JPH11143764A - Flash type memory, its management device, storage device and computer system - Google Patents

Abstract

PROBLEM TO BE SOLVED: To execute program on an extension storage device without loading an execution program to a main storage device by providing a block of a flash type memory with a header area which stores management information of the block and a storage area which stores a data file that is divided into a block unit. SOLUTION: This flash type memory is applied to a computer system which includes a processor 10, an extension or auxiliary storage device which includes flash memory 12 and its controller 14 and a RAM (main storage device) 16. A management unit of the memory 12 is one block that is a minimum erase (erasure) unit in order to manage an execution program area and a data area. The storage area of each block stores the execution program area or the data area, a header area which has information that manages the block and a data file which is divided into a block unit.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a flash memory, a management device thereof, a storage device, and a computer system. More specifically, the present invention relates to an improvement in an efficient flash memory management method.

[0002]

2. Description of the Related Art Generally, electrically erasable programmable read only memories (EEPROMs) including flash type floating gate transistors are readily available on the market today. These so-called flash memories are non-volatile memories similar in function and performance to EPROM memories, and further have a function of enabling in-circuit programmable operation for erasing blocks divided in the memory. In a flash memory, rewriting is performed by previously erasing a previously written area of the memory.

In a typical computer system, a program of an operating system (hereinafter simply referred to as "OS") manages data of a data storage device of the system. The attributes of the data storage device that are necessary and sufficient to achieve compatibility with the OS program are available from any location on the data storage device.
Data can be read and data can be written to it. However, in the case of a flash memory, data cannot be written to an area to which data has already been written unless data in that area is erased. For this reason, flash memory is not compatible with typical existing OS programs.

[0004] Focusing on such a point, software that enables a flash memory to be managed by an existing OS program has been proposed in the prior art.
In this prior art program, the flash memory is operated as a "write once read multiple times" device, or is operated as a "write multiple times read multiple times" device. The former are devices that cannot reuse previously written memory locations and can be used as auxiliary storage or extended storage. In the latter, an auxiliary storage device (semiconductor file storage device) having a control for making a previously written memory area reusable and reducing the number of times of rewriting of the flash memory is included in the area.

[0005]

However, the above background art has the following disadvantages. (1) According to the apparatus described in the prior art described above, an execution program is operated on an extended storage device using a flash memory without having a control structure for reducing the number of rewrites. That is, when newly rewriting data in the flash memory, it is necessary to erase all blocks at once and then store the data.
In the case where management is performed such as sharing or occupation of parallel execution programs and data typified by multitask, multithread, and multiprocess, a general MMU is used.
In a system in which a flash memory is managed for each block (page) by making full use of a (memory management unit), it is difficult to manage the number of rewrites and reduce the number of rewrites.

(2) An auxiliary storage device (semiconductor device) is a device having a control structure for reducing the number of times of rewriting of the flash memory, although there is no device capable of executing a program on an extended storage device using a flash memory. File storage device). In this method, information (rewrite number table) is stored in a different memory in order to reduce the number of rewrites in the auxiliary storage device. However, in a program that operates on a flash memory without loading an execution program into a main storage device, the method of simply storing the same program on the same flash memory completely and continuously stores data across blocks. It is difficult to achieve a realistic arrangement.

The present invention focuses on the above points,
An object of the present invention is to control a program on an extended storage device without loading an execution program in a main storage device when a control structure for reducing the number of times of rewriting of the flash memory on the extended storage device using the flash memory is provided. Is to be executed. Another object is to use the same flash memory and have a control structure to reduce the number of rewrites as an auxiliary storage device, and also to use the memory management device in an advanced program execution environment such as multi-processing, multi-thread, and multi-task. Another object of the present invention is to make it possible to safely manage the modules and data files of an execution program. Still another object is to provide a file management technique that enables an extended storage device and an auxiliary storage device to coexist on the same flash memory.

[0008]

A flash memory according to the present invention is divided into a block of a flash memory in which storage contents can be erased in block units, a header area for storing management information of the block, and a block unit. And a storage area for storing a data file. Alternatively, a flash memory capable of erasing stored contents in block units includes first, second, and third blocks, and a first block includes a header area for storing management information of the block, and a block unit. The second block has a storage area for storing an execution program file divided in block units, and the third block has an execution area for storing an execution program file divided in block units. It is characterized by comprising a header area for storing management information of a block in which a program file is stored, and a storage area for storing an execution program file divided in block units.

