous sectors, and copied to the built-in memory 603 of the device capable of reading and writing at high speed.

At this time, since the semiconductor flash memory 605 has the data structure shown in the above embodiment, necessary meta information can be copied to the built-in memory 603 in units of erase blocks.
In addition, since two FATs do not share the same erase block, only one FAT can be cut out in units of erase blocks and copied to the internal memory 603. Accordingly, the memory of the system can be saved. The file system configuration unit 602 acquires the size of the erase block by inquiring of the semiconductor flash memory 605 via the erase block unit acquisition unit 615.

  The meta information processing unit 606 manages the copied meta information in the same erase block unit as the erase block of the semiconductor flash memory 605 on the built-in memory 603, and associates a rewrite presence / absence management unit in erase block unit for each erase block. Here, the rewrite presence / absence management flag 608 associated with each erase block is used as an example of the rewrite presence / absence management means. All the rewrite presence / absence management flags 608 are set to “clean” immediately after mounting.

  When the application 601 requests to write data, the file system configuration unit 602 uses the user data area data processing unit 607 to transfer the data received from the application 601 to the user data area 614 of the semiconductor flash memory 605. Write. At this time, the meta information processing unit 606 creates meta information related to the data, and updates the meta information copied to the internal memory 603. In addition, the rewrite presence / absence management flag 608 is set to “dirty” for the erase block for which information has been updated.

  The meta information on the semiconductor flash memory 605 is updated at regular intervals. That is, the timer 610 notifies the meta information processing unit 606 of the update time of the meta information on the semiconductor flash memory 605 at regular intervals. When the notification is received, the meta information processing unit 606 refers to the rewrite presence / absence management flag 608 in the built-in memory 603 and detects the rewrite presence / absence management flag 608 that is “dirty”. This is because if the rewrite presence / absence management flag 608 is “dirty”, the contents of the corresponding meta information do not match between the internal memory 603 and the semiconductor flash memory 605.

For an erase block for which the rewrite presence / absence management flag 608 is determined to be “dirty” by the meta information processing unit 606, the erase block unit writing unit 604 reads the erase block from the built-in memory 603 and corresponds to the semiconductor flash memory 605. Write to the erase block in erase block units.
In this way, the meta information can always be written back in units of erase blocks, so that the number of rewrites of the semiconductor flash memory can be reduced, extending its life, and high-speed meta information writing. Can be realized.

Further, since data other than FAT is not mixed in the erase block copied to the internal memory 603, the rewrite presence / absence management flag 608 may be updated by paying attention only to the update of the FAT. As a result, the configuration of the apparatus related to writing back meta information can be simplified.
In this modification, the erase block size acquisition unit 615 acquires the size of the erase block. Alternatively, the user may manually notify the size of the erase block through the application 601. .

In this modification, all the meta information that is frequently rewritten at the time of mounting is copied to the built-in memory. Instead, only the part necessary during file recording is copied to the built-in memory in units of erase blocks. It is also good.
In this modification, the meta information write-back timing is notified to the meta information processing unit 606 by the timer 610, but instead, the application notifies the meta information to the file system configuration unit 602 at an appropriate timing. May be instructed to write back.

  The recording medium according to the present invention is a recording medium using a non-volatile memory having an erase block unit, and is useful as a technique that can extend the life of such a recording medium.

It is a figure showing the whole space which starts from the physical address 0 of the semiconductor flash memory based on embodiment of this invention. It is a table | surface which shows the structure of the partition boot sector 103 of FAT16 prescribed | regulated to JIS-X0605. It is a flowchart which shows the procedure which determines the number of sectors per FAT concerning embodiment of this invention. It is a figure showing the space which starts from the logical sector number 0 in the semiconductor flash memory concerning the modification (5) of this invention. It is a table | surface which shows the structure of the space bitmap 403 recorded on the semiconductor flash memory which concerns on the modification (5) of this invention. It is a figure for demonstrating the procedure for recording a file in the semiconductor flash memory which concerns on the modification (5) of this invention. FIG. 2 is a block configuration diagram when a semiconductor flash memory according to the prior art is initialized with a FAT16 file system, and particularly shows a volume area in the semiconductor flash memory. It is a figure which shows the data structure of the semiconductor flash memory (secondary storage) based on a prior art.

Explanation of symbols

1 ......... Semiconductor flash memory 101 ............... Master boot record and partition table 102, 104 ... Reserved area 103 ......... Partition boot sector 105, 106 ... FAT
107 ............ Root directory entry 108, 613... System area 109, 614... User data area 111, 112, 113, 114, 115, 609, 612. ............... Partition descriptor 403 ............... Space bitmap 601 ............... Application 602 ............... File system configuration means 603 ............... Built-in memory 604 ............... Erasing Block unit writing means 605... Semiconductor flash memory 606... Meta information processing unit 607... Data processing means 608. …………… Sector 615 …………… Erasing block acquisition means

Claims (3)

A formatting method in which a computer formats a flash memory with a FAT file system having a predetermined cluster size,
A first step of determining an address to write a partition boot sector, which is an address on the flash memory, which is a sector in which the configuration information of the FAT file system is written;
A second step in which the FAT start address for entering the file arrangement state is not shared with the partition boot sector and the erase block of the flash memory, and the erase block starts,
A third step in which the FAT size is an integral multiple of the erase block of the flash memory;
A fourth step of determining the number of effective clusters on the flash memory by dividing the cluster size from the total number of sectors that the flash memory has after the end of the FAT;
If the maximum number of clusters that can be managed by the FAT is smaller than the number of valid clusters, the FAT can be increased by an integer multiple of the erase block in the flash memory until the maximum number of clusters that can be managed by the FAT is equal to or greater than the number of valid clusters. A fifth step of increasing the size of
FAT maximum cluster number is valid clusters number and whether the same can be managed, is greater than the number of valid clusters, formatting method, which comprises a sixth step of writing the size of the FAT to partition boot record, the. A method in which a computer formats a flash memory with a FAT file system having a predetermined cluster size and having only one FAT,
A first step of determining an address to write a partition boot sector, which is an address on the flash memory, which is a sector in which the configuration information of the FAT file system is written;
A second step in which a FAT start address for entering a file arrangement state is set to an address not sharing the partition boot sector and the erase block of the flash memory;
A third step of determining the size of the FAT so that the last sector of the FAT matches the last sector of the erase block of the flash memory;
A third step of determining the number of effective clusters on the flash memory by dividing the cluster size from the total number of sectors that the flash memory has after the end of the FAT;
If the maximum number of clusters that can be managed by the FAT is smaller than the number of valid clusters, the FAT can be increased by an integer multiple of the erase block in the flash memory until the maximum number of clusters that can be managed by the FAT is equal to or greater than the number of valid clusters. A fifth step of increasing the size of
FAT maximum cluster number is valid clusters number and whether the same can be managed, is greater than the number of valid clusters, formatting method, which comprises a sixth step of writing the size of the FAT to partition boot record, the. 3. The formatting method according to claim 1, further comprising a step of writing the size of the root directory entry, which is a file list of the root directory, into the partition boot record as an integral multiple of the erase block of the flash memory.

Images