TWI385520B - Management methods and systems for storage units - Google Patents

Description

Storage unit management method and system

The present invention relates to a storage unit management method and system, and more particularly to a storage unit management method and system that can be used for non-volatile memory to classify and manage physical blocks of storage units.

Currently, non-volatile memory, such as flash memory, has been used very widely. It is characterized in that it is read and written in units of pages, deleted in units of blocks, and must be deleted before writing. Due to the above characteristics, the read/write instructions of the logical block address (LBA) used by the host or the application are very difficult to manage, and it is difficult to convert the logical address into a physical address to access the non-volatile The physical block of sexual memory. Therefore, a conversion layer is usually used between the host's file system and non-volatile memory to manage read and write commands from the host to access data in the memory. In the conventional storage management of non-volatile memory, the data is usually accessed and addressed according to modes such as page mapping or block mapping.

In the page mapping mode, the data in the storage unit is stored in units of pages and has a corresponding page mapping table. In the page mapping table, the logical page number is an index, and the address of the entity page is an entry (project content). When the page mapping storage system is used to process the read and write commands, the page mapping table can be directly searched and updated to complete the positioning of the related data page. In the block mapping mode, the data in the storage unit is stored in units of blocks and has a corresponding block mapping table. In the block mapping table, The logical block number is an index, and the address of the physical block is an entry. When the block mapping storage system is used to process the read and write commands, the block mapping table can be directly searched and updated to complete the positioning of the related data page.

In recent years, with the development of integrated circuit technology, the page/block size of the storage unit has been greatly improved. In addition, high-density storage units, such as individual pages of each block in NAND flash memory, must be written in order and cannot be randomly written. In addition, since NAND flash memory has the characteristics of reading and writing in units of pages, deleting in units of blocks, and having to be deleted before writing. Therefore, when a block-mapped storage system processes a write command, a block must first be allocated so that the requested page can be written to the allocated block. After that, the other pages in the logical block belonging to the corresponding request page are copied into the allocated block, and the original data in the allocated block must be deleted first. Therefore, especially for demand pages that may include hot data that is frequently accessed by the host, when the block size is increased, the write performance is greatly reduced. In the page mapping mode, the distribution of data is very complicated and confusing. When the free block is used up, the process of merging the data to acquire the new free block will be very complicated.

Therefore, some storage management systems propose a hybrid mapping mode that can combine block mapping with page mapping. In these systems, the storage unit may include individual areas of the corresponding block mapping and page mapping, and the use of the individual areas can only be switched through the data merge operation. Although the storage management system with hybrid mapping mode can provide greater flexibility to store data, in actual use, there are still many unused entities in the physical block of the corresponding block mapping and page mapping in the storage unit. page, In turn, the distribution of data in the storage unit is very complicated and confusing. Similarly, the consolidation of related data will also cause a systemic performance burden.

In view of this, the present invention provides a storage unit management method and system.

The storage unit management method of the embodiment of the present invention is configured to manage a storage unit including a plurality of physical blocks, wherein each physical block has one of a plurality of block type definitions. First, a sub-write command is obtained, wherein the sub-write command requires a data write operation for at least one logical page of the corresponding at least one logical block. Determining whether there is one candidate block in the storage unit having one of the block type definitions, wherein the logical page of the logical block according to the first block definition cannot be mapped to the candidate block. If the candidate block exists, the first block definition is converted to one of the block type definitions. A material is written to a specific page in the candidate block, and a logical page of the recording logical block is mapped to one of the specific pages of the candidate block.

The storage unit management system of the embodiment of the invention includes a storage unit and a processing module. The storage unit includes a plurality of physical blocks, each of which has one of a plurality of block type definitions. The processing module obtains a sub-write command, wherein the sub-write command requires a data write operation for at least one logical page of the corresponding at least one logical block. The processing module determines whether there is one candidate block in the storage unit having one of the block type definitions, wherein the logical page of the logical block according to the first block cannot be mapped to the candidate block. If the candidate block exists, the processing module converts the first block definition into one of the block type definitions, and the second block definition The data is written into one of the specific pages of the candidate block, and the logical page of the logical block is mapped to one of the specific pages of the candidate block.

The above method of the present invention can be recorded in physical media through code. When the code is loaded and executed by the machine, the machine becomes the means for practicing the invention.

The above described objects, features, and advantages of the invention will be apparent from the description and appended claims appended claims

Figure 1 shows a storage unit management system in accordance with an embodiment of the present invention.

The storage unit management system 100 according to the embodiment of the present invention includes a storage unit 110, a memory 120, and a processing module 130.

