JP2011175428A - Storage device - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a storage device allowing load reduction of a memory. <P>SOLUTION: A control unit 46 writes a first flag into a first page of a managing block when starting writing of user data into a block for the user data, and writes a second flag into a second page different from the first page when the writing of the user data ends. The control unit 46 decides that the user data corresponding to the first flag is normally written in the block for the user data when the first flag and the second flag are written in the first page and the second page, respectively, does not perform erasure processing to the block written with the user data, and estimates that the user data corresponding to the first flag is not normally written in the block for the user data when the second flag is not written in the second page though the first flag is written in the first page. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to a storage device having a rewritable memory.

  In recent years, flash memories have been distributed on the market as nonvolatile memories of electronic devices. The flash memory has been remarkably increased in capacity in recent years, and is used in, for example, a cellular phone (see, for example, Patent Document 1).

  In the flash memory, data is written and read in units of pages, and data is erased in units of blocks composed of a plurality of pages. By the way, in the flash memory, some pages may become unusable due to a physical defect or the like. Specifically, the flash memory has a problem that accurate data reading and writing cannot be performed if a part of a page is garbled.

  To deal with this problem, for example, when a bit-garbled page is generated, a technique for copying the block data having the page to a spare block and resetting and erasing the copy source block is widely known. .

JP-A-6-161867

  However, the above-described technique has a problem in that all the data of the block having the page that is garbled is copied to the spare block and consumes a large amount of the spare block.

  An object of the present invention is to provide a storage device that can reduce the burden on the memory.

  In order to solve the above-described problem, the storage device according to the present invention has a plurality of blocks each including a predetermined number of pages, accepts reading and writing of data in units of pages, and accepts erasing of data in units of blocks. A control unit for reading, writing, and erasing data from and to the nonvolatile memory, wherein the block of the nonvolatile memory includes a plurality of user data blocks for storing user data and the user data When the control unit starts writing user data to the user data block, a first page corresponding to the user data is displayed on the first page of the management block. When the flag is written and the writing of the user data is completed, it is different from the first page. When the second flag corresponding to the user data is written in the second page of the management block, and the first flag and the second flag are written in the first page and the second page, respectively, It is determined that the user data is normally written in the user data block, and the erase process is not performed on the block in which the user data is written, and the first flag is written in the first page. However, when the second flag is not written in the second page, it is estimated that the user data is not normally written in the user data block.

  In the storage device, it is preferable that the second page is a page next to the first page.

  In the storage device, the control unit may perform a process of moving data written in one block to another block based on a predetermined period or a use frequency of each block of the nonvolatile memory. preferable.

  In the storage device, a predetermined value is written in advance in the second page before data is written, and the control unit writes the second flag in a part of the second page. A value other than the specified value is written in an area other than the area where the second flag is written, and the first flag is written in the first page, but the second flag is written in the second page. If not, it is estimated that the user data corresponding to the first flag is not normally written in the user data block, and then a predetermined area other than the area where the second flag of the second page is written. If the specified value is written in the predetermined area, and if it is determined that the specified value is written in the predetermined area, the user data corresponding to the first flag If it is determined that the normal value is not written, and it is determined that the specified value is not written in the predetermined area, it is determined that the user data corresponding to the first flag is normally written, It is preferable not to perform an erasing process on a block in which user data is written.

  In the storage device, when the control unit determines that the specified value is written in the second page, the control unit determines that the user data corresponding to the first flag is not normally written. Then, when the user data is read out from the user data block a predetermined number of times and the user data can be read as significant data within the predetermined number of times, the user data is written to the block in which the user data is written. On the other hand, when the user data cannot be read as significant data within the predetermined number of times without performing the erase process, it is preferable to perform the erase process on the block in which the user data is written.

  According to the present invention, it is possible to reduce the memory load.

1 is an external perspective view of a mobile phone device. It is a functional block diagram which shows the function of a mobile telephone apparatus. It is a figure which shows the structure of the memory which a mobile telephone apparatus has. It is a figure with which it uses for description about the process which replaces a management block and a spare block. It is a figure with which it uses for the description about the process which replaces a block using a garbage collection function and a wear leveling function. It is a flowchart with which it uses for description about the writing process of data. It is a flowchart with which it uses for description about the process which reads the data of a page. It is a flowchart with which it uses for description about the discard (discard) process of a page. It is a flowchart with which it uses for description about the process after power activation. It is a figure which shows whether the erasure | elimination process is performed with respect to the block to which an error page belongs.

