TW200837562A - Non-volatile memory and method for class-based update block replacement rules - Google Patents

Abstract

In a nonvolatile memory with block management system, data are written to blocks and are erasable block by block. At any time a pool of blocks are open for storing data concurrently. The number of blocks in the pool is limited. A replacement system allows new blocks to be introduced into the pool without exceeding the limit. In particular, different classes of blocks in the pool each has its own replacement rule, such as closing a least active block before being replaced. In this way, possible inefficiency and premature closure of blocks in the pool can be avoided.

Description

200837562 IX. DESCRIPTION OF THE INVENTION: TECHNICAL FIELD The present invention generally relates to a non-volatile semiconductor memory, specifically a memory block management system, which employs an improved system for managing simultaneous opening of data for storage. Replacement of block pools. [Prior Art] A solid state memory capable of storing charges nonvolatilely

In the form of a small form factor card EEPROM and flash EEPR〇M, it has become a storage device of choice for various mobile and handheld devices (especially information appliances and consumer electronics). Unlike the Ram (p access memory), which is also a solid-state memory, the flash memory system is non-volatile, so it can retain its stored data even after the power is turned off. In addition to the ROM (read-only memory) system, the flash memory can be rewritten in the same way as the disk storage device. Although the cost is higher, it is more and more common to use flash memory as a mass storage application. Traditional mass storage, which is based on rotating magnetic media, such as hard drives: and floppy disks, is not suitable for mobile and handheld environments. This is because the volume of disc piracy tends to be bulky, I is prone to mechanical failure, and the delay time and high power half require these annoying properties to make the second storage cannot be used. Most mobile and portable applications have a card, on the other hand, flash memory (whether intrusive or detachable = two:) because of its small size, low power consumption, fast speed and "characteristics," Therefore, it is very suitable for mobile and handheld rings 124919.doc 200837562 Flash EEPROM and EEPROM (electrically erasable programmable read-only memory) are similarly non-volatile memory that can be erased And new data can be written or "programmed" into its memory unit. Both are in (iv) the crystal structure using the channel region between the source region and the drain region in the middle of a + conductor. a floating (unconnected) conductive pole. A control gate is then provided on the floating gate. The threshold voltage characteristic of the transistor is controlled by the amount of charge remaining on the floating gate. The floating gate is set In terms of charge level, it is necessary to apply a -corresponding (critical) voltage to the control gate before the transistor is "on" to enable conduction between its source region and the poleless region. Specifically, flash ee A flash memory such as ^rom allows the memory cells of the entire block to be erased at the same time. The floating gate can retain the electrical power of a certain Van Gogh cymbal U target, so it can be programmed into a threshold voltage. The voltage level of the S product in the cup inside the window. The size of the threshold voltage window is determined by the most inconspicuous position of the device and the maximum critical level.

To define, the level of dependence on A T should be a range of charges that can be programmed on the floating gate. The prospective location /αί V ® Φ depends on the characteristics, operating conditions and history of the memory device. People 》, ^ 上上' In this window, each of the different, resolvable threshold voltage levels can be used to represent the unit's clear memory state. When 蚱w + 广记... is divided into two different areas, each mouth is enough to store one owing IA ^ / 兀. Similarly, when the threshold voltage is divided into two areas, the memory unit can store more than 70 pieces of data. % usually uses one of the following two mechanisms to program the transistor as memory unit 124919.doc 200837562 into a "programmed, state." in "thermal electron emission,", is applied to the bungee The high voltage accelerates the electrons across the substrate channel region. At the same time, the high voltage applied to the control gate pulls the hot electrons onto the floating gate through a thin gate dielectric. In the 'wearing out', a high voltage is applied to the substrate relative to the substrate. In this way, electrons can be pulled from the substrate to the intermediate floating gate. Although the term 'programming' is used in the past to describe the writing of the memory by exposing electrons to the memory unit in the beginning to be erased in order to change the state of the memory, At present, it can be exchanged with more common words such as π write or π record. The following devices can be used to erase the memory device. For EEPROM, by applying with respect to the control gate A high voltage is applied to the substrate such that electrons can be induced in the floating gate to pass through a thin oxide into the substrate channel region (ie, F〇wler_N〇rdheim tunneling effect). A memory unit. In general, EEpR〇M can be erased on a byte by byte basis. For flash EEpR〇M, all memory can be erased every time C is erased or erased every time. One or more minimum erasable blocks, wherein a minimum erasable block may be composed of more than one segment and each segment may store 512 bytes or more of data. Body devices typically include a device that can be mounted on a card Or more memory chips. Each memory chip contains an array of memory cells supported by peripheral circuits such as decoders and erase circuits, write circuits, and read circuits. More complex memory devices A controller that implements intelligent and higher-order memory operation and interface. 124919.doc 200837562 Many commercially available non-volatile solid-state memory devices are currently in use. These memory devices may flash EeprOM. Alternatively, other types of non-volatile memory cells can be used. Flash memory and systems are provided in U.S. Patent Nos. 5, 5, 5, 5, 095, 344, 5, 315, 541, 5, 343, 063, and 5, 661, 053, 5, 313, 421 and 6, 222, 762. Examples and methods of making the same. A flash memory device having a NAND string structure is described in U.S. Patent Nos. 5,570,315, 5,903,495, and 6,046,935. a memory cell of the dielectric layer to fabricate a non-volatile memory device that utilizes a dielectric layer in place of the conductive floating gate described above Components. Eitan et al., IEEE Electron Device Letters, Vol. 21, No. 11, pp. 543-545, NR0M: a Novel Localized Trapping, 2_Bit Nonvolatile Memory Cell" A memory device of such a dielectric storage element. An ON dielectric layer extends across the channel between the source and drain diffusion regions. The charge for one data bit is localized in the dielectric layer adjacent to the drain and the charge system I for the other data bit is localized in the dielectric layer adjacent to the source. For example, U.S. Patent Nos. 5,768,192 and 6,011,725 disclose a non-volatile memory cell having a trapping dielectric sandwiched between two ceria layers. Multiple state data storage can be achieved by separately reading the binary state of the charge storage region separated by the space within the dielectric. To improve read and program performance, multiple charge storage elements or memory transistors in an array can be read or programmed in parallel. Therefore, the "page" of the memory component can be read or programmed simultaneously. In the existing memory architecture, 124919.doc 200837562 A column usually contains several interlaced pages or it may constitute one of the pages. All memory components on a page will be read or programmed simultaneously. In the flash memory system, the time taken to erase the operation; the length may be: the multiple of the read and program operation. Therefore, we hope to have a very large erase block. In this way, the erase time can be spread among a large number of memory cell aggregates. The characteristics of the flash memory will conclude that the data must be written to an erased = memory location. If the material from the specific logical address of the host is to be updated, the way is to rewrite the updated data in the same physical memory location. That is, the logical-to-physical address mapping has not changed. However, this means that the entire erase block will contain the physical location that must be erased before rewriting the data with j. This update method is inefficient because it needs to erase and rewrite the entire erase block, which is especially inefficient if the data to be updated only occupies a small portion of the erase block: it will also The erasing cycle that leads to more frequent memory blocks is not desirable in terms of the limited durability of this type of device. Another problem with the official flash memory system is the need for program system control and catalog information. The bead material is generated and accessed during each memory operation process. Therefore, effective procedures and convenient access will directly affect the validity month b. We want to keep this type of data in flash memory because the flash memory system was originally stored and non-volatile. However, using the intermediate file management system between the controller and the flash memory, the data cannot be accessed directly. In addition, system control and catalogue data tend to be very active and segmented, which is not conducive to storage in systems with large-scale 124,919.doc 200837562 inch block erase. By convention, this type of data is set in the controller RAM, allowing the controller to access it directly. After the power is supplied to the memory device, an initialization program enables scanning of the flash memory to compile the necessary system control and directory information for placement in the controller RAM. This program takes time and requires controller RAM capacity, so it is necessary to further increase the flash memory capacity. US 6,567,307 discloses a method of updating a program segment in a large erasure block, 纟 including recording the update data in a plurality of erase blocks as a scratch padding and finally in a logical order After reconfiguration, the valid sections between the various blocks are merged and the areas are rewritten. In this way, it is not necessary to erase and rewrite a block each time a slight update is made. WO 03/027828 and WO 00/49488 disclose a memory system for updating a program in a large erase block, which includes dividing the logical segment addresses into a plurality of zones. It retains a small logical address range for active system control data that is separated from the other zone for user data. In this way, manipulating the system control data in its own zone will not interact with user data associated with another zone. The update is at the logical sector level and a write indicator points to the corresponding physical segment in the block to be written. The mapping information is buffered in RAM and is eventually stored in the extent configuration table in the main memory. The latest version of a logical section will discard all previous versions of the existing block that have become partially obsolete. The implementation of the abandoned project collection also maintains a receivable portion of the abandoned block. 124919.doc -10- 200837562 The system of W technology #r tends to spread the update data in many blocks; or 'this update data may cause many existing blocks to become partially abandoned. The result is usually for the blocks that are partially discarded: = abandon: the collection is 'very efficient and will cause the memory to be raised: old. In addition, compared to non-sequential updates, there is no systematic way to program updates. There is a memory. Therefore, it is necessary to have a high-capacity and high-efficiency non-volatile memory. It is necessary to have a south-capacity non-volatile memory that can be recorded in a large block without any problems. . SUMMARY OF THE INVENTION - A non-volatile memory system is organized into a plurality of entity groups of a plurality of physical memory locations. Each entity group (meta block) can be erased as a unit and can be used to store a logical f group. The story management system allows for updating a logical data group by configuring a zone dedicated to recording the updated information for the logical group. The update metablock records the update" according to the order of the collection, and does not restrict the record to be (sequential) or not (disorderly) and the original metablock is closed, so that the update metablock is closed. In the further = will: = in the program, where m will eventually produce the - completely filled metablock in the correct (four) order, instead of the original metablock. In the case of chaos, the directory data is kept in multiple logical groups in a way that is frequently updated. One of the features of the invention that has been updated at the same time allows for the updating of lean materials in a logical group by group. Therefore, when a logical group is to be updated, the distribution of logical units is 124919.doc 200837562 (Γ! The dispersion of memory units with new waste units) is limited to a certain & This is especially true when the logical group is normally contained within the -f block. During the update period of the logical group, one or two blocks must typically be assigned to buffer the updated logical units. Therefore, it is only necessary to implement the collection of abandoned projects in a relatively small number of blocks. The merged or compressed method can be used to collect the discarded items of the %-disorganized block. * The economic performance of the update program will become more apparent in the general procedures of the update blocks compared to the sequential blocks. Thus, there is no need to configure any additional blocks for clutter (non-sequential) updates. All update blocks are assigned a sequential update area, and any update block can (4) be a more messy update block. More specifically, the _update block can be arbitrarily changed from sequential to cluttered. Effective use of system resources allows multiple logical groups to be updated simultaneously. This will further increase efficiency and reduce the additional burden. In accordance with one aspect of the present invention, in a non-volatile memory employing a block management system, an improved block replacement scheme is implemented for a system that supports simultaneous opening of a first predetermined maximum number of updated blocks for recording data. The update block is mainly a sequential update block in which data is recorded in a logical sequential order, but the second predetermined maximum number of update blocks are allowed to be disorderly updated, wherein the data is not recorded in a logical sequential order. Whenever the new configuration of the update block causes the update block to exceed the first or second predetermined maximum number, one of the existing update blocks in the set will be closed and removed to meet the limit. Before closing the update block, its data is merged into sequential blocks. The improved scheme avoids 124919.doc -12· 200837562 The out-of-order update can lead to the merger of an additional number of messy blocks. This is accomplished by separating the sequential update blocks and the cluttered update blocks into individual replacement or combination sets. Specifically, when the sequential update causes the configuration of the new update block to exceed the first predetermined maximum number, the least recently used sequence update block of the set area preferentially moves the space.

