JP5579135B2 - Data storage device, memory control device, and memory control method - Google Patents

Description

  Embodiments described herein relate generally to a data storage device, a memory control device, and a memory control method that use a nonvolatile memory.

  2. Description of the Related Art In recent years, development of an SSD (solid state drive) using a NAND flash memory (hereinafter sometimes simply referred to as a flash memory), which is a rewritable nonvolatile memory, has been promoted as a data storage device.

  In the flash memory, data erasure processing is necessary before the data is written into a storage area where data is written. In the flash memory, a writing unit (page unit) and an erasing unit (block unit) for erasing processing are different. Here, one block is a storage area corresponding to, for example, 64 pages.

  In SSD, when data rewrite to the flash memory proceeds, the ratio of storage areas in which valid data cannot be stored in the block increases due to invalid data (data that is not the latest). The SSD executes a compaction process in order to effectively use the storage area in the block. The compaction process is a process in which valid data is collected from a block whose effective data storage area has become low density and rewritten to another block. The compaction target block is reproduced as a storage area that can be effectively used by the erasing process after the rewriting process.

JP 2009-211119 A

  In order to effectively perform the compaction process, it is important to search for a block having a lower density as a compaction target block. The search process for the compaction target block greatly affects the performance of the compaction process. Conventionally, a search method using a management table that classifies all block IDs according to the number of valid tracks has been proposed. However, this is not necessarily an effective search process because a large dedicated memory area is required for the management table.

  An object of the present invention is to provide a data storage device that can improve the efficiency of compaction processing by realizing effective compaction target block search processing.

According to the embodiment, the data storage device includes a flash memory having a block as a data erasing unit and a controller. The controller includes compaction processing means for performing compaction processing on the flash memory . The compaction processing unit calculates a threshold value of the effective data amount in the block based on the usable storage capacity of the flash memory and the effective data amount in the block, and the threshold value is determined according to the change in the usable storage capacity. And a search unit for searching for a block having a relatively small effective data amount as a target block of the compaction process based on the latest threshold acquired from the threshold calculation unit . And have.

The block diagram for demonstrating the structure of SSD regarding embodiment. 1 is a block diagram for explaining a configuration of a flash memory controller according to an embodiment. FIG. The figure for demonstrating the effective cluster number threshold value table regarding embodiment. 7 is a flowchart for explaining threshold value calculation processing for the number of effective clusters according to the embodiment. The flowchart for demonstrating the search process of the compaction object block regarding embodiment. The flowchart for demonstrating the compaction process regarding embodiment.

  Embodiments will be described below with reference to the drawings.

[Configuration of data storage device]
FIG. 1 is a block diagram illustrating a configuration of a data storage device according to the embodiment.

  As shown in FIG. 1, the data storage device is an SSD (solid state drive), and includes an SSD controller 10 and a NAND flash memory (hereinafter, flash memory) 11 that is a nonvolatile memory. The flash memory 11 is an SSD data storage medium, and can be externally connected in units of channels or packages, for example.

  The SSD controller 10 includes a flash memory controller 12, a host interface (host I / F) 13, a buffer management module 14, a buffer memory (DRAM) 15, and a microprocessor (CPU) 16. The CPU 16 performs overall control of the SSD controller 10. The DRAM 15 operates as a buffer memory and holds a logical / physical address conversion table, data transferred from the host system 1, and the like.

  The host I / F 13 controls the transfer of data, commands, and addresses between the host system 1 and the SSD controller 10. The host system 1 is, for example, a computer including an interface of SATA (Serial ATA) standard. The buffer management module 14 controls a buffer memory (DRAM) 15 and controls data transfer between the host I / F 13 and the buffer memory 15. The buffer management module 14 controls data transfer between the flash memory controller 12 and the buffer memory 15.

  The flash memory controller 12 includes an input / output interface (simply an interface) 20 and a memory access controller 21. The interface 20 performs data input / output between the flash memory 11 and the memory access controller 21 and data input / output between the flash memory 11 and the compaction module 24. The memory access controller 21 controls a data read operation and a write operation with respect to the flash memory 11 via the interface 20.

  The memory access controller 21 includes a controller 22, a block management module 23, and a compaction module 24. The controller 22 controls the operation of each module such as the block management module 23 and the compaction module 24 connected via the internal bus. The block management module 23 has management information for managing a table for managing the effective cluster of each block, a free block (unused standby state block), and the like, and manages each block of the flash memory 11.

(Compact module)
As shown in FIG. 2, the compaction module 24 includes a block search module 30, an effective cluster number threshold value calculation module (hereinafter sometimes abbreviated as a threshold value calculation module) 31, and a compaction block list management module (hereinafter referred to as a block list). A cluster search module 33, a command list generation module 34, and a command list input module 35.