  Embodiments of the present invention will be described below. FIG. 1 is an external perspective view of a mobile phone device 1 incorporating a storage device 5 according to the present invention. 1 shows a form of a so-called foldable mobile phone device, the form of the mobile phone device according to the present invention is not particularly limited to this. For example, a sliding type in which one casing is slid in one direction from a state in which both casings are overlapped, or a rotary type in which one casing is rotated around an axis along the overlapping direction ( Turn type), or a type (straight type) in which the operation unit and the display unit are arranged in one housing and does not have a connecting unit.

The mobile phone device 1 includes an operation unit side body 2 and a display unit side body 3. The operation unit side body unit 2 includes an operation unit 11 and a microphone 12 to which a voice uttered by a user of the mobile phone device 1 is input on the surface unit 10. The operation unit 11 includes a function setting operation key 13 for operating various functions such as various settings, a telephone book function, a mail function, and the like, and an input operation key 14 for inputting a telephone number and characters such as mail. The operation key 15 includes a determination operation key 15 for performing determination and scrolling in various operations.
Further, the display unit side body unit 3 includes an LCD (Liquid Crystal Display) display unit 21 for displaying various types of information on the surface unit 20 and a speaker 22 for outputting the voice of the other party of the call. ing.

  Further, the upper end of the operation unit side body 2 and the lower end of the display unit side body 3 are connected via a hinge mechanism 4. In addition, the cellular phone device 1 relatively rotates the operation unit side body 2 and the display unit side body 3 connected via the hinge mechanism 4, thereby The display unit side body 3 can be opened (opened) with each other, or the operation unit side 2 and the display unit 3 can be folded (folded).

  FIG. 2 is a functional block diagram showing functions of the mobile phone device 1. As shown in FIG. 2, the mobile phone device 1 includes an operation unit 11, a microphone 12, a main antenna 40, an RF circuit unit 41, an LCD control unit 42, an audio processing unit 43, a memory driver 44, The memory 45 and the control unit 46 are provided in the operation unit side body 2, and the LCD display unit 21, the speaker 22, and the LCD driver 23 are provided in the display unit side body 3. Hereinafter, the memory driver 44, the memory 45, and the control unit 46 are collectively referred to as the storage device 5.

  The main antenna 40 communicates with a base station or the like in a first use frequency band (for example, 800 MHz), and can support a second use frequency band (for example, 1.5 GHz) for GPS communication. It is. In the present embodiment, the first frequency band used is 800 MHz, but other frequency bands may be used. Further, the main antenna 40 may be separately provided with an antenna that communicates with an external device in the first use frequency band and can correspond to the second use frequency band for GPS communication.

  The RF circuit unit 41 demodulates the signal received by the main antenna 40 and supplies the processed signal to the control unit 46. Then, the signal supplied from the control unit 46 is modulated and transmitted to an external device (base station) via the main antenna 40. On the other hand, the control unit 46 is notified of the strength of the signal received by the main antenna 40.

  The LCD control unit 42 performs predetermined image processing under the control of the control unit 46, and outputs the processed image data to the LCD driver 23. The LCD driver 23 stores the image data supplied from the LCD control unit 42 in the frame memory and outputs it to the LCD display unit 21 or the sub LCD display unit 30 at a predetermined timing.

  The sound processing unit 43 performs predetermined sound processing on the signal supplied from the RF circuit unit 41 under the control of the control unit 46, and outputs the processed signal to the speaker 22. The speaker 22 outputs the signal supplied from the sound processing unit 43 to the outside.

  In addition, the audio processing unit 43 processes the signal input from the microphone 12 according to the control of the control unit 46, and outputs the processed signal to the RF circuit unit 41. The RF circuit unit 41 performs a predetermined process on the signal supplied from the sound processing unit 43 and outputs the processed signal to the main antenna 40.

The memory driver 44 performs access (data writing, reading, and erasing) to the memory 45.
The memory 45 is composed of a nonvolatile memory (for example, a NAND flash memory). The memory 45 is accessed by the memory driver 44 and stores user data such as image data and document data.
The control unit 46 controls the entire mobile phone device 1 and is configured using a central processing unit (CPU) or the like.

Here, a difference between a NAND flash memory (hereinafter referred to as a NAND type) and a NOR flash memory (hereinafter referred to as a NOR type) will be described.
The common points of the NOR type and the NAND type are (1) only rewriting from bit “1” to bit “0” is performed at the time of data writing, and (2) block unit determined at the time of erasing data. It can only be erased.

  As for (1), the bit “1” is written as a specified value in the flash memory in a state where no data is written, and when writing the bit “1” when writing data, the specified value is used. Is maintained and bit “0” is written, a process of rewriting bit “1” to bit “0” is performed.

  Regarding (2), due to the circuit configuration of the flash memory, the erasing process can be performed in units of blocks. In the flash memory, erasing data means that all bits are “1” (specified value).