In the current system, there are generally two types of data: user data and control data. User data is typically transferred from the host to the memory system in a logical sequential order. The sequential update block is configured to optimally program sequential write operations from the host. User data can also be in a logical, non-sequential order, especially when the logic is subsequently updated. The messy update block is built to best program data in a non-sequential order. Another source of clutter or non-sequential data is control data maintained by the file system or memory system, such as file and directory information generated during the process of storing user data. The previous scheme that meets the maximum system number while supporting the maximum system number while opening the update block is the least recently used update block in the closed zone, regardless of its order or clutter. This scenario improves on the previous scenario, where essentially the updated block in the pool during the sequential write operation needs to be closed to move space for the new configuration, then the least recently used sequential update block in the cluster is closed. This ensures a variety of more:, block effective geocentric program sequential writing and random writing operations. The specific sequential write operation that avoids host execution can force invalidation of premature shutdown of the messy update block containing AT and directory information. It is possible to effectively create another _disorder area & to store the attached directory computer ’ ― once the large sequential write operation is completed, it is updated again. Improvement ^ 124919.doc -13- 200837562 Generation of policy mandatory substitution and separation of merged pools to prevent additional additional burdens of merged hetero-Iblocks during sequential writes, and potential to open sequential or open messy blocks The merger to manage subsequent FAT and catalog information updates. The generalization of the scheme updates the blocks according to a set of attributes, such as updating the block to store sequential or non-sequential data, or storing system data of a predefined type. When implementing a pool of a limited number of update blocks, the update blocks of each level will have their own replacement rules when the maximum number of levels supported by the level is exceeded. For example, sequential update blocks and non-sequential update blocks are two different levels. The replacement rules for each of these levels are the same, that is, the new block is substituted for the lowest block of $t. @此# When the cluster of the sequential update block is exceeded, the lowest-function block in the collection area will be closed and removed before the new block is introduced into the collection area. The same is true for the episodes of non-sequential update blocks. Each level of melon and beta has its own alternative rules that are independent of the other levels. The rule of the alternative rule (4) replaces the least recent access, the near access multiple access, the infrequent access, the most frequent access, etc. according to the corresponding level. Additional features and advantages of the present invention will become apparent from the following description of the preferred embodiments of the invention. An example of a system of ghosts, which can implement the aspect of the invention. Similar memory system is already in

The title of "Non-Volatile Memory and Method with Control Data Management" is disclosed in Gorobets et al., "Non-Volatile Memory and Method with Control Data Management", which is disclosed in US Patent Application Publication No. US-2005-0144365-124919.doc -14-200837562 A1. ". BRIEF DESCRIPTION OF THE DRAWINGS Figure 1 is a schematic illustration of a main hardware assembly suitable for implementing the memory system of the present invention. The memory system 20 typically operates with a host 10 via a host interface. The memory system is typically in the form of a memory card or embedded memory system. The memory system 20 includes a memory 200 that is controlled by the controller 1 . Memory 200 includes one or more arrays of non-volatile memory cells distributed over one or more integrated circuit wafers. The controller 1 includes an interface 110, a programmer 120, an optional coprocessor 121, a ROM 122 (read only memory), a RAM 130 (random access memory), and an optional programmable non-volatile memory 124. . The interface 11 has another component that interfaces the controller to one of the components and to the memory 200. The firmware stored in the non-volatile ROM 122 and/or the optional non-volatile memory 124 provides code for the programmer 120 to implement the functions of the controller. The program error correction code can be programmed by the programmer 120 or the optional co-programmer 121. Instead of the specific embodiment, the controller 1 is implemented by a state machine (not shown). In another embodiment, the controller 100 can be implemented within a host. Logic and physical block structure, FIG. 2 illustrates a memory according to a preferred embodiment of the present invention, which is woven, and woven into a plurality of physical segment groups (or metablocks) and is controlled by the controller j hidden body "living to manage. The memory is organized into metablocks, in which each meta-region burns oxygen V 〇, the ghosts in order to erase the physical segments s〇, ..., Sn_i group 0 Μ When running an application under the conditions of the file system or the system, the main 124919.doc -15-200837562 machine 1 access memory 200. Generally speaking, the host system is in logical sector units. Addressing information, for example, each segment can contain 5 12 bytes of data. Similarly, the host typically reads or writes to the memory system in units of logical clusters, each logical cluster consisting of one or more logical segments. In some host systems, there may be an optional host side memory manager to perform lower level memory management in the host. In the case of a large number of eves during a read or write operation, host 1 〇 essentially sends a command to the memory system 2〇

A segment is fetched or written containing a string of logical segments of data having consecutive addresses. The memory side memory manager is implemented in the controller 1 of the memory system 20 to manage the storage and retrieval of the host logical sector data in the metablock of the flash memory 200. In a preferred embodiment, the memory manager package 3 is used to manage a plurality of software groups of erase, read, and write operations of the metablock. The memory manager also maintains system control and directory information associated with its operation in flash memory 2 and control time RAM 130. "(〇到3A(iii) shows forbearance, 丨王 description is very according to the present issue%. Pingyi Wang Kaijie I Court case is between - logical group and unary block. The physical memory block has A physical segment is used to store the logical segments of the data of the logical group. Figure 3 (1) shows the data from the logical group, and the logical blocks in the complex logical sequence are 〇, 卜..., (10). 3A(ii) shows the same data stored in the metablock in logical order. When several blocks are missed in this way, 'it' is treated as "sequentially, the metablock: Has been stored in a different order _, in this case, the metablock is treated as, non-sequential, or, chaotic,. 124919.doc -16 - 200837562 at the lowest address of a logical group There may be an offset between the lowest addresses of the mapped metablocks. In this case, the logical sector addresses are wound back into the loop from the bottom of the logical group within the metablock to the top. For example, Figure 3A (iii) In the first position, the metablock begins to store the information of the logical block k. When the logical block N-丨, it wraps around to the segment 〇, and finally stores the data associated with the logical segment k-1 in its last physical segment. In a preferred embodiment, the page label is used. Identifying any offset, such as identifying the starting logical sector address of the data stored in the first physical segment of the metablock. When the two blocks differ by only one page label, the two blocks are considered Having a logical section 0 stored in the same order Figure 3B schematically illustrates a mapping between a plurality of logical groups and a plurality of metablocks. Each logical group is mapped - a unique metablock, wherein Except for a few logical groups that currently update the data. After updating - logical: group, the group can be mapped to a different metablock. The mapping information is maintained in a set of logical to physical directories, such The catalogue will be explained in more detail below. Other types of logical group-to-meta-block mappings are also contemplated. For example, 'Alan Sinclair's US patent application jointly owned by the same invention as the present invention, 俨月Tea' shouted for Adapt In tive Metablocks, the meta-blocks with variable sizes are revealed. The entire disclosure of the shared pending case is incorporated herein by reference. The split operation, and the same, the logical section of the entire logical address range of the body system 124919.doc -17- 200837562 group. For example, the distribution system can be distributed anywhere in the logical address space The section of the data and the section containing the user data. Unlike the prior art system, there is no special branch or bar of the system section (ie, the section related to the file configuration table, directory or subdirectory). In order to localize in the logical address space it is possible to include segments of data with high frequency and small size updates. Conversely, this scheme of updating the logical group of segments will effectively map the typical access patterns of the system segments to the typical access patterns of the archives. Figure 4 illustrates the alignment of a unitary block in physical memory and various structures. The flash memory includes a block of memory cells that can be erased together as a unit. This erase block is the minimum erase unit of the flash memory or the minimum erasable unit (MEU) of the memory. The minimum erase unit is the memory design parameter of the memory. Although in some memory systems that support multiple MEU erase, you can configure a “super MEU” that includes more than one MEU. For flash EEPROM, A MEU may comprise a segment, but preferably comprises a plurality of segments. In the illustrated example, it has M segments. In a preferred embodiment, each segment may store 512 bytes. The data has a user data portion and a header portion for storing the system or management data. If the metablock is composed of P MEUs and each MEU includes one district 4 and then the parent metablock will There are μ segments. The 兀 block represents a group of memory locations at the system level, such as a segment that can be erased together. The physical address space of the flash memory is treated as a set of metablocks, one dollar The block is the minimum erasing unit. Within this specification, the term metablock is synonymous with the block and is used to define the minimum erasing unit at the level of the media management for media management 124919.doc 200837562, and the term "Minimum erase unit MEU is used to indicate flash flash The minimum erase unit of the memory. $ Connect multiple minimum erase units (MEU) to form the - metablock to maximize the stylized speed and erase speed. Multiple information pages in multiple MEUs are configured to be side-by-side programmed, and multiple MEUs are configured to be erased side by side. Γ \ In flash memory, one page can be used in a single operation. A group of a memory cells. A page may include multiple segments. Similarly, an array of memory cells may be segmented into more than one plane, where each plane can be erased or erased in a plane. There is only one MEU. Finally, the plane can be distributed among one or more memory chips. In the flash memory, the MEU can include one or more pages. A MEU organization within a flash memory can be recorded. In the plane, since one MEU from each plane can be programmed at the same time, it is advantageous to form a plurality of MEU metablocks by selecting one MEU from each plane (see Figure 5B below). 5 A shows the meta-area formed by the minimum erasing unit connecting different planes Each metablock (eg, ΜΒ0, MB1, ...) is composed of MEUs of different planes from the memory system, in which different planes can be distributed among one or more wafers. The metablock chain shown in Figure 2. The way manager 170 manages the links of the MEUs for each metablock. Each metablock is configured during the initial formatter and maintains its constituent MEU throughout the life of the system, unless one of the MEUs is present Figure 5B illustrates a specific embodiment in which a minimum erase unit (MEU) is selected from each plane for concatenation into a unitary block. 124919.doc -19- 200837562 Figure 5C illustrates where each will Another embodiment in which more than one meu is selected in the plane for joining into a unitary block. In another embodiment, more than one MEU can be selected from each plane to form a super MEU. For example, two MEUs can be used to form a super meu. In this case, more than one pass may be required for a read or write operation. The linking and reconnection of the MEU into a metablock is also disclosed in the co-pending and co-owned U.S. patent application entitled "Adapted by Deterministic Grouping of Blocks int" Multi_m〇ck (7), by Gonzales et al. The person applying on the same day as the present application. The entire disclosure of the application pending is hereby incorporated by reference herein. And the unintentional block diagram in the flash memory. The metablock management system includes various functional modules implemented in the control M 100 and various control materials (including catalog materials)