According to one of the main modes, the management information of the block includes (1) information indicating whether a data file or an execution program file is stored in the block, and (2) erasure for the block. Count, (3)
At least one of an erase sequence number indicating the order in which blocks were erased, (4) time information of block erasure, and (5) information indicating whether the block is in a valid / transition / invalid / initialization state. One is included.

The memory management apparatus according to the present invention is characterized in that a data file divided in block units is assigned to each block, and means for storing management information of each block in the corresponding block is provided. Alternatively, there is provided means for allocating an execution program file to a plurality of adjacent second and third blocks and storing management information of a block to which the execution program file is allocated in the third block. It is characterized by.

According to one of the main modes, when selecting a candidate block for erasing a block, (1) if the number of erasures is the same, the order of erasure based on the erase sequence number Means for preferentially selecting a block, and (2) means for preferentially selecting a block having an earlier erased block based on the time information if the number of erases is the same. It is characterized by the following.

According to another aspect, at least one of (1) means for reducing the number of rewrites of data stored in each block, and (2) means for restoring the system and restoring data is provided. It is characterized by having.

[0013] A storage device according to the present invention includes any one of the flash memories and any one of the management devices. The computer system of the present invention
The storage device is provided, and the execution program file is operated on a flash memory.

The above and other objects, features, and advantages of the present invention will be apparent from the following detailed description and the accompanying drawings.

[0015]

Embodiments of the present invention will be described below in detail. In this embodiment, as shown in FIG. 1, for example, an extended or auxiliary storage device including a processor 10, a flash memory 12, and a control device 14,
The present invention is applied to a computer system including an M (main storage) 16. In such a system, using the same flash memory, the data of the execution program module and the data of the data file coexist, and a data storage method for reducing the number of times of rewriting data as much as possible. It is intended to obtain a device for management and control.

First, in order to manage the execution program area and the data area, the management unit of the flash memory 12 is one block which is the minimum erase (erase) unit. In this block, which is the minimum unit, no mixture of the execution program area and the data area is allowed. Each block has an execution program area or data area and a header area having information for managing the block.

Here, the block to be allocated to the data area is declared to have the data attribute in the header area of the block. Similarly, a block to be allocated to the execution program area is declared to have the attribute of the execution program in the header area of the block. However, regarding this execution program area, 1
One header area is not necessarily required for one block. That is, when it is desired to continuously secure a plurality of linked blocks as an execution program area, the number of blocks to be allocated is declared in the header area. by this,
Declared chained blocks need not have header information. That is, as the memory arrangement of the execution program, the execution program area is secured in a plurality of continuous blocks by one header area. On the other hand, the data area is 1
One block is mixed with the header area. In the data area, information on the corresponding block is added.

FIG. 2 shows an example of the basic structure of software in this embodiment. As shown in FIG.
0 and socket interface with application software 22
As a face, a flash management manager 26 is provided in a portion corresponding to the so-called driver layer 24. The flash management manager 26 can be considered as a system driver of the present embodiment. This flash management manager
The management of block information and the reading of data
De / write control is performed.

For example, as shown in FIG. 3A, block erasing is possible in units of i bytes and the number of blocks is n +
It is assumed that there is one (hexadecimal) number of flash memories. Each of these blocks will be referred to as a physical block. Then, a physical block number is assigned to each physical block from "0" to "n" in ascending order of address. In the illustrated example, the upper part of the figure has a lower address, and is a physical block 0, a physical block 1, a physical block 2, ..., a physical block n.

Here, if it is assumed that j bytes are secured as a header area of information for managing each physical block, the size of the execution program area or data file area allocated to each physical block is ij
It becomes bytes. And, in the storage area of the same block,
It is not allowed to mix execution program areas and data file areas. FIG. 3B shows this state. For example, the top physical block 0 has a header area 30A of j bytes and a storage area 30 of ij bytes.
B exists. In the next physical block 1, a j-byte header area 31A and an ij-byte storage area 31B exist. The same applies to the following physical blocks.