  The NAND type has a concept of pages, and data is basically written in units of pages. Further, a page has a limit on the number of times data is written (several thousands to several hundred thousand times). The NAND type includes a plurality of blocks each including a plurality of pages. For example, one page is 2 kbytes, and one block is composed of 64 pages (128 kbytes).

  Further, the NOR type uses a flag as a countermeasure against power interruption during data writing. In the NOR type, data can be written up to a bit unit because of the circuit configuration. For example, when data is written, a predetermined bit (management area) in an initial state is rewritten from “0” to “1”. At the end of writing, a process of rewriting the predetermined bit from “1” to “0” is performed. In the NOR type, the state of the predetermined bit is confirmed, and if the predetermined bit is “1”, it is determined that the data could not be normally written because the power was cut off during the data writing.

  However, since the NAND type can write data in units of pages as described above, simply adopting the NOR type processing method causes the management area to be in units of pages, and the user data area is insufficient. There is a risk of it.

  In the NOR type, the management area and the user data area are fixed in one-to-one correspondence. Therefore, in the NAND type, for example, if the number of flags to be written in the management area is set to m (m pages) due to a physical block size limitation, the number of user data may be limited.

  In addition, in the NAND type, if the NOR type processing method is simply adopted, when the data writing is completed after rewriting the predetermined page in the management area when the data writing is started, the predetermined processing is performed. The entire block to which the page belongs is erased. Since the NAND type has a limit on the number of rewrites as described above, the lifetime may be shortened if the entire block is erased each time data is written.

  In the NAND type, as in the NOR type processing method, the flag is read from the management area immediately before the user data is read from the user data area, and whether or not illegal data is written due to the power-off is determined. Checking every time user data is accessed is disadvantageous in terms of processing speed.

  In addition, the NAND type is a device in which a bit error is likely to occur due to the circuit configuration, and therefore an ECC (Error Check and Correct) code is generally added. The ECC can cope with (correctable) a 1-bit error, but cannot cope with (correctable) an error of 2 bits or more. The ECC can detect that an error has occurred for an error of 2 bits or more.

  In the NAND type, when data of a block is read out, ECC is performed and when an error of 2 bits or more is detected, the block can be reused unless the block is erased (specified value write processing). Prohibited.

  In the NAND type, one block is composed of a plurality of pages, and writing of user data is normally performed in units called pages. However, if a power interruption occurs during writing, the user data is written in the page being written. It is known that the data of the above becomes an indefinite value (not a specified value such as FFh but a random value). Here, the power interruption means a state in which the supply of power is interrupted due to the remaining battery level or the mobile phone device 1 being dropped, and the processing cannot be continued.

  Further, when the control unit 46 reads data of a page in which an indefinite value is written (hereinafter referred to as an indefinite page), an error of 2 bits or more is detected by ECC. However, in the system, it is necessary to discard the undefined page and not use it, that is, not to be accessed by the control unit 46.

  Here, it is necessary to distinguish between a 2-bit error that generally occurs and a 2-bit error due to an indefinite page that occurs due to a power failure. In the former 2-bit error, there is a case where the block cannot be reused physically. However, since the latter 2-bit error is an error at the time of writing, the block can be reused.

  The storage device 5 adopts a NAND type as the memory 45, ensures reliability when the memory 45 is controlled to read and write data by FFS (flash file system), and user data written in the memory 45 is stored in the memory 45. It has a function of determining whether or not the data has been normally written with the burden on the memory 45 minimized.

Hereinafter, the operation of the control unit 46 for exhibiting the above-described function will be described in detail.
The memory 45 includes a plurality of blocks each having a predetermined number of pages, receives data reading and writing in units of pages, and receives data erasing in units of blocks. As shown in FIG. 3, the memory 45 includes a user data area UA and a management area MA. The user data area UA includes a plurality of user data blocks UB1 to UBn that store user data. The management area MA includes a plurality of management blocks MB1 to MBn that manage user data blocks UB1 to UBn. The memory 45 is assigned a unique address.

  As described above, the block has a size of 128 kbytes, and the page has a size of 2 kbytes. In addition, it is assumed that one block includes 64 pages. The user data area UA is used for storing user data. The management area MA is used for user data writing management.

  Although the details will be described later, the physical positions of the blocks belonging to the user data area UA and the management area MA do not need to be fixed, and by using the garbage collection function or the wear leveling function, The structure by which a block is replaced suitably may be sufficient.

The control unit 46 reads, writes, and erases data with respect to the memory 45.
Here, a specific processing method by the control unit 46 will be described. When writing user data, a predetermined page in a predetermined user data block is designated based on a predetermined method. In the following, for convenience, the user data is written in the user data block UB2. This will be explained assuming that.