  The block search module 30 searches for a candidate for a compaction target block from an active block group in which valid data is recorded. The threshold calculation module 31 calculates a threshold for the number of effective clusters (referred to as an effective cluster number threshold) used as a condition in the block search processing executed by the block search module 30. The block list management module 32 manages the compaction target block candidates and the blocks being subjected to compaction processing, respectively, according to a list structure.

  The cluster search module 33 determines whether the valid data or invalid data in the cluster unit, which is the smallest readable unit for each block in the compaction process, in the list managed by the block list management module 32. To do. The cluster search module 33 searches for valid clusters that are valid data in cluster units.

  The command list generation module 34 generates a compaction processing request command (compaction request command) set for the valid clusters searched by the cluster search module 33. The command list input module 35 inputs the compaction request command received from the command list generation module 34 to the interface 20. Thereby, the compaction process for the flash memory 11 is executed.

[Compact processing]
Hereinafter, the compaction processing of this embodiment will be described with reference to the flowcharts of FIGS. 3 and 4 to 6.

  As will be described later, the SSD executes a compaction process in order to effectively use the storage area in the block (see FIG. 6). In the compaction module 24 of this embodiment, as shown in FIG. 2, the modules 30 to 35 cooperate to execute compaction processing at high speed and efficiently.

  First, as shown in the flowchart of FIG. 5, the block search module 30 searches for a candidate for a compaction target block from an active block group in which valid data is recorded. In this case, in order to effectively perform the compaction process, it is necessary to search a low-density block with as few effective clusters as possible as a compaction target block.

  Here, the cluster is a minimum access unit when reading data from the flash memory 11. One cluster is, for example, a data amount corresponding to half of one physical page. In the case of a multi-channel SSD, a logical page is composed of physical pages of the number of channels × the number of planes. Therefore, the effective cluster is effective data in units of clusters. The effective cluster number (CN) is the effective data amount in one block.

  As will be described later, the block search module 30 uses a threshold value related to the number of effective clusters (CN) (effective cluster number threshold value TH) as a condition for searching for a compaction target block candidate. The effective cluster number threshold TH is calculated by the threshold calculation module 31. Hereinafter, the processing of the threshold calculation module 31 will be described with reference to the flowchart of FIG.

  As illustrated in FIG. 3, the threshold calculation module 31 includes a management table 100 that manages the effective cluster number threshold. In the management table 100, the margin ratio (M) of the storage capacity of the flash memory 11 and the effective cluster density (CD) are associated with each other. The margin ratio (M) is the ratio of the storage capacity that can actually be used to the device notation capacity.

  The threshold calculation module 31 accepts the occurrence notification from the block management module 23 at the timing when the unusable block (BB: Bad Block) occurs (block 400). When the threshold calculation module 31 accepts the occurrence notification, the threshold calculation module 31 calculates (recalculates) a margin rate (M) of the storage capacity (block 401). Here, the margin ratio (M) is calculated by the equation “M = MC ÷ TC × 100 (%)”. MC is a spare capacity, which is a capacity obtained by subtracting TC from the total storage capacity available for storing user data. TC is a device notation capacity of the flash memory 11.

  Further, the threshold calculation module 31 refers to the management table 100 and calculates an effective cluster density (CD) corresponding to the calculated margin rate (M) (block 402). Here, one cluster is a data amount corresponding to half of one page (physical page), for example. Further, one logical page is composed of, for example, 32 physical pages, and one block is a data amount of, for example, 64 logical pages. From these relationships, information in which the margin ratio (M) and the effective cluster density (CD) are associated with each other is created in advance as the management table 100.

  Next, the threshold calculation module 31 calculates the effective cluster number threshold (TH) from the effective cluster density (CD) calculated from the management table 100 (block 403). The effective cluster number threshold (TH) is calculated by the formula “TH = TN × CD”. TN is the total number of clusters in the logical block. The threshold value calculation module 31 stores the calculated effective cluster number threshold value (TH) in the internal memory.

  The threshold value calculation module 31 compares the previous effective cluster number threshold value (TH) stored in the internal memory with the current value calculated this time (block 404). If the threshold value has changed, the threshold value calculation module 31 updates the existing value in the internal memory to the current value (latest value) of the effective cluster number threshold value (TH) (block 405).

  In short, the threshold calculation module 31 obtains the corresponding effective cluster density (CD) from the management table 100 according to the fluctuation of the margin ratio of the flash memory 11, and uses the latest effective cluster number threshold (from the effective cluster density (CD)). TH) is calculated. The threshold calculation module 31 updates the existing value in the internal memory to the latest effective cluster number threshold (TH). The threshold value calculation module 31 may calculate the average effective cluster number of the active block group, calculate the latest effective cluster number threshold value (TH) based on the average value, and store it in the internal memory.