The storage unit 110 can be a non-volatile memory such as NAND flash memory. The data in the storage unit 110 can be stored in units of blocks and pages in a block mapping mode and a page mapping mode, respectively. It should be noted that the physical blocks of the storage unit 110 in this case may have different block type definitions and are managed by classification. FIG. 2 shows a physical block of the storage unit 110. In this embodiment, the block type definition may include a data block definition, a contiguous block definition, a dedication block definition, and a shared block definition. Each of the physical blocks in the storage unit 110 may have one of the above-described block type definitions and may be converted. The physical block 111 having the data block definition stores data according to the block mapping mode, and all pages in the physical block have been written into the data. In the physical block 112 having the contiguous block definition, the plurality of pages starting from the first page are stored according to the block mapping mode, and the entity is There is at least one page of unwritten material in block 112. The data in the physical block 113 having the dedication block definition corresponds to the same logical block, and the plurality of pages starting from the first page are stored according to the block mapping mode, and at least one page is not based on The block mapping mode stores data. The data in the physical block 114 having the shared block definition stores data according to the page mapping mode and corresponds to at least two specific logical blocks. It should be noted that the storage unit 110 may also include an idle physical block (not shown) that does not specify a block type definition. The transition between the definitions of the block types will be described later.

In addition, the storage unit 110 also includes a mapping directory, a complex block mapping table, a complex page mapping block table, and a complex page mapping table. The memory 120 may be a random access memory (RAM) for storing a mapping directory, a block mapping table, and a page mapping loaded from the storage unit 110 when the storage unit management system 100 is operating normally. Block table, and page mapping table. The mapping directory, the block mapping table, the page mapping block table, and the page mapping table will be described below.

Figure 3 shows one of the destination formats of the mapping directory in accordance with an embodiment of the present invention. The mapping directory maps the root directory of user data. The mapping directory contains a plurality of items as shown in FIG. 3, and each item can point to a page containing a part of the complex block mapping table, the page contains a block mapping table and a user of a predetermined size, such as 256 MB. The data corresponds. As shown in FIG. 3, the mapping directory item 300 includes a 'MappingDirectoryNumber' field 340, a 'bInCache' field 310, a 'CachedBmtNumber, field 320, and A 'BmtPageLocation' field 330. The 'MappingDirectoryNumber' field 340 records a mapping directory number as an index to this mapping directory entry 300. The 'bInCache' field 310 indicates whether the corresponding block map table page has been loaded into the memory 120 by the storage unit 110. For example, if the block mapping table page is already loaded in the memory 120, the 'bInCache' field 310 records '1'; if the block mapping table page has not been loaded into the memory 120, then 'bInCache '0' is recorded in field 310. For the tile mapping table page already loaded into the memory 120, the 'CachedBmtNumber' field 320 records the address of the corresponding tile mapping table page in the memory 120. For the tile mapping table page that has not been loaded into the memory 120, the 'CachedBmtNumber' field 320 is 'Invalid', and the physical storage address of the tile mapping table page on the storage unit 110 is recorded in 'BmtPageLocation' In field 330.

Figure 4 shows one of the destination formats of the block map in accordance with an embodiment of the present invention. The block mapping table contains a plurality of items as shown in FIG. 4, wherein each item records storage location information of the data corresponding to the logical block. As shown in FIG. 4, the tile mapping table entry 400 includes a 'BlockOffset' field 430, a 'bPageMapped' field 410, and a 'BlockInfo' field 420. The 'BlockOffset' field 430 records a block displacement for use as an index to the block map entry 400. As previously mentioned, the data in the storage unit 110 can be stored in a block mapping mode or a page mapping mode. Therefore, the 'bPageMapped' field 410 is used to indicate the mapping mode employed by the logical block that owns one of the block mapping tables. for example. If this logical block is in block mapping mode, Then '0' is recorded in the 'bPageMapped' field 410: If the logical block is in the page mapping mode, '1' is recorded in the 'bPageMapped' field 410. The 'BlockInfo' field 420 records detailed mapping information. Wherein, for the logical block adopting the block mapping mode, the physical location of the corresponding data block, such as the physical block number, is recorded in the 'BlockInfo' field 420. For a logical block that uses the page mapping mode, a page mapping block number is recorded in the 'BlockInfo' field 420 to indicate which page mapping block this logical block is.

Figure 5 shows one of the destination formats of the page mapping block table in accordance with an embodiment of the present invention. The page mapping block table contains a plurality of items as shown in FIG. 5, wherein each item records a starting address in the memory 120 corresponding to the page mapping table of the logical block. As shown in FIG. 5, the page mapping block table item 500 includes a 'PageMappedBlockNum' field 510 and a 'PageMappingTableInfo' field 520. The 'PageMappedBlockNum' field 510 is used to record the page mapping block number. When the logical block uses the page mapping mode, the page mapping block number in the 'PageMappedBlockNum' field 510 corresponds to the block mapping table item 400. The data recorded in the 'BlockInfo' field 420. The 'PageMappingTableInfo' field 520 is used to record the start address of the page mapping table corresponding to the logical block in the memory 120.