The graphs and lists are distributed in a hierarchical manner in the flash memory 200 and the controller RAM 130. The functional modules implemented in the controller 1 include an interface module 110, a logical-to-physical address translation module 14A, an update block manager module 150, an erase block manager module 16 and a Block link manager 170. The ι面110 enables the metablock management system to interface with a host system. The logical-to-physical address translation module 14 maps the logical address from the host to a physical memory location. The update block manager module 15 manages the data update operation for a given logical data group in the memory. The erase block management module 16 官 administers the erase operations of the metablocks and manages their configuration to store the new 124919.doc -20- 200837562 information. The metablock block joins the crying & the minimum erasable block of the 170th administrative section: the group is connected to form a given metablock. These modules will be described in detail in their respective sections. i During operation, the metablock management system generates and uses control data (such as address, control, and status information). Since most of the control data tends to change small-sized data frequently, it cannot be easily stored and efficiently maintained in a fast-paced structure: itself. Use a hierarchical and distributed approach to store more static control data in non-volatile flash memory while positioning a smaller number of more variable control data in the controller RAM for more efficient Update and access. In the event of a power failure or failure, the control data in the volatile controller RAM of the scheme 2 can be quickly reconstructed by scanning a small set of control data in the non-volatile memory. This is possible because the present invention limits the number of blocks associated with a possible activity of a given logical group of one of the materials. In this way, scanning can be limited. In addition, some of the control data that needs to be persisted is stored in a non-volatile metablock that can be updated on a sector-by-segment basis, with each update generating a new segment that is recorded, which replaces a previous segment. A segment indexing scheme can be used for control data to track updates on a sector-by-segment basis in a metablock. The non-volatile flash memory 200 stores a large amount of control data with relative statics. This includes a group address table (GAT) 210, a hash block index (CBI) 220, an erase block list (EBL) 230, and a MAP 240. GAT 210 tracks the mapping between logical groups of blocks and their corresponding metablocks. The mapping does not change except for the updated map. CBI 220 Tracking Update Period 124919.doc -21- 200837562 Mapping of logical non-sequential sections. The EBL 230 tracks the pool of metablocks that have been erased. MAP 240 is a one-bit image that shows the erased state of all metablocks in the flash memory. Volatile controller RAM 130 stores a small portion of the control data that changes frequently and is accessed. This includes a configuration block list (ABL) 134 and a clear block list (CBL) i36. ABL 134 tracks the configuration of the block used to record the updated data, while CBL 136 tracks the metablocks that have been deconfigured and erased. In the preferred embodiment, RAM 130 acts as one of the cache memories for the control data stored in flash memory. Update Block Manager The Update Block Manager 15 (shown in Figure 2) updates the program logic group. In accordance with one aspect of the invention, a dedicated update metablock is configured for each logical group of segments undergoing updating to record updated material. In a preferred embodiment, any segment of one or more of the logical groups is recorded in the update block. An update block can be managed to receive updated data in sequential or non-sequential (also known as cluttered) order. A cluttered update block enables the segment data to be updated in any order within a logical group and in any repetition of individual segments. In particular, a sequential update block can become a cluttered update block without the need to locate any data segments. The predetermined configuration of the blocks for messy data updates is not required; it automatically adapts to non-sequential writes at any logical address. Therefore, unlike prior art systems, it is not specifically handled whether the updated segments of the logical group are in a logical or non-sequential order. The general update block is simply used to record various segments in the order in which the pieces are requested by the host. For example, even if the host system 124919.doc -22 - 200837562 material or system control feed tends to be updated in a messy manner, it is not necessary to process the area corresponding to the host system data differently than the area with the host user data. The area of the logical address space. In contrast, the data of the complete logical group of the segment is stored in a single 7L block in a logical sequential order. In this way, the index of the stored logical section can be predefined. When a 7L block stores all segments of a given logical group in a predefined order, it is called, complete ". As for an update block, when it is filled with the updated data in the most logical order, then the update block will be replaced by an updated block, which can easily replace the original block. On the other hand, if the update block is filled with update data in an order that is logically different from the order of the complete block, the update block is a non-sequential or messy update area: and the program sequence must be further reversed in order to ultimately Preferably, the update data of the logical group is stored in the same order as the complete block, which is preferably in a logical sequential order in the single metablock. The further processing involves merging the update section in the update block with the unchanged block in the original block into another update block. The merged update block will then be in logical sequential order and can be used to replace the original block. Under certain pre-conditions, the merged program is followed by one or more compression programs. The dust reduction process simply re-records the segments of the messy update block in an alternate milk update block while eliminating any duplicate logical segments that have been discarded by subsequent updates of the same logical segment. This update allows multiple update lines to run simultaneously until a predefined maximum is reached. Each update line is a logical group that undergoes an update using its dedicated update metablock. 124919.doc -23- 200837562 Sequential Data Update When the data belonging to a logical group is first updated, the one-summer area is 并^ and ^ is used as the update block of the update data of the logical group. ......,, A ^ When the field host receives a command, it configures the update block to write a logical group, one or more areas & one of the pieces, where an existing meta block is already storing all its complete District ί again. For a host write operation, the data i ^ B T + your brother a segment record '1 on the update block. Because & each host writes a segment of one or more segments with consecutive logical addresses, following the first update is always sequential. In subsequent host writes, updated segments within the same logical group are recorded in the update block in the order received from the host. A block continues to be = a progressive update block, while the logical sequence is maintained by the segments updated by the hosts in the associated logical group. All segments updated in this logical group are written to this sequential update block until the block is closed or converted to a cluttered update block. 7A illustrates an example of a plurality of segments in a logical group in a sequential update block that have been subjected to two separate host write operations in a sequential order, the original block in the logical group. The corresponding section will become obsolete. In the host write operation #; [in, update the data in the logical segment LS5 to LS8. The updated data (such as LS5' to LS8) is recorded in the newly configured dedicated update block. For convenience, the first segment to be updated in the logical group is recorded in a dedicated update block starting at the location of the first physical segment. In general, the first logical segment to be updated does not have to be the logical first segment of the group, so there may be 124919.doc -24-200837562 between the beginning of the logical group and the starting point of the updated block. shift. This offset is referred to as the page label as previously described in conjunction with Figure 3A. The subsequent sections are updated in logical sequential order. When the last segment of the logical group is written, the group address is wrapped and the write sequence continues with the first segment of the group. In host write operation #2, a segment of the data in the logical sections LS9 to LS12 is updated. The updated data (such as LS9, to LSI 2) is recorded in a dedicated update block directly at the end of the last write. It can be seen that the situation of two host writes has been recorded in the update block in the logical sequential order (ie LS5, -LS 12,). The update block is treated as a sequential update block 'because it has been populated in a logical sequential order. The update data recorded in the update block discards the corresponding data in the original block. Cluttered Data Updates When any segment updated by a host within an associated logical group has logical non-sequentiality, it is possible to initiate messy update block management for an existing sequential update block. A clutter update block is in the form of one of the data update blocks, wherein the logical segments within an associated logical V group can be updated in any order and in any number of iterations. When a segment written by a host is logically non-sequential to a previously written segment within the updated logical group, it is established by a conversion from a sequential update block. All of the regions # that are subsequently updated in this logical group are again written to the next available segment location in the hash update block, regardless of its logical sector address within the group. 7B illustrates an example of a plurality of sectors among a logical group in a hashed update block that are subscribed to five separate host write operations in a hashed order, while the original block of the logical group The section that has been replaced and the section that has been copied in the 124919.doc -25-200837562 messy update block becomes obsolete. In the host write operation #1, a given logical group stored in an original metablock is updated, and the LUNs are further LS10 to LSI1. The updated logical section lS1〇 is stored in the newly configured update block to the LS 11. At this time, the t new block is a sequential update block. In host write operation #2, the logical section ls5_LS6 is updated to LS5'-LS6 and recorded in the update block with the last written address. This converts the update block from a sequential update block to a messy update block. In the host write operation #3, the logical sector LSI0 is updated again and recorded in a position below the update block as Lsi,. At this time, the LSI 〇 '' in the update block replaces the Lsl 先前 in the previous record, which in turn replaces the LS 10 in the original block. In the host write operation #4, the data in the logical section L S1 0 is updated again and recorded in a position below the update block as LS10. Therefore, LSI 0'" is now the latest material and is the only valid material for logical segment LS 10. In the host write operation, the data in the logical section LS30 is updated and recorded in the update block as . Thus, the example illustrates that segments within a logical group can be written into a cluttered update block in any order and in any repetition. Forced sequential update Figure 8 illustrates an example of performing a separate host write operation with discontinuous logical addresses and sequentially writing the plurality of logical segments in the logical group in a sequential order. In host write #1, the update data in logical segment LS5-LS8 is recorded in a dedicated update block with LS5, _LS8. In host write #2, the update lean in the logical section lsi4_lsi6 is recorded as LSI4, -LS 16, which is recorded in the last update area 124919.doc -26-200837562. However, there is an address jump between LS8 and 1^14. However and the host writes

Apply the fill operation (#2A). In this way, the crack is merged into the update block to preserve the sequence characteristics of the update block. Figure 9 is a flow diagram illustrating the process of updating a logical data group by the update block manager in accordance with a general embodiment of the present invention. The update program includes the following steps: Step 260: The memory is organized into a plurality of blocks, each of which is divided into a plurality of memory units that can be erased together, and each memory unit is used for storage. A logical data unit. Step 262 · The data is organized into a plurality of logical groups, each of which is divided into a plurality of logical units. Step 264. In the case of the edge standard, all logical units of a logical group are stored in the memory units of a primitive block according to the first specified order 'better logical order. . In this way, the index used to access the individual logical units in the block can be known. Step 270: For a given logical data group (e.g., lgx), a request can be made to update a logical unit within the LGX. (Logical unit updates are given as an example. In general, 'this update will be a fragment of one or more consecutive logical units within LGx.) Step 272: The requested update logic unit will be stored in a In the second block, the block is dedicated to recording the updates of LGx. The record 124919.doc -27- 200837562 is based on the second order of the invention, and generally the order of the updates is required. It is commonly used to record data in a manner that initially sets an update block to depend on the =j order or clutter order. So, block. / brother ~ order, the second block may be a sequential block or messy area step 274 · · complex - " σ a block successor with the program returns to step 270 to record the request of the building unit. Go to φ field appears scheduled to close When the condition is met, the second block is closed to receive the update. In this case, the program proceeds to step 276: 〜~# a, 疋~The second block that has been closed is The same block as the original block | ^ ^ # 11 records its update logical unit. When the two block record logics are only 沬i differing by one page label, the two blocks are considered Having the same chuan g skin, 'one, order, as shown in Figure 3A. If the two blocks have the same order, the program will proceed to step 280, otherwise it must be in step 290. Implementing some sort of obsolete item collection. Step 280: Because the second block has an order similar to the first block, σ is used to replace the original first block of 5 hai. Then the update process ends at step 299. Step 290: From the second block (update block) and the first block (original Collecting the latest version of each logical unit of the given logical group, then writing the merged logical units of the given logical group in an order similar to the first block to In a third block, step 292·· because the third block (merged block) has a sequence similar to the first block, which can be used to replace the original first block. Then the more is 124919. Doc -28- 200837562 The new program will end at step 299. Step 299: When the ci〇seout program creates a complete update block, it will become the new standard block for the given logical group. The update thread of the logical group will then end. 囷10 is a flowchart illustrating a procedure for updating a logical data group by the update block manager in accordance with a preferred embodiment of the present invention. The method includes the following steps: Step 310: For a given logical data group (for example, a logical segment within the LGX may be required to be updated. (The sector update is given as an example. Generally, the update will be Is a continuous logic in LGX Fragment of the segment.) Step 312: If there is no update block of a dedicated KLGx, proceed to step 410 to start a new update thread for the logical group. This can be configured by a dedicated record. The update block of the update data of the logical group is completed. If an update block has been opened, proceed to step 314 to record the update segment in the update block. Step 314: Suppose the current update area If the block is already cluttered (ie, non-sequential), then the process proceeds to step 510 to record the requested update segment in the hash update block. If the current update block is sequential, proceed to step 316. Used to process a sequential update block. Step 316: One of the features of the present invention is to allow an update block to be initially set to be commonly used to record data in a logical sequential or cluttered manner. However, because the logical group will eventually store its data in a logically sequential order in the unary block, it is desirable to keep the 124919.doc -29-200837562 update block as possible. Then, when an update block is closed for further updates, fewer programs will be required because no obsolete item collection is required. Therefore, it is possible to determine whether the requested update will follow the sequential order of the updated block. If the update is followed sequentially, then step 510 is taken to implement a sequential update' and the update block will remain in order. On the other hand, if the update is not followed (scurry update), then if it is not taken f

The action H will convert the sequential update block into a messy update block. In the particular embodiment, no action is taken to save the situation, and the sequence will go directly to step 370 where the update is allowed to convert the updated block into a messy update block. The optional forced sequence program, in the specific example, will be implemented as much as possible according to the upcoming messy--forced sequence step step 32g, in order to retain the new block. There are two cases, both of which require a section that is lost from the original block to maintain the sequential order of the logical segments recorded in the updated block. The first type of 孙7孙访语虹么* ^ θ The system will establish a short-range address hop. In the case of the younger brother, the block is updated, and the block is updated to keep it in sequence. (4) The forced sequence procedure 320 includes the following sub-steps: (4) 330: If the update establishes a value that is not greater than the predetermined amount Cb If the logical address jumps, the program is bound to r, (4) will enter the mandatory sequential update private sequence in step 350, and if not (4), it will be called the 34 () test case completion. Ancient" has a mandatory cycle 124919.doc -30 - 200837562 Step 340: If the number of unfilled physical segments exceeds a predetermined design parameter Cc (typically typically half the size of the updated block), then the update The block is rarely used and will not be closed early. The program proceeds to step 370 and the update block will become cluttered. On the other hand, if a relatively large portion of the update block is filled, then It is considered to have been fully utilized, so step 360 can be entered to force the sequence to be closed. Step 350: Force sequential update to keep the current sequential update block 'sequential, as long as the address does not jump more than a predetermined amount CB Basically, the section in the associated original block of the update block is copied to fill the gap spanned by the address hop. Therefore, the data in the intermediate address will be utilized before proceeding to step 510. To fill the sequential update block, in order to record the current update sequentially. Step 360: If the current sequential update block has been mostly filled, the forced sequence will be closed. The current sequential update block, instead of #^ let the upcoming messy update convert it into a messy update block. The definition of "chaotic" or non-sequential update is an update with the following items: no exception for the above address jump Covered forward address transition, reverse address transition, or address repetition. To avoid a random update block being converted by a messy update, by copying the associated original portion of the updated block in the abandoned block The segment is filled with the location of the segment that is not written in the update block. Then, the original block is discarded and can be erased. Now, the current update block will have a set of logical segments that are fully loaded. Then it can be closed as a complete meta-ghost to replace the original meta-block. Then the program will proceed to steps 124919.doc -31 - 200837562 430 to configure a new update block in the correct location. A record to accept the upcoming segment update requested in step 310. Converting to a messy update block Step 370: When the upcoming update is not in a sequential order and is not satisfied as appropriate When the program is forced to enter the step 510, the upcoming update section (which has a non-sequential address) can be recorded in the update block for the sequence. The update block is converted into a cluttered update block. If there is a maximum number of cluttered update blocks, the least recently accessed messy update block must be closed before the conversion is allowed to prevent the maximum number of cluttered update blocks from being exceeded. The manner of identifying the least recently accessed messy update block is the same as the general case described in step 42, except that it is limited to a cluttered update block. At this point, the merge operation described in step 550 can be used to achieve the close. The purpose of a messy update block. Configuring a new update block under system condition constraints. Step 410: The process of configuring a wipe partition to an update block begins with a decision that exceeds a predetermined system limit. For the relationship between limited resources, the memory management system usually allows the maximum number of update blocks UMAX to exist at the same time. This limit is the sum of the sequential update block and the cluttered update block' and the system-item design parameters. In the preferred embodiment, for example, t Xu is limited to a maximum of 8 update blocks. In addition, due to the system resources, the relationship with higher demand, the maximum number of messy update blocks that can be simultaneously opened will also have corresponding predetermined limits (for example, four). This has been configured to UMAX update blocks, and it is only possible to meet the requirements of the next configuration after shutting down one of the existing configured update blocks 124919.doc -32- 200837562. The process continues to step 420. When the number of open update blocks is less than CA, the program will proceed to step 430. Step 420: If the predetermined maximum number of update blocks cA is exceeded, the least recently accessed update block is closed and the abandoned item collection is implemented. The least recently accessed update block is determined to be the update block associated with the least recently accessed logical block. The purpose of accessing the block that contains the least recently accessed block is to include the write and optionally read logical segments. A list of open update blocks can be saved in the order of access; at initialization, it is assumed that there is no access sequence. When the update block is sequential, the closing of an update block will follow the same procedure as that described in step 36 and step 53; when the update block is messy, it will follow and step 54 The same procedure as described above. This close operation creates space for the new update block configuration in step 430. Step 430: Configure a new meta zone to hold your inflammation#

The spitting block can satisfy this configuration requirement as an update block specific to the given logical group LGX. ^ ^ Receiver, the program will proceed to step 510. The update data is recorded in the update block. Step 510: The requested update segment merges C records in the next available physical location in the update block. Next, the current #广 a sequence will proceed to step 520 to determine whether the update block can be closed. The update block settlement space can accept additional updates, and step S22' closes the update area. Step 520: If the update block still has to proceed to step 570. Otherwise, it will enter 124919.doc 33 200837562 Dragon broadcast: When the write request requires more writes than the block = (4) space, there are two possible implementations to fill the update area. In an implementation, the write request can be divided into two parts, where the first knives are written until the last physical segment of the block. The block is then closed' and the second portion of the write will be treated as the next one, written in the implementation, retaining the required write, and the block will fill the remaining The section is then closed. This request write will be treated as the next request write.

