JP2010067284A - Control method of memory system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a control method of a memory system by which a service life can be prolonged by suppressing a concentration of write in a certain specific region in a nonvolatile memory. <P>SOLUTION: The control method of a memory system 1 includes a nonvolatile memory 10 comprising a plurality of blocks as a unit of data erasure. The method includes steps of: measuring erasure timing in which data of each block were erased; creating, for each block, a block table 30B in which a status value indicating a free state or an occupant state is associated with the erasure timing; detecting the block in which write is concentrated on a short-term basis; selecting a first block whose erasure timing is stale in the free state on the basis of information in the block table; selecting a second block, whose erasure timing is stale in the occupant state on the basis of information in the block table; and moving data of the second block to the first block in the case that the first block is included in the detected blocks. <P>COPYRIGHT: (C)2010,JPO&INPIT

Description

  The present invention relates to a method for controlling a memory system, for example, a method for controlling a memory system including a NAND flash memory.

  In recent years, nonvolatile semiconductor memories have been used in various applications such as large computers, personal computers, home appliances, and mobile phones. As one type of the nonvolatile semiconductor memory, a NAND flash memory that is electrically rewritable and can be nonvolatile, has a large capacity, and is highly integrated is known. In recent years, this NAND flash memory is also considered as a replacement for a hard disk drive (HDD).

  The NAND flash memory is a non-volatile semiconductor memory that needs to be erased before writing, and its lifetime depends on the number of rewrites. In writing / erasing data in the NAND flash memory, electrons are injected / released into / from a charge storage layer (for example, a floating gate) by applying a high voltage between the substrate and the control gate. If this is performed many times, the gate oxide film around the floating gate is deteriorated, and electrons injected into the floating gate are lost over time. As a result, data may be destroyed. That is, as the number of rewrites increases, the period during which data can be retained (retention) after rewriting is shortened (retention characteristics are degraded).

  In addition, data recorded by a personal computer or the like has both temporal locality and regional locality (see Non-Patent Document 1). For this reason, if data is recorded as it is at an address designated from the outside as it is, rewriting, that is, erasing processing concentrates in a specific area in a short time, and the bias of the number of erasures becomes large.

  On the other hand, the lifetime of the NAND flash memory also depends on the interval of the erasing process, and it has been confirmed that the longer the interval, the better the retention characteristics and the longer the lifetime (see Non-Patent Document 2). This also indicates that if the erasing interval is short, the retention characteristics are poor and the lifetime is impaired. It has also been confirmed that even if writing is performed at a short interval, the retention characteristic is restored unless an erasing process is performed for a corresponding long period of time (see Non-Patent Document 3).

David A. Patterson and John L. Hennessy, “Computer Organization and Design: The Hardware / Software Interface”, Morgan Kaufmann Pub, 31 August 2004 Neal Mielke et al., “Flash EEPROM Threshold Instabilities due to Charge Trapping During Program / Erase Cycling”, IEEE TRANSACTIONS ON DEVICE AND MATERIALS RELIABILITY, VOL. 4, NO. 3, SEPTEMBER 2004, PP.335-344 Neal Mielke et al., “Recovery Effects in the Distributed Cycling of Flash Memories”, 44th Annual International Reliability Physics Symposium, San Jose, 2006, PP.29-35

  The present invention provides a control method of a memory system that can extend the lifetime by suppressing rewriting from being concentrated on a specific area of a nonvolatile memory.

A memory system control method according to an aspect of the present invention is a memory system control method including a nonvolatile memory having a plurality of blocks, which are data erasing units. A step of measuring, and a step of creating a block table for associating a state value indicating an empty state or an in-use state with the erasing time for each block;
A step of detecting a block in which rewriting is concentrated in the short term, a step of selecting a first block having an empty state and an erasure time based on the information of the block table, and a use based on the information of the block table A step of selecting a second block that is in an intermediate state and has an erasing time; and if the detected first block is included in the detected block, the data of the second block is moved to the first block. A process.

  ADVANTAGE OF THE INVENTION According to this invention, the control method of the memory system which can extend a lifetime can be provided by suppressing that rewriting concentrates on a certain area | region of a non-volatile memory.

1 is a schematic diagram showing an example of a computer system including a memory system 1 according to a first embodiment of the present invention. 1 is a schematic diagram showing a configuration of a memory system 1 according to a first embodiment. 1 is a circuit diagram showing a configuration of one block included in a NAND flash memory 10. FIG. FIG. 2 is a block diagram showing an example of the configuration of a NAND controller 11. The block diagram which shows the structure of the block control part 30 and the erasure | elimination time measurement part 31. FIG. 5 is a flowchart showing a series of write operations by the NAND controller 11. 5 is a flowchart showing block release processing by the NAND controller 11. The block diagram which shows the structure of the allocation block selection part 32. FIG. 5 is a flowchart showing an allocation block selection process by an allocation block selection unit 32. 5 is a flowchart showing block allocation processing by the NAND controller 11. 7 is a flowchart showing block erase processing by the NAND controller 11; The block diagram which shows the structure of the short-term rewriting detection part 34. FIG. The flowchart which shows the short-term rewriting detection process by the short-term rewriting detection part 34. FIG. The figure explaining the specific example of a short-term rewriting detection process. FIG. 3 is a block diagram showing a configuration of a leveling unit 35. 5 is a flowchart showing leveling processing by the NAND controller 11. The block diagram which shows the structure of the replacement | exchange source block selection part 33. FIG. The flowchart which shows the replacement | exchange source block selection process by the replacement | exchange source block selection part 33. FIG. The block diagram which shows the structure of the allocation block selection part 32 which concerns on the 2nd Embodiment of this invention. The flowchart which shows the allocation block selection process by the allocation block selection part 32 which concerns on 2nd Embodiment. The block diagram which shows the structure of the replacement | exchange source block selection part 33 which concerns on the 3rd Embodiment of this invention. The flowchart which shows the replacement | exchange source block selection process by the replacement | exchange source block selection part 33 which concerns on 3rd Embodiment. The block diagram which shows the structure of the short-term rewrite detection part 34 which concerns on the 4th Embodiment of this invention. The flowchart which shows the short-term rewriting detection process by the short-term rewriting detection part 34. FIG. The flowchart which shows the short-term rewrite detection process following FIG. The figure explaining the specific example of a short-term rewriting detection process. The block diagram which shows an example of a structure of the NAND controller 11 which concerns on the 5th Embodiment of this invention. FIG. 2 is a block diagram showing a configuration of a block control unit 30. 7 is a flowchart showing a read operation by the NAND controller 11. The flowchart which shows the replacement | exchange source block selection process by the replacement | exchange source block selection part 33. FIG. The block diagram which shows the structure of SSD1 which concerns on an Example. FIG. 3 is a block diagram showing a configuration of a drive control circuit 102. FIG. 2 is a block diagram showing a configuration of a processor 108. The perspective view which shows an example of the portable computer carrying SSD100.

  Hereinafter, embodiments of the present invention will be described with reference to the drawings. In the following description, elements having the same function and configuration are denoted by the same reference numerals, and redundant description will be given only when necessary.

[First Embodiment]
The memory system of this embodiment is mounted on, for example, a printed circuit board on which a host device is mounted, and transfers data to and from the host device via a bus. Alternatively, the memory system of this embodiment is configured to be detachable from the host device, and transfers data to and from the host device via the bus while being connected to the host device. FIG. 1 is a schematic diagram illustrating an example of a computer system including a memory system 1 according to the present embodiment.

  The computer system includes a central processing unit (CPU) 2, a main memory 3 such as a DRAM (dynamic random access memory), a memory controller 4 that controls the main memory 3, and the memory system 1 of the present embodiment. ing. The CPU 2, the main memory 3, and the memory system 1 are connected to each other via an address bus 5 that handles addresses and a data bus 6 that handles data.

  In such a computer system, in response to a transfer request (read request or write request) from the CPU 2, if it is a write request, the data of the CPU 2 (including data input from the outside) or the main memory 3 Data is transferred to the memory system 1. If it is a read request, the data of the memory system 1 is transferred to the CPU 2 or the main memory 3.

  The memory system 1 includes a NAND flash memory 10 that is a kind of nonvolatile semiconductor memory, and a NAND controller 11 that controls the NAND flash memory 10. Hereinafter, an example of the configuration of the memory system 1 will be described.