Next, an example of securing an execution program area will be described with reference to FIG. In FIG. 4A, eight consecutive physical blocks 0 to 7 are allocated as an execution program area, and the other physical blocks 8 to n are data.
This is an example of allocation as a data file area. As described above, the execution program area secured in a plurality of continuous physical blocks is managed by one header area. On the other hand, the data file area is managed by the header area for each block even if it is secured in a plurality of continuous physical blocks. Therefore, in the example of FIG. 4A, the physical block header 3
2A exists only in physical block 0, and physical blocks 1 to
7 does not exist. Then, as described later, the management information and the like of the physical blocks 1 to 7 are collected in the header area 32A of the physical block 0. On the other hand, the data file areas 38B to 3nB are secured in the physical blocks 8 to n, respectively.
8A to 3nA, which are managed for each block.

Next, the example of FIG.
Is assigned as a data file area, and physical block 1
.. A are assigned as execution program areas, and thereafter are assigned as data file areas. FIG.
In the example of [C], physical blocks 0 to n-9 are allocated to a data file area, physical blocks n-8 to n-3 are allocated to an execution program area, and the subsequent blocks are allocated to a data file area. . In this manner, the continuously secured execution program area is managed by one header area. There is no restriction on the number of physical blocks or fixed address arrangement. However, even if the data file area is continuously secured, it is managed for each block.

In the example shown in FIG. 5, the execution program area is divided and arranged. First, in the example of FIG.
The execution program area (1) is allocated to the physical blocks 0 to 7, and the data file area is allocated to the physical block 8. Then, physical blocks 9 to F-
The execution program (2) is allocated to the data block 1, and the data after the physical block F are reserved for the data file area. Also in this example, a continuous execution program area is managed by one header area.

In FIG. 5B, physical block 0 is allocated for the data file area, and physical block 1 is allocated.
8 are assigned to the execution program area (1).
The physical blocks 9 to n-9 are allocated for the data file area, and the remaining blocks are allocated as the execution program area (2). FIG.
In [C], physical blocks 0 and 1 are allocated as data file areas, and physical blocks 2 to 4 are allocated to an execution program area (1). The subsequent physical blocks 5 to n-9 are allocated as data file areas, and the physical blocks n-8 to n-1 are allocated.
Is allocated as the execution program area (2),
The last physical block n is allocated to the data file area. FIG. 5 shows an example in which two areas (1) and (2) are designated as execution program areas. However, in the present embodiment, the execution program area can be further arranged in a plurality of areas.

Next, the header area of the physical block will be described. The physical block header has an execution program area and a data file area. In order to determine these two types of areas, it is desirable that the header has a common part structure. Therefore, the structure of the common part of the physical block header is assigned, for example, as follows.

(1) Flag indicating status (valid / invalid / transitioning / initialization completed) (2) Marker flag indicating execution program area / data file area (3) Number of consecutive reserved block areas ( 4) Number of block erases (5) Number indicating the order of block erase (erase
(6) Block initialization time (7) Block area name

First, the flag (1) will be described. Flash management manager 26 (see FIG. 2)
Reads the contents of each physical block header and grasps how the flash memory is managed. Bit data in a memory cell of a flash memory has a relatively short processing time for changing a logical value from "1" to "0".
The change from "0" to "1" requires a block erase or erase once, which takes a long time. That is, in order to rewrite 1-bit data, it is necessary to temporarily save the data in the block and write it back after the block erase, which is a long processing time.

Therefore, in practice, the state of the physical block is managed using the flag indicating the state. That is, in order to reduce the overhead associated with the above-described rewriting, the physical block header contains a valid state, an invalid state, an initialized state, or a valid state in the physical block header. State is recognized for each physical block, and operations such as writing data, invalidating data, erasing a physical block, and initializing a physical block are performed. .

Next, the marker flag (2) is a flag indicating whether the corresponding physical block has been allocated to the data file area or the execution program area. Thus, when the execution program area is designated, the number of blocks following the execution program is designated to the number of consecutive reserved blocks in (3). However, when it is designated in the data file area, the number of consecutive reserved blocks is set to zero.

Next, the number of erasures (block erase) of each block is recorded as the number of block erasures in (4). When performing a block erase, the previously recorded erase count is temporarily held, and after the block erase is completed, the temporarily erased count is incremented by 1 and data is written. Store.

Next, in (5), the numbers indicating the order of erasure of the blocks are given in any order at the same time.
This is for determining whether or not the operation has been performed. A block erase is performed on a certain physical block, and if another physical block has the same block erase count, the erase sequence number of the block with the erase count is incremented by one to erase the corresponding block header. -Store it in the sequence number. That is, when there are a plurality of blocks having the same number of erasures, the value of the largest erase sequence number among them is incremented by 1 and stored in the erase sequence number in the corresponding block header.