Each page mapping block has a page mapping table addressed by the 'PageMappingTableInfo' field 520. Figure 6 shows the format of an item of a page mapping table in accordance with an embodiment of the present invention. The page mapping table contains multiple The item shown in Figure 6, wherein each item records the physical storage location of the logical page in the corresponding logical block, such as the device number, the physical block number, and the physical page displacement of the logical page of one of the corresponding logical blocks of the page mapping Wait. As shown in Fig. 6, the page mapping table item 600 includes a 'PageOffset' field 610 and a 'PageInfo' field 620. The 'PageOffset' field 610 is used to record the page displacement as an index of the item, and the 'PageInfo' field 620 is used to record the physical storage location of the logical page in the corresponding logical block. Note that in some embodiments, the number of items in the page map is equal to the number of pages included in one logical block.

Please continue to refer to Figure 1. The processing module 130 can receive a write command from an upper layer, such as an application in the host, and decompose the write command into at least one sub-write command in units of pages, the sub-write command including a logical block address ( Logic Block Address, LBA). The processing module 130 converts the sub-write command into a physical write command according to the mapping directory, the block mapping table, the page mapping block table, and the page mapping table, that is, converting the logical block address (LBA) into one The physical address is used to access the physical page in the storage unit 110, and the write command and the block type definition of the converted physical block are executed to write the data into a specific page of the specific physical block in the storage unit 110.

7A to 7D are diagrams showing a storage unit management method according to an embodiment of the present invention.

First, as in step S702 (FIG. 7A), a write command is received by an upper layer, such as an application, and the received write command is decomposed as in step S704. Writes a command for at least one child in units of logical pages. The sub-write command may include a logical block address (LBA), wherein the logical block address includes a mapping directory number, a block offset (Block Offset), and a page shift. (Page Offset), a Device Number, and/or a Sector Offset. Figure 8 shows the format of a logical block address in a sub-write command in accordance with an embodiment of the present invention. As shown in FIG. 8, the Logic Block Address (LBA) 800 includes an 'MDN' field, a 'BO' field, a 'PO' field, a 'DN' field, and a 'SO' field. Among them, 'MDN' field record mapping directory number (Mapping Directory Number), 'BO' field record block shift (Block Offset), 'PO' field record page shift (Page Offset), 'DN' field record Device Number, and the 'SO' field records the sector offset (Sector Offset). It should be noted that the sub-write command can be used to know that the data is to be written to at least one logical page of at least one logical block. In step S706, a specific block mapping table is located in the complex block mapping table in the mapping directory according to the 'MDN' field in the sub-write command. In step S708, it is determined whether the specific block mapping table has been loaded in the memory. It is reminded that determining whether a specific block mapping table has been loaded into the memory may be based on the data of the 'bInCache' field 310 in the item 300 of the logical directory address 800 corresponding to the mapping directory number as shown in FIG. To judge. If the specific block mapping table is not in the memory (No in step S708), in step S710, a candidate block mapping table is selected from the block mapping table already loaded in the memory, and in step S712, the specific area is selected. The block map table is loaded into the memory to replace the candidate block map. Noteworthy In some embodiments, when the memory space is insufficient, the operation of selecting the candidate block mapping table is required, and the candidate block mapping table may be selected according to a longest unused (LRU) algorithm. If the memory still has space, you can directly load the specific block mapping table into the memory. Step S714 may be performed after loading the specific block mapping table.

If the specific block mapping table is already in the memory (YES in step S708), step S714 is performed. In step S714, a first specific item is located in the specific block mapping table according to the block displacement (recorded in the 'BO' field) in the logical block address 800 of the sub-write command (as shown in FIG. 4). A block mapping table item 400). The first specific item includes a mapping mode setting (data of the 'bPageMapped' field 410) and a block information (data of the 'BlockInfo' field 420). In step S716, it is determined according to the mapping mode setting whether the mapping mode of the corresponding block is a page mapping mode. If the mapping mode is the block mapping mode (NO in step S716), step S718 is performed. In step S718, it is determined whether there is a free item in the page mapping block table (such as the page mapping block table item 500 shown in FIG. 5). If there is an idle item in the page mapping block table (YES in step S718), the operation in step S722 is performed. If there is no free item in the page mapping block table (No in step S718), in step S720, the data merge job of the data page is performed to obtain at least one free item from the page mapping block table (like a page mapping). Block Table Item 500). In step S722, the corresponding idle item of the sub-write command is added to the page mapping block table, and the operation of step S724 is performed. If the mapping mode is the page mapping mode (YES in step S716), step S724 is performed. In step S724, according to the block information in the first specific item ('BlockInfo' field The page mapping block number in 420 is located by a page mapping block table to locate a second specific item (such as page mapping block table item 500), and according to step S726, positioning to a corresponding specific page mapping according to the second specific item. A specific page mapping table for one of the blocks.

