TWI646461B - Data storage device and data maintenance method thereof - Google Patents

Abstract

The invention provides a data storage device including a random access memory and a controller. The random access memory has a prefetch area. The controller loads a plurality of data link relationship sets in the data link relationship table into a plurality of sections in the prefetch area. Among the data link relationship sets, the number of data link read sets is less than a predetermined value. The knot relationship set belongs to one of the plural infrequent data link relationship sets, and the data link relationship set that has been read the number of times reaches the predetermined value belongs to one of the plural frequent data link relationship sets.

Description

   Data storage device and data maintenance method   

The present invention relates to a data storage device, and more particularly to a data storage device that maintains data in a prefetched area in a specific manner.

Flash memory is a universal non-volatile data storage device that is electrically erased and programmed. Take non-gate flash memory (namely NAND FLASH) as an example, it is often used as a memory card, a universal serial bus flash memory device (USB flash device), a solid-state hard disk (SSD), and an embedded flash memory. Memory module (eMMC) ... etc.

A flash memory (eg, NAND FLASH) storage array includes a plurality of blocks, and a floating gate transistor can be used to form the flash memory. The floating gate in the floating gate transistor can capture the charge to store data. In addition, the conversion relationship between the physical page in the flash memory and the logical page specified by the host device requires a physical to logical table record. However, a large capacity is required to load the entire physical logic table into the random access memory. How to effectively update and manage the data loaded in random memory is an important issue.

The data storage device and data maintenance method provided by the present invention can maintain data stored in random memory through three types of tables.

The invention provides a data storage device including a random access memory and a controller. The random access memory has a prefetch area. The controller loads a plurality of data link relationship sets in the data link relationship table into a plurality of sections in the prefetch area. Among the data link relationship sets, the number of data link read sets is less than a predetermined value. The knot relationship set belongs to one of the plural infrequent data link relationship sets, and the data link relationship set that has been read the number of times reaches the predetermined value belongs to one of the plural frequent data link relationship sets.

In an embodiment, the controller further establishes an infrequent index group to maintain the infrequent data link relationship set, and establishes a frequent index group to maintain the frequent data link relationship set, wherein the infrequent index group It consists of an infrequent head indicator and an infrequent tail indicator, and the frequent indicator group is composed of a frequent head indicator and a frequent tail indicator. Sectors have multiple shared indicators in order. The infrequent header indicator is a shared indicator used to store the last infrequent data link relationship set that is read. The infrequent tail indicator is used to store the longest unread. The common indicator of the segments of the infrequent data link relationship set. The frequent header indicator is a common indicator used to store the last read of the frequent data link relationship set. The frequent tail indicator is used to store the last A common indicator for the segments of the frequent data link relationship set being read.

The present invention also provides a data storage device including a random access memory having a prefetch area and a controller. The controller is used to load a part of the plurality of data link relationship sets stored in a data link relationship table into a plurality of sections in the prefetch area, wherein when the plurality of sections in the prefetch area are all filled and a first When the data link relationship set needs to be loaded into the prefetch area, the controller is further configured to select one of the first sections according to an infrequent index group to load the first data link relationship set. When the frequent indicator group does not have one of the corresponding sections, the controller further converts a frequent indicator group into an infrequent indicator group, and selects the first section according to the converted infrequent indicator group.

The invention also provides a data maintenance method applicable to a data storage device having a flash memory. The flash memory includes a plurality of pages, where each page has a logical address and a physical address. The data maintenance method includes: when the flash memory is powered on, according to at least one read command or at least one write command, a data chain is loaded into a plurality of sections of a prefetch area in a random memory. A part of a plurality of data link relationship sets in a link relationship table, wherein each data link relationship set has a correspondence relationship between at least one logical address and physical address of a page, and each data link relationship A set corresponds to a set of indicators, in which the data link relationship set stored in the prefetch area is read less than a predetermined number of data link relationship sets belongs to the infrequent data link relationship set, and the prefetch area is stored The data link relationship set in the data link relationship set that is read to reach a predetermined value belongs to the frequent data link relationship set.

In an embodiment, the data maintenance method further includes establishing an infrequent index group to maintain a plurality of infrequent data link relationship sets; and establishing a frequent index group to maintain a plurality of frequent data link relationship sets, wherein prefetching The data link relationship set that is read to reach a predetermined value in the data link relationship set stored in the area belongs to the frequent data link relationship set.

In one embodiment, the data maintenance method further includes: establishing a prefetch area correspondence table to record a set index corresponding to the data link relationship set in the prefetch area in the corresponding column of the plural section in the prefetch area correspondence table. ; Establishing an order table to record the order of frequent data link relationship sets being read from the prefetch area and the order of infrequent data link relationship sets being read from the domain fetch area; and establishing a reverse order table , Respectively, to record the reverse order of the frequent data link relationship set read from the prefetch area and the reverse order of the infrequent data link relationship set read from the prefetch area.

In still another embodiment, the data maintenance method further includes: when the prefetch area is full and a new data link relationship set needs to be loaded, determining whether the infrequent tail indicator is any of the common indicators; when When the infrequent tail indicator is any of the common indicators, the infrequent data link relationship set in the infrequent data link relationship set that has not been read for the longest time is determined according to the infrequent tail indicator; Data is written to the section of the infrequent data link relationship set that belongs to the infrequent data link relationship set that has not been read for the longest time; when the infrequent tail indicator is not any of the common indicators, the frequent tail indicator The stored common indicator is written into the infrequent tail indicator, the common indicator stored in the frequent head indicator is written into the infrequent header indicator, and the frequent tail indicator and the common indicator stored in the infrequent tail indicator are deleted.

100‧‧‧ electronic system

120‧‧‧Host

140‧‧‧data storage device

160‧‧‧controller

162‧‧‧ Computing Unit

164‧‧‧Permanent Memory

166‧‧‧RAM

180‧‧‧Flash memory

TB1‧‧‧Data Link Relation Table

MR_0 ~ MR_MX‧‧‧ Correspondence

TS_0 ~ TS_M‧‧‧ Data Link Relationship Collection

CIX_0 ~ CIX_N‧‧‧Shared indicators

CA‧‧‧ Prefetch Area

S_0 ~ S_N‧‧‧ section

SE_TB‧‧‧Order List

SE_0 ~ SE_N‧‧‧Order column

SMR_TB‧‧‧Pre-fetch area mapping table

SMR_0 ~ SMR_N‧‧‧ section corresponding column

RSE_TB‧‧‧Reverse order

RSE_0 ~ RSE_N‧‧‧Reverse column

HIX‧‧‧Head Indicator

TIX‧‧‧Tail indicators

FR_HIX‧‧‧ Frequent Head Indicator

FR_TIX‧‧‧ Frequent Tail Indicator

LRU_HIX‧‧‧Infrequent Head Indicator

LRU_TIX‧‧‧Infrequent tail indicator

S900 ~ S902, S1000 ~ S1090, S2100 ~ S2102, S2200 ~ S2299‧‧‧ steps

FIG. 1 is a block diagram of an embodiment of an electronic system provided by the present invention.

FIG. 2 is a block diagram of an embodiment of a data link relationship table provided by the present invention.

FIG. 3 is a schematic diagram of an embodiment of a complex number table in a random access memory provided by the present invention.

FIG. 4 is a schematic diagram of another embodiment of a complex number table in a random access memory provided by the present invention.

FIG. 5 is a schematic diagram of another embodiment of a complex number table in a random access memory provided by the present invention.

FIG. 6 is a schematic diagram of another embodiment of a complex number table in a random access memory provided by the present invention.

FIG. 7A is a schematic diagram of another embodiment of a complex number table in a random access memory provided by the present invention.

FIG. 7B is a schematic diagram of another embodiment of a complex number table in a random access memory provided by the present invention.

FIG. 8 is a schematic diagram of another embodiment of a complex number table in a random access memory provided by the present invention.

FIG. 9 is a flowchart of an embodiment of a data maintenance method provided by the present invention.

10A-10F are flowcharts of another embodiment of a data maintenance method provided by the present invention.

11 to 20 are block diagrams of another embodiment of a table in a random access memory provided by the present invention.

FIG. 21 is a flowchart of another embodiment of a data maintenance method provided by the present invention.

22A to 22H are flowcharts of another embodiment of a data maintenance method provided by the present invention.

The device and method of use of various embodiments of the present invention will be discussed in detail below. It is worth noting, however, that many of the possible inventive concepts provided by the present invention can be implemented in various specific ranges. These specific embodiments are only used to illustrate the device and method of use of the present invention, but not to limit the scope of the present invention.

FIG. 1 is a block diagram of an embodiment of an electronic system provided by the present invention. The electronic system 100 includes a host 120 and a data storage device 140. The data storage device 140 includes a flash memory 180 and a controller 160, and can be operated according to a command issued by the host 120. The controller 160 includes an operation unit 162, a permanent memory (for example, a read-only memory ROM) 164, and a random access memory (RAM) 166. The permanent memory 164 and the loaded code and data form firmware, which is executed by the operation unit 162, so that the controller 160 controls the flash memory 180 based on the firmware. The random access memory (RAM) 166 is used to load codes and parameters to provide the controller 160 to act according to the loaded codes and parameters. The flash memory 180 includes a plurality of blocks, and each block includes a plurality of pages. Each page has a logical address and a physical address. The physical address is a fixed address of the page in the flash memory 180. The logical address is the address defined by the host 120 and the controller 160 for the page. In addition, the flash memory 180 further includes a data link relationship table TB1 that can be dynamically updated by the controller 160, where the data link relationship table TB1 is used to record the complex correspondence relationship between logical addresses and physical addresses of all pages. It is worth noting that the flash memory 180 is erased with the block as the smallest unit, and the page is written with the smallest unit.

FIG. 2 is a block diagram of an embodiment of a data link relationship table provided by the present invention. As shown in FIG. 2, the data link relationship table TB1 includes the correspondences between the logical addresses and physical addresses of all the pages MR_P0 ~ MR_PMX, where each correspondence MR_P0 ~ MR_PMX corresponds to one of the flash memories 180 respectively. One page, and any two of the corresponding relationships MR_P0 ~ MR_PMX do not correspond to the same page. In addition, the data link relationship table TB1 is divided into a plurality of data link relationship sets TS_0 ~ TS_M, and each data link relationship set TS_0 ~ TS_M has more than one corresponding relationship. For example, the data link relationship set TS_0 has a corresponding relationship MR_P0 ~ MR_PX, the data link relationship set TS_1 has a corresponding relationship MR_P0 ~ MR_P2X, and so on. In this embodiment, each data link relationship set TS_0 ~ TS_M has X + 1 corresponding relationships, but the present invention is not limited thereto. In other embodiments, each data link relationship set may have a different number of corresponding relationships, and the number of link relationships included in the data link relationship set may be determined by the developer or dynamically adjusted by the controller 160. It is worth noting that each data link relationship set TS_0 ~ TS_M corresponds to a set index. In one embodiment, the set indicator may be a hexadecimal value such as 0xAA, 0xBB, 0xCC, 0x128, and the like.

After the data storage device 140 is powered on, the controller 160 can load the data link relationship table TB1 into one of the prefetch areas in the random access memory 166 for reading and updating. However, as the memory capacity of the flash memory 180 increases, the data link relationship table TB1 also becomes larger and larger. In the case that the prefetch area planned in the random access memory 166 is insufficient to store the entire data link relationship table TB1, the controller 160 links some data in the data link relationship table TB1 according to the currently required data Relation sets are loaded into the prefetch area. In detail, when there is still space in the prefetch area CA to load at least one of the data link relationship sets TS_0 ~ TS_M, the controller 160 performs a loading procedure to load the required data link relationship sets. Into the blank space of the prefetch area. When the prefetch area is full and a new data link relationship set TS_0 ~ TS_M needs to be loaded, the controller 160 performs a replacement procedure to find the data link relationship set that has not been read for the longest time in order to update the new data link relationship set TS_0 ~ TS_M. The set of data link relationships is overlaid on the most recently read data link relationship set. When the data link relationship set required by the controller 160 already exists in the prefetch area, the controller 160 performs an update procedure to update the latest reading status.

FIG. 3 is a block diagram of an embodiment of a table in a random access memory provided by the present invention. As shown in FIG. 3, after the data storage device 140 is powered on, the controller 160 creates a blank prefetch area correspondence table SMR_TB, a reverse order table RSE_TB, and a sequence table SE_TB in the random access memory 166. To record the status of the data in the prefetch area CA.

The prefetch area CA has a plurality of sections S_0 ~ S_N for storing the data link relationship set in the data link relationship table TB1, and one section is used for storing a data link relationship set. In other words, the number of segments S_0 ~ S_N is smaller than the number of data link relationship sets TS_0 ~ TS_M, that is, N <M.

The prefetched area correspondence table SMR_TB has a plurality of section corresponding columns SMR_0 ~ SMR_N to record the set index corresponding to the data link relationship set TS_0 ~ TS_M in the prefetched area CA, where the prefetched area correspondence table SMR_TB corresponds to the section The columns SMR_0 ~ SMR_N sequentially correspond to the sections S_0 ~ S_N in the prefetch area CA. It is worth noting that the initial values of the corresponding columns SMR_0 ~ SMR_N in each section of the prefetched region correspondence table SMR_TB are a specific value. In this embodiment, the specific value is 0xFFFF, but the present invention is not limited to this. In addition, the set indexes of the data link relationship sets TS_0 ~ TS_M are not equal to a specific value.

The sequence table SE_TB is used to record the order in which the data link relationship set TS_0 ~ TS_M is read from the prefetch area CA. The sequence table SE_TB has a plural sequence column SE_0 ~ SE_N corresponding to the section S_0 in the prefetch area CA in order. ~ S_N. Each sequence column SE_0 ~ SE_N in the sequence table SE_TB is used to store the common indicator of another sequence column to point to the section of the data link relationship set that was read before the section corresponding to the sequence column, respectively. The corresponding order column. It is worth noting that the initial value of each sequence column SE_0 ~ SE_N in the sequence table SE_TB is a specific value. In this embodiment, the specific value is 0xFFFF, but the present invention is not limited to this.

The reverse order table RSE_TB is used to record the reverse order of the data link relationship set TS_0 ~ TS_M read from the prefetch area CA. The reverse order table RSE_TB has a complex reverse order column RSE_0 ~ RSE_N corresponding to the prefetch area Sections S_0 ~ S_N in CA. Each reverse-order column RSE_0 ~ RSE_N in the reverse-order table RSE_TB is used to store the shared index of another reverse-order column, respectively, to point to the next data link relationship read after the section corresponding to the reverse-order column Reverse column corresponding to the section of the collection. It is worth noting that the initial value of each of the reverse order columns RSE_0 ~ RSE_N in the reverse order table RSE_TB is a specific value. In this embodiment, the specific value is 0xFFFF, but the present invention is not limited to this.

It is worth noting that, in an embodiment, the segments S_0 ~ S_N have specific common indexes CIX_0 ~ CIX_N in order. The section corresponding columns SMR_0 ~ SMR_N and the corresponding sections S_0 ~ S_N have the same common index CIX_0 ~ CIX_N. The sequence columns SE_0 ~ SE_N and their corresponding sections S_0 ~ S_N have the same common indicators CIX_0 ~ CIX_N. The reverse order columns RSE_0 ~ RSE_N and their corresponding sections S_0 ~ S_N have the same common index CIX_0 ~ CIX_N. In detail, the shared indicators of the segment S_0, the sequence column SE_0, the segment corresponding column SMR_0, and the reverse sequence column RSE_0 are all CIX_0. The shared index of segment S_1, sequence column SE_1, segment corresponding column SMR_1, and reverse sequence column RSE_1 are all CIX_1. The common indicators of the segment S_2, the sequence column SE_2, the segment corresponding column SMR_2, and the reverse sequence column RSE_2 are CIX_2, and so on. It is worth noting that each common indicator CIX_0 ~ CIX_N is not equal to a specific value. In an embodiment, the common indicators CIX_0 to CIX_N are sequentially hexadecimal values such as 0x0, 0x1, 0x2, and 0x3, but the present invention is not limited thereto.