[1. Configuration of Memory System 1]
FIG. 2 is a schematic diagram showing the configuration of the memory system 1. The memory system 1 includes a NAND flash memory 10 and a NAND controller 11. The NAND controller 11 includes a host interface circuit (host I / F) 21, a processing unit (MPU: micro processing unit) 22, a ROM (read only memory) 23, a RAM (random access memory) 24, and a NAND interface circuit (NAND I / F) 25.

  The host interface circuit 21 performs interface processing between the NAND controller 11 and the host device (CPU 2, main memory 3) according to a predetermined protocol.

  The MPU 22 controls the operation of the entire memory system 1. For example, when the memory system 1 is supplied with power, the MPU 22 reads the firmware (control program) stored in the ROM 23 onto the RAM 24 and executes predetermined processing to store various tables on the RAM 24. create. In addition, the MPU 22 receives a write request, a read request, and an erase request from the host device, and executes predetermined processing on the NAND flash memory 10 in response to these requests.

  The ROM 23 stores a control program and the like controlled by the MPU 22. The RAM 24 is used as a work area for the MPU 22 and stores control programs and various tables loaded from the ROM 23. The NAND interface circuit 25 performs interface processing between the NAND controller 11 and the NAND flash memory 10 according to a predetermined protocol.

  The NAND flash memory 10 is configured by arranging a plurality of blocks which are the minimum unit of data erasure. FIG. 3 is a circuit diagram showing a configuration of one block included in the NAND flash memory 10. Each block includes (m + 1) NAND strings arranged in order along the X direction (m is an integer of 0 or more). The select transistors ST1 included in each of (m + 1) NAND strings have drains connected to the bit lines BL0 to BLm and gates commonly connected to the select gate line SGD. The selection transistors ST2 included in each of the (m + 1) NAND strings have a source commonly connected to the source line SL and a gate commonly connected to the selection gate line SGS.

  Each memory cell transistor MT is composed of a MOSFET (metal oxide semiconductor field effect transistor) having a stacked gate structure formed on a semiconductor substrate. The stacked gate structure is configured by sequentially stacking a gate insulating film, a charge storage layer (floating gate electrode), an inter-gate insulating film, and a control gate electrode on a semiconductor substrate. In the memory cell transistor MT, the threshold voltage changes according to the number of electrons stored in the floating gate electrode, and data is stored according to the difference in threshold voltage. The memory cell transistor MT may be configured to store a binary value (1 bit), or may be configured to store a multi-value (data of 2 bits or more).

  In addition, the memory cell transistor MT is not limited to a structure having a floating gate electrode, and the threshold is adjusted by trapping electrons at the nitride film interface as a charge storage layer such as a MONOS (metal-oxide-nitride-oxide-silicon) type. Possible structures may be used. Similarly, the memory cell transistor MT having the MONOS structure may be configured to store 1 bit, or may be configured to store multiple values (data of 2 bits or more).

  In each NAND string, (n + 1) memory cell transistors MT are arranged such that their current paths are connected in series between the source of the select transistor ST1 and the drain of the select transistor ST2 (n is , An integer greater than or equal to 0). That is, (n + 1) memory cell transistors MT are connected in series in the Y direction so that adjacent ones share a diffusion region (source region or drain region).

  The control gate electrodes are connected to the word lines WL0 to WLn in order from the memory cell transistor MT located closest to the drain side. Therefore, the drain of the memory cell transistor MT connected to the word line WL0 is connected to the source of the selection transistor ST1, and the source of the memory cell transistor MT connected to the word line WLn is connected to the drain of the selection transistor ST2.

  The word lines WL0 to WLn connect the control gate electrodes of the memory cell transistors MT in common between the NAND strings in the block. That is, the control gate electrodes of the memory cell transistors MT in the same row in the block are connected to the same word line WL. The (m + 1) memory cell transistors MT connected to the same word line WL are handled as one page, and data writing and data reading are performed for each page.

  The bit lines BL0 to BLm commonly connect the drains of the selection transistors ST1 between the blocks. That is, NAND strings in the same column in a plurality of blocks are connected to the same bit line BL.

  Each functional block in each embodiment of the present invention can be realized as either hardware or software, or a combination of both. For this reason, each functional block is generally described below in terms of their function so that it is clear that they are any of these. Whether such a function is realized as hardware or software depends on a specific embodiment or a design constraint imposed on the entire system. Those skilled in the art can implement these functions in various ways for each specific embodiment, and determining such implementation is within the scope of the invention.

  Hereinafter, an example of a more specific configuration of the NAND controller 11 will be described. FIG. 4 is a block diagram showing an example of the configuration of the NAND controller 11 in the present embodiment.

  The NAND controller 11 includes a block control unit 30, an erase time measurement unit 31, an allocation block selection unit 32, a replacement source block selection unit 33, a short-term rewrite detection unit 34, a leveling unit 35, and a NAND interface circuit (NAND I / F). 25.

  Each time the data of each block included in the NAND flash memory 10 is erased, the erase time measuring unit 31 measures the erase time of the block. Then, the erase time measuring unit 31 sends the measured erase time to the block control unit 30.

  The block control unit 30 manages various types of information including the number of times of erasure and the erasure time for each block. The block control unit 30 issues a write request, a read request, and an erase request to the NAND flash memory 10 in response to a transfer request from the CPU 2. Specifically, the block control unit 30 includes an address table 30A and a block table 30B, which will be described later, and using these tables, a write request, a read request, and an erase request are sent to the NAND flash memory 10. Issue.

  The allocation block selection unit 32 selects a block (hereinafter referred to as an allocation block) to be allocated for writing when new data supplied from the outside (CPU 2 or main memory 3 or the like) is written in the NAND flash memory 10, for example. To do. That is, the allocation block selection unit 32 receives information on all blocks (all block information) stored in the block table 30B from the block control unit 30, and selects an allocation block according to a predetermined condition. Then, the allocation block selection unit 32 sends a block number (allocation block number) corresponding to the allocation block to the block control unit 30. Also, the allocation block selection unit 32 sends information (allocation block information) corresponding to the allocation block among all block information to the leveling unit 35. The data written to the NAND flash memory 10 includes system data and the like necessary for management inside the memory system 1 in addition to user data supplied from the outside of the memory system 1.

  The leveling unit 35 executes a leveling process to be described later. Then, the leveling unit 35 issues a write request, a read request, and an erase request to the NAND flash memory 10 in accordance with the leveling process. The leveling means that the number of block erasures is made uniform (so-called wear leveling). In this way, by aligning the number of block erases, it is possible to prevent the erase process from concentrating on a part of the blocks, so that the lifetime of the NAND flash memory 10 can be extended.

  The replacement source block selection unit 33 selects a data replacement source block (hereinafter referred to as a replacement source block) that is used in the leveling process by the leveling unit 35. That is, the replacement source block selection unit 33 receives all block information from the block control unit 30 and selects a replacement source block according to a predetermined condition. Then, the replacement source block selection unit 33 sends information corresponding to the replacement source block (replacement source block information) among all the block information to the leveling unit 35.

  The NAND interface circuit 25 receives a write request, a read request, and an erase request from the block control unit 30 and the leveling unit 35. The NAND interface circuit 25 instructs the NAND flash memory 10 to write, read, and erase data in response to these requests.

  The short-term rewrite detector 34 detects a block that is frequently rewritten (in a short period of time) (that is, a block with a short erase interval). The erasure time of blocks that have been rewritten in a short period is frequently updated (the erasure time becomes new), but the blocks that have not been rewritten in a short period of time remain old. In the present embodiment, a block in which rewriting is performed in a short period is detected using the difference between these erase times. For this reason, the short-term rewrite detection unit 34 sorts the blocks in use in order of the erasure time, and erases a certain search target block and a comparison target block having a new erasure time next to this search target block. Calculate the difference. If the difference in erase time exceeds a predetermined threshold, it is determined that the block whose erase time is newer than the comparison target block is a block that has been rewritten in a short period of time. This determination result is sent to the block control unit 30 as short-term rewrite information.

  FIG. 5 is a block diagram illustrating the configuration of the block control unit 30 and the erase time measurement unit 31. The block control unit 30 includes an address table 30A, a block table 30B, and a calculation unit 30C. The block control unit 30 receives an address and various information supplied from the outside, and updates the address table 30A and the block table 30B based on the information.