Step 522: If the update block is looped, it proceeds to step 530 to order the sequence to close. If the update block is messy, then the state is called the state to make a messy shutdown. Step-by-Step Update Block Closure Step 530 • Since the update block is sequential and fully populated, the logical group stored therein is complete. The metablock is complete and replaces the original metablock. At this point, the original block is completely discarded and can be erased. The program then proceeds to step 570 to end the update thread for the given logical group. The messy update block is closed. Step 540: Because the update block is non-sequentially populated, it may contain a logical section of the update portion multiple times, so the waste project collection is implemented to save the valid data. The messy update block will be compressed or merged. In step 542, it will be decided which program to implement. Step 542 · Implementing compression or merging will depend on the aging situation of the updated block. If a logical section is updated several times, its logical address will be highly aging 124919.doc -34- 200837562. Multiple versions of the same logical segment will be recorded in the update block, but only the last recorded version is a valid version of the logical segment. In an update block containing logical sections with multiple versions, the number of different logical segments will be much less than the number of logical groups. In the preferred embodiment, when the number of different logical segments in the update block exceeds a predetermined design parameter CD (which is typically half the size of a logical group), the settlement procedure will occur. The merge is implemented in step 55, otherwise the program will perform compression in step 56. r Step 55: If the messy update block is to be merged, a new standard metablock containing the merged data is used to replace the original block with the updated block. After the merge, the update thread will end in step 570. Step 560: If the messy update block is to be compressed, it will be replaced by a new update block carrying the compressed data. After the second C is shortened, the compressed update block program will be terminated in step 570. Alternatively, it may be postponed until the update block is again compressed by £^K, thereby removing the possibility that the compression operation is followed by a merge operation without an intermediate update. When the next request in step 5〇2 is to be updated in the LGX, then the new update block can be utilized to further update the given logical block. Step 570: When the settlement program establishes a complete update block, it becomes the new standard block for the given logical group. The update thread for this logical group will then end. When the settlement program creates a new update block to replace an existing update block, the new update block can be used to record 124919.doc •35- 200837562 for the given logical group, gray, Wenyong Next update. When an update block has not been closed, the program will continue when the step 310 shows that the next request is to be updated in the LGX. From the above procedure, Cangshan can be released. ¥ When closing a messy update block, the updated information recorded in f will be further processed. Specifically, the valid data will be collected through the obsolete project, may be collected into another messy block, or the original block associated with it will be merged to form a new standard sequence. Block.囷11A is a more detailed flow chart for closing the merge procedure of the messy update block shown in FIG. The messy update block merge is tied to one of the two possible programs implemented by the update block (eg, when the update block is written because its last physical segment location is written and the king P is filled) . The merging is selected when the number of different logical segments in the written block exceeds the predetermined design parameter Cd. The merge procedure step 550 shown in Figure 10 includes the following substeps: 51. When the hybrid IL update block is closed, a replaceable new metablock is configured. Step 552: & the latest version of the logical segment of the hashed update block and its associated original block, ignoring all discarded segments. 'Step 54 · Record the collected valid areas in the logical sequence in the "Hai Xinyuan block to form a complete block, that is, to form a logical group with a sequential record A block of all of these logical segments of the group. Step S56: whai new complete block to replace the original block. 124919.doc -36- 200837562 Step 558: Erasing the closed update block and the original block. Figure 11B is a more detailed flow chart illustrating the compression procedure for closing the cluttered update block shown in Figure 1. Compression is selected when the number of different logical segments written into the block is below the predetermined design parameter CD. The compression procedure step 560 shown in Figure 1A includes the following sub-steps: Step 561 - When a hashed update block is to be compressed, a new metablock is replaced. Step 562: Collect the latest version of each logical segment in the existing messy update block to be compressed. Step 564: Record the collected segments in the new update block to form a new update block having the compressed segments. Step 566: The new (five) update block with the compressed segment replaces the existing update block. Step 568: Erasing the updated block that has been closed. Logic and Metablock Status Figure 12A illustrates all possible states of a logical group and possible transitions between the states under various operations.囷12Β is - a table listing the possible states of a logical group. The definitions of these logical group states are as follows: 1. Complete. : All logical segments in this logical group are written in a single-70 block in a logical sequential order 'which may utilize page label winding. 2. Unwritten: No logical segments have been written in this logical group. The logical group marks a group of address blocks that are not written and have no configured metablocks. It will respond to host reads for each segment in the group. 124919.doc •37- 200837562 Send back a predefined profile. 3. Sequential update: The logical group bed addresses are f, the inner knife & the segment has been logically sequential in the singular block. 'It may use the page mark, so J, 曰 replace any of the 6 HM groups Corresponding logical block of the previous complete state 0 4········································································ ,

^. The Hierarchy equation replaces the corresponding logical block of any preceding complete state in the group. The segments in the group may be written more than once - and the latest version will replace all previous versions.囷UA describes all possible states of a unitary block and possible transitions between those states under various operations. A 13B is a table listing the possible states of the - metablock. The metablock states are defined as follows: • Erased · All sections in the metablock have been erased. In, its section rendering logic All sections are the same 2 · Sequential update: The metablock has been partially written in sequential order, which may utilize page labels. Logical group. 3. Scramble update: The metablock has been partially or completely written, and its sections appear in a logical, non-sequential order. Any section may be written once, on. All sections belong to the same logical group. 4: Complete: The metablock has been completely written in a logical sequential order, which may utilize page tags. 5 · Original: The metablock is the previous complete metablock, but at least its section of 124919.doc -38- 200837562 has been discarded due to host data updates. _ Figure 14 (sentence to 14〇1) is a state diagram showing the effect of the "and the operations performed by the block" on the logical group. 囷14(A) is displayed corresponding to the _th ^u &amp The person writes the logical group of the action and the meta-relationship of the meta-region rabbit. The host will logically order a previously unwritten logical group - one of the segments - new configuration Among the erased metablocks. The logic is in the middle of the fine-grained state. ^Group, read by the block will enter the orderly update map eggs) Display corresponding to the first _ complete operation (four) series group and meta-region The conversion-free status is closed. - The previously un-written person's sequential update logic group will become complete when the jin district & m machine sequentially writes the person. If the card uses a predefined data pattern to fill the remaining This conversion may also occur if the segment is written to fill the group. The metablock becomes complete. ί κ Figure 14(c) shows the logical group and metablock corresponding to the first-time messy operation. The shape of the conversion is sad. - The previously unwritten person's sequential update logic group will change when at least the section is written by the host in a non-sequential manner.囷14(D) does not correspond to the logical group and metablock conversion of the first compression operation. A valid segment in a previously unwritten random update logical group It will be copied from the old block to a new messy metablock, and then the old block will be erased. Figure 14(E) does not correspond to the logical group and metablock of the first merge operation. The transformation is shown in Figure 4. All valid segments in a previously unwritten random update logical group are removed from the old messy block for filling in a logically sequential order. Erase the block. The section that is not written by the host 124919.doc -39- 200837562 will be filled - the milky block. The Yibeke pattern. Then the old figure 14(F) will be erased... The state diagram of the logical group disk block conversion in a sequential write operation. The 哕砰, /, 兀 & 4 host will logically sequence a complete logical group or add a wide area ^ ^ area & is written into the erased metablock of a new configuration. The logical group "and the brother,, /, 4 7L block will enter In the sequential update state, the previous complete metablock becomes an original metablock. “Graphic” (G) does not correspond to the state diagram of the logical group fish metablock conversion of a sequential filling operation. The sequential update logical group becomes complete when all the segments are sequentially written by the host. In the case of using the product & true charge from the original block, the logical group is sequentially updated to make it become a complete abandoned project. This conversion will also occur during the collection period, after which the original block can be erased. Figure 14 (H) does not correspond to the state diagram of the logical group and metablock conversion of the non-sequential writer operation. The update logical group becomes messy when at least the section is written by the host in a non-sequential manner. The non-sequential section writes ^ makes the active section in the child update block or the corresponding original block become discarded 0 囷 14 (1) shows the state of the logical group and metablock conversion corresponding to a compression operation Figure. All valid segments in a messy update logical group are copied from the old block to a new messy metablock, and the old block is erased. This original block will not be affected. Fig. 14(J) shows a state diagram of logical group and metablock conversion corresponding to a merge operation. All valid sections in a messy update logical group will be 124919.doc -40- 200837562 copied from the old messy block and the original block, used to populate a new configuration in a logical sequential order Wipe the block. The old messy block and the original block are then erased. Update Block Tracking and Management 囷 15 illustrates a preferred embodiment of the structure of an Configuration Block List (ABL) for tracking open and closed update blocks and erased blocks for configuration. The configuration block list (ABL) 610 is saved in the controller RAM 130 to allow management of the configuration of the erased block, allocation of blocks, update blocks, associated blocks and control structures, and Enables correct logic to physical address translation. In the preferred embodiment, the ABL includes a list of erased blocks, an open update block list 614, and a closed update block list 616. The Open Update Block List 614 is a collection of block items in the ABL that have open update block attributes. The list of open update blocks has one item for each data update block currently open. Each project retains the following information. The LG system currently updates the logical group address specific to the metablock. < Sequential/chaotic state, which indicates that the update block is filled with sequential or cluttered update data. MB is the metablock address of the update block. The page label is recorded in the starting logical section at the first physical location of the update block. The number of segments written indicates the number of segments that have been written to the update block. MBg is the metablock address of the associated original block. The page Tag() is the page label of the relevant original block. The Closed Update Block List 616 is a subset of the Configuration Block List (ABL). It is the set of 124919.doc -41 - 200837562 in the ABL that has the block item with the updated block attribute turned off. The closed update block list has _ items for each data update block that has been closed, but its items are not updated in a logical to main entity directory. Each item retains the following information. Lg is the logical group address reserved for the current update block. MB is the metablock address of the update block. The page tag is recorded in the first physical bit of the update block: the starting logical segment at the location. hall. The original block address of the associated metablock. The hash block index - the sequential update block has data stored in a logical sequential order, so that any logical segment in the block can be easily located. A messy update block stores its logical sections in reverse order and can also be staggered - multiple update generations of logical sections. Additional information must be maintained to track the location of the parent valid logical segment in the messy update block. In a preferred embodiment, the hash block index data structure allows for tracking and fast access - all active segments in a messy block. The messy block index independently manages the logical location, the area' and the efficient program system, the hot area of the user data. The index data structure essentially allows the indexing information to be maintained in flash memory with little update requirements so that performance is not significantly affected. On the other hand, the month/month of the most recently written section in the messy block is saved in the list of messy sections in the controller RAM. Similarly, the cache of index information from the flash memory is saved in the controller RAM to maximize the number of flash sector accesses for address translation. It will be used in the chaotic block index (CBI) section of each flash block. 124919.doc -42- 200837562 Figure 16A shows the data field illustrating a messy block index (CBI) section. A hash block index section (CBI section) containing an index for each sector mapped to a cluttered update block within a logical group that defines within the cluttered update block or within its associated original block The location of each section of the logical group. A CBI section includes: a messy block index field for tracking valid sectors within the messy block; a messy block information field for tracking address parameters for the messy block; A segment index field that is used to track valid CBI segments within a metablock (CBI block) that stores CBI segments. Figure 16B illustrates an example of such a Scrambled Block Index (CBI) section to be recorded in a dedicated metablock. This exclusive metablock will be referred to as CBI block 620. When a CBI section is updated, it is written to the next available physical sector location in the CBI block 620. Multiple copies of a CBI segment may exist in the CBI block, but only the last written copy is valid. For example, the CBI section of the logical group LG! has been updated to the second most recent version to be a valid version. The location of each valid segment in the CBI block is identified by an index set in the last written CBI section in the block. In this example, the last CBI segment written in the block is the CBI segment of LGn6 and its index set replaces the valid index set for all previous index sets. When the CBI block eventually completely fills the CBI section, the block is compressed during a control write operation by rewriting all valid sections to a new block location. Then erase the full block. A messy block index field within a CBI section contains one of each logical section mapped to a cluttered update block within a logical group or subgroup. 124919.doc -43 - 200837562. Each index item represents the offset in which the valid data of the corresponding logical section is located within the messy update block. The reserved index value indicates that there is no valid data for the logical segment in the messy update block, and the corresponding segment in the associated original block is a valid segment. The cache of some messy block index block items is saved in the controller RAM. The messy block information block in a CBI section contains one of each of the messy update blocks present in the system, which records the address parameter information of the block. The information in this field is only valid in the last written section in the CBI block. This information also appears in the data structure in RAM. The project of the parent messy update block contains three address parameters. The first parameter is the logical address (or logical group number) of the logical group associated with the messy update block. The second parameter is the metablock address of the cluttered update block. The second parameter is the physical address offset of the last segment written in the hash update block. The offset information sets the starting point of the scan of the hash update block during initialization to reconstruct the data structure in RAM. The section index field contains items for each valid CBI section in the CBI block. It defines the offset within the CBI block that locates the most recently written CBI section associated with each approved messy update block. The reserved value of an offset in the index indicates that the approved cluttered update block does not exist. Figure 16C illustrates a flow diagram of accessing data for a logical segment in a given logical group that is undergoing a messy update. During the update process, the update data is recorded in the messy update block, while the unchanged data remains in the original metablock associated with the logical group. The procedure for accessing a logical section of a logical group under a messy update condition is as follows: 124919.doc -44- 200837562 Step 650: Start locating a given logical section of a given logical group. Step 652 · Locate the last written CBI section in the CBI block. Step 654: Locate the cluttered update block or the original block associated with the given logical group by looking up the messy block information block of the last written CBI section. This step can be implemented at any time prior to step 662. Step 658: If the last written CBI segment is for a given logical group, the CBI segment is located. Proceed to step 662. Otherwise, proceed to step 660. Step 660: Locate the CBI section of the given logical group by looking up the section index field of the last written CBI section. Step 662: locating a given logical segment among the hetero-IL block or the original block by looking up the hash block index field of the located CBI segment. Figure 16D illustrates a flow diagram of accessing a material of a logical segment in a given logical group that is being scrambled, according to an alternate embodiment, in which the logical group has been partitioned into a plurality of subgroups. The finite I capacity of a CBI segment can only track a predetermined maximum number of logical segments. When a logical group has more logical segments than a logical segment of a single CBI segment, the logical group is divided into a plurality of subgroups, and a CBI segment is assigned to each subgroup. In one example, each CBI section has sufficient capacity to track a logical group consisting of 256 sectors and up to 8 cluttered update blocks. If the logical group has a size of more than 256 segments, then there is a separate CBI segment for each 256 segment subgroup within the logical group. There may be a CBI section for up to 8 subgroups within a logical group to support a logical group of up to 2048 segments of size 124919.doc -45 - 200837562. In a preferred embodiment, an indirect indexing scheme can be used to facilitate index management. Each item of the section index has direct and indirect fields. The direct segment index defines the offset within the CBI block that locates all possible CBI segments associated with a particular hash update block. The information in this field is only valid in the last written CBI section associated with that particular messy update block. A reserved value of an offset in the index indicates that the CBI section does not exist because the corresponding logical subgroup associated with the cluttered update block does not exist or has not been updated because the updated block has been configured. The indirect section index defines the offset within the CBI block that locates the most recently written CBI section associated with each of the approved messy update blocks. The reserved value of an offset in the index indicates that the approved cluttered update block does not exist. Figure 16D shows the procedure for accessing a logical section of the logical group under a messy update. The steps are as follows: Step 670: Split each logical group into multiple subgroups and assign a cbi section to each subgroup. Step 680: Begin to locate a given logical segment of a given group of a given logical group. Step 682: Locate the last written CBI section in the CBI block. Step 684: Locate the cluttered update block or the original block associated with the given sub-group by looking up the messy block information block of the last written CBI section. This step can be performed at any time prior to step 696. Step 686 · If the last written CBI segment is for a given logical group, proceed to step 691. Otherwise proceed to step 690. 124919.doc -46- 200837562 Step 690: Locate the last written segment of multiple CBI segments of a given logical group by looking up the indirect segment index field of the last written CBI segment. Step 691: At least one CBI section associated with one of the subgroups of the given logical group has been located. carry on. Step 692: If the located CBI segment is related to a given subgroup, the CBI segment of the given subgroup is located. Proceed to step 696. Otherwise, proceed to step 694. Step 694: Locate the CBI section of the given subgroup by looking up the direct section index field of the currently located CBI section. Step 696: Locate the given logical segment among the hash block or the original block by looking up the messy block index field of the CBI section of the given subgroup. Figure 16E illustrates an example of a Scrambled Block Index (CBI) section and its functionality in a particular embodiment of partitioning each logical group into multiple subgroups. A logical group 700 initially stores its complete data in an original metablock 7 〇 2 . The logical group is then updated, which configures a dedicated messy update block 704. In these examples, the logical group 7〇〇 is divided into a plurality of subgroups (e.g., subgroups A, B, C, and D), each of which has 256 segments. First locate the CBI block 620