In summary, the controller 160 establishes a prefetch area correspondence table SMR_TB to record a set index corresponding to the data link relationship set TS_0 ~ TS_M in the prefetch area CA. The controller 160 establishes a reverse order table RSE_TB, and records the reverse order of the data set TS_0 ~ TS_M from the prefetch area CA. The controller 160 creates an order table SE_TB, and records the order in which the data sets TS_0 ~ TS_M are read from the prefetch area CA. In one embodiment, the controller 160 is further configured to establish a head index HIX and a tail index TIX in the random access memory 166, wherein the head index HIX and the tail index TIX can also be established in other memories or circuits. The invention is not limited to this. The head index HIX is the common index CIX_0 ~ CIX_N of the last read sector S_0 ~ S_N in the prefetch area CA to point to the last read sector S_0 ~ S_N in the prefetch area CA. The tail index TIX is the common index CIX_0 ~ CIX_N in the pre-fetched area CA which is the longest unread section S_0 ~ S_N to point to the longest unread section S_0 ~ S_N in the prefetched area CA. It is worth noting that since the corresponding sections, sequence columns, section corresponding columns, and reverse sequence columns have the same shared indicators, the head indicator HIX and the tail indicator TIX also point to the corresponding order columns, section corresponding columns, and Reverse order bar. In this embodiment, the controller 160 may select one of the plurality of sections S_0 ~ S_N in the prefetch area CA according to the tail index TIX. In other words, the controller 160 may select the most unread section (the least recently read data link relationship set) in the prefetch area CA according to the tail index TIX to load the new data link relationship set into the Selected section. In addition, the controller 160 may update the reverse sequence table RSE_TB and the sequence table SE_TB according to a head index HIX, data in the reverse sequence table RSE_TB, and data in the sequence table SE_TB. Next, the relationship between the prefetch area CA, the prefetch area correspondence table SMR_TB, the reverse order table RSE_TB, and the order table SE_TB will be explained in order from FIG. 3 to FIG. 8 in the following paragraphs.

In the embodiment of the initial state in FIG. 3, when the controller 160 receives a read command or a write command, the controller 160 finds out whether the read command or the write command includes the read command or A first data link relationship set of a first correspondence relationship of a page specified by the write command. For example, the correspondence between the logical address and the physical address of the page indicated by the read or write command is stored in the data link relationship set TS_2, and the controller 160 is based on the received read or write command. Learn the designated page and find the corresponding data link relationship set TS_2. Next, the controller 160 determines whether the data link relationship set TS_2 has been loaded into the prefetch area CA according to whether a set index corresponding to the first data link relationship set TS_2 exists in the prefetch area correspondence table SMR_TB. For example, the set index of the data link relationship set TS_2 is 0xAA. As shown in FIG. 3, if no pre-fetch area CA is written with any data link relationship set, the corresponding columns SMR_0 ~ SMR_N of all sections in the pre-fetch area correspondence table SMR_TB are 0xFFFF. Therefore, in this embodiment, the controller 160 determines that the set index 0xAA of the data link relationship set TS_2 does not exist in the prefetched region correspondence table SMR_TB. In other words, the controller 160 determines that the data link relationship set TS_2 does not exist in the prefetch area CA according to the set index 0xAA of the data link relationship set TS_2 that does not exist in the prefetch area correspondence table SMR_TB. Next, the controller 160 determines whether the prefetch area CA has blank sections S_0 ~ S_N. As shown in FIG. 3, the sections S_0 ~ S_N of the prefetch area CA are all blank, and the controller 160 selects a blank section. In this embodiment, the controller 160 selects from the bottommost segment, so the controller 160 selects the segment S_N. Next, the controller 160 reads the data link relationship set TS_2 from the data link relationship table TB1 in the flash memory 180 to load the data link relationship set TS_2 into the selected blank section S_N. Next, the controller 160 updates the prefetch area correspondence table SMR_TB, the order table SE_TB, the reverse order table RSE_TB, the head index HIX, and the tail index TIX to record the current state of the prefetch area CA. As shown in FIG. 4, the controller 160 writes the set index 0xAA of the data link relationship set TS_2 into the corresponding section SMR_N of the section S_N, and simultaneously defines the head indicator HIX and the tail indicator TIX as the sections S_N. Common indicator CIX_N. It is worth noting that in this embodiment, the order column SE_N and the reverse column RSE_N in the sequence table SE_TB and the reverse sequence table RSE_TB corresponding to the section S_N maintain a specific value of 0xFFFF. Finally, the controller 160 reads the data link relationship set TS_2 stored in the prefetch area CA to execute the received write command and read command.

Next, in the embodiment of FIG. 4, when the controller 160 receives a read command or a write command, the controller 160 similarly finds out a read command or a write command including the read command. A data link relationship set corresponding to one of the pages specified by the command or write command. For example, in this embodiment, a data link relationship set corresponding to one of the pages specified by the read command or the write command is TS_8. Next, the controller 160 determines whether the data link relationship set TS_8 has been loaded into the prefetch area CA according to whether a set index corresponding to the data link relationship set TS_8 exists in the prefetch area correspondence table SMR_TB. For example, the set index of the data link relationship set TS_8 is 0xBB. As shown in FIG. 4, the prefetch area CA is only written into the data link relationship set TS_2, and only the section corresponding column SMR_N in the prefetch area correspondence table SMR_TB is 0xAA, and the others are 0xFFFF. Therefore, in this embodiment, the controller 160 determines that the set index 0xBB of the data link relationship set TS_8 does not exist in the prefetch area correspondence table SMR_TB. In other words, the controller 160 determines that there is no data link relationship set TS_8 in the prefetch area CA according to the set index 0xBB of the data link relationship set TS_8 does not exist in the prefetch area correspondence table SMR_TB. Next, the controller 160 determines whether the prefetch area CA has a blank sector. As shown in FIG. 4, the sections S_0 ~ S_N-1 of the prefetch area CA are all blank, and the controller 160 selects a blank section. In the present embodiment, the controller 160 selects the section S_N-1. Then, the controller 160 reads the data link relationship set TS_8 from the data link relationship table TB1 in the flash memory 180 to load the data link relationship set TS_8 into the selected blank section S_N-1. Next, the controller 160 updates the prefetch area correspondence table SMR_TB, the order table SE_TB, the reverse order table RSE_TB, the head index HIX, and the tail index TIX to record the current state of the prefetch area CA. As shown in FIG. 4, the controller 160 obtains a segment S_N to which the last data link relationship set TS_2 read in the prefetch area CA according to the header index HIX, and performs the corresponding order of the segments S_N. The common index CIX_N in the column SE_N is written in the sequence column SE_N-1 corresponding to the current section S_N-1. Next, the controller 160 writes the common index CIX_N-1 corresponding to the reverse sequence column RSE_N-1 of the current sector S_N-1 into the reverse sequence column RSE_N corresponding to the sector S_N. Next, the controller 160 defines the head index HIX as the common index CIX_N-1 of the current section S_N-1, and maintains the tail index TIX at the common index CIX_N. The updated prefetch area correspondence table SMR_TB, sequence table SE_TB, reverse sequence table RSE_TB, head index HIX, and tail index TIX are shown in FIG. 5. Finally, the controller 160 reads the data link relationship set TS_8 stored in the prefetch area CA to execute the received write command and read command.

Assuming that the controller 160 successively receives write commands and read commands corresponding to different sets of data link relationships, the controller 160 repeats the above steps of converting from FIG. 4 to FIG. 5 and sequentially sets the data link relationships set Write to S_N-2 ~ S_0, as shown in Figure 6. In other words, the controller 160 repeatedly executes the loading procedure and fills all the prefetch areas CA. It is worth noting that in this embodiment, since the initial value of the sequence table SE_TB and the reverse sequence table RSE_TB is a specific value (0xFFFF), there is no need to write the specific value into the corresponding section in the loading program. The reverse order column and the order column corresponding to the most recently unread section in the prefetch area CA. In other embodiments, if the sequence table SE_TB and the reverse sequence table RSE_TB are different from the specified specific value, the controller 160 needs to write the specific value to the corresponding inverse of the current section in the last step of the loading procedure. The sequence column and the sequence column corresponding to the least recently read segment in the prefetch area CA.

The following is a description of the update procedure. In the embodiment of FIG. 6, when the controller 160 receives a read command or a write command, the controller 160 similarly finds out whether the read command or the write command includes the read command or the write command. A data link relationship set corresponding to one of the pages specified by the write command. For example, in this embodiment, a data link relationship set corresponding to one of the pages specified by the read command or the write command is TS_0. Next, the controller 160 determines whether the data link relationship set TS_0 has been loaded into the prefetch area CA according to whether a set index 0x5 corresponding to the data link relationship set TS_0 exists in the prefetch area correspondence table SMR_TB. In this embodiment, as shown in FIG. 6, the set index 0x5 of the data link relationship set TS_0 already exists in the region corresponding column SMR_2 in the prefetch region corresponding table SMR_TB. Therefore, in this embodiment, the controller 160 determines that the set index 0x5 of the data link relationship set TS_0 exists in the prefetched region correspondence table SMR_TB. In other words, the controller 160 determines that the data of the data link relationship set TS_0 already exists in the prefetch area CA according to the set index 0x5 of the data link relationship set TS_0 in the prefetch area correspondence table SMR_TB. Next, the controller 160 updates the prefetch area correspondence table SMR_TB, the order table SE_TB, the reverse order table RSE_TB, the head index HIX, and the tail index TIX to record the current state of the prefetch area CA. As shown in FIG. 6, the controller 160 reads the sequence column SE_2 of the current sector S_2 to obtain a sector corresponding to a sector written before the current sector S_2 as the sector S_3. Next, the controller 160 writes the shared index CIX_1 stored in the reverse order column RSE_2 corresponding to the current section S_2 into the reverse order column RSE_3 corresponding to the section S_3. In other words, the controller 160 points originally to the reverse-order column RSE_3 of the reverse-order column RSE_2, and instead points to the reverse-order column RSE_1. Next, the controller 160 reads the value in the reverse order column RSE_2 corresponding to the current sector S_2 in the reverse order column RSE_2 according to the common index CIX_2 of the current sector S_2, and finds the The common index CIX_1 of the next sector S_1 read after the current sector S_2. Then, the controller 160 writes the common indicator CIX_3 stored in the sequence column SE_2 corresponding to the current section S_2 into the sequence column SE_1 corresponding to the section S_1. In other words, the controller 160 points originally to the sequence column SE_1 of the sequence column SE_2, and instead points to the sequence column SE_3. Next, the controller 160 obtains a segment S_0 to which the last data link relationship set TS_1 read in the prefetch area CA belongs according to the header index HIX, and sets the shared index of the sequence column SE_0 corresponding to the segment S_0. CIX_0, write the sequence column SE_2 corresponding to the current section S_2. Then, the controller 160 writes the common index CIX_2 corresponding to the reverse sequence column RSE_2 of the current segment S_2 into the reverse sequence column RSE_0 corresponding to the current segment S_2. Finally, the controller 160 defines the head indicator HIX as the common indicator CIX_2 of the current section S_2, maintains the tail indicator TIX at the common indicator CIX_N, and writes a specific value into the reverse sequence column RSE_2 corresponding to the current section S_2. The updated prefetch region correspondence table SMR_TB, sequence table SE_TB, reverse sequence table RSE_TB, head index HIX, and tail index TIX are shown in FIG. 7A. Next, the controller 160 reads the corresponding segment S_2 in the prefetch area CA according to the common index CIX_2 of the segment corresponding column SMR_2 for storing the set index 0x5 to obtain the data link relationship set TS_0. In other words, the controller 160 reads the data link relationship set TS_0 stored in the prefetch area CA to execute the received write command and read command.

It is worth noting that, in another embodiment of the update program, when the data link relationship set corresponding to the write command and the read command is the longest read data link relationship set in the prefetch area CA The operation of the controller 160 is different from that shown in FIG. 7A. For example, in this embodiment, a data link relationship set corresponding to one of the pages specified by the read command or the write command is TS_2. Next, the controller 160 determines whether the data link relationship set TS_2 has been loaded into the prefetch area CA according to whether a set index 0xAA corresponding to the data link relationship set TS_2 exists in the prefetch area correspondence table SMR_TB. In this embodiment, as shown in FIG. 6, the set index 0xAA of the data link relationship set TS_2 already exists in the area corresponding column SMR_N in the prefetch area corresponding table SMR_TB. Therefore, in this embodiment, the controller 160 determines that the set index 0xAA of the data link relationship set TS_2 exists in the prefetched region correspondence table SMR_TB. In other words, the controller 160 determines that the data of the data link relationship set TS_2 already exists in the prefetch area CA according to the set index 0xAA of the data link relationship set TS_2 in the prefetch area correspondence table SMR_TB. It is worth noting that, in this embodiment, it can be known from the tail index TIX in FIG. 6 that the data link relationship set is TS_2, which is the longest read data link relationship set in the prefetch area CA. Next, the controller 160 updates the prefetch area correspondence table SMR_TB, the order table SE_TB, the reverse order table RSE_TB, the head index HIX, and the tail index TIX to record the current state of the prefetch area CA. As shown in FIG. 6, the controller 160 reads the value in the reverse order table RSE_TB corresponding to the current section S_N in the reverse order column RSE_N according to the common index CIX_N of the current section S_N, and according to the value in the reverse order column RSE_N The common indicator CIX_N-1 finds the common indicator CIX_N-1 of the sector S_N-1 that is read next after the current sector S_N. Next, the controller 160 writes the value stored in the sequence column SE_N corresponding to the current section S_N in the sequence column SE_N-1 corresponding to the section S_N-1, where the current section S_N is the longest unread Take the section, so the value stored in the sequence column SE_N is a specific value. Next, the controller 160 obtains a segment S_0 to which the last data link relationship set TS_1 read in the prefetch area CA belongs according to the header index HIX, and sets the shared index of the sequence column SE_0 corresponding to the segment S_0. CIX_0, write the sequence column SE_N corresponding to the current section S_N. Next, the controller 160 writes the common index CIX_N corresponding to the reverse column RSE_N of the current sector S_N into the reverse sequence column RSE_0 corresponding to the sector S_0. Finally, the controller 160 defines the head index HIX as the common index CIX_N of the current sector S_N, and the tail index TIX as the common index CIX_N-1 of the sector S_N-1 that is read next after the current sector S_N. And the specific value is written in the reverse order column RSE_N corresponding to the common indicator CIX_N indicated by the current head indicator. The updated prefetch area correspondence table SMR_TB, sequence table SE_TB, reverse sequence table RSE_TB, head index HIX, and tail index TIX are shown in FIG. 7B. Next, the controller 160 reads the corresponding segment S_N in the prefetch area CA according to the common index CIX_N of the segment corresponding column SMR_N used to store the set index 0xAA to obtain the data link relationship set TS_2. In other words, the controller 160 reads the data link relationship set TS_2 stored in the prefetch area CA to execute the received write command and read command.

