JP2008047155A - Batch erasable nonvolatile memory and mobile phone - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To obtain a batch erasable nonvolatile memory capable of storing data written with high frequency. <P>SOLUTION: The memory includes an address information storing means for storing address information about blank areas in a cluster consisting of a plurality of sectors; a data write means for executing, when receiving a write request of data, processing of writing the data into at least one of the blank areas referring to the address information about the blank areas; and an address information update means for updating the address information associated with the at least one of the blank areas after the data write means completes execution of writing the data. When the sector includes a plurality of data blocks, the data write means refers to a used state flag which is stored in a header area of the sector for indicating a used state of the data block, and executes the process of writing the data into a blank area of the data block. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a batch erasing nonvolatile memory such as a flash EEPROM (hereinafter referred to as a flash memory) for recording nonvolatile data and a mobile phone.

  In current embedded systems such as mobile phones, when storing system programs, a flash memory that can be randomly accessed is used, while data that is frequently rewritten or non-volatile data that has a relatively small amount of data is recorded. In this case, an EEPROM is used, and when a large amount of nonvolatile data is recorded, a flash memory by serial writing or an SRAM with a backup power source is used.

  Here, the EEPROM has a larger number of storage elements per bit than the flash memory, and the unit price per bit increases. In addition, the SRAM with a backup power supply increases the cost because it is powered. By storing the data stored separately in the flash memory, there is an advantage that the mounting area and cost of the device are reduced.

  The conventional flash memory in the embedded system cannot read other system programs during the execution of the system program writing process, and writes the program by performing a special process when writing the system program. However, recently, with the advent of a batch erasable nonvolatile memory (hereinafter referred to as a BGO flash memory) that can read data from other areas where data has not been written during the execution of data writing processing, It is now possible to load the flash memory that stores the program into the flash memory that stores the program.

  As a use example of storing data in a flash memory, there is a semiconductor disk as shown in Japanese Patent Publication No. 7-50558 (see FIG. 14). In this system, a flash memory 4 is used as a semiconductor disk, and a control device 1, a RAM 2, a bus control unit 3, and an internal bus 5 are provided to control the semiconductor disk. There are limitations that SRAM does not have.

  That is, data writing is one-way from 0 to 1 or 1 to 0. For this reason, when rewriting to a place where writing has been completed once, it is necessary to erase the entire block including the place to be written at once and set the entire block to 0 or 1 before writing data. For this reason, it is difficult to perform writing in byte units as in EEPROM and SRAM.

  Since the conventional batch erasing type nonvolatile memory is configured as described above, if the entire block including the place where the data writing process is executed is batch erased, the data is written again to the same location. However, the number of erasable times of the flash memory is only guaranteed about one-tenth that of an EEPROM, and there is a problem that it is difficult to cope with storing data with a high rewrite frequency. It was.

  The present invention has been made to solve the above-described problems, and an object of the present invention is to obtain a collective erasing nonvolatile memory that can cope with storage of data with high rewrite frequency.

  In order to solve the above-described problems and achieve the object, the present invention receives address information storage means for storing address information of unused areas in a cluster composed of a plurality of sectors, and a data write request. Referring to the address information of the unused area, the data writing means for executing the data writing process in the unused area, and when the data writing means executes the data writing process, the data writing means Address information updating means for updating address information of the use area, and when the sector is composed of a plurality of data blocks, the data writing means uses the data block stored in the header area of the sector A data write process is executed in an unused area of the data block with reference to the use status flag indicating “.”

  As described above, according to the present invention, when a sector is composed of a plurality of data blocks, the data is referred to the usage status flag indicating the usage status of the data block stored in the header area of the sector. Since the data writing process is executed in the unused area of the block, there is an effect that up to 254 types of data can be freely rewritten in units of 1 to 8 bytes.

An embodiment of the present invention will be described below.
Embodiment 1 FIG.
FIG. 1 is a block diagram showing a batch erase type nonvolatile memory according to Embodiment 1 of the present invention. In FIG. 1, reference numeral 11 denotes a BGO flash memory (address information storage means) composed of clusters having a plurality of sectors. In the first sector of the BGO flash memory 11, address information of unused areas is stored. Reference numeral 12 denotes a microcomputer that controls the BGO flash memory 11, reference numeral 13 denotes a RAM of the microcomputer 12, and reference numeral 14 denotes a CPU of the microcomputer 12. The CPU 14 stores address information of unused areas in the head sector of the BGO flash memory 11. When the address information storage means receives the data write request, the address information storage means refers to the address information of the unused area in the cluster, and the data writing means executes data write processing in the unused area. Address information updating means for updating the address information of the used area, ID management means for registering an ID in each cluster, and pointer construction means are configured. 1 is connected to peripheral devices such as a mobile phone using the system.