Next, as the time at which the block is initialized in (6), the time at which the corresponding physical block was erased is entered. This is not necessary for a system without a real-time clock.

Next, the block area name (7) is a unique name assigned to the block in the execution program or data file.

When allocating a new block, the above-mentioned management information is used to manage the entire block efficiently, that is, so as not to be biased to a specific block and to make the number of erasures uniform for all blocks. Used. First, when allocating a new physical block, the flash management manager 26 preferentially allocates a physical block with a smaller number of erases. If there is a block having the same number of erasures, the block with the smallest erase sequence number is assigned. The youngest elephant
The block with the sequence number has the oldest erased time. In a system having a real-time clock, the clock is used instead of the erase sequence number. That is, when there is a block having the same number of erasures, the block having the oldest time in the physical block among the blocks is allocated. In this way,
If one of the sequence number and the real-time clock is used, the block allocation can be evenly and efficiently equalized.

The execution program is located in the area after the physical block header and in the area.
By arranging the data in the same manner as in the case of the ROM. This is because the part that depends on the OS 20 is large.
This will be addressed by conventional methods.

On the other hand, the data file is allocated by the flash management manager 26. As shown in FIG. 6A, the data file has a physical block header 4
The data is stored in the storage area 40B excluding the area of 0A. As a data storage structure, a logical data record in a data file is blocked, and as shown in FIG. 6B, a data record header 42A, a data record
The data record 42B is attached in the form of attaching the data footer 42C.
Of the packet 42. Blocking methods include fixed-length records, variable-length records, and spanned records.
In this embodiment, any method may be used. Since this blocking often depends on the OS 20 and the application software 22, the flash management manager 26 determines and controls the blocking method in the system.

The data record header 42A and the footer 42C have a linear list structure.
The optimum use of the block area name differs depending on the data recording method. When designing the system, partition names, data file names,
Assign the best one, such as a directory name. The optimum block area name often depends on the data recording method.

The logical record 42B of the data file is composed of the data record packet 42 as described above.
Is stored in the data file area 40B of the allocated physical block. FIG. 6B shows this state. The data record packets (1) to (n) are i-
It is stored in the j-byte data file area 40B.
The flash management manager 26 performs dual management of management of the data record packet 42 and management by the physical block 40B.

Next, the procedure for initializing a physical block will be described. If the flash memory has not been initialized, the flash management manager 26 first initializes each physical block. In the case of initialization, all blocks are erased once, and the initialization completion flag, the number of erasures (once), and the erase are added to the block header of each physical block.
Information such as a sequence number or an initialization time is stored. Next, usually, an area of each physical block is declared (secured) in order to store an execution program and a data file. That is, it stores a marker flag indicating the execution program area and the data file area, and information on the number of areas of the continuous reserved block. At this time, a block area name may be assigned. If the area name of the block is to be assigned later, the flash management manager 26 takes care of the assignment.

Next, the procedure for storing a data file will be described. Turn on the valid flag of the physical block header
And then record the blocked data record.
Are sequentially stored in the data area. The data record is added to the data area, but when it is desired to rewrite the data record, rewriting that portion is a large overhead due to the characteristics of the flash memory.

In order to store data from the old logical data record to the new logical data record, the old data record is invalidated and a new data record is added. Do it by the way. The flash management manager 26 can determine that an invalid data record is invalid by providing an invalid flag in the data record header. When all of the data record packets in the data area of the physical block are invalidated, the flash management manager 26 turns on the invalid flag in the physical block header to turn this physical block on. Block erase can be performed. When the block erase is completed, the above-mentioned block header is initialized.

Next, a procedure for storing an execution program will be described. When a plurality of blocks are initially reserved as the execution program area, first, each information such as a valid flag in the physical block header, a marker flag indicating the execution program area, and the number of reserved area blocks is stored. As a result, the ROMized data of the execution program is stored in the execution program area.

The above is operated as a system, and as a result of the operation, the number of rewrites of the data file area is compared with the number of rewrites of the execution program area. If there is a large number of gaps, it is desirable to change the currently arranged area and perform defragmentation. Since the flash management manager 26 manages the respective areas, the flash management manager 26 can cope with the dynamic arrangement in the defragmentation operation, and there is no fear of causing a contradiction as a system. By performing these operations, an efficient area arrangement can be realized. Further, as a result, the operation time can be extended, and the rewrite life (time, not the number of times) of the flash memory can be extended.
It should be noted that some flash memories do not allow other operations during a write operation. In such a case, exclusive control is performed by a programming technique such as a task or a thread.