  As shown in FIG. 3, when the controller 46 starts writing user data to the user data block UB2, the first flag (during writing) corresponding to the user data in the first page P1 of the management block MB1. A process of writing F1 is performed. Then, when the writing of the user data is finished, the control unit 46 performs a process of writing a second flag (write end flag) F2 corresponding to the user data on the second page P2 different from the first page P1.

  Further, when writing the first flag F1 to the first page P1, the control unit 46 writes the address (write address) A1 of the page to which user data is written (a predetermined page in a predetermined user data block).

  When writing the second flag F2 to the second page P2, the control unit 46 writes the user data (a predetermined page in a predetermined user data block, and the user data is written over a plurality of pages). If it is, the address (write address) A1 of the first page) is written. As will be described in detail later, when writing the second flag F2 and the write address A1 to the second page P2, the control unit 46 writes a unique value (for example, 55h) in the remaining area of the second page P2.

  In addition, in the first page P1, the control unit 46 keeps the specified value (FFh) without writing other data for the areas other than the area where the write address A1 and the first flag F1 are written.

  Here, even when a 2-bit error is detected by ECC when the management area MA is accessed, it is a 2-bit error that generally occurs or due to an indefinite page that occurs due to a power failure. It is necessary to distinguish whether it is a 2-bit error.

  Therefore, as shown in FIG. 3, the control unit 46 writes a predetermined unique value (for example, 55h) for all areas other than the write address A1 and the second flag F2 in the second page P2. I do.

  According to such a configuration, when the control unit 46 accesses the second page P2 and detects a 2-bit error by ECC, the control unit 46 reads the value written in the second page P2, and the read value is A process is performed to determine whether or not the value matches a predetermined unique value.

  If the read value does not contain a unique value or contains a small number of unique values (if there are many mismatches in the unique value and the data is corrupted), the control unit 46 turns off the power during the writing of the unique value. It is determined that the second page P2 is an indefinite page, and if a certain number or more of unique values are included, it is determined that a 2-bit error generally occurs. To do.

  When the control unit 46 determines that the second page P2 is an indefinite page, after writing the user data normally and then writing the second flag F2 on the second page P2, It is considered that a power interruption has occurred, and the erase process is not performed on the block to which the second page P2 belongs.

  If the control unit 46 determines that a general 2-bit error has occurred in the second page P2, the control unit 46 discards (discards) all the pages of the management block, and uses the garbage collection function. Then, the erasure process of the block to which the second page P2 belongs is performed.

  When it is assumed that one page is composed of four sectors, the control unit 46 performs ECC for each sector. That is, the control unit 46 performs correction when a 1-bit error occurs, and detects an error when a 2-bit error occurs.

  Further, as a guideline, the control unit 46 determines that the page is an indefinite page when there are two or more sectors where the pattern of the unique value (55h) does not match 2 bits. This is because there is a possibility that a general 2-bit error has occurred if only one sector has a 2-bit mismatch.

  In this way, when the user data is normally written in the user data block, the storage device 5 generates a pair of the first page P1 and the second page P2 in which the same write address A is written. In addition, a process of writing the write address A1, the second flag F2, and a predetermined unique value is performed on the second page P2.

  Further, when the power is turned on, the control unit 46 accesses the management area MA, detects the pair of the first page P1 and the second page P2 based on the write address A1, and detects the first page P1 and the second page P2. When the first flag F1 and the second flag F2 are respectively written in the page P2, it is determined that the user data corresponding to the first flag F1 is normally written in the user data block UB2, The erasure process is not performed on the user data block UB2 in which the user data is written.

  Further, the control unit 46 accesses the management area MA when the power is turned on, and the first flag F1 is written in the first page P1, but the second flag F2 is written in the second page P2. If not, it is presumed that a power interruption has occurred during the writing of user data to the user data block UB2, and the user data corresponding to the first flag F1 has been normally written to the user data block UB2. Estimated not.

  When the power is turned on, the control unit 46 detects the pair of the first page P1 and the second page P2 based on the write address A1, and the first flag F1 is written in the first page P1. However, even when the unique value is not written in the second page P2, it is determined that the user data corresponding to the first flag F1 is normally written in the user data block UB2, and the user data is The configuration may be such that the erase process is not performed on the written user data block UB2.

  The fact that the write address A1 has been written to the second page P2 means that the user data has been written normally. In such a state, it is considered that the power interruption occurred while writing the user data normally and writing the unique value after writing the write address A1 and the second flag F2 to the second page P2. Therefore, it is determined that the user data corresponding to the first flag F1 is normally written in the user data block UB2.