  Next, the operation will be described. First, the physical address, cluster management area, and application area of the BGO flash memory 11 will be described with reference to FIG. However, it is assumed that the BGO flash memory 11 becomes FFh when data is erased, and the data rewrite direction is 1 to 0.

  The cluster ID of the cluster management area is expressed by 1 byte and can be expressed in 254 ways excluding 00h and FFh (a unique ID is registered in the cluster used by the application). Also, since FFh is in an initial state, FFh is an unused cluster ID (FFh indicates a writable cluster), 00h is an invalid cluster ID (00h indicates a non-writable cluster), and Become. The same value is stored in the three cluster IDs in FIG. 2, and when two or more cluster IDs match, it is determined that the cluster ID is valid.

  Next to the three cluster IDs, 1 byte of the copy flag is secured, and the area excluding the 4 bytes of the cluster management area of the erase block becomes the application area. This application area is an area that an application can use freely.

  Each application does not directly hold an address where its own data is stored, but holds its own application ID. As a restriction, this application ID must be unique for all applications using the system. When an application actually uses the BGO flash memory 11, it searches for a cluster ID that matches its own application ID, searches for a physical address used by itself, and further determines the data that it wants to read / write. Calculate the address.

  Since the BGO flash memory 11 needs to erase data in units of blocks, when erasing only an area in the block, it is necessary to save necessary data, erase the block, and write necessary data. Saving the necessary data to the RAM 13 is not desirable in terms of cost, meaning of using the BGO flash memory 11, and data reliability. For this reason, data is inevitably saved on the BGO flash memory 11, but for this purpose, an unused area in which no data is written, that is, a blank area is required. The above operation is called data reclaim and is a process that cannot be avoided in the BGO flash memory 11.

  Here, processing at the time of data reclaim will be described with reference to FIGS. 3 and 4. For example, when data reclaim for cluster A becomes necessary (step ST1), the application of cluster A searches for a blank area on the BGO flash memory 11 (step ST2). For example, if the cluster D is a blank area, the address of the blank area is acquired with reference to the cluster ID of the cluster D (step ST4). However, if there is no blank area, the invalid cluster is erased collectively to create a blank area (step ST3), and the address of the blank area is acquired (step ST4).

  When the address of the blank area is acquired, the application of cluster A sets the copy flag of cluster A currently in use to 01h (step ST5), and its own application ID (ID of 1h) is assigned to the blank area (cluster D). ) Is registered (step ST6). (Refer to the middle of Figure 4)

  When the application of cluster A registers its own application ID in the blank area (cluster D), the necessary data is copied to the blank area (step ST7). When the copying is completed, the cluster A which is the old area is erased or invalidated. And the process ends (steps ST8 and ST9). (See the lower part of Fig. 4)

  However, if a failure such as power interruption occurs during data reclaim, a cluster with the same cluster ID may exist. In such a case, there is a restriction that the cluster ID assigned to the application cannot be duplicated, so it is necessary to remove it when the system is started.

  For this reason, the presence of duplicate clusters is checked when the system is activated. If duplicate clusters are detected, those having a copy flag of FFh among the duplicate clusters are deleted. The reason is that it is difficult to determine how far the data reclaim processing has been completed if a system failure occurs before batch erasure or data invalidation. Therefore, the copy source is determined by looking at the copy flag, and the copy destination is deleted. This is because if the copy source is held, it is possible to avoid a problem that the data contents are broken and cannot be written.

  Next, when a device error (write error) occurs during data writing, it is necessary to execute data writing processing again. In order to cope with such a case, two or more blank areas are prepared. When a data write error occurs, the cluster ID in which the write error has occurred is changed to an invalid cluster, a new blank cluster is acquired, and reclaim processing is executed. Even in the invalid cluster, once the collective erasure is performed, the cluster state may return to the original normal state. Therefore, the invalid area is erased collectively when the power is turned on or the system is idle. This is an effective means against sudden errors in the device.

  FIG. 5 is a structural diagram showing internal data in the application. The cluster is composed of sectors of n pages, a header is secured on the 0th page, and a data area is secured on and after the first page. Of these, the first 4 bytes of page 0 are used as a cluster ID and a copy flag, followed by 32 bytes of stored page information. The length of data to be written is predetermined by the system and is not changed in the system.