After locating the specific page mapping table, first determining whether there is a candidate block in one of the block type definitions shown in FIG. 2 in the storage unit 110, that is, determining whether there is a candidate area defined by the first block. a block, wherein a logical page of a logical block to be written to the data cannot be mapped to the candidate block according to the first block definition. Where the map cannot be mapped, for example, when the first block is defined as a contiguous block definition 112 as shown in FIG. 2, and the logical page to be written contains the first page of the corresponding logical block. The situation at the time. Since the plurality of pages starting from the first page of the candidate block are stored according to the block mapping mode according to the contiguous block definition 112, the logical page to be written into the data cannot be mapped to the candidate block. For another example, the case where the map cannot be mapped is when the first block is defined as the contiguous block definition 112 as shown in FIG. 2, and the logical block address LBA of the logical page to be written and the candidate area are The address of the last page in the page where the data has been written in the block is not continuous. According to the contiguous block definition 112, the candidate block stores data according to the block mapping mode, that is, the logical address LBA of the logical page should be continuous, and therefore the data cannot be written to the remaining blank page of the candidate block. in.

When the logical page of the logical block to which the data is to be written cannot be mapped to the candidate block, the present invention converts the first block definition into one of the block types. definition. For example, It is to convert the contiguous block definition 112 as shown in FIG. 2 into the dedication block definition 113, or convert the dedication block definition 113 into the shared block definition 114, and write the data to be written into the candidate area. One of the specific pages in the block. For a more detailed description, reference may be made to the description of the following FIGS. 7A to 7D.

In step S728 (Fig. 7A-2), it is determined according to the sub-write command whether the logical page to be written contains the first page of the corresponding logical block. If so (YES in step S728), as in step S730 (Fig. 7B-1), it is judged whether or not there is one candidate block having a contiguous block definition in the storage unit. If there is no candidate block having the contiguous block definition in the storage unit (NO in step S730), in step S732, it is determined whether there is one candidate block having the dedication block definition in the storage unit. If there is a candidate block with a dedication block definition in the storage unit (YES in step S732), in step S733, the candidate block is converted from the dedication block definition to a shared block definition, and then step S738 is performed (step 7B- 2 picture). In step S738, a free block is reconfigured as a candidate block having a contiguous block definition, and in step S740, an entity write command of the corresponding storage unit is generated according to the sub-write command to write the data to have continuous The specific block in the candidate block defined by the block. Thereafter, the operation of step S774 (Fig. 7D) is performed. If there is no candidate block having the dedication block definition in the storage unit (No in step S732 of Fig. 7B-1), steps S738 and S740 (Fig. 7B-2) are directly executed.

If there is a candidate block having a contiguous block definition in the storage unit (Yes in step S730 of FIG. 7B-1), in step S742, it is determined whether the candidate block is already near or nearly full. If the candidate block is already nearly full (YES in step S742), in step S744, the candidate block is determined by the continuous block. The meaning is converted into a data block definition, and the operations of steps S738 and S740 (Fig. 7B-2) are performed. If the candidate block is not full or is not nearly full (No in step S742 of FIG. 7B-1), in step S746, the candidate block is converted from the contiguous block definition to the dedication block definition, and steps are as follows. S748 (Fig. 7B-2), generating a physical write command of the corresponding storage unit according to the sub-write command to write the data into a specific page in the candidate block having the dedication block definition. Thereafter, as in step S750, it is determined whether the candidate block having the dedication block definition is full. If the candidate block having the dedication block definition is already full (YES in step S750), the operations of steps S734 and S736 are performed. In step S734, the candidate block is converted from the dedication block definition to the shared block definition, and in step S736, one of the corresponding storage unit entity write commands is generated according to the sub-write command to write the data to have a share. The block is defined in a specific page in other candidate blocks. That is to say, the remaining logical pages to be written are written into other candidate blocks with shared block definitions, and the candidate blocks with shared block definitions are shared with pages belonging to other logical blocks. Entity storage space. Thereafter, the operation of step S774 (Fig. 7D) is performed. If the candidate block having the dedication block definition is not full (No in step S750 of FIG. 7B-2), the operation of step S774 (FIG. 7D) is directly performed. It is worth noting that since the number of logical pages to be written by the sub-write command may exceed the number of pages that the candidate block having the dedication block definition can provide, the judgment of step S750 must be performed.