With this configuration, when starting to write user data, the storage device 5 writes the first flag together with the write address to the first page, and when the user data has been written, the second page P2 Since the write address and the second flag are written in, when the power is turned on, the management area MA is accessed, the pair of the first page and the second page is detected, and the first flag and the second flag are written. If it is confirmed, it can be determined that the user data has been normally written, and the erasure process is not performed on the user data block in which the user data is written.
Therefore, the storage device 5 can determine whether or not the user data written in the memory 45 is normally written with the burden on the memory 45 minimized.

  The state of the management block MB is checked only once at the start-up after the power is turned on, and does not need to be checked every time user data is read.

  Further, if it is determined that a power interruption has occurred, the storage device 5 may perform a discarding (discarding) process so as not to use the block including the indefinite data. For example, by using a garbage collection function or the like, which will be described in detail later, an erase process (a process for writing a specified value) of a discarded block is performed, and the block can be reused.

  Further, the management block MB does not correspond one-to-one with the user data block UB. Therefore, the storage device 5 writes a flag or the like to the lower pages that are sequentially vacant independently of the user data block UB. Further, the first flag F1 and the second flag F2 that have been checked once become unnecessary, and are sequentially deleted by using a garbage collection function described later. Therefore, in the storage device 5, the management block MB apparently has no size limitation, and there is no limitation on the size of all data blocks.

The second page P2 is preferably the next page after the first page.
According to such a configuration, since the first page and the second page are continuous, the control unit 46 can easily search for a paired page when accessing the management area MA. Become.

Here, in the management block MB, the frequency of use of a specific block increases, and when the predetermined number of uses is reached, the block becomes unusable, and as a result, the capacity of the entire memory 45 is reduced.
Therefore, the control unit 46 preferably performs processing for moving data written in one block to another block based on a predetermined period or the use frequency of each block of the memory 45.

The control unit 46 has a garbage collection function and a wear leveling function.
The garbage collection function is a function for collecting memory areas that are no longer used by a program or memory areas in the gaps between programs to increase a continuously usable memory area. According to the garbage collection function, the memory 45 having a finite capacity can be used effectively.

  The wear leveling function is a function that distributes blocks (pages) for data rewriting as evenly as possible so as not to concentrate on specific blocks (pages). According to the wear leveling function, repeated writing and erasing on a specific page can be prevented, so that the life of the memory 45 can be extended.

  The control unit 46 performs a process of moving data written in one block to another block by using the garbage collection function on condition that a predetermined period has elapsed. Further, the control unit 46 may perform a process of moving data written in one block to another block using a wear leveling function based on the frequency of use of each block in the memory 45.

  With this configuration, the storage device 5 can increase the continuous usable memory area and can extend the life of the memory 45.

Further, as described above, when data is written to the management block MB, there is no free space to be written to the block. Here, the memory 45 can be erased only in block units due to the circuit configuration.
Therefore, the control unit 46 uses the garbage collection function or the wear leveling function to write management data to the management block, and when all the pages are used up, the control unit 46 is required for the spare block that exists separately. Copy the data and erase the management block. Therefore, the control unit 46 uses the spare block as a management block, and uses the block that was originally used for management as a spare block.

  The management block MB is a block that is frequently used and frequently used. Therefore, the control unit 46 controls the management block MB so as to be a target of wear leveling.

  The management data written to the management block MB is effectively invalid data because it is not used after it is used to determine whether or not user data has been normally written when the power is turned on. Therefore, strictly speaking, the control unit 46 does not need to copy the management data written in the current management block MB to the spare block. Therefore, when the management data is not copied to the spare block, the spare block is a management block whose entire area is empty from the beginning.

  The page data written in the past to the management block MB is data that does not need to be accessed again. The control unit 46 discards (discards) the system when it is in an idle state (the mobile phone device 1 is not in use). Thereafter, the control unit 46 uses the garbage collection function to reproduce a new management block MB.

  Note that if continuous writing of user data is performed without the system becoming idle, the management block MB may run out of free space. When there is no page for writing management data, the control unit 46 forcibly uses the garbage collection function and does not copy the page data even if it is not discarded (discarded). The storage device 5 also has a function of ensuring a free space for writing management data by forcibly using the garbage collection function.

  Further, it is assumed that only the management block has a high writing frequency in the storage device 5, and that the garbage collection function is used to replace the spare block. If it does so, the block which repeats replacement of data many times will become frequently used.

  Therefore, the storage device 5 is provided with an erase counter that counts the number of erases for each block. The control unit 46 checks the count value of the erase counter for all blocks of the memory 45 at the time of activation, and executes the wear leveling function when the difference between the maximum and minimum count values exceeds a certain value. . With this configuration, in the storage device 5, the static block with a small count value becomes the next spare block, and the entire memory 45 can be used uniformly.