The following is a description of the replacement procedure. In the embodiment of FIG. 7A, when the controller 160 receives a read command or a write command, the controller 160 similarly finds out whether the read command or the write command includes the read command or the write command. A data link relationship set corresponding to one of the pages specified by the write command. For example, in this embodiment, a data link relationship set corresponding to one of the pages specified by the read command or the write command is TS_77. Next, the controller 160 determines whether the data link relationship set TS_77 has been loaded into the prefetch area CA according to whether a set index 0x333 corresponding to the data link relationship set TS_77 exists in the prefetch area correspondence table SMR_TB. In this embodiment, it is assumed that the prefetch area CA does not have data of the data link relationship set TS_77, so the set index 0x333 does not exist in any of the area corresponding columns SMR_0 ~ SMR_N in the prefetch area corresponding table SMR_TB. The controller 160 determines that there is no data link relationship set TS_77 in the prefetch area CA according to the set index 0x333 of the data link relationship set TS_77 that does not exist in the prefetch area correspondence table SMR_TB. Next, the controller 160 determines whether the prefetch area CA has a blank area. As shown in FIG. 7A, the sections S_0 ~ S_N of the prefetch area CA are not blank. Next, the controller 160 judges the data link relationship set in the prefetch area CA that has not been read for the longest time according to the tail index TIX, so as to write the data in the data link relationship set TS_77 to the longest unread data in the prefetch area CA. The section to which the data link relationship collection belongs. As shown in FIG. 7A, the controller 160 judges the data link relationship set TS_2 that has not been read for the longest time in the prefetch area CA according to the common index CIX_N in the tail index TIX, and according to the common index CIX_N The sector with the longest unread is obtained as sector S_N. Next, the controller 160 reads the data link relationship set TS_77 from the data link relationship table TB1 in the flash memory 180 to load the data link relationship set TS_77 into the obtained section S_N to replace the original long-term The read data link relationship set TS_2. Next, the controller 160 updates the prefetch area correspondence table SMR_TB, the order table SE_TB, the reverse order table RSE_TB, the head index HIX, and the tail index TIX to record the current state of the prefetch area CA. First, the controller 160 writes a set index 0x333 corresponding to the data link relationship set TS_77 into a section corresponding column SMR_N corresponding to the section S_N in the prefetch area corresponding table SMR_TB. As shown in FIG. 7A, the controller 160 obtains a segment S_2 to which the last data link relationship set TS_0 read in the prefetch area CA according to the header index HIX, and sorts the segment S_2 in the corresponding order. The shared index CIX_2 of the column SE_2 is written in the sequence column SE_N corresponding to the current section S_N. Next, the controller 160 writes the common index CIX_N corresponding to the reverse column RSE_N of the current sector S_N into the reverse sequence column RSE_2 corresponding to the sector S_2. Next, the controller 160 reads the reverse order column RSE_N of the current section S_N to obtain the common index CIX_N-1 of the next section S_N-1 that is read after the current section S_N. Next, the controller 160 defines the head indicator HIX as the common indicator CIX_N of the current section S_N, and defines the tail indicator TIX as the common indicator CIX_N- of the next section S_N-1 that is read after the current section S_N. 1. Finally, the controller 160 writes the specific value into the sequence column SE_N-1 corresponding to the next read section S_N-1 after the current section S_N, and writes the specific value into the corresponding section S_N. In the reverse order column RSE_N. The updated prefetch area correspondence table SMR_TB, sequence table SE_TB, reverse sequence table RSE_TB, head index HIX, and tail index TIX are shown in FIG. 8. Finally, the controller 160 reads the data link relationship set TS_8 stored in the prefetch area CA to execute the received write command and read command.

FIG. 9 is a flowchart of an embodiment of a data maintenance method provided by the present invention. The data maintenance method shown in FIG. 9 is applicable to the data storage device 140 shown in FIG. 1. The process starts at step S900.

In step S900, after the flash memory 160 is powered on, the controller 160 loads one of the prefetch areas CA in the random memory 166 according to the received at least one read command or at least one write command. Part of a plurality of data link relationship sets TS_0 ~ TS_N in a data link relationship table TB1.

Next, in step S902, the controller 160 establishes a prefetched region correspondence table SMR_TB, a reverse sequence table RSE_TB, and a Sequence table SE_TB. In detail, the controller 160 establishes a prefetched region correspondence table SMR_TB in the random memory 166 to record a set index corresponding to a data link relationship set in the prefetched region CA. The controller 160 creates a reverse order table RSE_TB in the random memory 166 to record the reverse order of the data link relationship set in the prefetch area CA being read from the prefetch area CA. The controller 160 creates a sequence table SE_TB in the random memory 166 to record the order in which the data link relationship set in the prefetch area CA is read from the prefetch area CA. It should be noted that the order of reading in step S902 refers to the order in which the data link relationship table TB1 is read by the controller 160 from the prefetch area CA to execute the write command and the read command. In addition, in another embodiment, the controller 160 may also set a head index HIX and a tail index TIX according to the order in which the loaded data link relationship table TB1 is read. When the prefetch area CA is full and a new data link relationship set needs to be loaded, one of the sections S_0 ~ S_N in the prefetch area CA is selected according to the tail index TIX, so as to link the new data chain. The knot relationship set loads the selected section. After the new above-mentioned data link relationship set is loaded into the selected above-mentioned section, the controller 160 may update the upper-order table and the sequence table according to the header index, the data in the reverse-order table and the data in the sequence table SE_TB.

FIG. 10 is a flowchart of an embodiment of a data maintenance method provided by the present invention. The data maintenance method shown in FIG. 10 is applicable to the data storage device 140 shown in FIG. 1. The process starts at step S1000.

In step S1000, the controller 160 receives a read command or a write command. The read command and the write command are used to read or write a specific page in the flash memory 180. The read command and the write command may be commands received from the host 120 or the controller 160 itself. Write and read commands generated by the read and write operations required to maintain the data.

Next, in step S1002, the controller 160 determines, based on the read command or write command received in step S1000, one of the first data including a first correspondence relationship of the page specified by the read command or write command. Link relationship collection. For example, the correspondence between the logical address and the physical address of the page indicated by the read or write command is stored in the data link relationship set TS_2, and the controller 160 is based on the received read or write command. Learn the designated page and find the corresponding data link relationship set TS_2.

Next, in step S1004, the controller 160 determines whether a first data link relationship set has been loaded into the pre-set according to whether a first set index corresponding to the first data link relationship set exists in the prefetch area correspondence table SMR_TB. Take area CA. When the first data link relationship set is not loaded into the prefetch area CA, the flow proceeds to step S1006; otherwise, the flow proceeds to step S1060 to perform an update procedure.

In step S1006, the controller 160 determines whether there is a blank section in the prefetch area CA. It is worth noting that, in this embodiment, the blank section indicates a section that has not been loaded with any data link relationship set in the prefetch area CA. When there is a blank section in the prefetch area CA, the flow proceeds to step S1008 to execute a loading procedure; otherwise, the flow proceeds to step S1030 to execute a replacement procedure.

In step S1008, the controller 160 reads the first data link relationship set from one of the data link relationship tables TB1 in the flash memory 180 to load the first data link relationship set into the selected blank one. First section. For example, as shown in FIG. 3, the sections S_0 ~ S_N of the prefetch area CA are all blank, and the controller 160 selects a blank section. In this embodiment, the controller 160 selects from the bottommost segment, so the controller 160 selects the segment S_N. Next, the controller 160 reads the data link relationship set TS_2 from the data link relationship table TB1 in the flash memory 180 to load the data link relationship set TS_2 into the selected blank section S_N. In another embodiment, as shown in FIG. 4, the sections S_0 ~ S_N-1 of the prefetch area CA are all blank, and the controller 160 selects a blank section from the sections S_0 ~ S_N-1. . In this embodiment, since the controller 160 is used to select from the bottommost section, the controller 160 selects the section S_N and the controller 160 selects the section S_N-1, but the present invention is not limited thereto. Then, the controller 160 reads the data link relationship set TS_8 from the data link relationship table TB1 in the flash memory 180 to load the data link relationship set TS_8 into the selected blank section S_N-1.

Next, in step S1010, the controller 160 writes a first set index corresponding to the first data link relationship set into a section corresponding column corresponding to the first section in the prefetch area corresponding table SMR_TB. For example, in the embodiment of FIG. 4, the controller 160 writes the data link relationship set TS_2 whose set index is 0xAA into the section S_N. Therefore, the controller 160 writes the set index 0xAA of the data link relationship set TS_2 into the section corresponding column SMR_N corresponding to the section S_N in step S1010. In the embodiment of FIG. 5, the controller 160 writes the data link relationship set TS_8 whose set index is 0xBB into the section S_N-1. Therefore, the controller 160 writes the set index 0xBB of the data link relationship set TS_8 into the section corresponding column SMR_N-1 corresponding to the section S_N-1 in step S1010.

Next, in step S1012, the controller 160 determines whether the first section is the first loaded data link relationship set in the prefetch area CA. In other words, the controller 160 determines whether all other sections in the prefetch area CA are blank. When the first loaded data link relationship set in the prefetch area CA is set, the flow proceeds to step S1014; otherwise, the flow proceeds to step S1018. It is worth noting that the controller 160 can also determine whether the first section is the first set of data link relationships loaded in the prefetch area CA in the process of determining whether there is a blank section in step S1006. step.

In step S1014, the controller 160 simultaneously defines a head indicator HIX and a tail indicator TIX as one of the first common indicators in the first area. For example, in the embodiment of FIG. 4, the data link relationship set TS_2 is the first data link relationship set written in the prefetch area CA, where the data link relationship set TS_2 is written into the section S_N. Therefore, in step S1014, the controller 160 will simultaneously define the head index HIX and the tail index TIX as the common index CIX_N of the segment S_N. Then, the flow proceeds to step S1090.

In step S1018, the controller 160 obtains a second section to which the last data link relationship set read in the prefetch area CA according to the header index HIX. It is worth noting that, in this embodiment, the last set of data link relationships read in the prefetch area CA refers to the last prefetch area CA that was read by the controller 160 from the prefetch area CA. A set of data link relationships for receiving a read command or a write command.

Next, in step S1020, the controller 160 writes the common index of the second order column corresponding to the second section in the sequence table SE_TB into the first order column corresponding to the first section. For example, in the embodiment of the fifth figure, the data link relationship set TS_8 is not the first data link relationship set written in the prefetch area CA, where the data link relationship set TS_8 is written into the area Segment S_N-1 (first segment). Before updating the table in the random access memory 166, the controller 160 obtains the last read data chain in the prefetch area CA according to the current head index HIX (as shown in FIG. 4) in step S1018. A section S_N (second section) to which the relationship set TS_2 belongs. Next, in step S1020, the controller 160 writes the common index CIX_N of the sequence column SE_N (second sequence column) corresponding to the section S_N (second section) into the current section S_N-1 (the second section) A section) of the sequence column SE_N-1 (the first sequence is bad).

Next, in step S1022, the controller 160 writes the common index corresponding to the first reverse order column of the first section in the reverse order table RSE_TB into a second reverse order column corresponding to the second section. For example, in the embodiment of the fifth figure, the controller 160 in step S1022 corresponds to the reverse order column RSE_N-1 (first reverse order column) corresponding to the current section S_N-1 (first section). The common index CIX_N-1 is written in the reverse sequence column RSE_N corresponding to the sector S_N (second sector).

Next, in step S1024, the controller 160 defines the head index HIX as a first common index of one of the first sections. For example, in the embodiment of the fifth figure, the controller 160 defines the head indicator HIX as the common indicator CIX_N-1 (the first common indicator) of the current section S_N-1 (the first section) in step S1024. ), And the tail indicator TIX is maintained at the common indicator CIX_N (second common indicator). Then, the flow proceeds to step S1090.

In step S1030, the controller 160 determines, according to the tail index TIX, a third section to which the most recently unread data link relationship set in the prefetch area CA belongs. It is worth noting that, in this embodiment, the set of data link relationships in the prefetch area CA that has not been read for the longest time refers to the current prefetch area CA that has not been read by the controller 160 in the prefetch area CA for the longest A collection of data link relationships for reading to perform a received read command or write command.

Next, in step S1032, the controller 160 writes the data of the first data link relationship set into the third section. For example, in the embodiment of FIG. 8, a correspondence relationship of one of the pages specified by the read command or the write command is included in the data link relationship set TS_77 (the first data link relationship set). The controller 160 has determined in step S1004 that the set index 0x333 corresponding to the data link relationship set TS_77 does not exist in the prefetch area correspondence table SMR_TB, and the controller 160 has determined in step S1006 that there is no blank in the prefetch area CA. Section. Therefore, in step S1030, the controller 160 determines that the data link relationship set that has not been read for the longest time in the prefetch area CA is the data link relationship set TS_2 according to the tail index TIX shown in FIG. 7A before the change. The section corresponding to the link relationship set TS_2 is section S_N (third section). Therefore, in step S1032, the controller 160 reads the data link relationship set TS_77 (the first data link relationship set) from the data link relationship table TB1 in the flash memory 180 to collect the data link relationship set TS_77. Load the obtained segment S_N (third segment) to replace the data link relationship set TS_2 that has not been read for a long time. .

Next, in step S1034, the controller 160 writes the first set index corresponding to the first data link relationship set into the section corresponding column corresponding to the third section in the prefetch area corresponding table SMR_TB. In the embodiment of FIG. 8, the controller 160 then writes a set index 0x333 (first set index) corresponding to the data link relationship set TS_77 (first data link relationship set) in step S1034 to the prefetch In the area correspondence table SMR_TB, a section correspondence column SMR_N corresponding to the section S_N (third section).

Next, in step S1036, the controller 160 obtains a fourth section to which the last data link relationship set read in the prefetch area CA according to the header index HIX.

Next, in step S1038, the controller 160 writes the common index in the sequence table SE_TB corresponding to the fourth order column of one of the fourth sections into the third order column corresponding to one of the third sections in the order table SE_TB. . For example, in the embodiment of FIG. 8, the controller 160 obtains the last data link relationship read in the prefetch area CA according to the head index HIX shown in FIG. 7A before the change in step S1036. The set TS_0 belongs to the section S_2 (fourth section). Next, in step S1038, the controller 160 writes the common index CIX_2 of the sequence column SE_2 (the fourth sequence column) corresponding to the section S_2 (the fourth section) into the current section S_N (the third zone) Paragraph) sequence column SE_N (third sequence column).

Next, in step S1040, the controller 160 writes the common index corresponding to the third reverse order column in the reverse order table RSE_TB corresponding to one of the third sections into the reverse order table RSE_TB corresponding to one of the fourth sections. Reverse order bar. For example, in the embodiment of FIG. 8, the controller 160 writes a common index CIX_N corresponding to the reverse sequence column RSE_N (third reverse column) of the current section S_N (third section) to the area The reverse order column RSE_2 (fourth reverse order column) corresponding to the segment S_2 (fourth section).

Next, in step S1044, the controller 160 reads the third reverse-order column to obtain a common indicator of a fifth reverse-order column RSE_0 ~ RSE_N in the reverse-order column RSE_0 ~ RSE_N. In other words, the controller 160 reads the third reverse sequence column corresponding to the currently written third sector to obtain a fifth sector that was originally read next to the current third sector.

Next, in step S1044, the controller 160 defines the tail index TIX as a fifth common index corresponding to one of the fifth reverse order columns and the head index HIX as a third common index as one of the third order columns. For example, in the embodiment of FIG. 8, the controller 160 reads the reverse order column RSE_N (third reverse order column) of the current section S_N (third section) in step S1044 to obtain the original The common index CIX_N-1 (the common index in the fifth reverse column) of the current sector S_N (the third sector) is read next to the sector S_N-1. Next, the controller 160 defines the head indicator HIX as the common indicator CIX_N of the current section S_N (third section), and defines the tail indicator TIX as the section S_N that was originally read next after the current section S_N. A common index CIX_N-1 (the fifth common index) of -1 (the fifth sector).

Next, in step S1046, the controller 160 writes a specific value into the third reverse sequence column and the fifth sequence column corresponding to one of the fifth common indicators. In other words, the controller 160 writes the specific value into the reverse order column corresponding to the common indicator pointed to by the current head indicator, and writes the specific value into the sequence column corresponding to the common indicator pointed to by the current tail indicator. As shown in FIG. 8, the controller 160 writes a specific value into the reverse order column RSE_N and the common index CIX_N-1 (the fifth reverse order common index) corresponding to the current sector S_N (third sector). Sequence column SE_N-1 (the fifth sequence column). Then, the flow proceeds to step S1090.