  Next, the block search module 30 acquires the latest effective cluster number threshold (TH) stored in the internal memory of the threshold calculation module 31 as shown in the flowchart of FIG. 5 (block 500). Further, the block search module 30 acquires the number of valid clusters (CN) from each active block (a block in which valid data is recorded) (block 501). In this case, the block module 30 acquires the effective cluster number (CN) for each active block from the block management module 23.

  When the block search module 30 searches for a block whose acquired effective cluster number (CN) is 0, the block search module 30 notifies the block management module 23 to release the block as a free block (unused standby state block) ( YES in block 502, 505). The controller 21 executes an erasing process on the block. The block management module 23 registers the block after the erasure process as a free block.

  On the other hand, if the number of valid clusters (CN) is not 0, the block search module 30 compares the number of valid clusters (CN) with the number of valid clusters threshold (TH) (block 503). When the number of valid clusters (CN) is smaller than the threshold (TH), the block search module 30 registers the block of the number of valid clusters (CN) as a compaction target block in the candidate list (block 504).

  The block search module 30 repeats the search process for a predetermined range of active blocks, and adds blocks with the number of effective clusters (CN) smaller than the threshold (TH) to the candidate list (block 506). Here, the block search module 30 may select a predetermined number of blocks as the compaction target blocks in the order from the smallest effective cluster number (CN) among the blocks whose effective cluster number (CN) is smaller than the threshold (TH). .

  The block list management module 32 manages the compaction target block candidates and the blocks being subjected to compaction processing, respectively, according to a list structure. Specifically, the block list management module 32 manages the block ID (identification information) of the block group selected by the block search module 30 as a compaction target block candidate in a list structure.

  The block list management module 32 has, for example, three types of lists as the compaction target block candidate list according to the purpose, and manages the compaction target block. The first type list is for normal compaction processing, and is used for managing the block ID of the selected block. In other words, in normal compaction processing, the controller 22 checks the number of free blocks managed by the block management module 23, and if it is less than the specified number, in order to avoid a free block depletion state, the block search module 30 starts the compaction target block search process.

  The second list is for compaction processing for error recovery purposes. For example, the controller 22 periodically performs read confirmation, and in the case of a block in which the error bit has exceeded a specified value, the controller 22 uses the data in the block to move to another block for the purpose of saving. Such compaction processing is sometimes called patrol refresh processing. The third list is for compaction processing for the purpose of wear leveling for extending the life of the flash memory 11. Since the lifetime of the flash memory 11 depends on the number of block erases and the erase time interval, it is necessary to equalize the fatigue level of the blocks. This compaction process is started when it is determined that wear leveling is necessary by a wear leveling module (not shown).

  The block list management module 32 assigns priorities according to each purpose from the compaction target block candidate list, and generates a compaction target block list during the compaction process. Hereinafter, the compaction process will be described in detail with reference to the flowchart of FIG.

  The cluster search module 33 determines whether each compaction target block is valid data or invalid data for each compaction target block in the list managed by the block list management module 32, and determines valid clusters that are valid data. A search is made (block 600).

  The command list generation module 34 generates a compaction processing request command (compaction request command) set for the valid clusters searched by the cluster search module 33 (block 601). The command list input module 35 inputs the compaction request command received from the command list generation module 34 to the interface 20 (blocks 602 and 603). Thereby, the compaction process for the flash memory 11 is executed (block 604).

  Here, the compaction request command is a command for reading a valid cluster from the compaction target block and rewriting it in another block. In the present embodiment, it is assumed that the command list generation module 34 issues commands in units of pages, which is the smallest writable unit. In the case of an SSD having a multi-channel configuration, for example, a command corresponding to a number of clusters of “number of channels−number of parity channels−number of NULL blocks of destination block” × number of planes × number of clusters per page ”is set as one set. Generate.

  When the compaction process is completed, the controller 21 executes an erasure process on the compaction target block. The block management module 23 generates a list for registering the erased block as a free block (block 606).

  As described above, according to the present embodiment, when searching for a compaction target block at the time of performing the compaction process, a low-density block with few valid clusters is determined based on the valid cluster number threshold. This effective cluster number threshold is updated as the latest threshold according to the fluctuation of the margin ratio of the memory capacity of the flash memory 11.

  With such a search method of the present embodiment, it is possible to search a compaction target block quickly and efficiently, and as a result, it is possible to reduce the compaction processing time. That is, when searching for a compaction target block, it is desirable that the threshold number of effective clusters is small in order to realize block search with the lowest possible density. On the other hand, however, if the margin ratio of the flash memory 11 decreases, it becomes difficult to search for candidates for the compaction target block. Conversely, when the effective cluster number threshold is set to a large value in advance with a margin, the efficiency of the compaction process decreases.