FIG. 7 shows an example of storing a physical block header in a flash memory. The example in FIG. 7A is a flash memory in which 64 Kbytes are used as an erase block unit, and the system has n + 1 blocks. At the head of each physical block, as shown in FIG. 7B, 256 bytes are allocated as a header area which is management information of each physical block. As described above, in the case of a data file, a physical block header is always attached to a corresponding physical block. However, this is not the case for executable program files. Hereinafter, examples of the present embodiment will be described.

(1) Embodiment 1 Embodiment 1 is an example in which a block sequence number is used. In the file management device of this example, as shown in FIG. 8A, block erasing is possible in units of 64 Kbytes.
An M × 8 bit flash memory is used. FIG.
As shown in [A], “0” to “1F” are assigned to each block.
And number them in that order. If the execution program area is secured as a continuous area as shown in FIG. 5A for such a flash memory, the result is as shown in FIG. 8B. Naturally, there is no physical block header in the physical blocks 1 to 7 designated as the area in the execution program area (1). Similarly, the physical blocks 10 to E designated as areas in the execution program (2) have no physical block header.

Therefore, in this embodiment, as shown in FIG. 9, the first to eighteenth bytes have a common data structure. Thus, if this common part is scanned by the flash management manager 26, the entire structure can be grasped. Then, the contents assigned to the execution program area or the data file area may be added from the next 19th byte.

The contents of the common part will be described in order.
In the byte, as shown in FIG. 10A, there are flags of valid / in transition / invalid, initialization completed state, and reservation. Among these, the flag of valid / in transition / invalid indicates whether the block area is in a valid state, in a transition (transition period) from valid to invalid, or in a completely invalid state. The flag indicating that the initialization has been completed indicates that the initialization (format) of the block header is completed after the block is erased. The reservation flag is a spare flag to be used when expanding the system. An example of actual bit allocation is shown in Table 1 [A].

[0048]

[Table 1]

Next, the procedure of block erase will be described with reference to FIG. Flash Management Manager 2
When the block erase instruction is issued (50 in FIG. 10B) as a result of the determination in step 6, the number of erasures of the old block header and the like are temporarily stored as the next state (52), and the block erase of this block is performed. (54). Immediately after performing the block erase, the data values in the storage area of the erased block are all FFh (5).
6). Immediately after this, bit 4 is set to 0 by the initialization (format) operation (58) of the flash management manager 26, and the data value of this storage area transits from FFh to EFh (60).

Thereafter, a value obtained by adding +1 to the number of erasures temporarily stored before erasing is written in the header of this block, and the physical block is initialized. If it is assumed that the flash management manager 26 has newly allocated the block in response to a use request for this block (62),
Write 0 to bit 7 and the value of the data in this area is E
A transition is made from Fh to 6Fh (64). Further, when the flash management manager 26 recognizes the invalidation request of this block (66), 0 is written to the bit 6, and the data value of this area transits from 6Fh to 2Fh (68).

When performing the defragmentation, the function of the flash management manager 26 causes the physical block to have a transition state such as a copy source, a copy destination, and a move. To change from a valid state in a physical block as a data file area to a transitioning state (7
0), bit 5 is set to 0, and the data value of this area is 6
The transition may be made from Fh to 4Fh (72). Existence of this transitioning state means that the time required for the transition is not short, and that the operation becomes unstable due to battery exhaustion,
This is to enhance the security when the card type flash memory is removed. That is, in the present invention, the state of this transition stage is managed, and the loss of data due to a failure or accident is minimized and restored.

Next, the second byte of the common part shown in FIG. 9 will be described. In the second byte, as shown in FIG. 9, there are flags for the program / data and the number of continuous reserved block areas, as shown in FIG. 10 [C] and Table 1 [C]. Among them, the program / data marker flag is a flag for determining which of the execution program and the data file is assigned to the block. When the marker flag is 0, this block has been designated in the execution program area.

The flag of the number of continuous reserved block areas indicates the number of remaining blocks designated in the execution program area in hexadecimal. However, if the marker flag is 1
When this physical block is allocated as a data file area, the number of consecutive reserved blocks is 000000.
Write 0b.