  The address table 30A includes an address area (logical block address) including an address sent from the host device (CPU 2) via the address bus 5, and a block number (physical address) in the NAND flash memory 10 corresponding to the address area. Correspondence relationship with block address). By using this address table 30A, the block control unit 30 can specify which block the data in the address area including the address sent from the host device corresponds to. This address table 30A is updated, for example, with block release processing and block allocation processing described later.

  The block table 30B indicates, for each block number, whether the block number is registered in the address table 30A (whether it is empty) or whether the block number is registered in the address table 30A (whether it is in use). The state of the block shown, the number of times data has been erased (erase count), the erase time sent from the erase time measuring unit 31, and the short-term rewrite flag corresponding to the short-term rewrite information sent from the short-term rewrite detector 34 (FIG. 5) Is simply described as “flag”) as information. The block table 30B is updated with, for example, a block release process, a block allocation process, a block erase process, and a short-term rewrite detection process, which will be described later. All block information included in the block table 30B is sent to the allocation block selection unit 32, the replacement source block selection unit 33, and the short-term rewrite detection unit 34.

  If the block is in an empty state, the address area corresponding to the block is viewed from the host device regardless of whether the block is actually erased in the NAND flash memory 10 or not. Recognized as a free area that is not stored. On the other hand, if the block is in use, the address area corresponding to the block is recognized as a busy area in which data is stored when viewed from the host device.

  In practice, the address table 30A and the block table 30B are stored in the RAM 24 or both the RAM 24 and the NAND flash memory 10. However, the address table 30A and the block table 30B stored in the NAND flash memory 10 in a nonvolatile manner are not necessarily updated every time the RAM 24 is updated.

  Each time the data in the NAND flash memory 10 is erased, the arithmetic unit 30C increments the erase count of the erased block included in the block table 30B by one. In practice, the processing of the arithmetic unit 30C is performed by the MPU 22.

  The block control unit 30 receives an allocation block number from the allocation block selection unit 32. Then, the block number of the address table 30A is updated using the allocated block number. Further, the block control unit 30 issues a write request for this allocation block to the NAND flash memory 10. Further, the block control unit 30 issues a normal read request and an erase request corresponding to an external address to the NAND flash memory 10.

  The erase time measuring unit 31 includes an erase number counter 31A that counts the number of erases, and a calculation unit 31B that updates the count value of the erase number counter 31A. The erase time measuring unit 31 measures the number of times the erase process has been performed on the block in the NAND flash memory 10 and outputs the measured number as the erase time.

  Specifically, every time any block is erased, the arithmetic unit 31B increments the erase count counter 31A by one. The count value (erase number) of the erase count counter 31A is sent to the block controller 30 (specifically, the block table 30B) as the erase time. In the erase time measuring unit 31 of this example, the erase time becomes older as the erase number is smaller. Actually, the processing of the calculation unit 31B is performed by the MPU 22.

  In addition to this, the time at which the erasing process is performed (erasing time), the energization time of the NAND controller 11 and the like may be used. When using the erasing time as the erasing time, the erasing time measuring unit 31 includes a clock and outputs the time as the erasing time every time the erasing process is performed. When the energizing time is used as the erasing time, the erasing time measuring unit 31 includes a timer, and whenever the erasing process is performed, the energizing time until then is measured and the energizing time is output as the erasing time. As described above, it is possible to arbitrarily select any one of the number of times of erasing, the erasing time, and the energizing time as the erasing time. Note that it is possible to use information other than the above three types as long as it is information that can specify the erasing time.

[2. Write Operation of NAND Controller 11]
Next, the write operation of the NAND controller 11 will be described. FIG. 6 is a flowchart showing a series of write operations by the NAND controller 11.

  First, the NAND controller 11 receives a write request from the CPU 2 and starts a write operation (step S10). Subsequently, the block control unit 30 uses the address table 30A to determine whether or not a block number is registered (a block is allocated) for the address area including the address of the write request. (Step S11). When a block number is registered (a block is allocated), it is necessary to overwrite data in this address area, so the block control unit 30 releases a block corresponding to the block number. A release process is executed (step S12).

  When it is determined in step S11 that the block number is not registered (block is not allocated), or after the block release process is executed in step S12, the allocation block selecting unit 32 allocates blocks to the address area. An allocation block selection process for selecting (allocation block) is executed (step S13).

  Subsequently, the leveling unit 35 and the replacement source block selecting unit 33 execute a leveling process (step S14). Subsequently, the block control unit 30 determines whether or not the allocation block has been replaced in the leveling process (step S15). When the allocation block is exchanged, the process returns to step S13, and the allocation block selection unit 32 executes the allocation block selection process again.

  When it is determined in step S15 that the allocation block has not been replaced, the block control unit 30 uses the block number (allocation block number) corresponding to the allocation block to update the address table 30A and the block table 30B. Next, block allocation processing is executed (step S16). Subsequently, the block control unit 30 executes block erasure processing on the allocated block (step S17).

  Subsequently, the block control unit 30 writes new data to the allocation block from which the data has been erased (step S18). That is, the block control unit 30 issues a write request to the NAND interface circuit 25. Based on this write request, the NAND interface circuit 25 instructs the NAND flash memory 10 to write new data to the allocation block.

  Subsequently, the short-term rewrite detection unit 34 executes a short-term rewrite detection process (step S19). Then, the block control unit 30 updates the flag of the block table 30B using the short-term rewrite information sent from the short-term rewrite detection unit 34. In this way, a series of write operations are executed by the NAND controller 11.

Hereinafter, each process included in the write operation will be specifically described.
[2-1. Block release processing]
FIG. 7 is a flowchart showing block release processing by the NAND controller 11. The block control unit 30 sets the block number corresponding to the address area including the address of the write request included in the address table 30A to an unallocated state (unallocated state) (step S20). Subsequently, the block control unit 30 sets the state of the block number included in the block table 30B to an empty state (step S21). A block newly set in an empty state in the block release process is hereinafter referred to as a release block. Thereafter, the NAND controller 11 can write new data to the release block.

[2-2. Allocation block selection process]
FIG. 8 is a block diagram illustrating a configuration of the allocation block selection unit 32. FIG. 9 is a flowchart showing allocation block selection processing by the allocation block selection unit 32.

  The allocation block selection unit 32 includes two selectors 32A and 32B. The allocation block selection unit 32 receives all block information from the block control unit 30 (step S30). This all block information is sent to the selector 32A. Subsequently, the selector 32A confirms the state of all blocks, and extracts an empty block from all blocks (step S31). Then, the selector 32A sends block information (empty state block information) corresponding to the free block to the selector 32B.

  Subsequently, the selector 32B selects a block with the oldest erase time among the blocks extracted by the selector 32A as an allocation block (step S32). The allocation block number corresponding to this allocation block is sent to the block control unit 30. Also, the allocation block information corresponding to this allocation block is sent to the leveling unit 35.

[2-3. Block allocation process]
FIG. 10 is a flowchart showing block allocation processing by the NAND controller 11. The NAND controller 11 updates the information in the address table 30A and the block table 30B with respect to the allocation block selected by the allocation block selection unit 32.

  First, the block control unit 30 receives an allocation block number from the allocation block selection unit 32 (step S40). Subsequently, the block control unit 30 sets the block number of the allocation block for the address area including the address of the write request included in the address table 30A (step S41). Subsequently, the block control unit 30 sets the state of the allocation block included in the block table 30B to a busy state (step S42). In this way, the setting of the allocation block selected by the allocation block selection unit 32 is changed from the empty state to the in-use state.

[2-4. Block erase processing]
FIG. 11 is a flowchart showing block erase processing by the NAND controller 11. The block control unit 30 issues an erase request for the allocation block to the NAND interface circuit 25 (step S50). Based on this erase request, the NAND interface circuit 25 instructs the NAND flash memory 10 to erase the data in the allocation block (step S51). Subsequently, the erase time measurement unit 31 (specifically, the calculation unit 31B) increments the count value (erase number) of the erase count counter 31A by 1 (step S52).

  Subsequently, the block control unit 30 updates the number of times of erasure and the erasure time corresponding to the allocation block from which data has been erased, included in the block table 30B (step S53). Specifically, the block control unit 30 updates the erase time corresponding to the block number of the allocated block included in the block table 30B, using the erase time sent from the erase time measuring unit 31. The arithmetic unit 30C increments the number of erasures corresponding to the block number of the allocation block included in the block table 30B by one.