In step S1060, the controller 160 reads a sixth reverse order column corresponding to the sixth section in which one of the first data link relationship sets is stored, so as to obtain the next data link after the first data link relationship set. Read the seventh sector. In other words, the controller 160 reads a sixth reverse sequence column corresponding to a sixth sector in a reverse sequence table RSE_TB to obtain a seventh sector that is next read after the sixth sector. For example, in the embodiment of FIG. 7A, a correspondence relationship of one of the pages specified by the read command or the write command is included in the data link relationship set as TS_0. As shown in FIG. 6 before the change, the set index 0x5 of the data link relationship set TS_0 already exists in the region corresponding column SMR_2 in the prefetch region corresponding table SMR_TB. Therefore, in step S1004, the controller 160 has determined that the data link relationship set TS_0 has been loaded into the prefetch area CA according to the existence of the set index 0x5 corresponding to the data link relationship set TS_0 in the prefetch area correspondence table SMR_TB. Next, in step S1060, the controller 160 reads the reverse order column RSE_2 in the reverse order table RSE_TB corresponding to the current sector S_2 (sixth sector) according to the common index CIX_2 of the current sector S_2 (sixth sector). Value in the sixth reverse order column), and according to the shared index CIX_1 in the reverse order column RSE_2 (the sixth reverse order column), the next section read after the current section S_2 (the sixth section) is Section S_1. In the embodiment of FIG. 7B, a correspondence relationship of one of the pages specified by the read command or the write command is included in the data link relationship set as TS_2. As shown in FIG. 6 before the change, the set index 0xAA of the data link relationship set TS_2 already exists in the region corresponding column SMR_N in the prefetch region corresponding table SMR_TB. Therefore, in step S1004, the controller 160 has determined that the data link relationship set TS_2 has been loaded into the prefetch area CA according to the existence of the set index 0xAA corresponding to the data link relationship set TS_N in the prefetch area correspondence table SMR_TB. Next, in step S1060, the controller 160 reads the reverse order column RSE_N corresponding to the current section S_N (sixth section) in the reverse order table RSE_TB according to the common index CIX_N of the current section S_N (sixth section). Value in the sixth reverse order column), and find the next area to be read after the current section S_N (sixth section) according to the common index CIX_N-1 in the reverse order column RSE_N (sixth reverse order column). The segment is a segment S_N-1.

Next, in step S1062, the controller 160 writes the shared index stored in the sixth sequence column corresponding to the sixth sector in the seventh sequence column corresponding to one of the seventh sectors in the sequence table SE_TB. For example, in the embodiment of FIG. 7A, the controller 160 has obtained in step S1060 that the next sector read after the current sector S_2 (sixth sector) is sector S_1 (seventh sector). segment). Therefore, in step S1062, the controller 160 writes the sequence column corresponding to the current sector S_2 (sixth sector) in the sequence column SE_1 (seventh sequence column) corresponding to the sector S_1 (seventh sector). The shared indicator CIX_3 stored in SE_2 (sixth order column). In other words, the controller 160 points originally to the sequence column SE_1 (seventh sequence column) of the sequence column SE_2 (sixth sequence column), and points to the corresponding section S_3 that was read before the first data link relationship set. Sequence column SE_3. In the embodiment of FIG. 7B, the controller 160 has obtained in step S1060 that the next sector read after the current sector S_N (sixth sector) is sector S_N-1 (seventh sector). . Therefore, in step S1062, the controller 160 writes the current section S_N (sixth section) in the order column SE_N-1 (seventh order column) corresponding to the section S_N-1 (seventh section). The value stored in the corresponding sequence column SE_N (sixth sequence column).

Next, in step S1064, the controller 160 obtains an eighth section to which the last data link relationship set read in the prefetch area CA belongs according to the header index HIX.

Next, in step S1065, the controller 160 determines whether the sector to which the data link relationship set that has been read the least recently in the prefetch area CA belongs is the sixth sector according to the tail index TIX. When the section to which the data link relationship set that has not been read for the longest time belongs to the sixth section, the flow proceeds to step S1080; otherwise, the flow proceeds to step S1066. In other words, the controller 160 determines in this step whether the currently set data link relationship set to be read is the data link relationship set that has not been read for the longest time in the prefetch area CA.

In step S1066, the controller 160 reads a sixth sequence column corresponding to the sixth sector to obtain a ninth common index corresponding to one of the ninth sectors written before the sixth sector. In the embodiment of FIG. 7A, the controller 160 determines in step S1065 according to the tail index TIX that the data link relationship set TS_0 stored in the current section S_2 (sixth section) is not the longest time in the prefetch area CA. The collection of data link relationships read. Next, the controller 160 reads a sequence column SE_2 (sixth sequence column) corresponding to the current section S_2 (sixth section) in step S1066 to obtain a sequence corresponding to the current section S_2 (sixth section). One sector to be written is sector S_3 (ninth sector).

Next, in step S1068, the controller 160 writes the shared index stored in the sixth reverse order column into the ninth reverse order column corresponding to one of the ninth shared index. For example, in the embodiment of FIG. 7A, the controller 160 in step S1068 corresponds to the share stored in the reverse order column RSE_2 (sixth reverse order column) corresponding to the current section S_2 (sixth section). The index CIX_1 is written in the reverse order column RSE_3 (the ninth reverse order column) corresponding to the section S_3 (the ninth section).

Next, in step S1070, the controller 160 writes the sharing of the eighth sequence column corresponding to the eighth segment in the sequence table SE_TB in the sixth sequence column corresponding to the sixth segment in the sequence columns SE_0 ~ SE_N. index. For example, in the embodiment of FIG. 7A, the controller 160 obtains the last data link relationship set TS_1 read in the prefetch area CA according to the head index HIX in FIG. 6 in step S1064. One segment S_0 (eighth segment), and write the common index CIX_0 of the sequence column SE_0 (eighth sequence column) corresponding to the segment S_0 (eighth segment) into the current segment S_2 ( The sixth column) is a sequence column SE_2 (sixth sequence column).

Next, in step S1072, the controller 160 writes the common index corresponding to the sixth reverse order column of the sixth section in the reverse order table RSE_TB into the reverse order table RSE_TB corresponding to the eighth area that was last read. One of the eighth inverse columns. For example, in the embodiment of FIG. 7A, the controller 160 sets the common index CIX_2 of the reverse order column RSE_2 (the sixth reverse order column) corresponding to the current section S_2 (the sixth section) in step S1072. , Write the reverse sequence column RSE_0 (eighth reverse sequence column) corresponding to the sector S_0 (eighth sector).

Next, in step S1074, the controller 160 defines the head index HIX as a sixth common index in one of the sixth order columns. For example, in the embodiment of FIG. 7A, the controller 160 defines the head index HIX as the common index CIX_2 of the current section S_2 in step S1074, and maintains the tail index TIX at the common index CIX_N.

Next, in step S1076, the controller 160 writes a specific value into the sixth reverse order column. For example, in the embodiment of FIG. 7A, the controller 160 writes a specific value into the reverse sequence column RSE_2 (the sixth reverse sequence column) corresponding to the current section S_2 (the sixth section) in step S1076. )in. Next, the flow proceeds to step S1090.

In step S1080, the controller 160 writes the common index of the eighth sequence column corresponding to the eighth sector in the sequence table SE_TB in the sixth sequence column corresponding to the sixth sector. In the embodiment of FIG. 7B, the controller 160 determines in step S1065 according to the tail index TIX that the data link relationship set TS_2 stored in the current section S_N (sixth section) is not the oldest in the prefetch area CA. The collection of data link relationships read. Next, the controller 160 reads a sequence column SE_2 (sixth sequence column) corresponding to the current section S_2 (sixth section) in step S1080 to obtain a sequence corresponding to the current section S_2 (sixth section). One sector to be written is sector S_3 (ninth sector).

Next, in step S1082, the controller 160 writes the shared index corresponding to the sixth reverse order column of the sixth section in the reverse order table RSE_TB into the reverse order table RSE_TB corresponding to the eighth area that is read last. One of the eighth inverse columns. For example, in the embodiment of FIG. 7B, the controller 160 has obtained the last sector read in step S1064 as the sector S_0 (eighth sector). Therefore, the controller 160 writes the common index CIX_N in the reverse order column RSE_N (the sixth reverse order column) corresponding to the current section S_N (the sixth section) in step S1082 into the section S_0 (the eighth section) ) Corresponding to the reverse sequence column RSE_0 (eighth reverse sequence column).

Next, in step S1084, the controller 160 defines the head index HIX as a sixth common index in one of the sixth order columns, and defines the tail index TIX as a seventh common index in one of the seventh sections. For example, in the embodiment of FIG. 7B, the controller 160 defines the head indicator HIX as the common indicator CIX_2 of the current section S_2 (sixth section) in step S1084, and maintains the tail indicator TIX at the common indicator CIX_N. .

Next, in step S1086, the controller 160 writes a specific value into the sixth reverse order column. In other words, the controller 160 writes the specific value into the reverse order column RSE_N corresponding to the common indicator CIX_N indicated by the current head indicator. Then, the flow proceeds to step S1090.

In step S1090, the controller 160 reads the data in the first data link relationship set from the prefetch area CA to execute the write command or the read command received in step S1000. The flow ends in step S1090.

It is worth noting that generally reading and writing a file stored in the flash memory 180 will cause the related data link relationship set to be read repeatedly. In other words, some data link relationships related to commands that are not related to reading and writing files are usually read only once. In view of this, another embodiment of the present invention is provided to divide the frequently read data link relationship set and the not frequently read data link relationship set into two systems for maintenance, in which the prefetch area CA The data link relationship set that is read less than a predetermined value in the stored data link relationship set belongs to the infrequent data link relationship set, and is read from the data link relationship set stored in the prefetch area CA The set of data link relationships that reach a given value belongs to the set of frequent data link relationships. In one embodiment, the predetermined value is 1, but the present invention is not limited thereto. In other embodiments, the predetermined value can also be 2, 3, 4, 5, 6, 7, or 8. The developer can design based on the mode of reading or writing the flash memory 180 in different situations. Established value.

FIG. 11 is a block diagram of another embodiment of a table in a random access memory provided by the present invention. As shown in FIG. 11, after the data storage device 140 is powered on, the controller 160 creates a blank prefetch area correspondence table SMR_TB, a reverse sequence table RSE_TB, and a sequence table SE_TB in the random access memory 166. To record the status of the data in the prefetch area CA.

The prefetch area CA has a plurality of sections S_0 ~ S_N for storing the data link relationship set in the data link relationship table TB1, and one section is used for storing a data link relationship set. In other words, the number of segments S_0 ~ S_N is smaller than the number of data link relationship sets TS_0 ~ TS_M, that is, N <M.

The prefetched area correspondence table SMR_TB has a plurality of section corresponding columns SMR_0 ~ SMR_N to record the set index corresponding to the data link relationship set TS_0 ~ TS_M in the prefetched area CA, where the prefetched area correspondence table SMR_TB corresponds to the section The columns SMR_0 ~ SMR_N sequentially correspond to the sections S_0 ~ S_N in the prefetch area CA. It is worth noting that the initial values of the corresponding columns SMR_0 ~ SMR_N in each section of the prefetched region correspondence table SMR_TB are a specific value. In this embodiment, the specific value is 0xFFFF, but the present invention is not limited to this. In addition, the set indexes of the data link relationship sets TS_0 ~ TS_M are not equal to a specific value.

The sequence table SE_TB is used to record the order in which the frequent data link relationship set is read from the prefetch area CA and the order in which the infrequent data link relationship set is read from the domain fetch area. The sequence table SE_TB has a plural order. The columns SE_0 ~ SE_N sequentially correspond to the sections S_0 ~ S_N in the prefetch area CA. It is worth noting that in this embodiment, the order columns SE_0 ~ SE_N in the order table SE_TB can record two orders (the order in which the frequent data link relationship set is read from the prefetch area CA and the infrequent data link The order in which relation sets are read from the domain fetch area). In detail, the order column with the same common index as the section to which the frequent data link relationship set belongs belongs to the frequent order column, and the order column with the same common index to the section to which the frequent data link relationship set belongs belongs to the non- Frequent order bar. In other words, the attributes of the sequence column are dynamically determined according to the appearance (frequent or infrequent) of the complex data link relationship set stored in the corresponding section. Each frequent order column is used to store a common indicator of another frequent order column, which respectively points to the corresponding frequent corresponding to the section of the frequent data link relationship set that was read before the section corresponding to the frequent order column. Order bar. Each infrequently ordered column is used to store a common indicator of another infrequently ordered column to point to the section of the infrequent data link relationship set that was read before the section corresponding to the infrequently ordered column, respectively. The corresponding infrequent order column. It is worth noting that the initial value of each sequence column SE_0 ~ SE_N in the sequence table SE_TB is a -specific value. In this embodiment, the specific value is 0xFFFF, but the present invention is not limited to this.

The reverse order table RSE_TB is used to record the reverse order of the frequent data link relationship set read from the prefetch area CA and the reverse order of the infrequent data link relationship set read from the prefetch area CA. Among them, the reverse order table RSE_TB has a plurality of reverse order columns RSE_0 ~ RSE_N corresponding to the sections S_0 ~ S_N in the prefetch area CA in order. It is worth noting that in this embodiment, the reverse order columns RSE_0 ~ RSE_N in the reverse order table RSE_TB can record two orders (the reverse order of the frequent data link relationship set read from the prefetch area CA and The set of infrequent data link relationships is read in reverse order from the domain fetch area). In detail, the reverse order column with the same shared index as the section to which the frequent data link relationship set belongs belongs to the frequent reverse order column, and the reverse order column with the same shared index to the section where the frequent data link relationship set belongs to The columns are infrequent reverse columns. In other words, the attributes of the reverse sequence columns RSE_0 ~ RSE_N are dynamically determined according to the appearance (frequent or infrequent) of the complex data link relationship set stored in the corresponding section. Each frequent reverse order column RSE_0 ~ RSE_N is used to store a common indicator of another frequent reverse order column, which respectively points to the next frequent data link relationship set that is read after the section corresponding to the frequent reverse order column. Frequent reverse order column corresponding to the segment. Each infrequent reverse order column RSE_0 ~ RSE_N is used to store the common indicator of another infrequent reverse order column, respectively, to point to the next infrequent data chain that is read after the section corresponding to the infrequent reverse order column. Infrequent reverse columns corresponding to the sections of the relationship set. It is worth noting that the initial value of each of the reverse order columns RSE_0 ~ RSE_N in the reverse order table RSE_TB is a specific value. In this embodiment, the specific value is 0xFFFF, but the present invention is not limited to this.

It is worth noting that, in an embodiment, the segments S_0 ~ S_N have specific common indexes CIX_0 ~ CIX_N in order. The section corresponding columns SMR_0 ~ SMR_N and the corresponding sections S_0 ~ S_N have the same common index CIX_0 ~ CIX_N. The sequence columns SE_0 ~ SE_N and their corresponding sections S_0 ~ S_N have the same common indicators CIX_0 ~ CIX_N. The reverse order columns RSE_0 ~ RSE_N and their corresponding sections S_0 ~ S_N have the same common index CIX_0 ~ CIX_N. In detail, the shared indicators of the segment S_0, the sequence column SE_0, the segment corresponding column SMR_0, and the reverse sequence column RSE_0 are all CIX_0. The shared indicators of the segment S_1, the sequence column SE_1, the segment corresponding column SMR_1, and the reverse sequence column RSE_1 are all CIX_1. The common indicators of the segment S_2, the sequence column SE_2, the segment corresponding column SMR_2, and the reverse sequence column RSE_2 are CIX_2, and so on. It is worth noting that each common indicator CIX_0 ~ CIX_N is not equal to a specific value. In an embodiment, the common indicators CIX_0 to CIX_N are sequentially hexadecimal values such as 0x0, 0x1, 0x2, and 0x3, but the present invention is not limited thereto.