  FIG. 7 shows a data reading procedure. The application reads the storage page information with reference to the cluster ID, and acquires the page number that is the sector in which the data is stored. For example, when the application recognizes that the data is stored in the x page, the application refers to the storage position offset information stored in the head position of the x page, and where the data is stored in the x page. Recognize

  Here, the storage page information and the storage position offset information are represented by a bitmap field, both of which secure an area of 32 bytes and have 256 bits. The value of each bit is changed from 1 to 0 every time the use state of the sector changes.

  That is, for the stored page information, for example, when data is stored from the first page to the x page, the bit value is changed from 1 to 0 from LSB to x bit, and the total number of 0 is stored in the designated page. Correspond. On the other hand, as shown in FIG. 6, the storage position offset information corresponds to one byte of the data area and one bit of the storage position offset information, and is stored every time data is sequentially stored from the first byte of the data area. One bit is sequentially changed from 1 to 0 from the LSB of the position offset information.

  As a result, when the application recognizes the page number in which the data is stored from the total of 0 bits in the stored page information, the application sequentially searches for the value of the bit from the LSB of the storage position offset information, and from where it starts from the beginning of the data area. Recognizes whether data is stored. For example, when the storage position offset information from the LSB to 24 bits is 0, it is recognized that data is stored from the head of the data area to 24 bytes. Then, when the application recognizes the final position where the data is stored, since the position returned by the data length from that position is the start address of the data, the application executes a process of reading data from the start address.

  FIG. 8 is a flowchart showing a data writing procedure. First, when the application recognizes the final position where data is stored in the same procedure as when reading data, it checks whether or not a blank area (unused area) exists in the page (step ST11). .

  If there is a blank area, the next position after the last position where the data is stored is used as the head address, the data writing process related to the write request is executed, and the storage position offset information is updated (step ST12). ). That is, the bit corresponding to the position next to the final position where data is stored is changed from 1 to 0.

  On the other hand, if there is no blank area in the current page, it is investigated whether or not the next page exists in the cluster (step ST13). If the next page exists, the storage position offset information of the next page is referred to and it is checked whether or not the next page is blank (step ST14). This is because during the past data writing in the next page, if a failure such as a power interruption has occurred, the data being written may remain (the data being written is If it remains, there will be a problem that new data cannot be written.)

  If the next page is blank, the application executes the data write processing related to the write request at the start position of the data area and updates the storage position offset information (step ST15). The application also updates the stored page information (step ST16). That is, the bit corresponding to the next page is changed from 1 to 0.

  If the next page is not blank, it is further investigated whether the next page is blank, but if the next page does not exist, the above-mentioned data reclaim is executed and another unused cluster is checked. In step ST17, data write processing relating to the write request is executed at the head position of the data area of the first page.

  As is apparent from the above, according to the first embodiment, when a data write request is received, the address information of the unused area is referred to and the data write process is executed in the unused area. Thus, when an unused area exists in a cluster, data can be written without erasing the cluster at once, and as a result, the number of rewrites of the BGO flash memory 11 is increased. .

Embodiment 2. FIG.
In the first embodiment, when there is a blank area in the page, the latest data is written after the already written data. That is, a plurality of data is stored in the page. As shown in FIG. 9, a plurality of data blocks are configured in a sector (page) as shown in FIG. 9, and only the last stored data (latest data) is validated. Referring to the usage flag (in-page used block flag, in-page used block flag) indicating the usage status of the data block stored in the header area, data is written to the unused area of each data block. May be executed.

  FIG. 9 is a structural diagram showing internal data in the application. Create page headers and multiple data blocks for all clusters. In the page header, a cluster ID and a copy flag are placed at the head, followed by an in-page in-use block flag of 3 bytes and an in-page used block flag of 3 bytes. In the data block, the data NO and the data area usage status are set 1 byte each, and the subsequent 8 bytes are used as the data area.

  FIG. 10 is a correspondence diagram of the in-page in-use block flag, the in-page used block flag, and the data block. To search for a block in use, an exclusive OR of the in-page in-use block flag and the in-page in-use block flag is obtained, and the data block corresponding to the 1 bit becomes the in-use data block. (Data block P in the example of FIG. 10).

  Data is read and written by designating a data ID. Of these, 00h is used data ID, and FFh is blank data ID. In addition, 254 types of data IDs can be taken. The data size allocated to each ID can take a value from 1 byte to 8 bytes, and the data size is not changed in the program.

  In addition, the pointer construction means of the application creates a pointer for data retrieval (a pointer indicating the block position of the data block storing the data) in the RAM 13. The page number and the block number of the data area are stored inside the pointer. When the contents of the RAM 13 have disappeared, such as when the power is turned on, all data areas are searched and searched, and the contents are retained.