[2-5. Short-term rewrite detection process]
FIG. 12 is a block diagram illustrating a configuration of the short-term rewrite detection unit 34. The short-term rewrite detection unit 34 includes a selector 34A, an alignment unit 34B, a search list 34C, a determination unit 34D, a storage unit 34E that stores an erase time interval threshold, and an information output unit 34F.

  The selector 34A confirms the state of all blocks using all block information sent from the block control unit 30, and extracts a block in use from all blocks. The alignment unit 34B arranges (sorts) the blocks in use in order of erasure time. The search list 34C temporarily stores the sorted block information. In practice, the search list 34C is stored in the RAM 24.

  In the search list 34C, the determination unit 34D selects the block having the latest erase time as the “search target block” and the block having the newest erase time after the search target block as the “comparison target block”. Furthermore, the determination unit 34D calculates a difference in erase time between the search target block and the comparison target block, and determines whether or not this difference exceeds the erase time interval threshold. Based on this determination result, a boundary where the erasing time changes rapidly is searched.

  The information output unit 34F uses the determination result (number of the comparison target block) sent from the determination unit 34D to calculate a range of blocks whose erasure time is newer than the comparison target block among all the blocks. And the information output part 34F sends the said range to the block control part 30 as short-term rewriting information.

  The erase time interval threshold value stored in the storage unit 34E is used to specify the range of blocks that are frequently rewritten, and also determines the life (or data retention period) of the NAND flash memory 10. It is set on the basis of whether to extend the extent. Lowering the erase time interval threshold increases the possibility of short-term rewrite detection, while increasing the erase time interval threshold decreases the possibility of short-term rewrite detection. That is, when the erase time interval threshold is lowered, the number of leveling processes increases, while when the erase time interval threshold is raised, the number of leveling processes decreases. This leveling process is accompanied by a block data erasing process.

  Each time data is erased, the retention characteristics of the NAND flash memory 10 deteriorate (that is, the lifetime becomes shorter). On the other hand, as the erase interval becomes longer, the retention characteristic of the NAND flash memory 10 is restored. That is, when the erasing process is continuously performed at a short interval, the recovery time of the retention characteristic cannot be secured, so that the life of the NAND flash memory 10 is shortened. Therefore, in the present embodiment, these blocks are targeted for leveling by setting a short-term rewrite flag for blocks that are frequently rewritten. Then, taking into account the characteristics of the NAND flash memory 10 indicating how long the lifetime is restored by extending the erase interval by a predetermined time, the block to be leveled and the number of times of leveling are optimal. The erasure time interval threshold is determined so that

  FIG. 13 is a flowchart showing short-term rewrite detection processing by the short-term rewrite detection unit 34. First, the short-term rewrite detector 34 receives all block information from the block controller 30 (step S60). This all block information is sent to the selector 34A. Subsequently, the selector 34A confirms the state of all the blocks, and extracts a block in use among all the blocks (step S61). Then, the selector 34A sends block information (in-use state block information) corresponding to the in-use block to the alignment unit 34B.

  Subsequently, the aligning unit 34B confirms the erase time included in the in-use state block information, and sorts the in-use blocks in the order of the erase time (step S62). Then, the block information sorted by the sorting unit 34B is put in the search list 34C.

  Subsequently, in the search list 34C, the determination unit 34D selects a block with the latest erase time as a search target block, and selects a block with the newest erase time after the search target block as a comparison target block (step S63).

  Subsequently, the determination unit 34D calculates the difference in erasure time between the search target block and the comparison target block (step S64). Subsequently, the determination unit 34D determines whether or not the difference between the calculated erase times exceeds an erase time interval threshold (step S65). If the erase time interval threshold is not exceeded, the determination unit 34D determines whether or not two or more blocks remain in the search list 34C (step S66). When two or more blocks remain, the determination unit 34D excludes the search target block from the search list 34C, returns to step S63, and reselects the search target block and the comparison target block.

  When the erase time interval threshold is exceeded in step S65, the determination unit 34D sends the comparison target block number to the information output unit 34F. Using this comparison target block number, the information output unit 34F calculates a range of blocks that have a newer erasure time than the comparison target block among all the blocks (step S68). And the information output part 34F sends the said range to the block control part 30 as short-term rewriting information. Upon receipt of the short-term rewrite information, the block control unit 30 sets a short-term rewrite flag corresponding to the block included in the short-term rewrite information, and clears the short-term rewrite flag corresponding to the other blocks.

  If two or more blocks do not remain in step S66, it is determined that the short-term rewritten block cannot be detected, and the short-term rewrite detection process is terminated.

  FIG. 14 is a diagram illustrating a specific example of short-term rewrite detection processing by the short-term rewrite detection unit 34. There are 12 blocks (blocks # 0 to 11) included in the NAND flash memory 10. Of these, blocks # 0 to 7 are in-use blocks included in the search list 34C, and blocks # 8 to 11 are in use. Is empty. Blocks # 0 to # 7 are sorted in the order of erasure. The erase time for each block is as shown in the figure.

  First, in the search list 34C, the block # 0 with the newest erase time is selected as the search target block, and the block # 1 with the newest erase time is selected as the comparison target block. The difference in erase time between block # 0 and block # 1 is “10”. This difference “10” does not exceed the erase time interval threshold (for example, “100”). Therefore, the search target block # 0 is excluded from the search list 34C.

  Subsequently, in the search list 34C, the block # 1 with the newest erase time is selected as the search target block, and the block # 2 with the newest erase time is selected as the comparison target block. The difference in erase time between block # 1 and block # 2 is “110”. This difference “110” exceeds the erase time interval threshold. Therefore, the number of the comparison target block # 2 is sent from the determination unit 34D to the information output unit 34F.

  The information output unit 34F calculates blocks # 0, 1, and 8 to 11 whose erasure time is newer than the comparison target block # 2 among all the blocks including the vacant state blocks. Then, the information output unit 34F sends the blocks # 0, 1, and 8 to 11 to the block control unit 30 as short-term rewrite information. On the other hand, the block control unit 30 sets short-term rewrite flags corresponding to the blocks # 0, 1, 8 to 11, and clears short-term rewrite flags corresponding to other blocks. In this way, the information in the block table 30B is updated.

[2-6. Leveling process]
In order to extend the lifetime of the NAND flash memory 10, it is necessary to achieve both the equalization of the number of erases of each block and the avoidance of frequent rewriting in a short time in a specific area. If an attempt is made to extend the lifetime of the NAND flash memory 10 by paying attention only to equalizing the number of times of erasure, rewriting of a specific area may proceed in a short period of time depending on the write situation. If this correction (leveling) is frequently performed in a short period of time, the number of erasures is uniform, and writing concentration in a specific area in the short period can be avoided, but the number of erasing processes due to the correction increases, resulting in a long life. Can not be extended. On the other hand, if the correction interval is extended too much, erasure processing with a short interval will be concentrated in a specific area, and the lifetime will be shortened. Based on such knowledge, in the leveling process of this embodiment, the number of times of leveling is optimized and the number of times of erasing for each block is made uniform.

  The leveling process is executed by the leveling unit 35 and the replacement source block selecting unit 33. FIG. 15 is a block diagram illustrating a configuration of the leveling unit 35. The leveling unit 35 includes a leveling determination unit 35A and a storage unit 35B that stores a leveling threshold.

  The leveling determination unit 35A uses the leveling threshold, the allocation block information sent from the allocation block selection unit 32, and the replacement source block information sent from the replacement source block selection unit 33 to perform the leveling process during the current write operation. Determine whether to do it. When the leveling process is performed, the leveling determination unit 35A issues a read request, an erase request, and a write request accompanying the leveling process to the NAND interface circuit 25.

  FIG. 16 is a flowchart showing the leveling process performed by the NAND controller 11. First, the replacement source block selection unit 33 performs replacement source block selection processing (step S70). As a result of this selection process, replacement source block information is sent from the replacement source block selection unit 33 to the leveling unit 35.