Next, the third to fifth bytes in FIG.
As shown in [D], the number of erasures of the block is expressed in hexadecimal. The third byte is the lower digit of the number, 4
In the byte, the data is stored as the middle digit of the number, and in the fifth byte, the data is stored as the upper digit of the number. Therefore, FFFFFFh times is the maximum value of the block continuation. At this maximum value, no further increment is performed.

Next, the sixth byte in FIG. 9 is prepared for the flash management manager 26 to allocate an optimum physical block when the number of erasures of each physical block is the same. This is the erase sequence number described above. The erase sequence number records the order in which the physical blocks were initialized when the number of erasures of the physical block is the same as that of the other physical blocks. As a flow of the allocation of the erase sequence number, first, if there is no identical number of erasures in the corresponding physical block and other blocks,
The erase sequence number of the corresponding physical block is 0
x00.

If there is another block having the same number of erasures, the largest erase sequence number is detected, and the erase sequence number of the corresponding block is detected as the erase sequence number previously detected. The value obtained by incrementing the value of the kens number by 1 is recorded. The erase sequence number is assigned by such an algorithm. The erase sequence number is used as a criterion when the flash management manager 26 allocates a physical block.

Next, from the seventh byte to the eighteenth byte in FIG. 9, the area name of the corresponding block is recorded. When an executable program file is used for a partition name, a directory name, or the like, system operation is facilitated.
Further, in the case of a data file, it is conceivable to treat the block area name optimally according to the data record method. For example, when the data record method double-manages files such as fixed length, variable length record, and spanned record, a partition name, a directory name and the like are suitable. However, in the case where the files are centrally managed with fixed-length or variable-length records, it is desirable to treat them as data file names. In this embodiment, the block area name is in the 8.3 format, and uses an expression method such as MS-DOS.

The flash file management system is realized by using the physical block header having the above structure. The flash management manager 26 includes the OS 20 and the application software.
In response to a request from the application software 22 or the like, reading and writing of an execution program and data files are controlled.
First, at the time of writing, a characteristic of the flash memory is the influence of a write delay. For this reason, in order to perform writing efficiently with a reduced overhead, a method of additionally writing data in another data file area is adopted. To execute this method, the old data before rewriting and the new data after rewriting are logically changed, but the old data becomes physically invalid and the new data is changed. Data will be added.

That is, although the old data and the new data coexist physically, the old data is logically deleted by writing an invalid flag in the old data. If the flash file system repeats such an operation, there will be many logical invalid data. In some cases,
There is even the danger that the new data file area will be lost.
To avoid this, flash management manager 2
6 periodically detects invalidation of data and performs a defragmentation operation to efficiently secure a data file area.

The allocation of a physical block as a data file area is performed by the flash management manager 26. At this time, the OS 20 and the application software 2
Although the data file area is given by the request of No. 2, in the algorithm of this embodiment, it is preferentially assigned to a physical block with a small number of erasures. Also, when performing a defragmentation operation, for example, the data record packet is invalidated preferentially for a block having a small number of erasures, and the entire data file area of the physical block is invalidated. In addition, the number of times the blocks are used can be made uniform. In order to efficiently equalize the number of use of the physical block even in a small number of them, the physical block and the block erase to be allocated are managed by giving priority to the one having a smaller erase sequence number.

As the flash file system operates, as a result, the number of erasures of the execution program area and the data file area greatly differs. Then, when the number of times of erasure is increased to some extent, the flash management manager 26 replaces the physical block used in the execution program area with the physical block used in the data file area. The service life of the device is extended.

In the case of reading, the flash management manager
The controller 26 requests that the flash memory be read. After confirming permission, the blocked data record is read.
This is made possible by treating a data packet as a logical data record. In the case of writing, a request is made to the flash management manager 26 to perform writing to the flash memory, and after confirming permission, the logical data record is blocked, and the data record is packetized and designated. Store in area. Rewriting of data is realized by physically invalidating a data record packet and newly adding data. However, the memory area is physically reduced. The method of blocking these data is as described above, and is realized by the background art.

When all data is invalidated in the data file area of the physical block, the entire physical block is invalidated by setting the invalid flag of the physical block header to ON. Then, the flash management manager 26 detects this, erases the block, and starts the initialization operation. In addition, by the defragmentation operation, the transition flag of the physical block header to which the copy source data file belongs is turned ON.
To When all necessary data has been copied, the invalid flag of the physical block header is turned ON, and
The above-described block erase / initialization operation is performed. There is an abnormal state for some reason between the transition flag ON and the invalid flag ON. When the system is restarted, monitor and analyze these state flags to stop from any state. Alternatively, it can be detected whether an abnormality has occurred. And it is a valuable source of information for repairing and restoring the system.