In summary, the controller 160 establishes a prefetch area correspondence table SMR_TB to record a set index corresponding to the data link relationship set TS_0 ~ TS_M in the prefetch area CA. The controller 160 establishes a reverse order table RSE_TB to record the reverse order of the frequent data link relationship set read from the prefetch area CA and the non-frequent data link relationship set read from the prefetch area CA. Reverse order. The controller 160 establishes a sequence table SE_TB to record the order in which the frequent data link relationship set is read from the prefetch area CA and the order in which the infrequent data link relationship set is read from the domain fetch area. In an embodiment, the controller 160 is further configured to establish an infrequent index group and a frequent index group in the random access memory 166, wherein the infrequent index group is used to maintain an infrequent data link relationship set, and frequently The index group is used to maintain a collection of frequently linked data relationships. In one embodiment, the infrequent indicator group is composed of an infrequent header indicator LRU_HIX and an infrequent tail indicator LRU_TIX, and the frequent indicator group is composed of a frequent header indicator FR_HIX and a frequent tail indicator FR_TIX. In other embodiments, the infrequent header indicator LRU_HIX, the infrequent tail indicator LRU_TIX, the frequent header indicator FR_HIX, and the frequent tail indicator FR_TIX can also be established in other memories or circuits, and the present invention is not limited thereto. The infrequent header index LRU_HIX is a shared index for storing the last read infrequent data link relationship set segment to point to the last read infrequent data link relationship set in the prefetch area CA. The infrequent tail indicator LRU_TIX is a shared indicator for storing the segments of the infrequent data link relationship set that has not been read for the longest time, and points to the infrequent data link relationship set that has been read for the longest time in the prefetch area CA. The frequent header index FR_HIX is a shared index used to store the section of the last frequently read data link relationship set read to point to the last frequently read data link relationship set read in the prefetch area CA. The frequent tail index FR_TIX is a shared index used to store the segments of the frequent data link relationship set that have not been read for the longest time, so as to point to the most frequently read data link relationship set in the prefetch area CA. It is worth noting that because the corresponding sections, sequence columns, section corresponding columns, and reverse sequence columns have the same common indicators, the infrequent header indicator LRU_HIX, the infrequent tail indicator LRU_TIX, the frequent header indicator FR_HIX, and the frequent tail indicator FR_TIX also points to the corresponding order column, section corresponding column, and reverse order column. It is worth noting that, in an embodiment, the initial values of the infrequent header indicator LRU_HIX and the infrequent tail indicator LRU_TIX may be the common indicator CIX_N, and the initial values of the frequent header indicator FR_HIX and the frequent tail indicator FR_TIX may be specific values. This specific value is different from all common indicators CIX_0 ~ CIX_N. In another embodiment, the initial values of the infrequent header indicator LRU_HIX, the infrequent tail indicator LRU_TIX, the frequent header indicator FR_HIX, and the frequent tail indicator FR_TIX are all a specific value, which is different from all common indicators CIX_0 ~ CIX_N.

In this embodiment, after any data link relationship set TS_0 ~ TS_M is uploaded to the prefetch area CA, it belongs to the infrequent data link relationship set until another command causes the controller 160 to upload the data to the prefetch area CA. The infrequent data link relationship set is read, and the read infrequent data link relationship set belongs to the frequent data link relationship set. In addition, the controller 160 may select one of the plural sections S_0 ~ S_N in the prefetch area CA according to the infrequent tail index LRU_TIX. In other words, the controller 160 may select an infrequent data link relationship set that has not been read for the longest time in the prefetch area CA according to the infrequent tail index LRU_TIX to load a new data link relationship set to store the selected A section of a collection of frequent data link relationships. When all the data link relationship sets in the prefetch area CA belong to the frequent data link relationship set and the controller 160 needs to upload other new data link relationship sets, the controller 160 starts reading another file. Therefore, when the infrequent tail indicator LRU_TIX does not point to any common indicator, the controller 160 transfers the frequent indicator group to the infrequent indicator group and deletes the value in the frequent indicator group.

In detail, when the first collection index of the first data link relationship set that the controller 160 currently needs to read does not exist in the prefetch area correspondence table SMR_TB, the controller 160 determines whether the prefetch area CA has a blank area segment. When the controller 160 determines that the sections S_0 ~ S_N of the prefetch area CA have a blank first section, the controller 160 reads the first data link relationship from the data link relationship table TB1 in the flash memory 180 Set to load the first data link relationship set into the blank first section. When the controller 160 determines that the prefetch area CA does not have a blank sector, the controller 160 determines whether the infrequent tail index LRU_TIX is any of the common indexes. When the infrequent tail indicator LRU_TIX is any of the common indicators, the controller 160 determines the infrequent data link relationship set that has not been read for the longest time in the infrequent data link relationship set according to the infrequent tail indicator LRU_TIX, so that The data of the first data link relationship set is written into the section of the infrequent data link relationship set that belongs to the infrequent data link relationship set that has not been read for the longest time. When the infrequent tail indicator LRU_TIX is not any of the common indicators, the controller 160 writes the common indicator stored in the frequent tail indicator FR_TIX into the infrequent tail indicator LRU_TIX and writes the common indicator stored in the frequent head indicator FR_HIX. Enter the infrequent header indicator LRU_HIX, and delete the common indicator in the frequent tail indicator FR_TIX and the infrequent tail indicator LRU_TIX. In another embodiment, the controller 160 may also write the common indicator stored in the frequent header indicator FR_HIX into the non-frequent header indicator LRU_HIX, and write a specific value different from the common indicator into the frequent tail indicator FR_TIX and non- Frequent tail indicator LRU_TIX. In addition, the controller 160 is further configured to update the sequence table SE_TB and the reverse sequence table RSE_TB according to the infrequent indicator group and the frequent indicator group.

In the embodiment of the initial state in FIG. 11, when the controller 160 receives a read command or a write command, the controller 160 finds out whether the read command or the write command includes the read command or A first data link relationship set of a first correspondence relationship of a page specified by the write command. For example, the correspondence between the logical address and the physical address of the page indicated by the read or write command is stored in the data link relationship set TS_2, and the controller 160 is based on the received read or write command. Learn the designated page and find the corresponding data link relationship set TS_2. Next, the controller 160 determines whether the data link relationship set TS_2 has been loaded into the prefetch area CA according to whether a set index corresponding to the first data link relationship set TS_2 exists in the prefetch area correspondence table SMR_TB. Similar to the embodiment in FIG. 4, for example, the set index of the data link relationship set TS_2 is 0xAA. As shown in FIG. 11, if no pre-fetch area CA is written with any data link relationship set, the corresponding columns SMR_0 to SMR_N of all sections in the pre-fetch area correspondence table SMR_TB are 0xFFFF. Therefore, in this embodiment, the controller 160 determines that the set index 0xAA of the data link relationship set TS_2 does not exist in the prefetched region correspondence table SMR_TB. In other words, the controller 160 determines that the data link relationship set TS_2 does not exist in the prefetch area CA according to the set index 0xAA of the data link relationship set TS_2 that does not exist in the prefetch area correspondence table SMR_TB. Next, similar to the embodiment of FIG. 4, the controller 160 selects a blank sector S_N after judging that the prefetch area CA has a blank sector S_0 ~ S_N, and links the data in the flash memory 180 The relation table TB1 reads the data link relation set TS_2 to load the data link relation set TS_2 into the selected blank section S_N. Next, the controller 160 updates the prefetch area correspondence table SMR_TB, the order table SE_TB, the reverse order table RSE_TB, the infrequent indicator group, and the frequent indicator group to record the current state of the prefetch area CA. As shown in FIG. 12, the controller 160 writes the set index 0xAA of the data link relationship set TS_2 into the section corresponding column SMR_N corresponding to the section S_N. It is worth noting that since the data link relationship set TS_2 is newly loaded data, the data link relationship set TS_2 belongs to the infrequent data link relationship set. Therefore, the controller 160 defines both the infrequent head indicator LRU_HIX and the infrequent tail indicator LRU_TIX as the common indicator CIX_N of the section S_N. It is worth noting that, in this embodiment, the sequence column SE_N and the reverse column RSE_N of the sequence table SE_TB and the reverse sequence table RSE_TB maintain a specific value of 0xFFFF, and the frequent head indicator FR_HIX and the frequent tail indicator FR_TIX are maintained. The original state. Finally, the controller 160 reads the data link relationship set TS_2 stored in the prefetch area CA to execute the received write command and read command. It is worth noting that, since the sequence column SE_N and the reverse sequence column RSE_N and the section S_N with the infrequent data link relationship set TS_2 have the same common index CIX_N. Therefore, in this embodiment, the order column SE_N belongs to the infrequent order column, and the reverse order column RSE_N belongs to the infrequent order column.

Next, in the embodiment of FIG. 12, when the controller 160 receives a read command or a write command, the controller 160 similarly finds out a read command or a write command including the read command. A data link relationship set corresponding to one of the pages specified by the command or write command. For example, in this embodiment, a data link relationship set corresponding to one of the pages specified by the read command or the write command is TS_8. Next, the controller 160 determines whether the data link relationship set TS_8 has been loaded into the prefetch area CA according to whether a set index corresponding to the data link relationship set TS_8 exists in the prefetch area correspondence table SMR_TB. For example, the set index of the data link relationship set TS_8 is 0xBB. As shown in FIG. 12, the prefetch area CA is only written into the data link relationship set TS_2, and only the section corresponding column SMR_N in the prefetch area correspondence table SMR_TB is 0xAA, and the others are 0xFFFF. Similar to the embodiment in FIG. 5, the controller 160 determines that the set index 0xBB of the data link relationship set TS_8 does not exist in the prefetch area correspondence table SMR_TB, and selects a blank section S_N-1. Then, the controller 160 reads the data link relationship set TS_8 from the data link relationship table TB1 in the flash memory 180 to load the data link relationship set TS_8 into the selected blank section S_N-1. Next, the controller 160 updates the prefetch area correspondence table SMR_TB, the sequence table SE_TB, the reverse sequence table RSE_TB, the infrequent indicator group, and the frequent indicator group to record the current state of the prefetch area CA. It is worth noting that because the data link relationship set TS_8 is newly loaded data, the data link relationship set TS_8 belongs to the infrequent data link relationship set. Therefore, as shown in FIG. 12, the controller 160 obtains a section S_N to which the last infrequent data link relationship set TS_2 read in the prefetch area CA according to the infrequent header index LRU_HIX, and changes the area The common index CIX_N of the sequence column SE_N corresponding to the segment S_N is written into the sequence column SE_N-1 corresponding to the current section S_N-1. Next, the controller 160 writes the common index CIX_N-1 corresponding to the reverse sequence column RSE_N-1 of the current sector S_N-1 into the reverse sequence column RSE_N corresponding to the sector S_N. Next, the controller 160 defines the infrequent head indicator LRU_HIX as the common indicator CIX_N-1 of the current section S_N-1, maintains the infrequent tail indicator LRU_TIX at the common indicator CIX_N, and maintains the frequency of the frequent head indicator FR_HIX and the frequent tail indicator FR_TIX. value. The updated prefetch area correspondence table SMR_TB, sequence table SE_TB, reverse sequence table RSE_TB, infrequent header indicator LRU_HIX, and infrequent tail indicator LRU_TIX are shown in FIG. 13. Finally, the controller 160 reads the data link relationship set TS_8 stored in the prefetch area CA to execute the received write command and read command. It is worth noting that, since the sequence column SE_N-1 and the reverse sequence column RSE_N-1 and the section S_N-1 having the infrequent data link relationship set TS_8 have the same common index CIX_N-1. Therefore, in this embodiment, the order column SE_N-1 belongs to the infrequent order column, and the reverse order column RSE_N-1 belongs to the infrequent order column.

Assuming that the controller 160 successively receives a write command and a read command corresponding to the new data link relationship set, the controller 160 repeats the above steps of converting from FIG. 12 to FIG. 13 and sequentially sets the data link relationship set. Write to S_N-2 ~ S_0, as shown in Figure 14. In other words, the controller 160 repeatedly executes the loading procedure and fills all the prefetch areas CA. It is worth noting that in this embodiment, since the initial value of the sequence table SE_TB and the reverse sequence table RSE_TB is a specific value (0xFFFF), there is no need to write the specific value into the corresponding section in the loading program. The reverse order column and the order column corresponding to the most recently unread section in the prefetch area CA. In other embodiments, if the sequence table SE_TB and the reverse sequence table RSE_TB are different from the specified specific value, the controller 160 needs to write the specific value to the corresponding inverse of the current section in the last step of the loading procedure. The sequence column and the sequence column corresponding to the least recently read segment in the prefetch area CA. It is worth noting that in the above process, since no data link relationship set that has been uploaded to the prefetch area CA needs to be read, the frequent head indicator FR_HIX and the frequent tail indicator FR_TIX remain unchanged. In addition, in the embodiment of FIG. 14, all the order columns SE_0 ~ SE_N belong to the infrequent order column, and all the reverse order columns RSE_0 to RSE_N belong to the infrequent order column. For the method of frequent data link collection, please refer to the following description. In addition, regarding the update procedure of the infrequent data link set, the infrequent head indicator LRU_HIX can be regarded as the head indicator HIX, and the infrequent tail indicator LRU_TIX can be regarded as the tail indicator TIX. Refer to the description of Figs. The indicator FR_HIX and the frequent tail indicator FR_TIX remain unchanged.

The following is a description of the update procedure. In the embodiment of FIG. 14, when the controller 160 receives a read command or a write command, the controller 160 similarly finds out whether the read command or the write command includes the read command or the write command. A data link relationship set corresponding to one of the pages specified by the write command. For example, in this embodiment, a data link relationship set corresponding to one of the pages specified by the read command or the write command is TS_12. Next, the controller 160 determines whether the data link relationship set TS_12 has been loaded into the prefetch area CA according to whether a set index 0xABC corresponding to the data link relationship set TS_12 exists in the prefetch area correspondence table SMR_TB. In this embodiment, as shown in FIG. 14, the set index 0xABC of the data link relationship set TS_12 already exists in the region corresponding column SMR_3 in the prefetched region corresponding table SMR_TB. Therefore, in this embodiment, the controller 160 determines that the set index 0xABC of the data link relationship set TS_12 exists in the prefetched region correspondence table SMR_TB. In other words, the controller 160 judges that the data of the data link relationship set TS_12 already exists in the prefetch area CA according to the set index 0xABC of the data link relationship set TS_12 in the prefetch area correspondence table SMR_TB, which has the data link relationship set Section S_3 of TS_12 is the current section. Due to the data link relationship set TS_12 has been uploaded to the prefetch area CA. In other words, the number of times the data link relationship set TS_12 in the prefetch area CA has been read has reached the predetermined value "1". Therefore, in this embodiment, the data link relationship set TS_12 belongs to the frequency data link relationship set. Next, the controller 160 updates the prefetch area correspondence table SMR_TB, the sequence table SE_TB, the reverse sequence table RSE_TB, the non-frequency indicator group, and the frequency indicator group to record the current state of the prefetch area CA. As shown in FIG. 14, the controller 160 reads the sequence column SE_3 of the current sector S_3 to obtain a sector corresponding to a sector written before the current sector S_3 as the sector S_4. Next, the controller 160 writes the shared index CIX_2 stored in the reverse order column RSE_3 corresponding to the current section S_3 into the reverse order column RSE_4 corresponding to the section S_4. In other words, the controller 160 points originally to the reverse-order column RSE_4 of the reverse-order column RSE_3, and instead points to the reverse-order column RSE_2. Next, the controller 160 reads the value in the reverse order table RSE_TB corresponding to the current section S_3 in the reverse order column RSE_3 according to the common index CIX_3 of the current section S_3, and finds the value in the reverse order column RSE_3. The common index CIX_2 of the next sector S_2 read after the current sector S_3. Next, the controller 160 writes the common index CIX_4 stored in the sequence column SE_2 corresponding to the current sector S_2 into the sequence column SE_2 corresponding to the sector S_2. In other words, the controller 160 points originally to the sequence column SE_2 of the sequence column SE_3, and instead points to the sequence column SE_4. It is worth noting that by changing the sequence column SE_2 originally pointing to the sequence column SE_3 to the sequence column SE_4 and the sequence column RSE_4 originally pointing to the reverse sequence column RSE_3 to the reverse sequence column RSE_2, the sequence column SE_3 and the reverse sequence Column RSE_3 has been removed from the infrequent system. In other words, the order of reading the infrequent data link relationship set recorded in the sequence table SE_TB and the reverse sequence table RSE_TB does not include the data link relationship set TS_12 corresponding to the sequence column SE_3 and the reverse sequence column RSE_3. Next, the controller 160 writes the specific value (0xFFFF) into the sequence column SE_3 having the same common index CIX_3 as the current section S_3 as the start of the frequent sequence column in the sequence table SE_TB, and writes the specific value (0xFFFF) The reverse sequence column RSE_3, which has the same common index CIX_3 as the current section S_3, is used as the start of the frequent sequence column in the sequence table RSE_TB. Finally, the controller 160 defines the frequent head indicator FR_HIX as the common indicator CIX_3 of the current section S_3, and defines the frequent tail indicator FR_TIX as the common indicator CIX_3 of the current section S_3 and maintains the value of the non-frequent indicator group. The updated prefetch area correspondence table SMR_TB, sequence table SE_TB, reverse sequence table RSE_TB, frequent indicator group and infrequent indicator group, as shown in FIG. 15. Next, the controller 160 reads the corresponding segment S_3 in the prefetch area CA according to the common index CIX_3 of the segment corresponding column SMR_3 for storing the set index 0xABC to obtain the data link relationship set TS_12. In other words, the controller 160 reads the data link relationship set TS_12 stored in the prefetch area CA to execute the received write command and read command. It is worth noting that, because the sequence column SE_3 and the reverse sequence column RSE_3 have the same common index CIX_3 as the section S_3 with the infrequent data link relationship set TS_12. Therefore, in this embodiment, the order column SE_3 belongs to the infrequent order column, and the reverse order column RSE_3 belongs to the infrequent order column.