  Subsequently, the leveling determination unit 35A confirms the allocation block information sent from the allocation block selection unit 32, and determines whether or not the short-term rewrite flag included in the allocation block information is set (step S71). If the short-term rewrite flag is not set, that is, if the allocation block is not rewritten frequently, the leveling determination unit 35A determines that the difference in the number of erasures between the replacement source block and the allocation block exceeds the leveling threshold. It is determined whether or not there is (step S72). If the leveling threshold is not exceeded, the leveling unit 35 does not perform leveling in the current write operation.

  This leveling threshold is used to determine whether or not to perform the leveling process, and is set based on a standard of how much the lifetime (or data retention period) of the NAND flash memory 10 is extended. Is done. As described above, if an attempt is made to extend the lifetime of the NAND flash memory 10 by focusing only on equalizing the number of times of erasing, erasing processing accompanying leveling frequently occurs, resulting in an increase in the number of times of erasing to a specific area. End up. Therefore, in the present embodiment, the leveling process is executed only when the difference in the number of erasures between the replacement source block and the allocation block exceeds the leveling threshold.

  When the short-term rewrite flag is set in step S71 or when the leveling threshold is exceeded in step S72, the block control unit 30 executes block release processing for the replacement source block (step S73). This block release processing is the same as in FIG.

  Subsequently, the block control unit 30 executes a block allocation process for allocating the allocation block selected by the allocation block selection unit 32 as a block for moving the data of the replacement source block (step S74). This block allocation process is the same as in FIG.

  Subsequently, the NAND controller 11 reads the data of the replacement source block included in the NAND flash memory 10 (step S75). Specifically, the leveling determination unit 35A issues a read request to the NAND interface circuit 25 using the replacement source block information sent from the replacement source block selection unit 33. Based on this read request, the NAND interface circuit 25 instructs the NAND flash memory 10 to read data from the replacement source block. The read data is temporarily stored in the RAM 24 or the like.

  Subsequently, the NAND controller 11 erases the data of the allocation block (step S76). Specifically, the leveling determination unit 35A issues an erasure request to the NAND interface circuit 25 using the allocation block information sent from the allocation block selection unit 32. Based on this erase request, the NAND interface circuit 25 instructs the NAND flash memory 10 to erase the data in the allocated block. At this time, the block control unit 30 updates the erase count and erase time of the allocated block (see FIG. 11).

  Subsequently, the NAND controller 11 writes the data read from the replacement source block into the allocation block included in the NAND flash memory 10 (step S77). Specifically, the leveling determination unit 35A issues a write request to the NAND interface circuit 25 using the allocation block information. Based on this write request, the NAND interface circuit 25 instructs the NAND flash memory 10 to write data to the allocation block.

  As described above, since the data of the replacement source block can be transferred to the allocation block and the replacement source block can be made empty, the replacement source block that is considered not to be rewritten so much must be used again as the allocation block. Is possible. In addition, it is possible to prevent the number of erasures of the allocation block from increasing in the future by transferring the data that is not rewritten so much to the allocation block in which the number of erasures has already increased.

[2-6-1. Replacement source block selection process]
FIG. 17 is a block diagram illustrating a configuration of the replacement source block selection unit 33. FIG. 18 is a flowchart showing replacement source block selection processing by the replacement source block selection unit 33.

  The replacement source block selector 33 includes two selectors 33A and 33B. The replacement source block selector 33 receives all block information from the block controller 30 (step S80). This all block information is sent to the selector 33A. Subsequently, the selector 33A confirms the state of all the blocks, and extracts a block in use among all the blocks (step S81). Then, the selector 33A sends block information (in-use state block information) corresponding to the in-use block to the selector 33B.

  Subsequently, the selector 33B selects a block with the oldest erase time among the blocks extracted by the selector 33A as a replacement source block (step S82). The replacement source block information corresponding to the replacement source block is sent to the leveling unit 35. By such replacement source block selection processing, it is possible to select a block with good retention characteristics that stores data that has not been frequently rewritten as the replacement source block.

  As described above in detail, in this embodiment, the erase time when each block is erased is measured, and each block and the erase time are associated with each other and stored in the block table 30B. When data supplied from the outside is written to the NAND flash memory 10, the block with the oldest erase time is selected as an allocation block among the free blocks, and the data is written to this allocation block. .

  Therefore, according to the present embodiment, the erase interval can be extended for each block. For this reason, it is possible to reduce the deterioration of the retention characteristic of each block by utilizing the characteristic of the memory cell transistor that the retention characteristic is recovered by increasing the erase interval. As a result, the lifetime of the NAND flash memory 10 can be extended.

  Further, in the present embodiment, a block that is frequently rewritten by the short-term rewrite detection unit 34 is specified, and when this block is selected as an allocation block, a replacement source block and an allocation block having an erasure time old are determined. I try to replace it. Then, data that has been stored in the replacement source block and not frequently rewritten is transferred to the allocation block. As a result, the replacement source block that has been allocated once written and is not released for a long period of time is also released by the leveling process, so the replacement source block that is considered to have not been rewritten so much is used again as the allocation block It becomes possible to do. In addition, since data that has not been frequently rewritten is stored in the allocation block, it is possible to reduce the number of subsequent data erasures and, in turn, reduce deterioration of the retention characteristics of the allocation block. It becomes possible.

  Further, the leveling process is performed when the difference in the number of erasures between the replacement source block and the allocation block exceeds a threshold value. Thereby, it is possible to optimize the number of times of leveling without frequently performing leveling and to equalize the number of times of erasing for each block. As a result, it is possible to prevent the entire life of the NAND flash memory 10 from being shortened by increasing the number of times of erasure of some blocks.

[Second Embodiment]
The second embodiment shows another example of allocation block selection processing by the allocation block selection unit 32, and when an allocation block is selected, a block with an erasure time old and a small number of erasures is selected as the allocation block. Like to do.

  FIG. 19 is a block diagram illustrating a configuration of the allocation block selection unit 32 according to the second embodiment. FIG. 20 is a flowchart showing allocation block selection processing by the allocation block selection unit 32. The configuration of the NAND controller 11 other than the allocation block selection unit 32 is the same as that of the first embodiment.

  The allocation block selection unit 32 includes three selectors 32A to 32C and a storage unit 32D that stores an allocation block setting value. The allocation block selection unit 32 receives all block information from the block control unit 30 (step S90). This all block information is sent to the selector 32A. Subsequently, the selector 32A confirms the state of all blocks, and extracts an empty block from all blocks (step S91). Then, the selector 32A sends block information (empty state block information) corresponding to the free block to the selector 32B.

Subsequently, the selector 32B extracts block information of the condition set by the allocation block setting value from the free state block information (step S92). Here, the assigned block setting value includes
(A) A fixed number of blocks from the oldest erase time (B) A certain percentage of blocks from the oldest erase time (C) Any block whose erase time is older than the fixed time is set. Which of the conditions (A) to (C) is used as the allocation block set value can be arbitrarily selected. For example, when the condition (A) is used as the allocation block setting value, the selector 32B extracts a certain number of blocks from the oldest erased block among the empty blocks. Then, the selector 32B sends block information corresponding to the extracted block to the selector 32C.

  Conditions (A) and (B) sort empty blocks in order of erasure time, and search for a certain number / constant ratio of blocks from the oldest one. Therefore, since the number of blocks selected in step S92 can be increased, it is possible to increase the probability of selecting a block with a smaller number of erases in step S93. The condition (C) has the least processing load because it is only necessary to sort free blocks whose erase time exceeds a certain time. In addition, conditions (A) and (B) always extract a fixed number / percentage of blocks from the oldest erase time, whereas in condition (C), blocks whose erase time is more than a certain period are extracted. Is done. For this reason, in the condition (C), the erasure interval can always be kept constant, and the number of erasure comparison candidates can be reduced accordingly.

  Subsequently, the selector 32C selects, as an allocation block, a block having the smallest erase count among the blocks extracted by the selector 32B (step S93). The allocation block number corresponding to this allocation block is sent to the block control unit 30. Also, the allocation block information corresponding to this allocation block is sent to the leveling unit 35.

  As described above in detail, according to the present embodiment, when selecting an allocation block for writing data supplied from the outside, among the blocks in the empty state, allocation is performed for blocks having an erasure time that is old and having a small number of erasures. It becomes possible to select as a block. As a result, data from the outside can be written in a block with better retention characteristics, so that the lifetime of the NAND flash memory 10 can be extended.