If the logical page to be written by the sub-write command does not include the first page of the corresponding logical block (No in step S728 of FIG. 7A-2), as in step S752 (FIG. 7C-1), the storage unit is determined. Whether there is a continuous block Define one of the candidate blocks. If there is no candidate block having a contiguous block definition in the storage unit (NO in step S752), in step S754, it is determined whether there is one candidate block having a dedication block definition in the storage unit. If there is no candidate block having the dedication block definition in the storage unit (NO in step S754), the operation in step S762 (Fig. 7C-2) is performed. If there is a candidate block having a dedication block definition in the storage unit (Yes in step S754 of FIG. 7C-1), in step S756 (FIG. 7C-2), one entity of the corresponding storage unit is generated according to the sub-write command. A command is written to write the data to a particular page in the candidate block with the dedication block definition. Thereafter, as in step S758, it is determined whether the candidate block having the dedication block definition is full. If the candidate block having the dedication block definition is full (YES in step S758), in step S760, the candidate block is converted from the dedication block definition to the shared block definition, and according to step S762, the sub-write is performed according to step S762. The command generates an entity write command for the corresponding storage unit to write the data to a particular page in other candidate blocks with shared block definitions. That is to say, the remaining logical pages to be written are written into other candidate blocks with shared block definitions, and the candidate blocks with shared block definitions are shared with pages belonging to other logical blocks. Entity storage space. Thereafter, the operation of step S774 (Fig. 7D) is performed. If the candidate block having the dedication block definition is not full (NO in step S758 of FIG. 7C-2), the operation of step S774 is directly performed.

If there is a candidate block with a contiguous block definition in the storage unit (Yes in step S752 of FIG. 7C-1), in step S764, it is determined that the logical page to be written by the sub-write command has the logical page in the candidate block. Whether the last page of the data is continuous. If the phases are continuous (YES in step S764), as steps S766. Generate a physical write command of the corresponding storage unit according to the sub-write command to write the data into a specific page in the candidate block with the contiguous block definition. If it is not continuous (No in step S764), in step S768, it is determined whether the logical page to be written by the sub-write command and the address of the last page having the data in the candidate block having the contiguous block definition have a burst ( In other words, it is determined whether the logical block address LBA of the logical page to be written is smaller than the address of the last page in the page of the candidate block having the data. If the burst does not occur (NO in step S768), the operation in step S762 (Fig. 7C-2) is performed. Since the current high-density storage unit, such as NAND (non-) flash memory, each page of each block must be written in order, can not be randomly written, so when the LBA does not burst, that is, when you want to write When the logical block address LBA of the incoming logical page is larger than the address of the last page in the page with the data in the candidate block, the data should be written to the specific candidate block having the shared block definition. In the page, step S762 is performed. If a burst occurs (Yes in step S768 of FIG. 7C-1), in step S770, the candidate block is converted from the contiguous block definition to the dedication block definition, and in step S772, corresponding to the sub-write command is generated. One of the storage units entity writes a command to write the data to a particular page in the candidate block with the dedication block definition. Thereafter, the operation of step S758 (Fig. 7C-2) is performed.

Thereafter, as in step S774 (Fig. 7D), the generated physical write command is executed to write the data to a specific page in the storage unit. It is worth noting that, as previously mentioned, the sub-write command may include device number and/or magnetic zone displacement. When the sub-write command includes the device number and/or the magnetic zone displacement, it is even more The data is written into the storage unit according to the device number and/or the magnetic zone displacement. In step S776, the page mapping table and/or the specific block mapping table are updated according to the mapping relationship between the logical page of the logical block to be written by the corresponding sub-write command and the specific page of the candidate block in the actual storage unit. In step S778, it is determined whether the mapping table update of step S776 involves a block level mapping table update. If not (NO in step S778), the entire writing process is ended. If so (YES in step S778), in step S780, the updated specific block mapping table is written to the storage unit.

In view of this, the storage unit management method and system of the present invention can classify and manage the physical blocks of the storage unit to utilize the storage unit more efficiently.

The method of the present invention, or a specific type or part thereof, may be included in a physical medium such as a floppy disk, in the form of a software conversion layer (such as a code) used between a host file system and a non-volatile memory. A storage medium that can be read (eg, computer readable) by a disc, hard disk, or any other machine, wherein the device becomes a device for participating in the present invention when the code is loaded and executed by a machine, such as a computer. . The method and apparatus of the present invention can also be transmitted in a code format through some transmission medium such as a wire or cable, an optical fiber, or any transmission type, wherein the code is received, loaded, and executed by a machine such as a computer. At this time, the machine becomes a device for participating in the present invention. When implemented in a general purpose processor, the code in conjunction with the processor provides a unique means of operation similar to application specific logic.

Although the present invention has been disclosed in the above preferred embodiments, it is not intended to limit the invention, and any one skilled in the art can In the context of God and the scope, the scope of protection of the present invention is defined by the scope of the appended claims.