The flash management manager 26 monitors and manages the state of each physical block. A corresponding physical block is allocated in response to a request from the OS 20 or the application software 22, and valid → (in transition) → invalid → monitoring and management of the state of initialization are executed. And when you operate the flash file system,
Changes in the state of each physical block include: initialization → valid, valid → invalid, valid → transit, migrating → invalid, invalid → initialize,
Initialization → invalid.

[Initialization → valid] means that when a physical block is allocated from a initialized state to a data file area or the like, the valid state is turned on by a valid flag ON.

[Valid → Invalid] means that the physical block is valid as a data file and at least one logically valid data exists in the data file area. The data becomes logically invalid, and the flash management manager 26 then turns on the invalid flag of this physical block to invalidate the physical block.

[Valid → Transitioning] means that when at least one logically valid data record packet is copied to the data file area of a different physical block, the physical The transition flag is turned on for the block. In this state, the data record packet has been copied.

[Transitioning → Invalid] indicates that the above-mentioned transitional state has ended and the physical block has been invalidated. That is, the copying of the data record packet is completed, and the flash management manager 26 turns on the invalid flag.

[Invalid → Initialize] means that the physical block is erased from the invalid state, and when the physical block header is initialized, the flash management manager 26 sets the initialization flag. Turn ON. At this time, the initialization of the physical block header is completed, and the physical block header can be used as a data file area.

[Initialization → Invalid] indicates a state in which the physical block has been invalidated from the state in which the physical block has been initialized. This may exist as a state, but should not be overused by the system.

The flash management manager 26 manages and monitors the state of each physical block, and after escaping from an abnormal state due to any cause, sets the valid / invalid / transitioning / initialization flag to the data before the abnormal state. Data source for restoring and restoring data. That is, by considering these change states, it is possible to detect from which state the state becomes abnormal when the system is stopped or the system in which the abnormality has occurred is restored or the data record packet is restored.

(2) Embodiment 2 This embodiment 2 is a method of recording an initialization time in place of the erase sequence number of the embodiment 1. That is, as shown in FIG.
The year / month / day / hour / minute / second that was initialized from the 9th byte to the 21st byte is recorded. The erase sequence number in the first embodiment indicates the order of initialization when the number of erasures is the same. However, in the second embodiment, in the algorithm for equalizing the number of use of blocks, In order to improve the number of times of use even slightly, if the number of times of erasure is the same, the oldest time is prioritized to manage the physical blocks and block erase to be allocated.

(3) Embodiment 3 Embodiment 3 is a method of combining Embodiment 1 and Embodiment 2 and recording both the erase sequence number and the initialization time. As shown in FIG. 12, the erase sequence number is recorded in the sixth byte, and the year / month / day / hour / minute / second of initialization is recorded in the 19th to 21st bytes. Almost the same as the first and second embodiments, but in the algorithm for equalizing the number of use of the block,
In order to improve the number of times of use of all the physical blocks as efficiently as possible, when the number of times of erasure is the same, priority is given to the erase sequence number having a smaller number. Also, when there is a physical block with the oldest initialization time, if it can be determined that the number of erasures is more efficient than others, the physical block or block erase to be allocated is performed by an algorithm that preferentially allocates the physical block.
Is managed.

There are many embodiments of the present invention, and various modifications can be made based on the above disclosure.
For example, the following is also included. (1) In the above embodiment, the present invention is applied to a flash memory. However, another similar memory has the same write / read function as the flash memory, and has a block erase characteristic before writing. If so, it is equally applicable. (2) The physical block which is the erase unit of the flash memory can be similarly applied to any data unit other than the byte unit and the word unit.

[0075]

As described above, according to the present invention, the following effects can be obtained. (1) On an extended storage device using flash memory,
It has a control structure to reduce the number of rewrites of the flash memory, and loads the execution program into the main memory.
The program can be executed on the flash memory without loading. Further, even with an auxiliary storage device having a control structure for reducing the number of rewrites, storage can be performed with a data structure with little waste.

(2) Since an area for an information table such as the number of times of rewriting of each block is provided in the block, there is no need to provide a separate memory, and the cost can be reduced. In particular, the extended storage device and the auxiliary storage device can be mixed on one flash memory, and the space and cost of the system can be reduced.