Next, in the embodiment of FIG. 15, when the controller 160 receives a read command or a write command, the controller 160 similarly finds out that the read includes a read command according to the received read command or write command. A data link relationship set corresponding to one of the pages specified by the command or write command. For example, in this embodiment, a data link relationship set corresponding to one of the pages specified by the read command or the write command is TS_10. Next, the controller 160 determines whether the data link relationship set TS_10 has been loaded into the prefetch area CA according to whether a set index 0xCC corresponding to the data link relationship set TS_10 exists in the prefetch area correspondence table SMR_TB. In this embodiment, as shown in FIG. 15, the set index 0xCC of the data link relationship set TS_10 already exists in the region corresponding column SMR_N-2 in the prefetch region corresponding table SMR_TB. Therefore, in this embodiment, the controller 160 determines that the set index 0xCC of the data link relationship set TS_10 exists in the prefetched region correspondence table SMR_TB. In other words, the controller 160 judges that the data of the data link relationship set TS_10 already exists in the prefetch area CA according to the set index 0xCC of the data link relationship set TS_10 in the prefetch area correspondence table SMR_TB, and the data link relationship set The section S_N-2 of TS_10 is the current section. The collection TS_10 has been uploaded to the prefetch area CA due to the data link relationship. In other words, the number of times the data link relationship set TS_10 in the prefetch area CA has been read has reached the predetermined value "1". Therefore, in this embodiment, the data link relationship set TS_10 belongs to the frequency data link relationship set. Next, the controller 160 updates the prefetch area correspondence table SMR_TB, the sequence table SE_TB, the reverse sequence table RSE_TB, the non-frequency indicator group, and the frequency indicator group to record the current state of the prefetch area CA. As shown in FIG. 15, the controller 160 reads the sequence column SE_N-2 of the current section S_N-2 to obtain a section corresponding to the one written before the current section S_N-2 as the section S_N-1. . Next, the controller 160 writes the shared index CIX_N-3 stored in the reverse order column RSE_N-2 corresponding to the current section S_N-2 into the reverse order column RSE_N-1 corresponding to the section S_N-1. In other words, the controller 160 points originally to the reverse-order column RSE_N-1 of the reverse-order column RSE_N-2, and instead points to the reverse-order column RSE_N-3. Next, the controller 160 reads the value in the reverse order column RSE_N-2 in the reverse order table RSE_TB corresponding to the current sector S_N-2 according to the common index CIX_N-2 of the current sector S_N-2, and according to the reverse order column The common indicator CIX_N-3 in RSE_N-2 finds the common indicator CIX_N-3 of the sector S_N-3 that is read next after the current sector S_N-2. Next, the controller 160 writes the shared index CIX_N-1 stored in the sequence column SE_N-2 corresponding to the current sector S_N-2 into the sequence column SE_N-3 corresponding to the sector S_N-3. In other words, the controller 160 points the sequence column SE_N-3, which is originally the sequence column SE_N-2, to the sequence column SE_N-1. It is worth noting that by changing the original order column SE_N-3 to the order column SE_N-3 to the order column SE_N-1 and the original order column RSE_N-2 to the reverse order column RSE_N-1 to the reverse order column The steps of RSE_N-3, the sequence column SE_N-2 and the reverse sequence column RSE_N-2 have been removed from the infrequent system. In other words, the order of reading the infrequent data link relationship sets recorded in the sequence table SE_TB and the reverse sequence table RSE_TB does not include the data link relationship set TS_12 corresponding to the sequence column SE_3 and the reverse sequence column RSE_3. Next, the controller 160 obtains a segment S_3 to which the last frequently read data link relationship set TS_12 belongs in the prefetch area CA according to the frequent header index FR_HIX, and sets the sequence column SE_3 corresponding to the segment S_3. The common indicator CIX_3 is written in the sequence column SE_N-2 corresponding to the current section S_N-2. Next, the controller 160 obtains a segment S_3 to which the frequent data link relationship set TS_12 belongs which has not been read for the longest time in the prefetch area CA according to the frequent tail index FR_TIX, and will correspond to the current segment S_N-2. The common index CIX_N-2 of the reverse sequence column RSE_N-2 is written in the reverse sequence column RSE_3 corresponding to the sector S_3. It is worth noting that in this embodiment, there is only one frequent data link relationship TS_12 in the prefetch area CA, so the frequent data link relationship TS_12 is the last and most frequently read data link relationship that has not been read for a long time. Finally, the controller 160 defines the frequent head indicator FR_HIX as the common indicator CIX_N-2 of the current section S_N-2, maintains the frequent tail indicator TIX at the common indicator CIX_3, and writes a specific value into the current section S_N-2. Corresponding reverse order column RSE_N-2. In addition, the controller 160 also maintains the values in the infrequent indicator group. The updated prefetch area correspondence table SMR_TB, sequence table SE_TB, reverse sequence table RSE_TB, frequent indicator group and infrequent indicator group, as shown in FIG. 16. Next, the controller 160 reads the corresponding segment S_N-2 in the prefetch area CA according to the common index CIX_N-2 of the segment corresponding column SMR_N-2 for storing the collection index 0xCC to obtain the data link relationship set TS_10. In other words, the controller 160 reads the data link relationship set TS_10 stored in the prefetch area CA to execute the received write command and read command. It is worth noting that since the sequence column SE_N-2 and the reverse sequence column RSE_N-2 and the section S_N-2 with the infrequent data link relationship set TS_10 have the same common index CIX_N-2. Therefore, in this embodiment, the order column SE_N-2 belongs to the infrequent order column, and the reverse order column RSE_N-2 belongs to the infrequent order column.

Next, in the embodiment of FIG. 16, when the controller 160 receives a read command or a write command, the controller 160 similarly finds out that the read includes a read command according to the received read command or write command. A data link relationship set corresponding to one of the pages specified by the command or write command. For example, in this embodiment, a data link relationship set corresponding to one of the pages specified by the read command or the write command is TS_77. Next, the controller 160 determines whether the data link relationship set TS_77 has been loaded into the prefetch area CA according to whether a set index 0x333 corresponding to the data link relationship set TS_77 exists in the prefetch area correspondence table SMR_TB. In this embodiment, it is assumed that the prefetch area CA does not have data of the data link relationship set TS_77, so the set index 0x333 does not exist in any of the area corresponding columns SMR_0 ~ SMR_N in the prefetch area corresponding table SMR_TB. The controller 160 determines that there is no data link relationship set TS_77 in the prefetch area CA according to the set index 0x333 of the data link relationship set TS_77 that does not exist in the prefetch area correspondence table SMR_TB. Next, the controller 160 determines whether the prefetch area CA has a blank sector. As shown in FIG. 16, the sections S_0 ~ S_N of the prefetch area CA are not blank. Next, the controller 160 determines whether the infrequent tail indicator LRU_TIX is any one of the common indicators CIX_0 to CIX_N. As shown in Figure 16, the non-frequent tail indicator LRU_TIX is the common indicator CIX_N. Next, the controller 160 judges the infrequent data link relationship set that has not been read for the longest time in the prefetch area CA according to the infrequent tail index LRU_TIX, so as to write the data of the data link relationship set TS_77 into the longest time in the prefetch area CA The section to which the infrequent data link relationship set being read belongs. As shown in FIG. 16, the controller 160 determines the infrequent data link relationship set in the prefetch area CA that has not been read for the longest time is the data link relationship set TS_2 according to the common index CIX_N in the infrequent tail index LRU_TIX, The segment having the infrequent data link relationship set that has not been read for the longest time according to the common index CIX_N is the segment S_N. Next, the controller 160 reads the data link relationship set TS_77 from the data link relationship table TB1 in the flash memory 180 to load the data link relationship set TS_77 into the obtained section S_N to replace the original long-term The read infrequent data link relationship set TS_2. Next, the controller 160 updates the prefetch area correspondence table SMR_TB, the order table SE_TB, the reverse order table RSE_TB, the frequent indicator group, and the infrequent indicator group to record the current state of the prefetch area CA. First, the controller 160 writes a set index 0x333 corresponding to the data link relationship set TS_77 into a section corresponding column SMR_N corresponding to the section S_N in the prefetch area corresponding table SMR_TB. As shown in FIG. 16, the controller 160 obtains a segment S_0 to which the infrequent data link relationship set TS_1 last read in the prefetch area CA according to the infrequent header index LRU_HIX, and changes the segment S_0 The corresponding common index CIX_0 of the sequence column SE_0 is written in the sequence column SE_N corresponding to the current sector S_N. Next, the controller 160 writes the common index CIX_N corresponding to the reverse column RSE_N of the current sector S_N into the reverse sequence column RSE_0 corresponding to the sector S_0. Next, the controller 160 reads the reverse order column RSE_N of the current section S_N to obtain the common index CIX_N-1 of the next section S_N-1 that is read after the current section S_N. Next, the controller 160 defines the infrequent head indicator LRU_HIX as the common indicator CIX_N of the current sector S_N, and the infrequent tail index LRU_TIX as the common of the next sector S_N-1 that is read after the current sector S_N The indicator CIX_N-1, and the value of the frequent indicator group is maintained. Finally, the controller 160 writes the specific value into the sequence column SE_N-1 corresponding to the next read section S_N-1 after the current section S_N, and writes the specific value into the corresponding section S_N. In the reverse order column RSE_N. The updated prefetch area correspondence table SMR_TB, sequence table SE_TB, reverse sequence table RSE_TB, infrequent indicator group and frequent indicator group are shown in FIG. 17. Finally, the controller 160 reads the data link relationship set TS_77 stored in the prefetch area CA to execute the received write command and read command.

Assume that the controller 160 successively receives the write command and the read command corresponding to the frequent data link relationship set that has been uploaded to the prefetch area CA. The controller 160 repeats the steps of the above-mentioned conversion from FIG. 15 to FIG. Update the sequence table SE_TB, the reverse sequence table RSE_TB, and the frequent indicator group until all the data in the prefetch area CA are frequent data link relationship sets, as shown in FIG. 18. It is worth noting that when the last non-frequent data link relationship set TS_8 is left in the pre-fetch area CA, the order columns SE_N-1 and the counter-listing columns RSE_N-1 corresponding to the infrequent data link relationship set TS_8 will both Is a specific value (0xFFFF). In other words, when the controller 160 reads the infrequent data link relationship set TS_8 and finds that the corresponding sequence column SE_N-1 and the reverse column RSE_N-1 are both specific values (0xFFFF), the controller 160 judges The infrequent data link relationship set TS_8, which is about to be converted into a frequent data link relationship set, is the last infrequent data link relationship set. Therefore, the controller 160 deletes the common indicators in the infrequent head indicator LRU_HIX and the infrequent tail indicator LRU_TIX after the corresponding data of the data link relationship set TS_8 is updated to the frequent data system, as shown in FIG. 18.

Next, in the embodiment of FIG. 18, when the controller 160 receives a read command or a write command, the controller 160 similarly finds out a read command or a write command including the read command. A data link relationship set corresponding to one of the pages specified by the command or write command. For example, in this embodiment, a data link relationship set corresponding to one of the pages specified by the read command or the write command is TS_88. Next, the controller 160 determines whether the data link relationship set TS_88 has been loaded into the prefetch area CA according to whether a set index 0x444 corresponding to the data link relationship set TS_88 exists in the prefetch area correspondence table SMR_TB. In this embodiment, the prefetch area CA does not have data of the data link relationship set TS_88, so the set index 0x444 does not exist in any of the area corresponding columns SMR_0 ~ SMR_N in the prefetch area corresponding table SMR_TB. The controller 160 determines that there is no data link relationship set TS_88 in the prefetch area CA according to the set index 0x444 of the data link relationship set TS_88 that does not exist in the prefetch area correspondence table SMR_TB. Next, the controller 160 determines whether the prefetch area CA has a blank sector. As shown in FIG. 18, the sections S_0 ~ S_N of the prefetch area CA are not blank. Next, the controller 160 determines whether the infrequent tail indicator LRU_TIX is any one of the common indicators CIX_0 to CIX_N. As shown in Figure 18, the infrequent tail indicator LRU_TIX is blank or a preset value. Therefore, the controller 160 writes the common index CIX_N-1 in the frequent tail index FR_TIX into the non-frequent tail index LRU_TIX, writes the common index CIX_N-2 in the frequent head index FR_HIX into the frequent tail index LRU_HIX, and deletes the frequent tail The common indicators in the indicator FR_TIX and the frequent head indicator FR_HIX are shown in Figure 19. Next, the controller 160 judges the infrequent data link relationship set that has not been read for the longest time in the prefetch area CA according to the infrequent tail index LRU_TIX, so as to write the data of the data link relationship set TS_88 into the longest time in the prefetch area CA. The section to which the infrequent data link relationship set being read belongs. As shown in FIG. 19, the controller 160 determines the infrequent data link relationship set in the prefetch area CA that has not been read for the longest time according to the common index CIX_N-1 in the infrequent tail index LRU_TIX as the data link relationship set TS_8 , And according to the common index CIX_N-1, the section having the infrequent data link relationship set that has not been read the longest is the section S_N-1. Then, the controller 160 reads the data link relationship set TS_88 from the data link relationship table TB1 in the flash memory 180 to load the data link relationship set TS_88 into the obtained section S_N-1 to replace the original section S_N-1. The infrequent data link relationship set TS_8 that has not been read for a long time. Next, the controller 160 updates the prefetch area correspondence table SMR_TB, the order table SE_TB, the reverse order table RSE_TB, the frequent indicator group, and the infrequent indicator group to record the current state of the prefetch area CA. For detailed steps, reference may be made to the embodiments in FIG. 16 and FIG. 17, and details are not described herein again. The updated prefetch area correspondence table SMR_TB, sequence table SE_TB, reverse sequence table RSE_TB, infrequent indicator group and frequent indicator group are shown in FIG. 20. Finally, the controller 160 reads the data link relationship set TS_88 stored in the prefetch area CA to execute the received write command and read command.