124919.doc is the last written CBI section of the i-th section in subgroup B. Finally _ -47- 200837562 If the last written CBI section becomes one of the four CBI sections of a given logical group, it will be further determined whether it is indeed a given subgroup B containing the ith logical section CBI section. If so, the clutter block index of the CBI section will point to the metablock location to store the data for the i th logical segment. The segment location may be in the messy update block 704 or the original block 702. If the last written CBI section is one of the four CBI sections of the given logical group but is not actually targeted to the subgroup b, then its direct section index is looked up to locate the subgroup B. The CBI section. Once the positive CBI segment is located, the hash block index can be searched for the CBI to be located after the messy update block 704 and the original block 7〇2 are located. A section is not one of the four CBI sections of a given logical group' then looks up its indirect section index to locate one of the four CBI sections. In the example shown in Figure 16E, the CBI section of subgroup c is located. Next, the direct segment index of this cbi segment of subgroup c can be found to locate the exact CBI segment of the subgroup B. This example shows that when § finds its messy block index, it finds that the third logical segment has not changed and its valid data will be in the original block. The same considerations apply to locating the first logical segment in subgroup c of a given logical group. This example shows that the last written cm segment is not part of the four CBI segments of a given logical group.纟 (4) Segment Index Points to the four c-segments of a given logical group. The last of the four referred to is also the CBI section of the subgroup C. When the hash block index is found, it is found that the ">th logical segment is located at the specified position in the mutated update block 704 of the 124919.doc -48 - 200837562. A list of miscellaneous IL segments is present in the controller RAM for each cluttered update block in the system. Each list contains a record of the extents written in the messy update block since the last update of the relevant CBI section in the flash memory. The number of logical sector addresses of a particular hash update block (which may be stored in a hashed list) is a design parameter with a typical value of 8 to 16. The optimal size of the list is determined by its trade-off between the extra burden of messy data write operations and the segment scan time during initialization. During system initialization, each cluttered update block is scanned as needed to identify the valid segment written since the previous update of one of its associated CBI segments. A list of messy sections in the controller RAM for each cluttered update block is constructed. Each block only needs to be scanned from the last sector address defined in the messy block information field in its last written CBI section. When a messy update block is configured, a CBI section is written to correspond to all updated logical subgroups. The logical and physical address of the messy update block is written into an available messy block information field in the section, wherein the messy block index block has an invalid item. A list of messy sections is opened in the controller RAM. When a messy update block is closed, a CBI section is written in which the logical and physical addresses of the block are removed from the messy block information block in the block. The corresponding cluttered section in the RAM becomes unused. A list of corresponding hashes in the controller RAM is modified to include a record of the extents in which a messy update block is written. When the list of cluttered sections in the controller RAM has no free space for writing another section of a cluttered update block 124919.doc -49- 200837562 Recording 'Write logic related to the section in the list The subgroup updates the Cb I section and clears the list. When CBI block 620 becomes full, the valid CBI section is copied into a configured erase block and the previous CBI block is erased. Address Table The logical-to-physical address translation module 14 shown in Figure 2 is responsible for associating the logical address of a host with the corresponding physical address in the flash memory. The mapping between logical groups and entity groups (metablocks) is stored in a set of tables dispersed among the non-volatile flash memory 200 and the volatile but relatively fast ram 130 (see Figure 1). With the list. The address table is maintained in flash memory, which contains the metablock address of each logical group in the memory system. In addition, the logical-to-physical address record of the most recently written sector is temporarily saved in RAM. When the system is initialized after power up, the volatile records can be reconstructed from the block list and data section headers in the flash memory. Therefore, it is only necessary to report the address table %% in the flash memory less frequently, resulting in a lower percentage of the write operation of the control data. The hierarchy of the address records of the logical group includes a list of open update blocks, a list of closed update blocks in RAM, and a group address table (GAT) maintained in the flash memory. The open update block list is the list of the data update block that is enabled for writing the updated host segment data between the current open and the next month. When a block is closed, the project for the block, Jiang Yuexi, is moved to the list of closed update blocks. The list of closed update blocks is the controller of the block update block that has been closed. j.1 124919.doc -50- 200837562