[Third Embodiment]
The third embodiment shows another example of replacement source block selection processing by the replacement source block selection unit 33. When a replacement source block is selected, a block whose erasure time is old and the number of erasures is small is replaced. The original block is selected.

  FIG. 21 is a block diagram illustrating a configuration of the replacement source block selection unit 33 according to the third embodiment. FIG. 22 is a flowchart showing replacement source block selection processing by the replacement source block selection unit 33.

  The replacement source block selection unit 33 includes three selectors 33A to 33C and a storage unit 33D that stores replacement source block setting values. The replacement source block selector 33 receives all block information from the block controller 30 (step S100). This all block information is sent to the selector 33A. Subsequently, the selector 33A confirms the state of all blocks, and extracts a block in use from all blocks (step S101). Then, the selector 33A sends block information (in-use state block information) corresponding to the in-use block to the selector 33B.

Subsequently, the selector 33B extracts block information of the condition set by the replacement source block setting value from the in-use block information (step S102). Here, the replacement source block setting value includes
(A) A fixed number of blocks from the oldest erase time (B) A certain percentage of blocks from the oldest erase time (C) Any block whose erase time is older than the fixed time is set. Which of the conditions (A) to (C) is used as the replacement-source block setting value can be arbitrarily selected. For example, when the condition (A) is used as the replacement source block setting value, the selector 33B extracts a certain number of blocks from the oldest erased block among the blocks in use. Then, the selector 33B sends block information corresponding to the extracted block to the selector 33C.

  Conditions (A) and (B) sort empty blocks in order of erasure time, and search for a certain number / constant ratio of blocks from the oldest one. Therefore, since the number of blocks selected in step S102 can be increased, it is possible to increase the probability of selecting a block with a smaller number of erases in step S103. The condition (C) has the least processing load because it is only necessary to sort free blocks whose erase time exceeds a certain time. In addition, conditions (A) and (B) always extract a fixed number / percentage of blocks from the oldest erase time, whereas in condition (C), blocks whose erase time is more than a certain period are extracted. Is done. For this reason, in the condition (C), the erasure interval can always be kept constant, and the number of erasure comparison candidates can be reduced accordingly.

  Subsequently, the selector 33C selects the block with the smallest erase count among the blocks extracted by the selector 33B as the replacement source block (step S103). The replacement source block information corresponding to the replacement source block is sent to the leveling unit 35.

  As described above in detail, according to the present embodiment, when selecting a replacement source block to be used at the time of leveling, among blocks in use, a block with an erasure time that is old and has a small number of erasures is replaced. The original block can be selected. As a result, a block having good retention characteristics storing data that has not been frequently rewritten can be selected as a replacement source block. As a result, the NAND flash memory 10 can have a long lifetime.

[Fourth Embodiment]
The fourth embodiment shows another example of short-term rewrite detection processing, and short-term rewrite detection processing is performed using a binary search method.

  FIG. 23 is a block diagram illustrating a configuration of the short-term rewrite detection unit 34 according to the fourth embodiment. The short-term rewrite detection unit 34 includes a selector 34A, an alignment unit 34B, a search list 34C, a determination unit 34D, a storage unit 34E that stores a concentration factor threshold value, and an information output unit 34F.

  The determination unit 34D selects a search target block using the binary search method in the search list 34C. Then, a concentration factor used for short-term rewriting detection is calculated for this search target block, and it is determined whether or not this concentration factor exceeds a concentration factor threshold. Based on the determination result, blocks are excluded from the search list 34C, and the blocks in the search list 34C are narrowed down to one search target block. Then, using the last remaining search target block, a boundary where the erasure time changes rapidly is searched.

  The concentration factor threshold value stored in the storage unit 34E is used to specify a range of blocks that are frequently rewritten, and how long the lifetime (or data retention period) of the NAND flash memory 10 is. It is set on the basis of whether to extend. Lowering the concentration factor threshold increases the possibility of short-term rewrite detection, while increasing the concentration factor threshold decreases the possibility of short-term rewrite detection.

  24 and 25 are flowcharts showing the short-term rewrite detection process by the short-term rewrite detection unit 34.

  First, the short-term rewrite detector 34 receives all block information from the block controller 30 (step S110). This all block information is sent to the selector 34A. Subsequently, the selector 34A confirms the state of all the blocks, and extracts a block in use among all the blocks (step S111). Then, the selector 34A sends block information (in-use state block information) corresponding to the in-use block to the alignment unit 34B.

  Subsequently, the aligning unit 34B confirms the erase time included in the in-use state block information, and sorts the in-use blocks in the order of the erase time (step S112). Then, the block information sorted by the sorting unit 34B is put in the search list 34C, and a list number is given in order from the block with the newest erase time.

  Subsequently, the determination unit 34D sets “start” for the block with the newest erase time and “end” for the block with the oldest erase time in the search list 34C (step S113). Subsequently, the determination unit 34D selects a block whose list number is “(start + end) / 2” as a search target block (step S114).

Subsequently, the determination unit 34D calculates a concentration factor for the search target block (step S115). This concentration factor is given by:
Concentration factor = Current time−List number−Erase time The current time used to calculate the concentration factor is supplied from the erase time measurement unit 31 to the short-term rewrite detection unit 34, for example. As the current time, for example, the count value of the erase count counter 31A may be used.

  Subsequently, the determination unit 34D determines whether or not the calculated concentration coefficient exceeds a concentration coefficient threshold value (step S116). If the concentration factor threshold value is exceeded, the determination unit 34D sets the search target block to “end” (step S117). Subsequently, the determination unit 34D excludes from the search list 34C blocks having an erasing time older than “end” (step S117).

  On the other hand, when it is determined in step S116 that the concentration factor threshold value has not been exceeded, the determination unit 34D sets “start” as the block whose erasure time is next to the search target block (step S119). Subsequently, the determination unit 34D excludes from the search list 34C a block whose erase time is newer than “start” (step S120).

  After a predetermined block is excluded from the search list 34C in step S118 or step S120, the determination unit 34D determines whether two or more blocks remain in the search list 34C (step S121). When two or more blocks remain in the search list 34C, the process returns to step S114, and the search target block is selected again.

  When it is determined in step S121 that two or more blocks do not remain in the search list 34C, the determination unit 34D selects the last remaining block in the search list 34C as a search target block. Then, a concentration factor (= current time−list number−erasing time) is calculated for this search target block (step S122).

  Subsequently, the determination unit 34D determines whether or not the calculated concentration factor exceeds a concentration factor threshold value (step S123). When the concentration factor threshold value is exceeded, the determination unit 34D sends the search target block number to the information output unit 34F. Using this search target block number, the information output unit 34F calculates a range of blocks that have a newer erase time than the search target block among all the blocks (step S124). And the information output part 34F sends the said range to the block control part 30 as short-term rewriting information. Upon receipt of the short-term rewrite information, the block control unit 30 sets a short-term rewrite flag corresponding to the block included in the short-term rewrite information, and clears the short-term rewrite flag corresponding to the other blocks.

  If it is determined in step S123 that the concentration factor threshold is not exceeded, it is determined that the block subjected to the short-term rewrite cannot be detected, and the short-term rewrite detection process is terminated.

  FIG. 26 is a diagram for explaining a specific example of short-term rewrite detection processing by the short-term rewrite detection unit 34. There are 12 blocks (blocks # 0 to 11) included in the NAND flash memory 10. Of these, blocks # 0 to 7 are in-use blocks included in the search list 34C, and blocks # 8 to 11 are in use. Is empty. Blocks # 0 to # 7 are sorted in the order of erasure. The erase time for each block is as shown in the figure. The block numbers 0 to 7 included in the search list 34C are given list numbers from the newest erase time.

  First, in the search list 34C, the block # 0 with the newest erase time is set to “start”, and the block # 7 with the oldest erase time is set to “end”. Subsequently, block # 3 whose list number is “(start + end) / 2” is selected as a search target block, and a concentration factor (= current time−list number−erasure time) is calculated for search target block # 3. Then, it is determined whether or not the concentration factor exceeds a concentration factor threshold value. Here, it is assumed that the current time is “106” and the concentration factor threshold is “90”. The concentration factor of the search target block # 3 is “97”, which exceeds the threshold value. Therefore, the search target block # 3 is set to “end”, and the blocks # 4 to # 7 whose erase time is older than the block # 3 are excluded from the search list 34C.