FIG. 21 is a flowchart of an embodiment of a data maintenance method provided by the present invention. The data maintenance method shown in FIG. 21 is applicable to the data storage device 140 shown in FIG. 1. The process starts at step S2100.

In step S2100, after the flash memory 160 is powered on, the controller 160 loads one of the prefetch areas CA in the random memory 166 according to the received at least one read command or at least one write command. Part of a plurality of data link relationship sets TS_0 ~ TS_N in a data link relationship table TB1.

Next, in step S2102, the controller 160 establishes a frequency indicator group and a non-frequency indicator group in the random memory 166 according to the read order and number of times of the loaded data link relationship set to maintain the pre- Take the information in the area CA. In detail, the controller 160 establishes an infrequent index group and a frequent index group in the random access memory 166, wherein the infrequent index group is used to maintain an infrequent data link relationship set, and the frequent index group is used to maintain A collection of frequently linked data links. In one embodiment, the infrequent indicator group is composed of an infrequent header indicator LRU_HIX and an infrequent tail indicator LRU_TIX, and the frequent indicator group is composed of a frequent header indicator FR_HIX and a frequent tail indicator FR_TIX. The infrequent header index LRU_HIX is a shared index for storing the last read infrequent data link relationship set segment to point to the last read infrequent data link relationship set in the prefetch area CA. The infrequent tail indicator LRU_TIX is a shared indicator for storing the segments of the infrequent data link relationship set that has not been read for the longest time, and points to the infrequent data link relationship set that has been read for the longest time in the prefetch area CA. The frequent header index FR_HIX is a shared index used to store the section of the last frequently read data link relationship set read to point to the last frequently read data link relationship set read in the prefetch area CA. The frequent tail index FR_TIX is a shared index used to store the segments of the frequent data link relationship set that have not been read for the longest time, so as to point to the most frequently read data link relationship set in the prefetch area CA.

In another embodiment, in step S2102, the controller 160 further establishes a prefetched region correspondence table SMR_TB, a reverse sequence table RSE_TB, and a sequence table SE_TB according to the loaded data link relationship set. In detail, the controller 160 establishes a prefetch area correspondence table SMR_TB in the random memory 166 to record a set index corresponding to a data link relationship set in the prefetch area CA. The controller 160 establishes a reverse order table RSE_TB in the random memory 166 to record the reverse order of the frequent data link relationship set read from the pre-fetch area CA and the non-frequent data link relationship set self-fetch area. The order in which the CA is read is reversed. It is worth noting that in this embodiment, the reverse order columns RSE_0 ~ RSE_N in the reverse order table RSE_TB can record two orders (the reverse order of the frequent data link relationship set read from the prefetch area CA and The set of infrequent data link relationships is read in reverse order from the domain fetch area). Similarly, the controller 160 establishes a sequence table SE_TB in the random memory 166 to record the order in which the frequent data link relationship set is read from the prefetch area CA and the infrequent data link relationship set from the domain fetch area. The order in which they are read. When the prefetch area CA is full and a new data link relationship set needs to be loaded, one of the segments S_0 ~ S_N in the prefetch area CA is selected according to the infrequent tail index LRU_TIX, so that the new The data link relationship set loads the selected section. After the new above-mentioned data link relationship set is loaded into the selected above-mentioned section or the old data link relationship set is read, the controller 160 may according to the infrequent indicator group and / or the frequent indicator group and the reverse order table RSE_TB The data in the table and the data in the sequence table SE_TB update the table and indicators in the random memory to record the data status in the prefetch area CA.

22A to 22H are flowcharts of an embodiment of a data maintenance method provided by the present invention. The data maintenance method shown in FIGS. 22A to 22H is applicable to the data storage device 140 shown in FIG. 1. The process starts at step S2200. Steps S2200 ~ S2212, S2260 ~ S2262, S2266 ~ S2272, S2276 ~ S2282 are similar to steps S1000 ~ S10112, S1060 ~ S1062, S1066 ~ S1072, S1076 ~ S1082, please refer to the description in Figure 10, More details.

In step S2214, the controller 160 simultaneously defines an infrequent head indicator LRU_HIX and an infrequent tail indicator LRU_TIX as one of the first common indicators in the first area. For example, in the embodiment of FIG. 12, the data link relationship set TS_2 is the first data link relationship set written in the prefetch area CA, where the data link relationship set TS_2 is written into the section S_N. Therefore, in step S2214, the controller 160 defines both the infrequent head indicator LRU_HIX and the infrequent tail indicator LRU_TIX as the common indicator CIX_N of the section S_N. Then, the flow proceeds to step S2299.

In step S2218, the controller 160 obtains a second section to which the last infrequent data link relationship set read in the prefetch area CA belongs according to the infrequent header index LRU_HIX. It is worth noting that, in this embodiment, the last set of data link relationships read in the prefetch area CA refers to the last prefetch area CA that was read by the controller 160 from the prefetch area CA. It is used to perform the infrequent data link relationship set of the received read command or write command.

Next, in step S2220, the controller 160 writes the common index of the second sequence column corresponding to the second section in the sequence table SE_TB into the first sequence column corresponding to the first section. For example, in the embodiment of the fifth figure, the data link relationship set TS_8 is not the first data link relationship set written in the prefetch area CA, where the data link relationship set TS_8 is written into the area Segment S_N-1 (first segment). Before updating the table in the random access memory 166, the controller 160 obtains the last one read in the prefetch area CA according to the current infrequent header index LRU_HIX (as shown in FIG. 12) in step S1018. One of the segments S_N (second segment) to which the infrequent data link relationship set TS_2 belongs. Next, in step S2220, the controller 160 writes the common index CIX_N of the sequence column SE_N (second sequence column) corresponding to the section S_N (second section) into the current section S_N-1 (the second section) A section) of the sequence column SE_N-1 (the first sequence is bad).

Next, in step S2222, the controller 160 writes the common index corresponding to the first reverse order column of the first section in the reverse order table RSE_TB into a second reverse order column corresponding to the second section. For example, in the embodiment of FIG. 13, in step S2222, the controller 160 sets the reverse order column RSE_N-1 (first reverse order column) corresponding to the current section S_N-1 (first section). The common index CIX_N-1 is written in the reverse sequence column RSE_N corresponding to the sector S_N (second sector).

Next, in step S2224, the controller 160 defines the infrequent header indicator LRU_HIX as a first common indicator of one of the first sections. For example, in the embodiment of FIG. 13, in step S2224, the controller 160 defines the infrequent header indicator LRU_HIX as the common indicator CIX_N-1 (the first section) of the current section S_N-1 (the first section). Common indicator), and the non-frequent tail indicator LRU_TIX is maintained at the common indicator CIX_N (second common indicator), wherein the frequent indicator group is also maintained at the original value. Then, the flow proceeds to step S1090.

In step S2229, the controller 160 determines whether the infrequent tail indicator LRU_TIX or the infrequent head indicator LRU_HIX is any one of the common indicators CIX_0 to CIX_N. When the infrequent tail indicator LRU_TIX or the infrequent head indicator LRU_HIX is one of the common indicators CIX_0 ~ CIX_N, the flow proceeds to step S2230; otherwise, the flow proceeds to step S2291.

In step S2230, the controller 160 determines, according to the infrequent tail index LRU_TIX, a third section to which the infrequent data link relationship set that has not been read for the longest time belongs. It is worth noting that in this embodiment, the set of infrequent data link relationships that have not been read for the longest time in the prefetch area CA refers to the current prefetch area CA that has not been self-fetched by the controller 160 for the longest time. A collection of infrequent data link relationships that are read by the CA for the received read command or write command.

Next, in step S2232, the controller 160 writes the data of the first data link relationship set into the third section. For example, in the embodiment of FIG. 17, a correspondence relationship of one of the pages specified by the read command or the write command is included in the data link relationship set TS_77 (the first data link relationship set). The controller 160 has determined in step S2204 that the set index 0x333 corresponding to the data link relationship set TS_77 does not exist in the prefetch area correspondence table SMR_TB, and the controller 160 has determined in step S2206 that there is no blank in the prefetch area CA Section. Therefore, in step S2230, the controller 160 judges the infrequent data link relationship set that has not been read for the longest time in the prefetch area CA as the data link relationship set according to the infrequent tail index LRU_TIX shown in FIG. 16 before the change. TS_2, and the section corresponding to the data link relationship set TS_2 is section S_N (third section). Therefore, in step S2232, the controller 160 reads the data link relationship set TS_77 (the first data link relationship set) from the data link relationship table TB1 in the flash memory 180 to set the data link relationship set TS_77. Load the obtained segment S_N (third segment) to replace the data link relationship set TS_2 that has not been read for a long time.

Next, in step S2234, the controller 160 writes the first set index corresponding to the first data link relationship set into the section corresponding column corresponding to the third section in the prefetch area corresponding table SMR_TB. In the embodiment of FIG. 17, the controller 160 then writes a set index 0x333 (first set index) corresponding to the data link relationship set TS_77 (the first data link relationship set) to the prefetch in step S2234. In the area correspondence table SMR_TB, a section correspondence column SMR_N corresponding to the section S_N (third section).

Next, in step S2236, the controller 160 obtains a fourth section to which the last infrequent data link relationship set read in the prefetch area CA belongs according to the infrequent header index LRU_HIX.

Next, in step S2238, the controller 160 writes the common index in the sequence table SE_TB corresponding to one of the fourth sections in the fourth order column into the third table in the order table SE_TB corresponding to one of the third sections. . For example, in the embodiment of FIG. 17, in step S2236, the controller 160 obtains the last data chain read in the prefetch area CA according to the infrequent header index LRU_HIX shown in FIG. 16 before the change. The tie relationship set TS_0 belongs to the section S_2 (fourth section). Next, in step S1038, the controller 160 writes the common index CIX_2 of the sequence column SE_2 (the fourth sequence column) corresponding to the section S_2 (the fourth section) into the current section S_N (the third zone) Paragraph) sequence column SE_N (third sequence column).

Next, in step S2240, the controller 160 writes the common index corresponding to the third reverse order column in the reverse order table RSE_TB corresponding to one of the third sectors into the reverse order table RSE_TB corresponding to one of the fourth sectors Reverse order bar. For example, in the embodiment of FIG. 17, the controller 160 writes the common index CIX_N corresponding to the reverse sequence column RSE_N (third reverse column) of the current section S_N (third section) into the area The reverse order column RSE_2 (fourth reverse order column) corresponding to the segment S_2 (fourth section).

Next, in step S2242, the controller 160 reads the third reverse-order column to obtain a common indicator of a fifth reverse-order column RSE_0 ~ RSE_N in the reverse-order column RSE_0 ~ RSE_N. In other words, the controller 160 reads the third reverse sequence column corresponding to the currently written third sector to obtain a fifth sector that was originally read next to the current third sector.

Next, in step S2244, the controller 160 defines the infrequent tail indicator LRU_TIX as a fifth common indicator corresponding to one of the fifth reverse order columns and defines the infrequent head indicator LRU_HIX as a third common indicator in a third order column. . For example, in the embodiment of FIG. 17, the controller 160 reads the reverse order column RSE_N (third reverse order column) of the current section S_N (third section) in step S2244 to obtain the original The common index CIX_N-1 (the common index in the fifth reverse column) of the current sector S_N (the third sector) is read next to the sector S_N-1. Next, the controller 160 defines the infrequent head indicator LRU_HIX as the common indicator CIX_N of the current section S_N (third section), and defines the infrequent tail indicator LRU_TIX as being originally read next after the current section S_N The common index CIX_N-1 (the fifth common index) of the sector S_N-1 (the fifth sector).

Next, in step S2246, the controller 160 writes a specific value into the third reverse sequence column and the fifth sequence column corresponding to one of the fifth common indicators. In other words, the controller 160 writes the specific value into the reverse order column corresponding to the common indicator pointed to by the current head indicator, and writes the specific value into the sequence column corresponding to the common indicator pointed to by the current tail indicator. As shown in FIG. 17, the controller 160 writes a specific value into the reverse sequence column RSE_N and the common index CIX_N-1 (the fifth reverse sequence common index) corresponding to the current sector S_N (third sector). Sequence column SE_N-1 (the fifth sequence column). Then, the flow proceeds to step S2299.

In step S2259, the controller 160 determines whether the frequent head indicator FR_HIX or the frequent tail indicator FR_TIX is any one of the common indicators CIX_0 to CIX_N. When the frequent head indicator FR_HIX is one of the common indicators CIX_0 ~ CIX_N, the flow proceeds to step S2260; otherwise, the flow proceeds to step S2293.

Next, in step S2264, the controller 160 obtains one of the eighth sections to which the last frequent data link relationship set read in the prefetch area CA according to the frequent header index FR_HIX.

Next, in step S2265, the controller 160 determines whether the sector to which the data link relationship set that has not been read for the longest time in the prefetch area CA belongs to the sixth sector according to the frequent tail index FR_TIX. When the section to which the data link relationship set that has not been read for the longest time belongs to the sixth section, the flow proceeds to step S2280; otherwise, the flow proceeds to step S2266. In other words, the controller 160 determines in this step whether the currently set data link relationship set to be read is the data link relationship set that has not been read for the longest time in the prefetch area CA.

Next, in step S2274, the controller 160 defines the frequent head indicator FR_HIX as a sixth common indicator in one of the sixth sequence columns.

Next, in step S2284, the controller 160 defines the frequent head indicator FR_HIX as a sixth common indicator in a sixth sequence column, and defines the frequent tail indicator FR_TIX as a seventh common indicator in a seventh section.

Next, in step S2286, the controller 160 writes a specific value into the sixth reverse order column. In other words, the controller 160 writes the specific value into the reverse order column RSE_N corresponding to the common indicator CIX_N indicated by the current head indicator. Then, the flow proceeds to step S2299.

In step S2291, the controller 160 writes the common indicator stored in the frequent-tail indicator FR_TIX into the non-frequent-tail indicator LRU_TIX, and writes the shared indicator stored in the frequent-head indicator FR_HIX into the non-frequent-head indicator LRU_HIX.

Next, in step S2292, the controller 160 deletes the common indicator in the frequent tail indicator FR_TIX and the non-frequent tail indicator LRU_TIX. In another embodiment, the controller 160 may also write the common indicator stored in the frequent head indicator FR_HIX into the infrequent head indicator LRU_HIX, and then write the preset value into the frequent tail indicator FR_TIX and the infrequent tail indicator LRU_TIX. For example, in step S2229, the controller 160 determines that either the infrequent tail indicator LRU_TIX or the infrequent head indicator LRU_HIX is not any of the common indicators CIX_0 to CIX_N, as shown in FIG. 18. Next, the controller 160 writes the common indicator CIX_N-1 stored in the frequent tail indicator FR_TIX to the non-frequent tail indicator LRU_TIX and writes the common indicator CIX_N-2 stored in the frequent header indicator FR_HIX in steps S2291 ~ S2292. Enter the infrequent header indicator LRU_HIX, and delete the common indicator in the frequent tail indicator FR_TIX and the infrequent tail indicator LRU_TIX. Then, the flow proceeds to step S2230.