List in Ram. During the control write operation, the subset of items in the list is moved to a section in the group address table. The Group Address Table (GAT) is a list of metablock addresses for all logical groups of host data in the system. The GAT contains one item for each logical group arranged in order according to logical addresses. The nth item in the GAT contains the metablock address of the logical group of the address η. In the preferred embodiment, it is a table in flash memory that includes a set of segments (referred to as GAT segments), wherein the items define each logical group in the memory system. Metablock address. The GAT segment is located in one or more dedicated control blocks (referred to as G AT blocks) in the flash memory. Figure 17A illustrates the data block of a group of address table (GAT) segments. A QAT segment may have, for example, sufficient capacity to contain one of 128 consecutive logical groups of G AT items. The parent G AT segment contains two components, a gAT project set for the metablock address of each logical group within a certain range, and a GA Ding segment index. The first component contains information for locating metablocks associated with logical addresses. The second component contains information for locating all valid GAT segments within the GAT block. Each GAT project has three barriers, which are replaced by: metablock number; page label, as previously matched with 囷3A(in), and flag, which indicates whether the metablock is Has been reconnected. The GAT sector index lists the location of a valid GAT segment in a GAT block. This index is in each G AT segment, but is replaced by the version of the next write GAT segment in the GAT block. Therefore, only the version in the last written GAT section is valid. Figure 17B illustrates an example of such a group location 124919.doc - 51 - 200837562 address table (GAT) section to be recorded in one or more GAT blocks. The GAT block is a metablock dedicated to recording the gat section. When a GAT segment is updated, it is written to the next available physical segment location in the GAT block 720. Therefore, there may be multiple copies of a gat section in the gat block, and only the copy that was last written is valid. For example, the GAT segment 255 (which contains the metrics for logical groups LG3968 through LG4 〇 98) has been updated at least twice, with the latest version being a valid version. The location of each active segment in the GAT block can be identified by an index set in the last written GAT region 4 in the block. In this example, the last written gaT segment in the block is gAT segment 236 and its index set is a valid index set that replaces all previous index sets. When the G AT block is finally fully filled with the GAT segment, the block is compressed during a control write operation by rewriting all valid segments to a new block location. Then erase the full block. As explained above, a GAT block contains a group of logical address spaces, and one logically contiguous set of items. The GAT segment within a GAT block is used for logical-to-entity mapping of 128 consecutive logical groups. The number of GAT segments required to store an item of all logical groups within the address range traversed by a GAT block occupies only a portion of the total segment location in the block. Therefore, the section can be updated by writing a GAT section to the block and using the section location. The index of all valid gat segments and their locations in the gat block is maintained in an index block in the most recently written GAT segment. The portion of the total segment occupied by a valid gAT segment in a GAT block is a system design parameter, which is typically 25/〇. However, there are up to 64 valid GAT zones in each GAT block 124919.doc -52- 200837562. In systems with large logical capacity, it may be desirable to store the gat section in more than one GAT block. In this case, each GAT block is associated with a fixed range of logical groups. The GAT update controls a portion of the write operation that is triggered when the abl is used to configure the block (see Figure 18). It is implemented simultaneously with the ABL filling and CBL emptying operations. During the GAT update operation, a GAT segment has items that are updated with the greed of the corresponding item in the self-closing update block list. When updating a GAT project, any corresponding project can be removed from the Closed Update Block List (CUBL). For example, the G AT segment to be updated may be selected according to the first item in the closed update block list. The updated section is written to the next available location in the gat block. When no segment location is available for the updated GAT segment, the GAT rewrite operation occurs during a control write operation. A new GAT block is configured, and the valid GAT segment as defined by the GAT index is copied from the full GAT block in a sequential order. Then erase the full GAT block. A GAT cache is a copy of the controller ARM 130 of the item in the breakdown of 128 items in the GAT segment. The number of GAT cache items is a system design parameter with a typical value of 32. Each time a project is read from a GAT session, a GAT cache is created for that section. Maintain multiple G AT caches. This number is a design parameter with a typical value of four. Based on the least recently used method, a GAT cache is rewritten using a different section of the project. Metablock Management has been erased 124919.doc -53- 200837562 The erase block manager 1 60 shown in Figure 2 uses a list of directories to store directory and system control information for managing erase blocks. The lists are distributed among the controller RAM 130 and the flash memory 200. When it is necessary to configure an erased metablock to store user data or store the system control data structure, it is selected in the configuration block list (ABL) (see Figure 15) that is retained in the controller RAM. Available metablock numbers. Similarly, when a unary block is erased after it has been decommissioned, its number is added to the list of cleared blocks (cbl) that are also retained in the controller RAM. The control data will be stored in the Flash. The list includes the erased block list, and a bit map listing the erased states of all the metablocks in the flash. The erased block list and MAP are stored in individual segments and recorded in a dedicated metablock (called a MAP metablock). These lists are distributed among the controller RAM and flash memory, which provides an erased block record hierarchy to effectively manage the use of erased metablocks.囷18 is a schematic block diagram illustrating the distribution and flow of control and catalog information for the use and recycling of erased blocks. The control and directory data will be stored in a plurality of lists that are retained in the controller RAM 130 or retained in the MAP block 750 resident in the flash memory 200. . In the preferred embodiment, the controller RAM 130 retains the configuration block list (ABL) 610 and a cleared block list (CBL) 740. As previously described in connection with Figure 15, the configuration block list (ABL) 124919.doc -54· 200837562 tracks the metablocks that were recently configured to store user data or storage system control lean structures. When a new erased metablock needs to be configured, the next available metablock number is selected in the configuration block list (ABL). Similarly, the cleared block list (cbl) can be used to track updated metablocks that have been deconfigured and erased. Both the ABL and CBL are retained in the Control RAM 13 0 (see 囷1) for quick access and easy manipulation while tracking these very active update blocks. The configuration block list (ABL) tracks the pool of an erased metablock and configures the erased metablocks into an update block. Thus, each of the metablocks that can be described by an attribute indicates whether it is an erased block, an open update block, or a closed update in the upcoming configuration of the ABL. Block. It is shown in Fig. 18 that the Abl contains an erased ABL list 6 12, an open update block list 614, and a closed update block list 6 16 . In addition, associated with the Open Update Block List 6.4 is the associated original block list 615. Similarly, the associated erased block list 6丨7 is associated with the list of closed update blocks. As shown in the previous section 15, the related lists are a subset of the open update block list 614 and the closed update block list 616, respectively. The erased ABL block list 612, the open update block list 614, and the closed update block list 61 6 are all a subset of the configured block list (abl) 6 10, in each Projects each have corresponding attributes. The MAP block 750 is a metablock for flash memory 200 dedicated to storing erase management records. The MAP block stores a plurality of consecutive MAP block segments, each of which is an erase block management (ebm) 124919.doc -55-200837562 segment 760 or a MAP segment 780 . When the erased block has been configured and looped when the unary block is decommissioned, the related control and directory information is preferably included in a logical segment, and the logical segment is available in the MAP. The block is updated, and each updated data instance is recorded into a new block segment. There may be multiple copies of the plurality of EBM segments 760 and the plurality of MAP segments 780 among the MAP blocks 750, only the latest version is valid. An index of the location of the valid MAP segments is included in a field of the EMB block. During a control write operation, a valid EMB section must be written to the MAP block at the end. When the MAP block 75 0 is fully loaded, it can be compressed during control of the write operation by overwriting all active sectors to a new block location. Then erase the full block. Each EBM section 760 contains an erased block list (EBL) 770, which is a list of the addresses of the subset of the erased blocks. The erased block list (EBL) 770 can serve as a buffer containing the erased metablock number from which the metablock number can be periodically re-populated to refill the ABL, and periodically The metablock number is added to re-empt the CBL. The EBL 770 can act as a buffer for use as an available block buffer (ABB) 772, an erased block buffer (EBB) 774, and a cleared block buffer (CBB) 776. The Available Block Buffer (ABB) 772 contains a copy of the items located in the ABL 610 immediately following the previous ABL fill operation. In fact, it is a backup copy of the ABL after the ABL fill operation. The erased block buffer (EBB) 774 contains a plurality of erased block bits 124919.doc - 56 - 200837562, which were previously transmitted from the MAP segment 78 or the CBB list 776 (description As follows, and can be transferred to the ABL 610 during an ABL fill operation. The beta divide block buffer (CBB) 776 contains the address of the erased block, which is during a CBL flush operation. It is transmitted from the CBL 74〇 and will be transmitted to the MAP section 78 or the EBB list 774. Each of the MAP segments 780 contains a bit map structure called MAp. The MAP uses a bit for each metablock in the flash memory to indicate the erase status of each block. Bits corresponding to the block addresses listed in the ABL, CBL, or erased block list in the EBM section are not set to the erased state in the MAP. Any block that does not contain a valid data structure and is not designated as one of the erased block, erased block list, ABL or CBL in the MAP will not be used by the block configuration algorithm, so it cannot be saved. Take to store the host or control the structure. This provides a simple mechanism for excluding blocks with defective locations from the accessible flash memory address space. The hierarchy shown in Figure 18 effectively manages the erased block records and provides complete security for the list of block addresses stored in the controller's RAM. The erased block items can be exchanged between the block address list and one or more MAP segments 780 in an infrequent manner. After the power is turned off, during the system initialization, the information in the erased block list and the address translation table stored in the plurality of sectors in the flash memory is 'limitedly scanned in the flash memory. A small number of referenced data blocks can be used to reconstruct these lists. 124919.doc -57- 200837562 The algorithms used to update the hierarchy of erased metablock records can be configured in the following order using the erased block: the region from the MAP block 7 50 The block plexes are interleaved in the address sequence from the chunk address of the cBL 704, which reflects the order in which the tiers are updated by the host. For most metablock sizes and system memory capacities, a single MAP segment can provide a one-bit map for all metablocks in that system. In this case, the erased block must be configured for use in the same address order as recorded in this MAP segment. Erase Block Management Operations As mentioned earlier, the ABL 6 10 has a list of address locations that can be configured to use erased metablocks and metablocks that have recently been configured as data update blocks. The actual number of block addresses in the ABL is between the upper and lower limits. Both upper and lower limits are system design variables. The number of ABL items formatted during manufacturing is a function of card type and capacity. In addition, the number of items in the ABL may decrease near the end of the life of the system' because the number of available erased blocks is reduced due to the failure of the block during its lifetime. For example, after the fill operation, the items in the ABL may indicate the blocks available for the following purposes. The items of the data update block that are partially written are each block_item, which does not exceed the system limit of the maximum number of update blocks that are simultaneously open. One of the blocks has been erased to twenty items for configurable data update blocks. The four items of the erased block are used to be configurable as control blocks. ABL Fill Operation 124919.doc -58- 200837562 When the abl 610 is configured to run out, it must be refilled. The one-of-item operation that populates the ABL occurs during the control write operation. This operation is triggered when a block must be configured, but the erased block item contained in the ABL is not sufficient to be configured as a poor block update block or a specific other control data update block. During a control write, the ABL fill operation is performed concurrently with the GAT update operation. The following actions occur during an ABL fill operation: 1. Keep the ABL project with the attributes of the current data update block. 2. Keep the abl item with the attributes of the closed data update block, unless an item of the block is being written in the simultaneous (10) update operation, in which case it will be removed from the ABL. this project. 3. Keep the abl item of the unconfigured erase block. 4. Compress the ABL to remove the gap created by the item removal, thus maintaining the order of the items. 5. Completely fill the ABL by appending the next available item from the deleted list. 6. The ABB list will be rewritten with the current project in ABL. The list of block addresses for the CBL clear operation, the same. The CBL system that empties the CBL is erased by the controller RAM. The number of erased block items and the one of the abL phases occur during the control write operation. Therefore, it will be performed simultaneously with the ABL fill/GAT update operation or (10) block write operation. In a CBL flush operation, the project will be from the CBl and written to the CBB list 776. Removed from 740 124919.doc -59- 200837562 map exchange operation When the EBB list 774 has been emptied, the MAP segments 780 and the EBM segments 760 are erased during a control write operation. The MAP exchange operation is periodically performed between the block information. If all erased metablocks in the system are recorded in the EBM section 760, then no MAP section 780 exists and no MAP exchange is required. During a MAP exchange operation, a MAP segment for entering the erased block into EBB 774 is considered a source MAP segment 782. Conversely, the MAP segment used to receive the erased block from the CBB 776 is considered a destination MAP segment 784. If only one MAP segment exists, it acts as both the source and destination MAP segments, as defined below. The following actions are performed during a MAP exchange. 1 • Select a source MAP segment based on the incremental indicator. 2. Select a destination MAP segment based on the block address in the first CBB project that is not in the source MAP segment. 3. Update the destination map section in the manner defined by the relevant project in the CBB and remove the items from the CBB. 4. Write the updated destination MAP segment to the MAP block unless there is no separate source MAP segment. 5. Update the source map section in the manner defined by the relevant project in the CBB and remove the items from the CBB. 6_Add the remaining items in the CBB to the EBB. 7. Fill the EBB as much as possible with the erased sector address defined by the source MAP section. 124919.doc •60- 200837562 8·Write the updated source MAP segment into the MAP block. 9. Write an updated EBM section into the MAP block. Inventory Management 圏18 displays the distribution and flow of this control and catalog information between these various lists. For the sake of convenience, the operations for moving items between components of the list or changing the attributes of the items in Fig. 18 are indicated as follows [A] to [〇]. [A] When an erased block is configured as an update block for host data, the attributes of its items in the ABL will be changed from the erased AB1 block to the open update block. [B] When an erased block is configured as a control block, its items in the abl are removed. [c] When an ABL project with open update block attributes is created, the associated original block block is added to the project to record the original metablock of the logical group to be updated. site. This information can be obtained from this GAT. [D] When the block is closed-updated, the properties of its project located in the file change from the open update block to the closed update block. Rich off · * I ^ \*fr4 -Ί7Γ 口Μ /Ψ I η. Block, and located in the ABL·“口η< the attributes of the related metablocks in the project will become erased [F] In an ABL fill operation period p, _ between the pieces / month, in the same control write operation in the GAT address is updated 'book closed closed update block will be from the ABL Remove the item. Stomach 124919.doc -61 - 200837562 [G] During an ABL fill operation, when an item that has closed the update block is removed from the ABL, its associated erased original is The block project is moved to the CBL. [H] When a control block is erased, its project is added to the CBL. [I] Blocks that have been erased during an ABL fill operation. The item will be moved from the EBB list to the ABL and assigned the attributes of the erased ABL block. [J] After modifying all relevant ABL items during an ABL fill operation, the blocks in the ABL The address will replace the block address in the ABB list. [K] The CBL is controlled at the same time as an ABL fill operation. Items in the erased block will be moved to the CBB list. [L] During a MAP exchange operation, all related items will be moved from the CBB list to the MAP destination section. [M] Yu Yi During the MAP exchange operation, all related items will be moved from the CBB list to the MAP source section. [N] Continue [L] and [M], during the MAP exchange operation, all remaining items will be from this The CBB list is moved to the EBB list. [O] Next [N], during a MAP exchange operation, if possible, items other than the item moved in [M] will be moved from the MAP source section. Populate the EBB list. The logical-to-real address is converted to the physical location of a logical segment in the flash memory. Figure 2: 124919.doc -62- 200837562 shows the logical-to-physical address translation module 1 4 Logic-to-physical address translation is implemented. In addition to those logical groups that have been recently updated, the group address table (GAT) resident in the flash memory 200 or the GAT in the controller RAM 130 can be utilized. Take the implementation of the batch conversion. Address translation of the recently updated logical group will It is necessary to look up the address list of the update block that is mainly resident in the controller RAM 130. Therefore, the logic-to-physical address conversion procedure for a logical sector address depends on the logical group in which the sector is located. Related block types: The types of blocks are as follows: • Complete block, sequence data update block, messy data update block, closed data update block. Figure 19 shows the flow of the program showing logical to physical address translation. Figure. Basically, the corresponding metablock and the physical section are located by first using the logical sector address to find each update directory (e.g., the open update block list and the closed update block list). If the associated metablock is not part of an update procedure, the catalog information will be provided by the GAT. The procedure for logical to physical address translation is as follows: Step 800: Provide a logical sector address. Step 810: Look up the given logical address in the open update block list 614 (see Figures 15 and 18) in the controller ram. If the search fails, then step 820 is entered, otherwise step 830 is entered. Step 820: Look up the given logical address in the closed block list 616. If the search fails, the given logical address is not part of any update procedure; go to step 87 and perform a GAT address translation. Otherwise, proceeding to step 860, the updated update block address translation is performed. 124919.doc -63- 200837562 Step 830: If the update block containing the given logical address is sequential, proceed to step 840 to perform sequential update block address translation. Otherwise, proceed to step 850 to perform a messy update block address translation. Step 840 ·• Use the sequential update block address translation to obtain the metablock address. Proceed to step 880. Step 850: Acquire the metablock address by using a hash update block address translation. Proceed to step 880. Step 860: Acquire the metablock address by using the closed update block address translation. Enter the step called 880. Step 870: The group address table (GAT) conversion is used to obtain the metablock address. Proceed to step 880. Step 880: Convert the metablock address into a physical address. The conversion method will depend on whether the metablock has been rejoined. Step 890: The physical sector address has been obtained. The various address translation procedures are described in more detail below: Sequential Update Block Address Translation (Step 840) From the Open Update Block List 614 (see 囷 15 and 18), the information is directly converted into a sequence. The address translation of the target logical sector address in the logical group associated with the block is updated in the following manner. 1 · From the list of ''page tags' and 'number of segments written, and other fields determine whether the target logical segment is in the update block or its associated original block Among them. 2. From the side list, retrieve the metablock address that is appropriate for the target logical section. 3. Determine the sector 124919.doc -64-200837562 address in the metablock from the appropriate 'Page Label' block. The messy update block address translation (step 850) and a miscellaneous IL update The address translation sequence of the target logical sector address in the logical group associated with the block is as follows: 1. If it is determined from the list of messy sectors in the RAM that the sector is a recently written sector, The location in this list can directly complete the address translation. 2. The most recently written segment in the far CBI block will contain the target logical segment address in its messy block data corruption. The physical address of the hash update block. It also contains an offset in the cbi block of the last written CBI segment associated with the hash update block in its indirect segment index block (see figure) 16A-16E) 3. The information in these fields will be cached in the ram without reading the segment during the subsequent address conversion. The indirect section index field identifies the CBI section. 5 · The recently accessed messy update subgroup The direct segment index field is cached in RAM without the need to perform the read at step 4 to repeatedly access the same cluttered update block. 6. Direct segment index read at step 4 or step 5. The field can then identify the C BI section associated with the logical subgroup containing the target logical sector address. 7. Read the target logical sector address from the CBI section identified in step 6. The messy block index item 124919.doc -65- 200837562 8 · The recently read messy block index field can be cached in the controller RAM without performing the steps 4 and 7 The same logical subgroup is repeatedly accessed. 9. The hash block index item defines the location of the target logical segment in the hash update block or the associated original block. If the target logical segment The valid copy is located in the original block, and the original metablock and page tag information can be used to locate the valid copy. The updated block address translation is closed (step 860). The list is updated from the closed block. (See Figure 18) Directly completing the address translation of the target logical sector address in the logical group associated with the closed update block, as follows: 1) Reading the metablock assigned to the target logical group from the list The address of the sector in the metablock is determined in the address 0 2 ·>[the π page label of the list]. GAT Address Translation (Step 87) A project in the GAT is valid if a logical group is not referenced by the Open Update List or the Closed Update List. The address translation sequence of the target logical sector address in the logical group referenced by the GAT is as follows. 1. Estimate the range of such available gat caches in the RAM to determine if an item of the target logical group is contained within a gat cache. 2. If the target logical group is found in step 1, the GAT cache contains the complete group address information, which includes both the metablock address and the page label 124919.doc -66 - 200837562. The conversion of the target logical group address is allowed. 3. If the target address is not located in a G AT cache, the G AT index of the target G AT block must be read to identify the G AT segment associated with the target logical group address. position. 4. The GA-butyl index of the last accessed GAT block is stored in the controller RAM and can be accessed without having to read a segment from the flash memory. 5. A list of the metablock addresses of each GAT block and the number of extents written into each of the gat blocks is stored in the controller RAM. If the necessary GAT index cannot be obtained at step 4, it can be read from the flash memory immediately. 6. Read the GAT segment associated with the target logical group address from the segment location in the G AT block defined by the GAT index obtained at step 4 or step 6. A GAT cache is updated with a portion of the section containing the target item. 7. Obtain the target segment address from the metablock address block and the "page tag" block in the target GAT project. Metablock to physical address translation (step 880). If the flag associated with the metablock address indicates that the metablock has been rejoined, the associated LT segment is read from the BLM region to determine The erase block address of the target sector address. Otherwise, the erase block address is directly determined from the metablock address. Control Data Management 囷20 describes the implementation of the knowledge structure of the control data structure in the process of the memory management operation - the singularity of the white paper. Data update management