100‧‧‧Storage Unit Management System

110‧‧‧ storage unit

111‧‧‧ Blocks defined by the data block

112‧‧‧ Blocks defined by contiguous blocks

113‧‧‧Dedicated block definition block

114‧‧‧ Blocks defined by shared blocks

120‧‧‧ memory

130‧‧‧Processing module

300‧‧‧Map Directory Project

310‧‧‧’bInCache’ field

320‧‧‧’CachedBmtNumber’ field

330‧‧‧’BmtPageLocation’ field

340‧‧‧’MappingDirectoryNumber’ field

400‧‧‧block mapping table project

410‧‧‧’bPageMapped’ field

420‧‧‧’BlockInfo’ field

430‧‧‧’BlockOffset’ field

500‧‧‧Page Mapping Block Table Item

510‧‧‧’PageMappedBlockNum’ field

520‧‧‧’PageMappingTableInfo’ field

600‧‧‧ page mapping table project

610‧‧‧’PageOffset’ field

620‧‧‧’PageInfo’ field

S702, S704, ..., S780‧‧‧ steps

800‧‧‧ Logical Block Address

MDN‧‧‧ mapping directory number field

BO‧‧‧ block displacement field

PO‧‧‧ page shift field

DN‧‧‧ device number field

SO‧‧‧ magnetic zone displacement field

Figure 1 is a schematic diagram showing a storage unit management system in accordance with an embodiment of the present invention.

Figure 2 is a schematic diagram showing physical blocks in a storage unit in accordance with an embodiment of the present invention.

Figure 3 is a schematic diagram showing the format of a mapping directory entry in accordance with an embodiment of the present invention.

Figure 4 is a schematic diagram showing the format of a block map entry in accordance with an embodiment of the present invention.

Figure 5 is a schematic diagram showing the format of a page mapping block table entry in accordance with an embodiment of the present invention.

Figure 6 is a schematic diagram showing the format of a page mapping table entry in accordance with an embodiment of the present invention.

7A to 7D are flowcharts showing a storage unit management method according to an embodiment of the present invention.

Figure 8 is a diagram showing the format of a logical block address (LBA) included in a write command according to an embodiment of the present invention.

110‧‧‧ storage unit

111‧‧‧ Blocks defined by the data block

112‧‧‧ Blocks defined by contiguous blocks

113‧‧‧Dedicated block definition block

114‧‧‧ Blocks defined by shared blocks

Claims (28)