In step S2293, the controller 160 reads a sixth reverse order column corresponding to the sixth section storing one of the first data link relationship sets, so as to obtain the next data link after the first data link relationship set. Read the seventh sector. In other words, the controller 160 reads a sixth reverse sequence column corresponding to a sixth sector in a reverse sequence table RSE_TB to obtain a seventh sector that is next read after the sixth sector. For example, in the embodiment of FIG. 15, a correspondence relationship of one of the pages specified by the read command or the write command is included in the data link relationship set as TS_12. As shown in FIG. 14 before the change, the set index 0xABC of the data link relationship set TS_12 already exists in the region correspondence column SMR_3 in the prefetch region correspondence table SMR_TB. Therefore, in step S2204, the controller 160 has determined that the data link relationship set TS_12 has been loaded into the prefetch area CA according to the existence of the set index 0xABC corresponding to the data link relationship set TS_3 in the prefetch area correspondence table SMR_TB. Next, in step S2293, the controller 160 reads the reverse order column RSE_3 in the reverse order table RSE_TB corresponding to the current section S_3 (sixth section) according to the common index CIX_3 of the current section S_3 (sixth section) ( Value in the sixth reverse order column), and according to the common indicator CIX_2 in the reverse order column RSE_3 (sixth reverse order column), the next section read after the current section S_3 (sixth section) is Section S_2.

Next, in step S2294, the controller 160 writes the common indicator stored in the sixth sequence column corresponding to the sixth sector in the seventh sequence column corresponding to one of the seventh sectors in the sequence table SE_TB. For example, in the embodiment of FIG. 14, the controller 160 has obtained in step S1093 that the next sector read after the current sector S_3 (sixth sector) is sector S_2 (seventh sector) segment). Therefore, in step S1094, the controller 160 writes the sequence column corresponding to the current sector S_3 (sixth sector) in the sequence column SE_2 (seventh sequence column) corresponding to the sector S_2 (seventh sector). The shared indicator CIX_4 stored in SE_3 (sixth order column). In other words, the controller 160 points originally to the sequence column SE_2 (seventh sequence column) of the sequence column SE_3 (sixth sequence column), and points to the corresponding section S_4 that was read before the first data link relationship set. Sequence column SE_4.

Next, in step S2295, the controller 160 reads the sixth sequence column corresponding to the sixth sector to obtain a ninth common index corresponding to one of the ninth sectors written before the sixth sector. . In the embodiment of FIG. 15, the controller 160 reads the sequence column SE_3 (sixth sequence column) corresponding to the current section S_3 (sixth section) in step S2295 to obtain the sequence column S_4 corresponding to the current section S_3 (sixth section). (Sixth sector) The previously written sector is sector S_4 (ninth sector).

Next, in step S2296, the controller 160 writes the shared index stored in the sixth reverse order column into the ninth reverse order column corresponding to one of the ninth shared index. For example, in the embodiment of FIG. 15, the controller 160 in step S2295 corresponds to the share stored in the reverse order column RSE_3 (the sixth reverse order column) corresponding to the current section S_3 (the sixth section). The indicator CIX_2 is written in the reverse order column RSE_4 (the ninth reverse order column) corresponding to the section S_4 (the ninth section).

Next, in step S2297, the controller 160 writes a specific value into the sixth sequence column and the sixth reverse sequence column. For example, in the embodiment of FIG. 15, the controller 160 writes a specific value of 0xFFFF to the reverse order column RSE_3 (sixth reverse order column) corresponding to the current section S_3 (sixth section) in step S2297. And the sequence column SE_3 (the first and the third sequence columns).

Next, in step S2298, the controller 160 defines the frequent head indicator FR_HIX and the frequent tail indicator FR_TIX as one of the sixth common indicators in the sixth sequence column. For example, in the embodiment of FIG. 15, the controller 160 defines the frequent head indicator FR_HIX and the frequent tail indicator FR_TIX as the common indicator CIX_3 of the current section S_3 in step S2297. Then, the flow proceeds to step S2299.

In step S2299, the controller 160 reads the data in the first data link relationship set from the prefetch area CA to execute the write command or the read command received in step S2200. The process ends in step S2299.

The data storage device 140 and data maintenance method provided by the present invention can update the data in the prefetched area by using the index, reverse order table, sequence table, and a prefetched area correspondence table to record the status of the data in the prefetched area. Among them, the index, reverse order table, and order table can make the prefetched area correspondence table need to modify the number of data not more than five columns in each update process.

The method of the present invention, or a specific form or part thereof, may exist in the form of a code. The code can be stored in physical media, such as floppy disks, CD-ROMs, hard disks, or any other machine-readable (such as computer-readable) storage media, or is not limited to external forms of computer program products. Among them, When the code is loaded and executed by a machine, such as a computer, the machine becomes a device for participating in the invention. The code can also be transmitted through some transmission media, such as wire or cable, optical fiber, or any transmission type. Where the code is received, loaded, and executed by a machine, such as a computer, the machine becomes used to participate in the Invented device. When implemented in a general-purpose processing unit, the code in combination with the processing unit provides a unique device that operates similar to an application-specific logic circuit.

However, the above are only the preferred embodiments of the present invention. When the scope of implementation of the present invention cannot be limited by this, that is, the simple equivalent changes and modifications made according to the scope of the patent application and the description of the invention, All are still within the scope of the invention patent. In addition, any embodiment of the present invention or the scope of patent application does not need to achieve all the purposes or advantages or features disclosed by the invention. In addition, the abstract and the title are only used to assist the search of patent documents, and are not intended to limit the scope of rights of the present invention.

Claims (16)

A data storage device includes: a random access memory with a prefetch area; a flash memory including a plurality of pages, each of which has a logical address and a physical address, the flash memory The memory has the data link relationship table for recording the logical correspondences of the plurality of pages and the plural correspondences of the entity addresses. The data link relationship table is divided into the data link relationship sets. Each of the data link relationship sets has at least one of the corresponding relationships, and each of the data link relationship sets corresponds to a set index; and a controller is configured to link a data link relationship table into a data link relationship table. A part of the plurality of data link relationship sets is loaded into the plurality of sections in the prefetch area. Among the data link relationship sets, the data link relationship sets that have been read less than a predetermined value belong to the plural infrequent data links. A set of data link relationships that has been read for a predetermined number of times and belongs to a complex set of frequent data link relationships, where the controller establishes a non-frequent The target group maintains the infrequent data link relationship set, and establishes a frequent indicator group to maintain the frequent data link relationship set, wherein the infrequent indicator group is composed of an infrequent head indicator and an infrequent tail indicator. And the frequent indicator group is composed of a frequent head indicator and a frequent tail indicator, wherein the sections have a plurality of shared indicators in order, and the infrequent header indicator is used to store the last infrequent data read. The common indicator of the section of the link relationship set, the non-frequent tail indicator is used to store the common indicator of the section of the infrequent data link relationship set that has not been read for the longest, and the frequent header indicator is used to Store the shared indicator of the sector of the last frequent data link relationship set that was read last. The frequent tail indicator is used to store the sector of the frequent data link relationship set that has not been read for the longest time. Shared index, wherein the random access memory further includes a sequence table, a reverse sequence table, and a prefetched region correspondence table, wherein the prefetched region correspondence table has The corresponding column of the data section is used to record the set indicators corresponding to the data link relationship sets in the prefetch area, and the sequence table is used to record the frequent data link relationship sets from the prefetch area respectively. The order of reading and the order of reading the infrequent data link relationship set from the domain fetch area, and the reverse order table is used to record the frequent data link relationship set from the prefetch area, respectively The reversed order of reads and the reversed order of the collection of the infrequent data link relationships from the prefetched region. The data storage device according to item 1 of the scope of the patent application, wherein the corresponding columns of the sections in the prefetch area correspondence table correspond to the sections in the prefetch area in sequence, and the sequence table has a plurality of order columns according to The order corresponds to the sections in the prefetch area, the reverse order table has a plurality of reverse order columns corresponding to the sections in the prefetch area, and the section corresponding column, the order column, the The reverse order column has the same shared indicators as those corresponding sections. The data storage device according to item 2 of the scope of patent application, wherein the sequence columns having the same shared indicators with the sections to which these frequent data link relationship sets belong belong to the multiple frequent sequence columns, and the non- Frequent data link relationship sets The sections in which the sections belong to the same shared indicators belong to a plurality of non-frequent sequence columns, each of which is used to store another frequent sequence column The common indicator points to the frequent order column corresponding to the section of the frequent data link relationship set that was previously read before the section corresponding to the frequent order column, respectively. The frequent order column is used to store the common index of another infrequent order column to point to the section of the infrequent data link relationship set that was read before the section corresponding to the infrequent order column, respectively. The corresponding infrequent order column. The data storage device according to item 2 of the scope of patent application, wherein the reverse order columns that have the same shared indicators as the sections to which these frequent data link relationship sets belong are plural frequent reverse order columns, and the The non-frequent data link relationship set belongs to a plurality of non-frequent reverse order columns in which the segments belonging to the same set of shared indicators belong to a plurality of non-frequent reverse order columns, each of which is used to store another The common index of the frequent reverse order column respectively points to the frequent reverse order corresponding to the section of the frequent data link relationship set that is read next after the section corresponding to the frequent reverse order column. Column, each such non-frequent reverse order column is used to store the shared index of another infrequent reverse order column, respectively, to point to the next non-frequently read non-reverse order column after the section corresponding to the infrequent reverse order column The non-frequent reverse order column corresponding to the section of the frequent data link relationship set. The data storage device according to item 2 of the scope of patent application, wherein the controller is configured to find a first correspondence of one of the pages specified by the read command or the write command according to a read command or a write command. A first data link relationship set of the relationship, and according to whether there is a first set index corresponding to the first data link relationship set in the prefetch area correspondence table, determine whether the first data link relationship set has been Is loaded into the prefetch area, wherein the first data link relationship set is one of the data link relationship sets, the first set indicator is one of the set indicators, and the first The corresponding relationship is one of these corresponding relationships. The data storage device according to item 5 of the scope of patent application, wherein when there is an index corresponding to the first set in the prefetched area correspondence table, the controller is configured to store the corresponding column of the segment corresponding to the first set index. The shared indicator reads the corresponding section in the prefetch area to obtain the first data link relationship set, and reads the first data link relationship set stored in the random access memory. Data in order to perform the write command or the read command. The data storage device according to item 6 of the scope of patent application, wherein when the first set index does not exist in the prefetch area correspondence table, the controller determines whether the prefetch area has a blank section, and when the control When the processor determines that the segment of the prefetch area has a blank first segment, the controller reads the first data link relationship set from the data link relationship table in the flash memory, and The first data link relationship set loads the blank first section. The data storage device according to item 7 of the scope of patent application, wherein when the controller judges that the prefetch area does not have a blank section, the controller judges whether the infrequent tail indicator is any of the common indicators Wherein, when the infrequent tail indicator is any of the common indicators, the controller judges the infrequent data link relationship in the infrequent data link relationship set that has not been read for the longest time according to the infrequent tail indicator. A set to write the data of the first data link relationship set into the non-frequent data link relationship set that has not been read for the longest time and belongs to the section to which the infrequent data link relationship set belongs. The data storage device according to item 8 of the scope of patent application, wherein when the infrequent tail indicator is not any of the common indicators, the controller writes the common indicator stored in the frequent tail indicator into the infrequent tail. Indicator, write the common indicator stored in the frequent-head indicator into the infrequent-head indicator, and delete the frequent-tail indicator and the common indicators stored in the infrequent-tail indicator. According to the data storage device of the second patent application scope, the controller is further configured to update the sequence table and the reverse sequence table according to the infrequent indicator group and the frequent indicator group. A data maintenance method is applicable to a data storage device having a flash memory, wherein the flash memory includes a plurality of pages, each of which has a logical address and a physical address, and the data maintenance The method includes: when the flash memory is powered on, according to at least one read command or at least one write command, loading a data link into a plurality of sections of a prefetch area in a random memory. A plurality of data link relationship sets in the relationship table, each of which has a corresponding relationship between the logical address and the physical address of at least one of the pages, each of these The data link relationship set corresponds to a set of indicators. The data link relationship set in which the data link relationship set is read less than a predetermined value belongs to a plurality of infrequent data link relationship sets, and the data The data link relationship set in the link relationship set that is read to reach the predetermined value belongs to the plural frequent data link relationship set; establish a non-frequent index group to dimension The infrequent data link relationship set; establishing a frequent indicator group to maintain the frequent data link relationship set, wherein the infrequent indicator group is composed of an infrequent head indicator and an infrequent tail indicator, and the The frequent indicator group is composed of a frequent header indicator and a frequent tail indicator, in which the sections have a plurality of shared indicators in sequence, and the infrequent header indicator is a set of infrequent data link relationships used to store the last read The common indicator of the sector, the non-frequent tail indicator is used to store the common index of the sector of the infrequent data link relationship set that has not been read for the longest, and the frequent header indicator is used to store the last Read the common indicator of the section of the frequent data link relationship set, and the frequent tail indicator is used to store the common indicator of the section of the frequent data link relationship set that has not been read for the longest time; A prefetching area correspondence table, in which the corresponding fields in the plural section corresponding columns in the prefetching area correspondence table correspond to the data link relationship sets in the prefetching area Set the index of these; establish a sequence table to record the order in which the frequent data link relationship sets are read from the prefetch area and the infrequent data link relationship sets are read from the domain fetch area The order of reading; and establishing a reverse order table to record the reverse order of the frequent data link relationship sets being read from the prefetch area and the non-frequent data link relationship sets from the prefetch The order in which the regions are read is reversed. The data maintenance method according to item 11 of the scope of patent application, wherein the corresponding columns of the sections in the prefetched area correspondence table correspond to the sections in the prefetched area in sequence, and the order table has a plurality of order columns The order corresponds to the sections in the prefetch area, the reverse order table has a plurality of reverse order columns corresponding to the sections in the prefetch area, and the section corresponding column, the order column, the The reverse order column has the same shared indicators as those corresponding sections. The data maintenance method according to item 12 of the scope of patent application, wherein the sequence columns having the same shared indicators with the sections to which these frequent data link relationship sets belong are plural frequent sequence columns, and the non- Frequent data link relationship sets The sections in which the sections belong to the same shared indicators belong to a plurality of non-frequent sequence columns, each of which is used to store another frequent sequence column The common indicator points to the frequent order column corresponding to the section of the frequent data link relationship set that was previously read before the section corresponding to the frequent order column, respectively. The frequent order column is used to store the common index of another infrequent order column to point to the section of the infrequent data link relationship set that was read before the section corresponding to the infrequent order column, respectively. The corresponding infrequent order column. The data maintenance method according to item 12 of the scope of the patent application, wherein the reverse order columns having the same shared indicators with the sections to which the frequent data link relationship sets belong belong to the plural frequent reverse order columns, and the The non-frequent data link relationship set belongs to a plurality of non-frequent reverse order columns in which the segments belonging to the same set of shared indicators belong to a plurality of non-frequent reverse order columns, each of which is used to store another The common index of the frequent reverse order column respectively points to the frequent reverse order corresponding to the section of the frequent data link relationship set that is read next after the section corresponding to the frequent reverse order column. Column, each such non-frequent reverse order column is used to store the shared index of another infrequent reverse order column, respectively, to point to the next non-frequently read non-reverse order column after the section corresponding to the infrequent reverse order column The non-frequent reverse order column corresponding to the section of the frequent data link relationship set. The data maintenance method according to item 11 of the patent application scope further includes: when the prefetch area is full and a new collection of these data link relationships needs to be loaded, determining whether the infrequent tail indicator is the shared indicator Any one of them; when the infrequent tail indicator is any of the common indicators, the infrequent data link in the infrequent data link relationship set that has not been read for the longest time is judged according to the infrequent tail indicator A relationship set; and writing the data of the first data link relationship set into the infrequent data link relationship set that has not been read for the longest time in the section to which the infrequent data link relationship set belongs. The data maintenance method according to item 15 of the scope of patent application, further comprising: when the infrequent tail indicator is not any of the common indicators, writing the common indicator stored in the frequent tail indicator into the infrequent tail indicator , Write the common indicator stored in the frequent-head indicator into the infrequent-head indicator, and delete the frequent-tail indicator and the common indicators stored in the infrequent-tail indicator.