  Therefore, according to the present embodiment, it is possible to search for the compaction target block at high speed and efficiently by updating the effective cluster number threshold according to the fluctuation of the margin ratio. Further, the dedicated memory area required for the search process for the compaction target block is only the memory area for storing the effective cluster number threshold value table and the threshold value that the threshold value calculation module 31 has. For this reason, it is possible to minimize the dedicated memory area. In short, according to the present embodiment, an effective compaction target block search process suitable for the use state (margin ratio) of the flash memory can be realized with a minimum dedicated memory area.

  Although several embodiments of the present invention have been described, these embodiments are presented by way of example and are not intended to limit the scope of the invention. These novel embodiments can be implemented in various other forms, and various omissions, replacements, and changes can be made without departing from the scope of the invention. These embodiments and modifications thereof are included in the scope and gist of the invention, and are included in the invention described in the claims and the equivalents thereof.

1 ... Host system, 10 ... SSD controller, 11 ... NAND flash memory,
12: Flash memory controller,
13: Host interface (host I / F),
14 ... Buffer management module, 15 ... Buffer memory (DRAM),
16 ... Microprocessor (CPU), 20 ... I / O interface,
21 ... Memory access controller, 22 ... Controller,
23 ... Block management module, 24 ... Compaction module,
30 ... Block search module, 31 ... Valid cluster count threshold calculation module,
32 ... Compaction block list management module,
33 ... Cluster search module, 34 ... Command list generation module,
35 ... Command list input module.

Claims (9)

Flash memory with blocks as data erasure units;
A controller including compaction processing means for performing compaction processing on the flash memory,
The compaction processing means includes
The threshold value of the effective data amount in the block is calculated based on the usable storage capacity of the flash memory and the effective data amount in the block, and the threshold value is updated according to the change in the usable storage capacity. Threshold calculation means for calculating
A data storage device comprising: search means for searching for a block having a relatively small effective data amount as a target block for the compaction processing based on the latest threshold value acquired from the threshold value calculation means . The threshold calculation means includes
Table information in which a margin ratio indicating a ratio of the usable storage capacity is associated with an effective data density for calculating the threshold;
The data storage device according to claim 1, wherein the latest threshold value is calculated from the effective data density according to the fluctuation of the margin ratio with reference to the table information . The threshold value calculation means includes a memory means for storing the threshold value, and updates the threshold value stored in the memory means in accordance with a change in the margin ratio,
3. The data storage device according to claim 2, wherein the retrieval unit obtains a latest threshold value stored in the memory unit from the threshold value calculation unit, and retrieves a target block for the compaction process based on the latest threshold value. . The threshold calculation means includes
The data storage device according to claim 1, wherein an average effective data amount of a usable block group of the flash memory is calculated, and the latest threshold value is calculated based on the average value . The search means includes
5. The block according to claim 1, wherein a predetermined number of blocks are selected as target blocks of the compaction process in order of increasing effective data amount among the blocks having an effective data amount smaller than the latest threshold value. The data storage device described in 1. The controller is
A compaction process executing means for executing the compaction process on a target block of the compaction process searched by the compaction processing means;
The compaction process execution means includes:
Search valid data from the target block of the compaction process,
The data storage device according to any one of claims 1 to 5, wherein the valid data is rewritten in another block . The controller is
Means for erasing the target block of the compaction process after completion of the compaction process by the compaction process execution means;
The data storage device according to claim 6 , further comprising: a block management unit that manages the target block after the erasure process as a free block . A memory control device applied to a data storage device using a flash memory whose block is a data erasure unit,
A controller including compaction processing means for performing compaction processing on the flash memory;
The compaction processing means includes
The threshold value of the effective data amount in the block is calculated based on the usable storage capacity of the flash memory and the effective data amount in the block, and the threshold value is updated according to the change in the usable storage capacity. Threshold calculation means for calculating
Search means for searching a block having a relatively small effective data amount as a target block of the compaction process based on the latest threshold acquired from the threshold calculation means;
A memory control device . A memory control method applied to a data storage device including a flash memory having a block as a data erasing unit and a controller for performing compaction processing on the flash memory,
Based on the usable storage capacity of the flash memory and the effective data amount in the block, the threshold of the effective data amount in the block is calculated,
Calculate the latest threshold updated the threshold according to the change in the usable storage capacity,
A memory control method for retrieving a block having a relatively small effective data amount as a target block of the compaction process based on the latest threshold .

Images