  Moreover, the schematic diagram at the time of utilizing a garbage collection function is shown in FIG. FIG. 4 shows a case where the management block MB is processed using the garbage collection function. After the garbage collection function is used, the memory 45 has an image in which used block positions are interchanged as shown in FIG.

  In order to use the garbage collection function, a spare block (spare block SB) for moving data is required. Here, in order to make a block a new spare block SB, it is necessary to perform an erasing process on the block. Accordingly, when a block is newly set as a spare block SB, the control unit 46 sets all bits of the block to a specified value (bit “1”).

FIG. 5 shows a schematic diagram when the garbage collection function and the wear leveling function are used.
The data block B1 of the memory 45 is a data block in which data with a low rewrite frequency is stored, and the number of times of block erasure is reduced (hereinafter, the data block B1 is referred to as a static data block B1). The control unit 46 pays attention to the static data block B1 when using the wear leveling function.

  The control unit 46 performs a process of exchanging the static data block B1 and the spare block SB using the wear leveling function (see FIGS. 5A and 5B). Note that the control unit 46 may perform the replacement process using a garbage collection function.

  As a result, the control unit 46 stores static data by using the wear leveling function or the garbage collection function for the frequently used blocks that existed as the spare block SB before the wear leveling function is exhibited. Blocks can be made into less frequently used blocks.

  Next, the control unit 46 performs processing for exchanging the management block MB and the spare block SB using the wear leveling function or the garbage collection function (see FIGS. 5B and 5C).

  In this way, in the memory 45, the block position that has become the management block MB is a block in which static data was initially stored before using the wear leveling function, and a block that has been used less frequently until now. Can be changed to a management block. Therefore, the storage device 5 uses the entire memory 45 evenly and can extend the life.

Further, as described above, when no data is written in the block of the memory 45, a specified value (for example, bit “1”) is written in advance.
When writing the second flag to a part of the second page, the control unit 46 writes a value other than the specified value in an area other than the area in which the second flag is written, and the first flag is written in the first page. However, if the second flag is not written in the second page, it is estimated that the user data corresponding to the first flag is not normally written in the user data block UB. It may be configured to determine whether a prescribed value (for example, FFh) is written in a predetermined area other than the area where the second flag of 2 pages is written.

  In such a configuration, when the control unit 46 determines that the specified value is written in the predetermined area of the second page, the user data corresponding to the first flag is not normally written. Judge. This is because the first flag is written on the first page, but the second page remains at the specified value, so that power is cut off while user data is being written, and the user data is for user data. This is because it can be determined that the block could not be written normally.

  If the control unit 46 determines that the specified value is not written in the predetermined area of the second page, the control unit 46 determines that the user data corresponding to the first flag is normally written, and the user data The erase process is not performed on the block in which is written. If the second flag is not written on the second page and does not reach the specified value, it is considered that a power failure occurred while the unique value was written on the second page. Therefore, in such a case, it can be determined that the user data can be normally written in the user data block.

With this configuration, when the power is turned on, the storage device 5 accesses the management area MA, confirms that the first flag is written on the first page, and stores the second flag on the second page. If it is confirmed that the flag has not been written and has not reached the specified value, it is determined that the user data has been written normally, and the user data block to which the user data has been written is determined. Do not erase.
Therefore, the storage device 5 can determine whether or not the user data written in the memory 45 is normally written with the burden on the memory 45 minimized.

  Further, when it is determined that the specified value is written in the second page, the control unit 46 determines that the user data corresponding to the first flag F1 is not normally written, and thereafter, for the user data The configuration may be such that user data is read out a predetermined number of times from the block UB.

In the case of such a configuration, when the user data can be read as significant data within a predetermined number of times, the control unit 46 does not perform an erasure process on the block in which the user data is written.
In addition, when the user data cannot be read as significant data within a predetermined number of times, the control unit 46 performs an erasing process on the block in which the user data is written.

  Here, the first flag is written in the first page, but the second page remains at the specified value. This means that the power is cut off while the user data is being written to the user data block UB. Although it is thought that it occurred, theoretically, it is conceivable that the power interruption occurred immediately after the user data was written in the user data block UB.

Therefore, if the control unit 46 accesses the user data a predetermined number of times and can read it as significant data, the control unit 46 determines that the power interruption has occurred after the user data has been written, and the user data is written. The erase process is not performed on the existing block.
In addition, the control unit 46 accesses the user data a predetermined number of times, and if it cannot be read as significant data, a power interruption occurs while the user data is being written to the user data block UB. (That is, it is determined that the user data is broken), and an erasure process is performed on the block in which the user data is written.