(3) Valid / in transition / invalid /
Having a status flag of "initialization" is effective for restoring from a stop due to an abnormality in the computer system and restoring data.

[Brief description of the drawings]

FIG. 1 is a diagram showing a basic configuration of a system using one embodiment of the present invention.

FIG. 2 is a diagram illustrating a software structure according to one embodiment.

FIG. 3 is a diagram showing a memory map in which physical blocks are allocated in a flash memory according to one embodiment.

FIG. 4 is a diagram showing an example in which one execution program area and a plurality of data file areas are allocated to physical blocks.

FIG. 5 is a diagram showing an example in which a plurality of execution program areas and a plurality of data file areas are allocated to physical blocks.

FIG. 6 is a diagram showing an example of a data record packet in a data file area and an example of a storage method.

FIG. 7 is a diagram showing an example in which a flash memory having a block erase of 64 Kbytes is divided into physical blocks to secure a block header of 256 bytes.

FIG. 8 is a diagram showing an example of a memory map obtained by dividing a flash memory used in the first embodiment into an execution program area and a data file area.

FIG. 9 is a diagram illustrating an example of a memory map of a physical block header used in the first embodiment.

FIG. 10 is a diagram illustrating a part of a physical block header and a sequence according to the first embodiment.

FIG. 11 is a diagram illustrating an example of a memory map of a physical block header used in the second embodiment.

FIG. 12 is a diagram illustrating an example of a memory map of a physical block header used in the second embodiment.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 10 ... Processor 12 ... Flash memory 14 ... Flash memory control device 16 ... RAM 20 ... Operating system 22 ... Application software 24 ... Driver layer 26 ... Flash management manager 30A, 31A, 38A, 39A, 3 (n-1) A, 3
nA, 40A header area 30B, 31B, 38B, 39B, 3 (n-1) B, 3
nB, 40B storage area 32 execution program area 42 data record packet 42A data record header 42B data record 42C data record footer

Claims (15)

[Claims] 1. A block of a flash memory capable of erasing storage contents in units of blocks, comprising a header area for storing management information of the blocks and a storage area for storing data files divided in units of blocks. Flash type memory characterized by the above-mentioned. 2. A flash memory capable of erasing storage contents in block units includes first, second, and third blocks, and a first block includes a header area for storing management information of the block. , A storage area for storing a data file divided in units of blocks, a second block having a storage area for storing an execution program file divided in units of blocks, and a third block. A flash memory comprising: a header area for storing management information of a block in which an execution program file is stored; and a storage area for storing an execution program file divided in block units. 3. The flash memory according to claim 1, wherein the block management information includes information indicating which of a data file and an execution program file is stored in the block. 4. The flash memory according to claim 1, wherein the management information of the block includes an erase count for the block. 5. The flash memory according to claim 4, wherein the block management information includes an erase sequence number indicating an order of erasing the blocks. 6. The flash memory according to claim 4, wherein the management information of the block includes time information at which the block was erased. 7. The management information of the block includes information indicating whether the block is in a valid / transition / invalid / initialized state. 7. The flash memory according to any one of 4, 5, and 6. 8. The flash-type flash memory according to claim 1, further comprising means for allocating a data file divided in block units to each block and storing management information of each block in the corresponding block. Memory management device. 9. A means for allocating an execution program file to a plurality of adjacent second and third blocks and storing management information of a block to which the execution program file is allocated in the third block. The flash memory management device according to claim 2, further comprising: 10. A means for preferentially selecting blocks in the order of erasure based on the erase sequence number if the number of erasures is the same when selecting a candidate block for erasing the block. 6. The flash memory management device according to claim 5, wherein: 11. A means for preferentially selecting a block having an earlier erased block time based on said time information if the number of erases is the same when selecting a candidate block for erasing the block. 7. The flash memory management device according to claim 6, further comprising: 12. The flash memory management device according to claim 8, further comprising means for reducing the number of times data stored in each block is rewritten. 13. The system according to claim 8, further comprising means for restoring the system or restoring data.
13. The flash memory management device according to any one of 11 and 12. 14. The flash memory according to claim 1, 2, 3, 4, 5, 6, or 7, and the flash memory according to claim 8, 9, 10, 11, 12, or 13. A storage device, comprising: the management device according to any one of the preceding claims. 15. A computer system comprising the storage device according to claim 14, wherein the execution program file is operated on a flash memory.