Each of the Ram's towels is used for disposal, etc.; h: w for disposal (4) in the memory, the temple controls the lean section of the temple to produce ~u X and exclusive & block 'and also with the #h in RAM Early father changed the information. A male update management operation is implemented for the ABL, CBL, and the list of messy sections. The ABL is updated when the erased block is configured as an "update block or control block" or when the block is closed. The CBL is updated when the control block is erased, or when an item of a closed update block is written into the GAT. The updated hashed list is updated when a sector is written into a messy update block. A control write operation causes information from the control data structure in the Ram to be written into the control data structure in the flash memory, and if necessary, the flash memory and other supported control data in the RAM are updated accordingly structure. This operation is triggered when the ABL does not contain any items of the erased block that are to be configured to update the block, or when the CBI block is overwritten. In the preferred embodiment, the ABL fill operation, the CBL clear operation, and the EBM sector update operation are all performed during each control write operation. When the MAP block containing the EBM section becomes full, the valid EBM and MAP segments are copied into a configured erased block and the previous MAP block is erased. A GAT zone is written during each control write operation and the closed update block list is modified accordingly. When a GAT block becomes full, a GAT rewrite operation is performed. A 124919.doc -68-200837562 CBI section is written as described previously for the 'special hash sector write operation'. When the CBI block becomes full, the valid cbj segment is copied into a configured erased block and the previous CBI block is erased. As previously described, a MAP exchange operation is performed when there are no other erased block items in the EBB list in the EBM section. At each timing when the MAP block is rewritten, a MAp address (MApA) sector (whose A records the current address of the MAP block) is written into a dedicated MAPA block. When the MAPA block becomes full, the valid MAPA segment is copied into a configured erased block 1 and the previous MAPA block is erased. At each timing when the MAPA block is rewritten, a boot (b squad section is written to the current boot block. When the boot block becomes full, a valid boot area is used. The segment is copied from the current version of the boot block to the backup version, and then the backup version becomes the current version. The previous current version will be erased and become the backup version, and the valid boot sector will be Write back to it. / K Control data integrity and management control data examples related to the memory block management system - a small, Τ « open; / day mussel, and district: ghost configuration information, such as combined Round 2〇. As mentioned earlier, reading: the bait material will be retained in the high-speed (four) fine and slower non-volatile ^fe, body block. Any frequently changed control data will be In the case of retaining f:, it will periodically control the writing core to update the stored in a non-volatile meta-area. "In this way, the control ^ θ is expected to be non-volatile but slower. Among the (four) memories, 124919.doc •69- 200837562