  Therefore, the storage device 5 can determine whether or not the user data written in the memory 45 is normally written, with the burden on the memory 45 being minimized, and the user data can be written normally. Since the erasure process is executed on the user data block UB only when the process cannot be performed, the erasure process can be performed with the burden on the memory 45 minimized.

Here, the data writing process will be described with reference to the flowchart shown in FIG.
In step ST1, the control unit 46 writes the write address A1 and the first flag (during writing flag) F1 to a predetermined page of the management block MB.
In step ST2, the control unit 46 writes user data to the page of the user data block UB.

  In step ST3, when the writing of the user data in step ST2 is completed, the control unit 46 is the page (first page P1) next to the page (first page P1) in which the first flag F1 is written in step ST1. The write address A1 and the second flag (write end flag) F2 are written to the second page P2). Then, the control unit 46 writes the unique value in the remaining area of the second page P2 simultaneously with the writing of the write address A1 and the second flag F2 to the second page.

Further, the process of reading page data will be described with reference to the flowchart shown in FIG.
In step ST11, the control unit 46 reads the designated user data from the user data block UB on condition that the user data has been normally confirmed to be written in the user data block UB.

  Specifically, the control unit 46 refers to a table in which user data and an address indicating the location where the user data is written are referred to, thereby specifying the specified user data in a predetermined user data block. A process of reading from a predetermined page is performed.

  In this way, the control unit 46 receives the specified page data from the data block without being aware of the management block MB on condition that the user data has been successfully written to the user data block UB. Can be read.

  The page discarding (discarding) process will be described with reference to the flowchart shown in FIG. As described above, the second page P2 is written with the write address A1 and the second flag F2 (write end flag) when the write process is normally completed, and a unique value ( 55h) is written. Note that the following description focuses on the confirmation of the second page P2 and the discard process.

  In step ST21, the control unit 46 accesses the second page P2, and determines whether or not a 2-bit error has occurred due to the ECC. If a 2-bit error has not occurred (No), the process proceeds to step ST22. If a 2-bit error has occurred (Yes), the process proceeds to step ST23.

  In step ST22, the control unit 46 performs normal processing. Specifically, the control unit 46 confirms the presence / absence of the write address A1 and the second flag F2.

  In step ST23, the control unit 46 checks another area of the second page P2, and performs a comparison process as to whether or not it matches the unique value.

  In step ST24, the control unit 46 determines whether or not the page is an indefinite page from the result of the comparison process in the process of step ST23. The control unit 46 determines that the unique value is inconsistent and the data is corrupted, and determines that the page is an indefinite page. If the data is not an indefinite page, the control unit 46 determines that a 2-bit error generally occurs. . If it is determined that the page is indefinite (Yes), the process proceeds to step ST25. If it is determined that the page is not indefinite (No), the process proceeds to step ST26.

  In step ST25, when the control unit 46 determines that the second page P2 is an indefinite page, the user data is normally written, and then the second flag F2 is being written to the second page P2. Since it is considered that the power is cut off, the erasure process is not performed on the management block MB.

  In step ST26, when the control unit 46 determines that the second page P2 is not an indefinite page, it is considered that a general 2-bit error has occurred in the second page P2. All are discarded (discarded), and the management block MB is erased by using the garbage collection function.

  In this way, the storage device 5 determines whether or not the second page P2 is an indefinite page when it determines that a 2-bit error has occurred due to the ECC. In the case of an indefinite page, an erasure process is performed. Therefore, an unnecessary erasure process does not occur, and a 2-bit error caused by ECC can be dealt with with a minimum load on the memory 45.

  Next, processing after power-on will be described with reference to a flowchart shown in FIG. Here, the processing after power-on is processing after power-on when the storage device 5 is normally activated, and processing after power-on when the storage device 5 is restarted due to power-off.

  In step ST31, the control unit 46 checks the management block MB and confirms whether or not the second page P2, which is a pair of the first page P1, ends with a page including the second flag F2 (write end flag). To do. When the second flag F2 is written (Yes), it is considered that the user data is normally written in the user data block UB, and thus the processing is terminated. If the second flag F2 has not been written (No), the process proceeds to step ST32. In the process of step ST31, it cannot be determined whether the second flag F2 is written in the second page P2 or whether the second flag F2 is not written due to a 2-bit error. In the latter case, it is considered that the user data is normally written in the user data block UB.

  In step ST32, the control unit 46 confirms whether or not all the second pages P2 are the specified value (FFh). If all are specified values (Yes), the process proceeds to step ST33, and if all are not specified values (No), the process proceeds to step ST34. Note that the fact that all the second pages are not specified values is considered to have been written to the second page P2, and the user data is normally written to the user data block UB.

  In step ST33, since all of the second page P2 are specified values, the control unit 46 is considered to have lost power while writing user data to one user data block UB, and has an indefinite value. The discarding process is performed on the one user data block UB, and the process ends. Here, the discarded block may be reused, or reuse may be prohibited.