  Subsequently, the block # 1 whose list number is “(start + end) / 2” is selected as a search target block, and a concentration factor is calculated for the search target block # 1. The concentration factor of the search target block # 1 is “0”, which does not exceed the threshold value. Therefore, block # 2, which has the oldest erase time next to search target block # 1, is set to “start”, and blocks # 0, 1 whose erase time is newer than block # 2 are excluded from search list 34C.

  Subsequently, the block # 2 whose list number is “(start + end) / 2” is selected as a search target block, and a concentration factor is calculated for the search target block # 2. The concentration factor of the search target block # 2 is “97”, which exceeds the threshold value. Therefore, the search target block # 2 is set to “end”, and the block # 3 having an erasure time older than the block # 2 is excluded from the search list 34C.

  Subsequently, the block # 2 remaining last in the search list 34C is selected as a search target block, and a concentration factor is calculated for the search target block # 2. The concentration factor of search target block # 2 exceeds the threshold value. Therefore, the number of the search target block # 2 is sent from the determination unit 34D to the information output unit 34F.

  The information output unit 34F calculates blocks # 0, 1, and 8 to 11 whose erasure time is newer than the search target block # 2 among all the blocks including the empty state blocks. Then, the information output unit 34F sends the blocks # 0, 1, and 8 to 11 to the block control unit 30 as short-term rewrite information. On the other hand, the block control unit 30 sets short-term rewrite flags corresponding to the blocks # 0, 1, 8 to 11, and clears short-term rewrite flags corresponding to other blocks. In this way, the information in the block table 30B is updated.

As described above in detail, according to the present embodiment, it is possible to identify a block that is frequently rewritten by the short-term rewrite detection unit 34. Furthermore, in the short-term rewrite detection process, the search is completed with “log ₂ N” threshold checks, so the efficiency of short-term rewrite detection can be improved.

[Fifth Embodiment]
In the fifth embodiment, the NAND controller 11 includes an error correction circuit (ECC (Error Check and Correct) circuit), and an error is detected and corrected by the ECC circuit when data is read. In general, the number of errors tends to increase as time elapses from data writing. Therefore, it can be said that a block with a large number of errors is a block whose erasure time is old. Therefore, in this embodiment, the number of errors by the ECC circuit is used as an index for selecting the replacement source block.

  FIG. 27 is a block diagram illustrating an example of the configuration of the NAND controller 11 according to the fifth embodiment. The NAND controller 11 includes an ECC circuit 36. The ECC circuit 36 detects and corrects an error when reading data from the NAND flash memory 10. Further, the ECC circuit 36 sends the number of errors detected at the time of error detection to the block control unit 30 for each block.

  FIG. 28 is a block diagram illustrating a configuration of the block control unit 30. The block table 30B included in the block control unit 30 stores the number of errors for each block number. The number of errors included in the block table 30B is updated every time data is read from the NAND flash memory 10 (each time an error is corrected by the ECC circuit 36).

  Next, the read operation of the NAND controller 11 will be described. FIG. 29 is a flowchart showing a read operation by the NAND controller 11.

  First, the NAND controller 11 starts a read operation by receiving a read request from the CPU 2 (step S130). Subsequently, the block control unit 30 searches the block number corresponding to the address area including the address of the read request using the address table 30A (step S131).

  Subsequently, the block control unit 30 reads data from the block requested to be read (step S132). That is, the block control unit 30 issues a read request to the NAND interface circuit 25. Based on this read request, the NAND interface circuit 25 instructs the NAND flash memory 10 to read data from the read request block.

  Subsequently, the ECC circuit 36 performs error detection and correction on the data read from the NAND flash memory 10 (step S133). The error-corrected read data is sent to the CPU 2 and the like via the data bus 6. At the time of this error correction, the ECC circuit 36 calculates the number of errors and sends this number of errors to the block control unit 30. Upon receiving the number of errors from the ECC circuit 36, the block control unit 30 records this number of errors in the block table 30B. In this way, the number of errors included in the block table 30B is updated.

  Next, replacement source block selection processing by the replacement source block selection unit 33 will be described. FIG. 30 is a flowchart showing replacement source block selection processing by the replacement source block selector 33. The configuration of the replacement source block selection unit 33 is the same as that in FIG.

  The replacement source block selector 33 receives all block information from the block controller 30 (step S140). This all block information is sent to the selector 33A. Subsequently, the selector 33A confirms the state of all the blocks, and extracts a block in use among all the blocks (step S141). Then, the selector 33A sends block information (in-use state block information) corresponding to the in-use block to the selector 33B.

  Subsequently, the selector 33B extracts block information of the condition set by the replacement source block setting value from the in-use block information (step S142). Here, “blocks with more errors than a certain number” is set as the replacement source block setting value.

  Subsequently, the selector 33C selects the block with the smallest erase count among the blocks extracted by the selector 33B as the replacement source block (step S143). The replacement source block information corresponding to the replacement source block is sent to the leveling unit 35.

  As described above in detail, according to the present embodiment, in the writing process of the memory system 1, it is possible to replace a block in which rewriting is concentrated in the short term with a block that has been allocated by writing and is not released for a long period of time. As a result, it is possible to suppress the block consumption due to the short interval of the erasing time and to average the block consumption over the entire NAND flash memory 10.

  Furthermore, a block with a large number of errors due to ECC is selected as the replacement source block. As a result, the data of the block having a large number of errors is rewritten, so that the refresh processing (reading data stored in the NAND flash memory 10 and performing error correction at the same time as the leveling processing, and then again NAND flash memory 10 To write back.) As a result, since the number of refresh processes can be reduced, an effect of reducing the data write amount accompanying the refresh can be obtained.

[Example]
An example when the memory system 1 of each of the above embodiments is configured as an SSD (Solid State Drive) will be described. FIG. 31 is a block diagram showing the configuration of the SSD 100. As shown in FIG.

  The SSD 100 includes a plurality of NAND flash memories (NAND memories) 10 for data storage, a DRAM 101 for data transfer or work area, a drive control circuit 102 for controlling these, and a power supply circuit 103. The drive control circuit 102 outputs a control signal for controlling a status display LED provided outside the SSD 100. Instead of the DRAM 101, FeRAM (ferroelectric random access memory) may be used.

  The SSD 100 transmits / receives data to / from a host device such as a personal computer via an ATA interface (ATA I / F). Further, the SSD 100 transmits / receives data to / from the debugging device via the RS232C interface (RS232C I / F).

  The power supply circuit 103 receives an external power supply and generates a plurality of internal power supplies using the external power supply. These internal power supplies are supplied to each unit in the SSD 100. Further, the power supply circuit 103 detects the rise of the external power supply and generates a power-on reset signal. The power-on reset signal is sent to the drive control circuit 102.

  FIG. 32 is a block diagram showing a configuration of the drive control circuit 102. The drive control circuit 102 includes a data access bus 104, a first circuit control bus 105, and a second circuit control bus 106.

  A processor 107 that controls the entire drive control circuit 102 is connected to the first circuit control bus 105. A boot ROM 108 storing a boot program for each management program (FW: firmware) is connected to the first circuit control bus 105 via a ROM controller 109. The first circuit control bus 105 is connected to a clock controller 110 that receives a power-on reset signal from the power supply circuit 103 and supplies a reset signal and a clock signal to each unit.

  The second circuit control bus 106 is connected to the first circuit control bus 105. Connected to the second circuit control bus 106 are a parallel IO (PIO) circuit 111 that supplies a status display signal to the status display LED and a serial IO (SIO) circuit 112 that controls the RS232C interface.

  An ATA interface controller (ATA controller) 113, a first ECC (Error Check and Correct) circuit 114, a NAND controller 115, and a DRAM controller 119 are provided on both the data access bus 104 and the first circuit control bus 105. It is connected. The ATA controller 113 transmits and receives data to and from the host device via the ATA interface. An SRAM 120 used as a data work area is connected to the data access bus 104 via an SRAM controller 121.

  The NAND controller 115 includes a NAND interface circuit (NAND I / F) 118 that performs interface processing with the four NAND memories 10, a second ECC circuit 117, and a DMA transfer control DMA that performs access control between the NAND memory and the DRAM. A controller 116 is provided.