  At the time of reading, the application searches for the data block indicated by the pointer of the data ID, and executes the data reading process with reference to the data area usage status (see FIG. 11). That is, 8 bits in the data area usage status correspond to the usage status of 8 bytes in the data area, and a location where the data area usage status bit is 0 is a used location.

  FIG. 12 is a flowchart showing an initial routine for a data search pointer. In an initial routine such as when the power is turned on, the application searches for the address of the cluster to be used, and if there is data in the page, the pointer data is updated (steps ST21 to ST23). This is a search method for the used data in the page, but if all the blocks are searched, the number of data is very large. Refer to and search for usage data. Then, the above process is repeated until the search for all pages is completed (steps ST24 and ST25), and the creation of the pointer is completed when the search for the last page is completed.

  FIG. 13 is a flowchart showing a data writing procedure. First, the application searches for a blank area of a data block in which data is stored. If there is a blank area in the data area where data can be written, the application stores the data block (blank area) in the data block. The data writing process is executed (steps ST31 and ST32).

  On the other hand, if there is no blank area in the data area where data can be written, it is determined whether another new data block can be acquired (step ST33). When a new data block can be acquired, a process of writing the data NO, the data area, and the in-page in-use block flag to the new data block is executed (steps ST34 and ST35). The in-page used block flag and the data block ID are simultaneously written (steps ST36 and ST37). If a new data block cannot be acquired, data reclaim is executed (step ST38).

  As is apparent from the above, according to the second embodiment, when a sector is composed of a plurality of data blocks, a usage status flag (page) indicating the usage status of the data block stored in the header area of the sector. (In-use block flag, in-page used block flag), etc., so that the data write processing is executed in the unused area of the data block, so in units of 1 to 8 bytes, There is an effect that a maximum of 254 types of data can be freely rewritten.

Embodiment 3 FIG.
In the first embodiment and the second embodiment, the data write to the BGO flash memory 11 has been described. However, the present invention is not limited to this, and it is needless to say that the present invention can also be used in other batch erasable nonvolatile memories. Yes.

BRIEF DESCRIPTION OF THE DRAWINGS FIG. 1 is a configuration diagram showing a batch erase nonvolatile memory according to a first embodiment of the present invention. It is explanatory drawing which shows the internal structure and cluster management area of a BGO flash memory. It is a flowchart which shows the process of a data reclaim. It is explanatory drawing which shows allocation of cluster ID at the time of a data reclaim. It is a structural diagram which shows the internal data in an application. It is explanatory drawing which shows the correspondence of storage position offset information and data. It is explanatory drawing which shows the reading procedure of data. It is a flowchart which shows the writing procedure of data. It is a structural diagram which shows the internal data in an application. FIG. 4 is a correspondence diagram of an in-page in-use block flag and an in-page used block flag and a data block. It is a correspondence figure of a data area use condition and a data area. It is a flowchart which shows the initial routine of the pointer for a data search. It is a flowchart which shows the writing procedure of data. It is a block diagram which shows the semiconductor disk which stores data in flash memory.

Explanation of symbols

  11 BGO flash memory (address information storage means), 14 CPU (address information storage means, data writing means, address information update means, ID management means, pointer construction means).

Claims (4)

Address information storage means for storing address information of unused areas in a cluster composed of a plurality of sectors;
When receiving a data write request, referring to the address information of the unused area, data writing means for executing the data writing process in the unused area,
An address information updating unit that updates address information of the unused area when the data writing unit executes data writing processing;
When the sector is composed of a plurality of data blocks, the data writing means refers to the usage status flag indicating the usage status of the data block stored in the header area of the sector, and the data block is not used. A batch erasing nonvolatile memory characterized by executing a data writing process in an area.   2. The collective erasing type nonvolatile memory according to claim 1, wherein the data writing means executes a data writing process in an unused area of another data block when there is no unused area in the data block. Sex memory.   3. The batch erase type nonvolatile memory according to claim 2, further comprising pointer construction means for constructing a pointer indicating a block position of a data block storing data in the volatile memory. A mobile phone equipped with a batch erasing nonvolatile memory,
The batch erasing nonvolatile memory is
Address information storage means for storing address information of unused areas in a cluster composed of a plurality of sectors;
When receiving a data write request, referring to the address information of the unused area, data writing means for executing the data writing process in the unused area,
An address information updating unit that updates address information of the unused area when the data writing unit executes data writing processing;
When the sector is composed of a plurality of data blocks, the data writing means refers to the usage status flag indicating the usage status of the data block stored in the header area of the sector, and the data block is not used. A cellular phone characterized by executing a data writing process in an area.

Images