  In step ST34, the control part 46 confirms whether the unique value of the 2nd page P2 has become many. If the number of unique values is large, that is, if a certain number or more of unique values are not included (Yes), it is assumed that a power failure has occurred while writing the second flag F2 etc. on the second page P2. It is conceivable that the erasure process is not performed and the process is terminated. If the number of unique values is not increased, that is, if a certain number or more of unique values are included (No), the process proceeds to step ST35.

  In step ST35, the control unit 46 considers that a general 2-bit error has occurred in the second page P2, and discards (discards) all pages of the management block. When the management block is reused, the control unit 46 performs block erasure processing by using the garbage collection function, and ends the processing.

In addition, FIG. 10 shows a list of whether or not the erasure process is performed on the block to which the error page belongs.
When the read page is an error page due to power failure, the control unit 46 discards the read page and does not erase the block.

  Further, when the read page is an error page due to a defect in the memory 45, the control unit 46 retries the read a predetermined number of times (for example, 10 times) and succeeds in reading within the predetermined number of times. In this case, the page is treated as a normal page and the block is not erased.

  In addition, when the read page is an error page due to a defect in the memory 45, the control unit 46 retries reading a predetermined number of times (for example, 10 times). The block to which the error page belongs is copied to the spare block except the page where the error has occurred, and the block is erased. Alternatively, the control unit 46 may not use the block again (bad block formation).

  In this way, the storage device 5 can take measures against power interruption with a minimum burden on the NAND type memory 45. Further, the storage device 5 can recognize illegal page data in the middle of data writing and prevent a problem from occurring. Further, the storage device 5 can not only determine whether or not a power interruption has occurred during the writing of user data, but can also determine whether or not a power interruption has occurred during the writing of management page data. In addition, the storage device 5 can set a frequently used management block at an arbitrary block position in the memory 45, thereby extending the life of the memory 45.

  The management block MB is allocated as one block, but it is necessary not to overlap with the user data block UB. Therefore, the storage device 5 assigns the user data block UB to the position where the logical address is small, and assigns the management block MB to the position where the logical address is large. Since it is necessary to match the boundary of the block size, it is necessary to pay attention to this point when assigning logical addresses.

1 Mobile phone device 5 Storage device 44 Memory driver 45 Memory 46 Control unit UA User data area MA Management area UB User data block MB Management block

Claims (5)

A non-volatile memory having a plurality of blocks each having a predetermined number of pages, accepting reading and writing of data in units of pages, and accepting erasing of data in units of blocks;
A control unit that reads, writes, and erases data with respect to the nonvolatile memory,
The non-volatile memory block includes a plurality of user data blocks for storing user data and a plurality of management blocks for managing the user data blocks.
When the controller starts writing user data to the user data block, the controller writes a first flag corresponding to the user data to the first page of the management block, and the writing of the user data is completed. And writing a second flag corresponding to the user data on a second page of the management block different from the first page,
When the first flag and the second flag are respectively written in the first page and the second page, it is determined that the user data is normally written in the user data block, and the user Do not erase the block where data is written,
When the first flag is written in the first page, but the second flag is not written in the second page, the user data is not normally written in the user data block. Storage device that estimates. The storage device according to claim 1,
The storage device, wherein the second page is the next page of the first page. The storage device according to claim 1,
The said control part is a memory | storage device which performs the process which moves the data currently written in one block to another block based on the predetermined period or the usage frequency of each block of the said non-volatile memory. The storage device according to claim 1,
Before the data is written to the second page, a specified value is written in advance.
When the control unit writes the second flag in a part of the second page, the control unit writes a value other than the specified value in an area other than the area in which the second flag is written, If the flag is written but the second flag is not written in the second page, it is estimated that the user data corresponding to the first flag is not normally written in the user data block. And determining whether the specified value is written in a predetermined area other than the area where the second flag of the second page is written;
If it is determined that the specified value is written in the predetermined area, it is determined that the user data corresponding to the first flag is not normally written;
When it is determined that the specified value is not written in the predetermined area, it is determined that the user data corresponding to the first flag is normally written, and the block in which the user data is written is determined. A storage device that does not perform erasure processing. The storage device according to claim 4,
When it is determined that the specified value is written in the second page, the control unit determines that the user data corresponding to the first flag is not normally written, and thereafter, for the user data Read the user data a predetermined number of times to the block,
If the user data can be read as significant data within the predetermined number of times, the erasure process is not performed on the block in which the user data is written,
A storage device that performs an erasure process on a block in which the user data is written when the user data cannot be read as significant data within the predetermined number of times.

Images