  FIG. 33 is a block diagram illustrating a configuration of the processor 107. The processor 107 includes a data management unit 122, an ATA command processing unit 123, a security management unit 124, a boot loader 125, an initialization management unit 126, and a debug support unit 127.

  The data management unit 122 controls data transfer between the NAND memory and the DRAM and various functions related to the NAND chip via the NAND controller 115 and the first ECC circuit 114.

  The ATA command processing unit 123 performs data transfer processing in cooperation with the data management unit 122 via the ATA controller 113 and the DRAM controller 119. The security management unit 124 manages various types of security information in cooperation with the data management unit 122 and the ATA command processing unit 123. The boot loader 125 loads each management program (FW) from the NAND memory 10 to the SRAM 120 when the power is turned on.

  The initialization manager 126 initializes each controller / circuit in the drive control circuit 102. The debug support unit 127 processes debug data supplied from the outside via the RS232C interface.

  FIG. 34 is a perspective view showing an example of a portable computer 200 on which the SSD 100 is mounted. The portable computer 200 includes a main body 201 and a display unit 202. The display unit 202 includes a display housing 203 and a display device 204 accommodated in the display housing 203.

  The main body 201 includes a housing 205, a keyboard 206, and a touch pad 207 that is a pointing device. The housing 205 accommodates a main circuit board, an ODD (optical disk device) unit, a card slot, an SSD 100, and the like.

  The card slot is provided adjacent to the peripheral wall of the housing 205. An opening 208 facing the card slot is provided on the peripheral wall. The user can insert / remove an additional device into / from the card slot from the outside of the housing 205 through the opening 208.

  The SSD 100 may be used as a state of being mounted inside the portable computer 200 as a replacement for a conventional HDD, or may be used as an additional device while being inserted into a card slot included in the portable computer 200.

The memory system 1 of the above embodiment is not limited to the SSD, for example, it can be configured as a memory card represented by SD ^TM card. When the memory system 1 is configured as a memory card, it is applicable not only to portable computers but also to various electronic devices such as mobile phones, PDAs, digital still cameras, and digital video cameras.

  The present invention is not limited to the above-described embodiment, and can be embodied by modifying the components without departing from the scope of the invention. In addition, various inventions can be configured by appropriately combining a plurality of constituent elements disclosed in the embodiments. For example, some constituent elements may be deleted from all the constituent elements disclosed in the embodiments, or constituent elements of different embodiments may be appropriately combined.

  DESCRIPTION OF SYMBOLS 1 ... Memory system, 2 ... CPU, 3 ... Main memory, 4 ... Memory controller, 10 ... NAND type flash memory, 11 ... NAND controller, 21 ... Host interface circuit, 22 ... MPU, 23 ... ROM, 24 ... RAM, 25 ... NAND interface circuit, 30 ... Block control unit, 30A ... Address table, 30B ... Block table, 30C, 31B ... Calculation unit, 31 ... Erase time measurement unit, 31A ... Erase count counter, 32 ... Allocation block selection unit, 32A- 32C, 33A to 33C ... selector, 32D, 33D, 34E, 35B ... storage unit, 33 ... replacement source block selection unit, 34 ... short-term rewrite detection unit, 34A ... selector, 34B ... alignment unit, 34C ... search list, 34D ... Determination unit, 34F ... information output unit, 35 ... leveling unit, 35 Reference level determining unit 36: ECC circuit 100 ... SSD 101: DRAM 102 ... Drive control circuit 103 ... Power supply circuit 104 ... Data access bus 105 ... First circuit control bus 106 ... No. 2 circuit control buses 107: processor 108 boot ROM 109 ROM controller 110 clock controller 111 PIO circuit 112 SIO circuit 113 ATA interface controller 114 first ECC circuit DESCRIPTION OF SYMBOLS 115 ... NAND controller, 116 ... DMA controller, 117 ... 2nd ECC circuit, 118 ... NAND interface circuit, 119 ... DRAM controller, 120 ... SRAM, 121 ... SRAM controller, 122 ... Data management part, 123 ... ATA command processing part , 24: Security management unit, 125 ... Boot loader, 126 ... Initialization management unit, 127 ... Debug support unit, 200 ... Portable computer, 201 ... Main body, 202 ... Display unit, 203 ... Display housing, 204 ... Display device, 205 ... Housing Body, 206 ... keyboard, 207 ... touch pad, 208 ... opening.

Claims (22)

A control method of a memory system including a non-volatile memory having a plurality of blocks as data erasing units,
A step of measuring the erase time when the data of each block was erased;
For each block, creating a block table that associates a state value indicating an empty state or a busy state with the erase time;
Detecting a block in which rewriting is concentrated in the short term;
Selecting a first block that is free and has an older erase time based on the information in the block table;
Selecting a second block that is in use and has an older erase time based on the information in the block table;
Moving the data of the second block to the first block when the first block is included in the detected block;
A method for controlling a memory system, comprising:   2. The method of controlling a memory system according to claim 1, wherein the first block is selected again when data is moved from the second block to the first block. Further comprising the step of counting the number of times each block is erased,
3. The memory system control method according to claim 1, wherein the block table stores the number of times of erasure.   When the first block is included in the detected block, or when the difference in the number of erasures between the first block and the second block exceeds a threshold, the data of the second block is 4. The method of controlling a memory system according to claim 3, wherein the memory system moves to the first block.   5. The detecting step detects a boundary having a large difference in erase time between blocks, and detects a block whose erase time is newer than the boundary as a block in which rewriting is concentrated in a short time. A method for controlling a memory system according to any one of the above.   6. The memory system control method according to claim 1, wherein the block table stores a flag indicating whether or not rewriting is concentrated in a short period of time. The detecting step includes
A step of calculating a difference in erase time between an arbitrary third block among the blocks in use and a fourth block having the next erase time after the third block;
7. The memory according to claim 1, further comprising a step of detecting a block whose erase time is newer than that of the fourth block among all the blocks when the difference is larger than a predetermined interval. How to control the system.   The calculating step arranges information on blocks in use in order of erasing time, and calculates a difference in erasing time between the third block and the fourth block based on the arranged information. The method of controlling a memory system according to claim 7. The detecting step includes
Calculating the difference between the current time and the erase time for a block in use;
Determining a fifth block having the latest erase time among the blocks in which the difference exceeds the threshold;
7. The method of controlling a memory system according to claim 1, further comprising a step of detecting a block whose erase time is newer than the fifth block among all blocks.   The calculating step arranges information on blocks in use in order of erasure time, and determines whether a difference between the current time and the erasure time of the block exceeds a threshold based on the arranged information. The method of controlling a memory system according to claim 9.   11. The memory system control method according to claim 1, wherein the first block is free and has the oldest erase time.   11. The memory system control according to claim 1, wherein the first block has the smallest number of times of erasing among a predetermined number of blocks from the vacant state and the oldest erasing time. Method.   11. The memory system according to claim 1, wherein the first block has the smallest number of times of erasure among a predetermined number of blocks from the vacant state and the oldest erase time. Control method.   11. The method of controlling a memory system according to claim 1, wherein the first block has the smallest number of times of erasure among blocks which are free and have an erase time older than a certain time.   15. The method according to claim 1, wherein the second block is in use and has the oldest erase time.   15. The memory system according to claim 1, wherein the second block has the smallest number of times of erasing among a predetermined number of blocks in the in-use state and the oldest erasing time. Control method.   15. The memory system according to claim 1, wherein the second block has the smallest number of times of erasure among a certain number of blocks from the one in use and the oldest erase time. Control method.   15. The method of controlling a memory system according to claim 1, wherein the second block has the smallest erase count among blocks in use and whose erase time is older than a certain time. Further comprising calculating the number of errors in the data read from the non-volatile memory,
The block table stores the number of errors for each block;
15. The method of controlling a memory system according to claim 1, wherein the second block has the smallest number of erasures among blocks in use and having a larger number of errors than a certain number. The measuring step counts the total number of erasures executed in all blocks,
20. The method of controlling a memory system according to claim 1, wherein the erase time corresponds to the total number of erase times. The step of measuring measures the time when each block was erased,
20. The memory system control method according to claim 1, wherein the erasing time corresponds to the time. The measuring step measures the energization time of the memory system when each block is erased,
20. The memory system control method according to claim 1, wherein the erasing time corresponds to the energization time.

Images