A storage unit management method, comprising the steps of: providing a storage unit, wherein the storage unit comprises a plurality of physical blocks, each of the physical blocks having one of a plurality of block type definitions; obtaining at least one sub-write command The sub-write command is configured to perform a data write operation on at least one logical page of the at least one logical block; and determine whether there is a first block definition in the storage unit having one of the block type definitions. a candidate block, wherein the logical page of the logical block cannot be mapped to the candidate block according to the first block definition; if the candidate block exists, the first block definition is converted into the block Determining, in the category definition, a second block definition; and writing a data store to a specific page of the candidate block, and recording the logical page of the logical block to map with one of the specific pages of the candidate block a relationship, wherein the block type definition includes a contiguous block definition and a dedication block definition, wherein the contiguous block definition of the physical block from the first page The plurality of pages are stored according to a block mapping mode, and at least one page of unwritten data exists in the physical block, and the data in the physical block defined by the dedication block corresponds to one For a specific logical block, a plurality of pages starting from the first page store data according to the block mapping mode, and at least one page does not store data according to the block mapping mode. For example, the storage unit management method described in claim 1 of the patent scope, The sub-write command includes at least a mapping directory number, a block displacement, and a page offset, and the method further includes the following steps: positioning the specific block mapping from the complex block mapping table according to the mapping directory number. a first specific item is located in the specific block mapping table according to the block displacement, wherein the first specific item includes a mapping mode setting and a block information; when the mapping mode is set to a page mapping mode Locating a second specific item from a page mapping block table according to the block information, and locating a specific page mapping table corresponding to a specific page mapping block by the plurality of page mapping tables according to the second specific item; Updating the specific block mapping table or the page mapping table according to the mapping relationship. The storage unit management method according to claim 2, further comprising the steps of: determining, according to the mapping directory number, whether the specific block mapping table is already loaded in a memory; and if the specific block mapping table Not loaded in the memory, selecting a candidate block mapping table from the block mapping tables already loaded into the memory; and loading the specific block mapping table into the memory to replace The candidate block mapping table. The storage unit management method according to claim 2, wherein when the mapping mode is set to a block mapping mode, a new idle is added. Project to this page maps the block table. The storage unit management method of claim 1, further comprising the steps of: converting the sub-write command to at least one entity write command corresponding to the storage unit according to the mapping relationship; and performing the entity write a command to write the data to the particular page in the candidate block in the storage unit. The storage unit management method according to claim 1, wherein the block type definition further includes a data block definition and a shared block definition, wherein the physical block system having the data block definition The data is stored according to the block mapping mode, and all the pages in the physical block have been written into the data, and the data in the physical block defined by the shared block is stored according to a page mapping mode, and corresponding Up to at least two specific logical blocks. The storage unit management method of claim 6, further comprising the steps of: determining whether the at least one logical page includes the first page of the logical block; and if the at least one logical page includes the logical block a first page, determining whether the candidate block having the contiguous block definition exists in the storage unit; if the candidate block having the contiguous block definition exists in the storage unit, the candidate block is from the contiguous block The block definition is converted to the dedication block definition to store the data in the particular page in the candidate block; And if the candidate block having the contiguous block definition does not exist in the storage unit, configuring a free block as the candidate block having the contiguous block definition to store the data to have the contiguous block definition In the particular page in the candidate block. The storage unit management method of claim 7, wherein when there is no candidate block having the contiguous block definition in the storage unit, the method further comprises the step of: determining whether the storage unit exists The candidate block having the dedication block definition; and if the candidate block having the dedication block definition exists in the storage unit, converting the candidate block from the dedication block definition to the shared block definition. The storage unit management method of claim 7, wherein the candidate block having the contiguous block definition exists in the storage unit, and converting the candidate block from the contiguous block definition to the offering After the block definition, the method further comprises the steps of: determining whether the candidate block having the dedication block definition is full; and if the candidate block having the dedication block definition is full, the candidate area is The block is converted to the shared block definition by the dedication block definition to store the data in the particular page in the other candidate block having the shared block definition. The storage unit management method described in claim 6 of the patent application scope, The method further includes the following steps: determining whether the at least one logical page includes the first page of the logical block; if the at least one logical page does not include the first page of the logical block, determining whether the storage unit has a Determining the candidate block by the contiguous block; if the candidate block having the contiguous block definition is not present in the storage unit, determining whether the candidate block having the dedication block definition exists in the storage unit; Having the candidate block with the dedication block definition in the storage unit, directly storing the data in the specific page in the candidate block; and if the storage unit does not have the definition of the dedication block The candidate block stores the data in the specific page of the other candidate blocks in the storage unit having a shared block definition. The storage unit management method of claim 10, wherein when there is the candidate block having the dedication block definition in the storage unit, the method further comprises the step of: determining that the dedication block definition is Whether the candidate block is full; and if the candidate block having the dedication block definition is full, converting the candidate block from the dedication block definition to the shared block definition to store the data To this particular page in other candidate blocks with shared block definitions. The storage unit management method of claim 10, wherein when there is the candidate block having the contiguous block definition in the storage unit, the method further comprises the step of: determining that the logical page has the continuity Whether the last page of the candidate block in the block definition is continuous; if the logical page is continuous with the last page of the candidate block having the contiguous block definition, the data is directly stored Going to the specific page in the candidate block; if the logical page is not consecutive with the last page having the data in the candidate block having the contiguous block definition, determining the logical page and having the contiguous block definition Whether the address of the last page of the candidate block has a burst in the candidate block; if the logical page does not protrude from the address of the last page of the candidate block having the contiguous block definition The data is stored in the particular page in the other candidate blocks having the shared block definition; and if the logical page has the contiguous area Defining an address of the last page of the candidate block with the data in the candidate block, and converting the candidate block from the contiguous block definition to the dedication block definition to store the data to the dedication block definition In the particular page in the candidate block. The storage unit management method of claim 12, wherein after the step of storing the data to the specific page in the candidate block having the dedication block definition, the method further comprises the step of: determining that Whether the candidate block defined by the dedication block is full And if the candidate block having the dedication block definition is full, converting the candidate block from the dedication block definition to the shared block definition to store the data to have a shared block definition Among the other candidate blocks in this particular page. The storage unit management method of claim 1, further comprising writing the updated specific block mapping table to the storage unit. A storage unit management system, comprising: a storage unit, comprising a plurality of physical blocks, each of the physical blocks having one of a plurality of block type definitions; and a processing module that obtains at least one sub-write command, The sub-write command is configured to perform a data write operation on at least one logical page of the at least one logical block, and determine whether there is one of the first block definitions in the storage unit having one of the block type definitions. a candidate block, wherein the logical page defining the logical block according to the first block cannot be mapped to the candidate block, and if the candidate block exists, converting the first block definition to the block type definition One of the second blocks defines that a material is written to a specific page of the candidate block, and the logical page of the logical block is recorded and mapped to one of the specific pages of the candidate block, wherein the The equal block type definition includes a contiguous block definition and a dedication block definition, wherein the plurality of page pages starting from the first page in the physical block defined by the contiguous block According to a data storage block mapping mode, and there is at least one unwritten page of the entity data block, The data in the physical block having the dedication block definition corresponds to a specific logical block, and the plurality of pages starting from the first page are stored according to the block mapping mode, and at least one page exists. Data is not stored according to the block mapping mode. The storage unit management system of claim 15, wherein the sub-write command comprises at least a mapping directory number, a block displacement, and a page displacement, and the processing module is further configured according to the mapping directory number. Positioning a specific block mapping table in the complex block mapping table, and positioning a first specific item in the specific block mapping table according to the block displacement, wherein the first specific item includes a mapping mode setting and a block Information, when the mapping mode is set to a page mapping mode, a second specific item is located in a page mapping block table according to the block information, and a specific one is located in the plurality of page mapping table according to the second specific item. A page mapping block is a specific page mapping table, and the specific block mapping table or the page mapping table is updated according to the mapping relationship. The storage unit management system of claim 16, wherein the processing module further determines, according to the mapping directory number, whether the specific block mapping table has been loaded into a memory, if the specific block mapping table Not loaded in the memory, a candidate block mapping table is selected from the block mapping tables already loaded in the memory, and the specific block mapping table is loaded into the memory to replace The candidate block mapping table. The storage unit management system of claim 16, wherein the processing module adds a free item to the page mapping block table when the mapping mode is set to a block mapping mode. The storage unit management system of claim 15, wherein the processing module further converts the sub-write command into at least one entity write command corresponding to the storage unit according to the mapping relationship, and performs the entity write Entering a command to write the data to the particular page in the candidate block in the storage unit. The storage unit management system of claim 15, wherein the block type definition further comprises a data block definition and a shared block definition, wherein the physical block definition having the data block definition The data is stored according to the block mapping mode, and all the pages in the physical block have been written into the data, and the data in the physical block defined by the shared block is stored according to a page mapping mode, and corresponding Up to at least two specific logical blocks. The storage unit management system of claim 20, wherein the processing module further determines whether the at least one logical page includes a first page of the logical block, and if the at least one logical page includes the logical block a first page, determining whether the candidate block having the contiguous block definition exists in the storage unit, and if the candidate block having the contiguous block definition exists in the storage unit, the candidate block is Converting the contiguous block definition to the dedication block definition to store the data in the specific page in the candidate block, if there is no candidate block having the contiguous block definition in the storage unit, configuring one The free block is the candidate block defined by the contiguous block to store the data in the particular page in the candidate block having the contiguous block definition. The storage unit management system as described in claim 21 The processing module further determines whether the candidate block having the dedication block definition exists in the storage unit if the candidate block having the contiguous block definition does not exist in the storage unit, if the storage block exists The candidate block having the dedication block definition is present in the unit, and the candidate block is converted from the dedication block definition to the shared block definition. The storage unit management system of claim 21, wherein the candidate block having the contiguous block definition exists in the storage unit, and converting the candidate block from the contiguous block definition to the offering After the block is defined, the processing module further determines whether the candidate block having the dedication block definition is full. If the candidate block having the dedication block definition is full, the candidate block is used by the candidate block. The dedication block definition is converted to the shared block definition and the data is stored in the particular page in the other candidate blocks with the shared block definition. The storage unit management system of claim 20, wherein the processing module further determines whether the at least one logical page includes the first page of the logical block, and when the logical page does not include the logical block Determining, in the first page, whether the candidate block having the contiguous block definition exists in the storage unit, and if the candidate block having the contiguous block definition is not present in the storage unit, determining whether the storage unit is in the storage unit There is the candidate block having the definition of the dedication block. If the candidate block having the dedication block definition exists in the storage unit, the data is directly stored in the specific page in the candidate block, if The candidate block having the dedication block definition does not exist in the storage unit, and the data is stored in the storage unit to the specific page in another candidate block having a shared block definition. in. The storage unit management system of claim 24, wherein when the candidate unit having the dedication block definition exists in the storage unit, the processing module further determines the candidate having the dedication block definition Whether the block is full, if the candidate block having the dedication block definition is full, converting the candidate block from the dedication block definition to the shared block definition to store the data to have a share The block is defined in the other specific page of the candidate block. The storage unit management system of claim 24, wherein when the candidate block having the contiguous block definition exists in the storage unit, the processing module further determines the logical page and has the contiguous block. Determining whether the last page of the candidate block in the candidate block is continuous. If the logical page is continuous with the last page of the candidate block having the contiguous block definition, the data is directly stored in the candidate block. In the specific page in the candidate block, if the logical page is not consecutive with the last page of the candidate block having the contiguous block definition, the logical page is determined to have the contiguous block definition Whether the address of the last page of the candidate block in the candidate block has a burst, and if the logical page does not have a burst with the address of the last page of the candidate block having the contiguous block definition, Storing the data in the particular page in the other candidate block having the shared block definition, if the logical page and the candidate having the contiguous block definition Last address of a page of data having red block projection occurs, the candidate block of the consecutive block is defined by the switching block is defined for dedication to the dedicated data storage block having to Defined in that particular page in the candidate block. The storage unit management system of claim 26, wherein after processing the data to the specific page in the candidate block having the dedication block definition, the processing module further determines that the dedication block is Determining whether the candidate block is full, and if the candidate block having the dedication block definition is full, converting the candidate block from the dedication block definition to the shared block definition to use the data Stored in that particular page in other candidate blocks with shared block definitions. The storage unit management system of claim 15, wherein the processing module further writes the updated specific block mapping table to the storage list.