No frequent access is required. The hierarchy of a plurality of control data structures (e.g., gat, CBI, MAP, and MAPA not included in 囷2〇) is retained in the flash memory. Thus, a control write operation causes information from the control data structure in RAM to update the equivalent control data structure in the flash memory. As described in 囷2G, the block management system will maintain a control data set in the Shaanxi flash memory during its operation. This control data set is stored in the metablocks in a similar host data. In this regard, the control data itself will be managed in blocks and will be updated to allow for the collection of obsolete items. It is also mentioned above that there is a - control data hierarchy, in which the control data in the lower hierarchy is updated higher than the upper layer. For example, assuming each control block has N control segments to write, then the following control update and control block reconfiguration sequences typically occur. Referring again to 囷2〇, (4): The CBI update will fill the CBI block and trigger a cbi reconfiguration (rewrite) with a MAP update. If the messy block is closed, it will also trigger a catch update. A MAP update is triggered each time the GAT update is made. Each n&gAT update fills the block and triggers a GAT block reconfiguration. In addition, when a MAP block is fully loaded, it also triggers an MAp block reconfiguration and a MAPA block (if present, the boot block will point directly to the map) to update. In addition, when the MAPA block is fully loaded, it also triggers a MAPA block reconfiguration, a boot block update, and a _ MAp update. In addition, when an open block is fully loaded, it will trigger the reconfiguration of an active boot block into another boot block. 124919.doc -70- 200837562 Update Block Alternative According to another aspect of the present invention, in the non-volatile memory using the block management system, at most the first predetermined maximum number for simultaneously enabling the recording of data is supported. The system of the updated block implements an improved block management scheme. The update block is mainly a sequential update block in which data is recorded in a logical sequential order 'but allows at most a second predetermined maximum number of update blocks to be messy update blocks' where the data is not recorded in a logical sequential order. Whenever the new configuration of the update block can cause the set of update blocks to exceed the first or second predetermined maximum number, one of the existing update blocks in the set will be closed and removed to meet the limit. Before closing the update block, its data is merged into sequential blocks. Improvements are made to avoid sequential updates that can result in an additional number of cluttered blocks being merged. This is accomplished by separating the sequential update blocks and the cluttered update blocks into individual replacement or combination sets. Specifically, when the sequential update causes the configuration of the new update block to exceed the first predetermined maximum number, the most recently used sequential update block of the set area preferentially moves the space. In the current system, there are generally two types of data: user data and control data. User data is typically transferred from the host to the memory system in a logical sequential order. The sequential update block is configured to optimally program sequential write operations from the host. User data can also be in a logical, non-sequential order, especially when the logic is subsequently updated. The messy update block is built to best program data in a non-sequential order. Another source of clutter or non-sequential data is control data maintained by the file system or memory system, such as file and directory information generated during the process of storing user data. The actual system that supports the maximum number of simultaneous updates while opening the update block 124919.doc -71 - 200837562 The restricted previous plan is to close the least recent block in the set, whether it is sequential or cluttered. This scenario improves the previous scenario, where essentially the updated block in the pool during the sequential write operation needs to be closed to move the space to the new configuration, the least recently used sequential update block in the collection. This ensures a variety of more: = block is effectively used for program-by-sequence writer operations and random writer operations. In particular, the large sequential write operation of the host to prevent the host from executing may force the premature closure of the FAT and directory information to be prematurely closed. Another messy block can be created very quickly to store fat and directory information: once the large sequential write operation is completed, it is updated again. Improvement = generation of policy-based mandatory substitution and separation of merged pools to prevent new additions to the cluttered blocks during the sequential writes, and the potential to open sequential or open mixed I blocks. Subsequent fat and catalog information updates. The practical system limitations for enabling the update block at the same time as the maximum number of maximums have been previously described. For example, in conjunction with FIG. 1(), in step specific embodiment, step 410 tests whether the new configuration exceeds the maximum number of update blocks Umax, which can be turned on to receive updates. If it exceeds lx, the update zone A attacker's 'on-the-fly or messy update block' will be closed within step 420 to keep the system within the specified limits. Circle 21 schematically illustrates the number of update blocks for the block management system / two specified limits. The total number of update blocks ('Αχ) has an overall limit, which is given by the sum of the number of messy update blocks Ng and the number of sequential update blocks Ns. Because the resource of the chaotic update block is more dense, it needs a messy block 124919 .doc •72· 200837562 Indexing (CBI) Additional Maintenance 'The maximum number of messy block indexes ("uCMAX', there are limits to the good system. For this reason, the first limit needs to update the total number of blocks Ne+NS<;=UMAX. The second limitation requires a random update block number nc<=ucmax. Figure 22 illustrates a typical consumption of a combination of two limitations optimized for various memory devices. A given combination is specified by Umax_Uc_. n Specify the block management system, which allows up to three update areas in the update set area, and only one of them blocks the update block. Similarly, "7_3, specifies the block that supports the seven most updated blocks in the hometown. Management systems, where up to three can be as messy as update blocks, and simpler memory systems with smaller memory capacity will have more restrictions, resulting in smaller maximum numbers. 囷23A, 囷23B, and 囷23C Signal Describe the sequence of events, which is used to introduce the new update block into the episode of the update block according to the first case of the first case. 囷23Α Unintentionally stated as shown in FIG. 22, 5_2 , the configured update pool In this example, the 'update pool is filled with the maximum five allowed update blocks. The update pool is further divided into sequential sets 丨2〇〇, which contains three sequential update blocks SI, S2 And S3, and the cluttered area 13〇〇, which contains the largest two cluttered or non-sequential update blocks, and C5. This example shows the first case where the lowest active block happens to be a sequential update block, such as S3 1201时 When a new update block needs to be configured, one of the existing update blocks in the update set needs to be closed to move out of space. For example, when the host targets a logical segment group that is not served by an existing update block in the set. The group writes the sequential data, 124919.doc -73 - 200837562 will need to configure a new update block to record the data. Figure 23B unintentionally illustrates the closure of the update zone in the lowest role according to the previous scenario, in order to move the space to the new New block. In this case, the lowest active update block happens to be S3 1201 and is closed and removed from the update block pool. 囷23C schematically illustrates the removal of the closed update block to remove After the space, the newly configured update block is introduced into the pool. In this case, the newly configured update block S6 m2 is introduced into the sequence set area 1200 to record the data in a logical sequential order. In this way, the maximum allowed update is not exceeded. The number of blocks UMAX 圏 24 Α, 囷 24 Β and Fig. 24C schematically illustrate a sequence of events for introducing a new update block into the set of update blocks according to the previous alternative in accordance with the second case.囷24Α Schematically illustrates the use of the 5_2" configuration update as described in Figure 22. In this example, the update pool is filled with up to five allowed update blocks. The update pool is further divided into a sequenced area 丨2〇〇, which includes three sequential update blocks SI, S2 and S3, and a cluttered area 13〇〇, which contains the largest two chaotic or non-sequential update blocks. And C5. This example shows a second case in which the lowest active block happens to be a cluttered update block, such as C4 1 3 0 1 Figure 24B schematically illustrates the closing of the lowest active update pool according to the previous scheme in order to move out of space Give the new update block. In this case, the lowest active update block is S3 13〇1 and is closed and removed from the update block aggregate. 124919.doc -74- 200837562 Figure 2 4C schematically illustrates the introduction of a newly configured update block into the pool after the closed update block has been removed to move the space. In this case, the newly configured update block S6 1212 is introduced into the sequence set area 1200 to record the data in a logical sequential order. In this manner, the maximum number of update blocks allowed is not exceeded. Umax 0 Figures 25A and 25B illustrate the maintenance of Umax and UCMAX in the schemes of FIG 10, FIG 23B, and FIG 24B, respectively. These two restrictions are usually applied simultaneously. Figure 25 illustrates the previous circle 1, step 41 and block 23, and the scheme described in Figure 24, where the least recently accessed update block is closed whenever the new configuration exceeds the predetermined limit. Step 1252: Organizing a non-volatile memory into blocks, each block for storing data that can be erased together. Step 1254: Configure up to a first predetermined number of update blocks, which are simultaneously opened to store updates of logical data units. / (4) 1256: Whenever the new update block is configured to exceed the predetermined number, close the update block that is close to the least access to move the space to the new update block. Figure 25Β illustrates that the previous step is less The scheme described in the township 370, in which the number of messy update blocks exceeds the old one, and the pre-emptive limit is closed, and the recent least-accessed messy (non-sequential) update block is closed. Step 1354: In the block 0 logic to store the logical data unit in the exclusive open update block, configure the logical data unit to configure a maximum second predetermined number of update areas 124919.doc -75- 200837562 Step 1356 ··Whenever non-sequence The number of update blocks will exceed the second predetermined number, and one of the least recently accessed non-sequential update blocks is closed so as not to exceed the second predetermined number. One of the disadvantages of this prior scheme is that it can lead to excessive shutdown of cluttered update blocks in certain situations. This scheme is particularly inefficient in the case of sequential writes that cause a messy block containing control data to be prematurely closed to move space within the pool to a newly configured sequential block. For example, if the closed messy update block C4 13 10 records FAT and directory information, an alternate configuration will be required immediately to run the function as soon as possible after completing the sequential write. This requires another round of closing of the update block in the current minimum usage in the collection area to move the space to the alternate update block for recording control data. In accordance with the current aspect of the present invention, closing the update block so as not to exceed the predetermined maximum number of update blocks further refines from only selecting the least recently used update block to reducing the number of excessively closed instances of the cluttered update block. Preferably, in the case of the embodiment, if the update block is configured to record the sequential data and one of the update block sets needs to be closed to move the space, the sequential update block in the lowest role in the set is closed. 26A, 26B, and 26C schematically illustrate an event sequence for introducing a new update block into a set of update blocks in accordance with the improved alternative. FIG. 26A schematically illustrates the use of FIG. ,, 5-2" configured update pool. In this example, the update pool is filled with up to five allowed update blocks. The update pool is further divided into a sequenced area 12〇〇, which includes three sequential update blocks SI, S2 and S3, and a cluttered area 13〇〇, which contains a maximum of two 124919.doc -76-200837562 clutter or Non-sequential update block 'and. . This example shows a situation similar to 囫24A, where the lowest active block happens to be a clutter update block, such as C4 1301. $, which is shown as a sequential block S3 1202 as the lowest active player in the sequenced area 12〇〇.囷26B schematically illustrates the closing of an update block in the update block pool in accordance with the present modification to move the space to the new update block. If the new configuration associated with the sequential update and the number of updated blocks in the set is already at the maximum value of ^, one of the updated blocks in the closed set area must be moved to the newly configured 'update block'. However, in the case of the lowest active block C4 1301, it is ignored because it is a cluttered block. Instead, the block will be updated in the lowest role in the sequence set 1200. In this example, the system S3 1202 will be closed and removed from the update block pool.囷 26C schematically illustrates the introduction of the newly configured update block into the pool after the closed update block has been removed to move the space. In this case, the newly configured update block S6 1212 is introduced into the sequence set area 1200 to record the data in a logical sequential order. In this way, the maximum number of allowed update blocks is not exceeded (UMAX. Figures 27A, 27B and 27C schematically illustrate sequence of events for introducing new clutter update blocks into the update area in accordance with the improved alternative Blocks of blocks 囷27A schematically illustrates the use of the “Hong 2" configured update pool as described in 囷22. In this example, the update pool is filled with a maximum of five allowed update blocks. Divided into a sequential set area 12〇〇, which includes three sequential update blocks SI, S2 and S3, and a cluttered set area 13〇〇, which contains a maximum of two 124919.doc -77-200837562 chaotic or non-sequential update blocks , C4 and C5. This example shows that the lowest-acting block happens to be a sequential update block, such as S6 1201. In addition, it shows the clutter block C4 1302 〇囷 27B as the lowest-acting person in the cluttered area 13〇〇 Optionally, the closing of an updated block in the update block set according to the improvement scheme is performed, so as to move out the space to the newly updated block. If the new messy update block is introduced into the already filled messy area 1300, the messy set must be closed. District One of the update blocks is to move out the space. In this example, the messy pool area 13 00 already contains the largest two messy update blocks. When another messy update block is created, for example, when the existing sequential update block S 1 1220 Converted to a messy block 'will exceed the maximum number of cluttered blocks, unless one of the blocks has been removed from this moon's case and the minimum active messy block C4 1302 is removed from the cluttered zone 130 0 0, Figure 27C schematically illustrates the introduction of a new clutter update block into the pool after another messy update block has been closed and removed to move the space. In this case, S1 has been taken from the sequence set area. The sequential update block 122 in 2〇〇 is switched to the clutter update block C6 1320 in the messy set £1300. In this way, the maximum number of allowed random update blocks UCMAX is not exceeded. A modification of a set of limited update blocks during a sequential update according to the first embodiment is illustrated. Step 1400: Organizing a non-volatile memory into blocks, each block being used for storage and erasing together Information. Step 1402: Configure the most And a first predetermined number of update blocks, which are simultaneously opened to store the update of the logical data unit. 124919.doc -78- 200837562 Step 1406 ··Respond to the write command for writing the sequential data, the logic is sequentially executed The data unit is written to the update block. Step 1408: Respond to a predetermined condition that the update block to be closed satisfies the further write of the sequential alpha-bubble units, and configure a new update block to continue Write, and if the new configuration exceeds the first predetermined number 'priority to any least recently accessed update block in a non-sequential order, closes a least recently accessed update block in the sequential order. Figure 29 is a flow chart illustrating the present modification of managing a limited set of limited update blocks having two predetermined limits in accordance with a second embodiment. Step 1410: Organize a non-volatile memory into blocks, each of which is used to store data that can be erased together. Step 1412: Configure up to a first predetermined number of update blocks, which are simultaneously opened to store updates of logical data units. Step 1416: Configure a maximum of a second predetermined number of update blocks in the open update blocks for storing logical data units in a logical, non-sequential order. Step 1418: whenever the introduction of the update block for storing the data in the logical sequential order may exceed the first predetermined number, the latest least-accessed update block containing the data is closed in a logical sequential order to move the space for introduction. Update block. Step 1420: whenever the introduction of the update block for storing the data in a logical non-sequential order may exceed the second predetermined number, the update area of the least recently accessed one of the included data is closed in a logical non-sequential order. The space is given to the introduced update block. 124919.doc -79- 200837562 The generalization of this scheme is to update blocks based on a set of attributes, such as updating block files to store sequential or non-sequential data, or storing a predetermined type of system data. When implementing a pool of a limited number of updated blocks, the updated blocks of each level will have their own alternative rules when they exceed the maximum number of levels supported by the level. ^ For example, sequential update blocks and non-sequential update blocks are two different levels. The replacement rules for each of these levels are the same, that is, the new block is used to replace the block in the lowest role. Therefore, when the cluster { of the sequential update block is exceeded, the lowest-function block in the collection area will be closed and removed before the new block is introduced into the collection area. The same is true for the episodes of non-sequential update blocks. In general, each level has its own alternative rules that are independent of the other levels. An example of an alternative rule is to replace the least recent access, the most recent access, the least frequency access, the maximum frequency access, etc., according to the corresponding level. Figure 30 is a flow diagram illustrating the improvement of a set of limited update blocks that have a hierarchically based alternative rule. ^ Step 1430: Organize a non-volatile memory into blocks, each block storing % of the data that can be erased together. Step 1432: • Provide an aggregate of up to the first predetermined maximum number of update blocks simultaneously opened for storing updates of logical data units. Step 1436: A set of predefined levels is provided for updating the block based on a set of attributes. Each level supports a subset of the updated blocks of the associated maximum number of associated blocks. Step 1438: Provide a set of corresponding substitution rules to the group of predefined levels to specify the updated blocks within the individual subsets to be replaced. 124919.doc 200837562 Step 1438 - Grouping the update blocks within the set into the corresponding subset by level. Step 14401 whenever the other level of the other-update block 'off' is introduced, and the lowest active update block within the subset area containing the associated predetermined maximum number of updated blocks is removed. All patents, patents, papers, books, books, other publications, documents and contents cited herein are hereby incorporated by reference in their entirety for all purposes. In the event of any contradiction or conflict between the definitions or terms used in any incorporated publications, documents or events and the text of this document, the definition or terminology of this document shall prevail. Although the various aspects of the invention have been described in terms of specific embodiments, it should be understood that the present invention is protected by the full scope of the appended claims. BRIEF DESCRIPTION OF THE DRAWINGS A housing 1 schematically illustrates a main hardware component suitable for implementing the memory system of the present invention. % 囷 2 illustrates a memory according to a preferred embodiment of the present invention, which is organized into a plurality of physical segment groups (or meta-blocks) and is memoryd by the controller. Body manager to manage. A preferred embodiment of the plurality of metablocks Figures 3A(1) through 3A(iii) schematically illustrate the mapping between a logical group and a metablock in accordance with an embodiment of the present invention.囷 3B schematically illustrates the mapping between a plurality of logical groups. Figure 4 illustrates the alignment of a unitary block in physical memory and various structures. 124919.doc -81 - 200837562 囷5A shows a schematic diagram of a metablock composed of the minimum erasing units connecting different planes.囷s B illustrates a specific embodiment in which a minimum erase unit (MEU) is selected from each plane for joining into a unitary block.囷5C illustrates another embodiment in which one or more of the planes are selected to be joined into a unitary block.囷6 is a schematic block diagram of the metablock management system being designed in the controller and the flash memory. ° " 囷7A illustrates an example of a plurality of segments that are sequentially written into a logical group in a sequential update block. Figure 7B illustrates an example of a plurality of segments among a logical group that are written in a cluttered update block in a cluttered order.囷8 illustrates an example of performing a separate host write operation with discontinuous logical addresses and sequentially writing a plurality of sectors among a logical group in a sequential update block. Figure 9 is a flow diagram showing a procedure for updating a logical data group implemented by the update area and block manager in accordance with a general embodiment of the present invention. 10 is a flow chart showing a procedure for updating a logical data group implemented by the update block manager in accordance with a preferred embodiment of the present invention. Figure 11A is a more detailed flow chart illustrating the merge procedure for closing the cluttered update block shown by 囷10. Figure 11B is a more detailed flow chart illustrating the compression procedure for closing the cluttered update block shown by 囷10. 124919.doc •82- 200837562 Figure u Qianming—all possible states of a logical group, and possible transitions between those states under various operations. The graph is a table that lists the possible states of a logical group. Figure 13A illustrates all possible states of the metablock and possible transitions between the states under various operations. - metablock system - corresponds to an entity group of the logical group. Figure 13B shows a table listing the possible states of a metablock. f Figures 14(A) through 14(J) show the effect of the state of the logical group and the operations performed on the physical block. Figure 15 illustrates a preferred embodiment of a structure for a configuration block list (ABL) for tracking open and closed update blocks and erased blocks for configuration.囷16A illustrates the data field of a messy block index (CBI) section. Figure 16B illustrates an example of such a Scrambled Block Index (CBI) section to be recorded in a dedicated metablock. Figure 16C is a flow diagram illustrating the access to a logical section of a logical group v of a hypothetical logical group v that is undergoing a messy update. Figure 16D is a flow diagram illustrating the material of a logical segment in a hypothetical logical group that is being scrambled to update, in accordance with an alternate embodiment, in which a logical group has been partitioned into a plurality of subgroups. Figure 16E illustrates an example of a Scrambled Block Index (CBI) section and its functionality in a particular embodiment of partitioning each logical group into multiple subgroups.囷17A illustrates the data fields of a group of address table (GAT) sections. Figure 17B illustrates an example of the group address table 124919.doc -83 - 200837562 (GAT) section to be recorded in a GAT block. Figure 18 is a schematic block diagram of the distribution and flow of this control and directory education for the use and recycling of erased blocks. Circle 19 is a flow chart showing the procedure for logical to physical address translation.囷20 shows the hierarchy of operations performed on the control data structure during the memory management operation.囷21 schematically illustrates two specified limits for the number of update blocks for the block management system.囷22 illustrates a typical example of a combination of two limitations for optimizing various memory devices.囷23Α schematically illustrates the use of the “5_2" configured update pool as described in 囷22. Figure 23 Β Unexplained to illustrate the closure of the minimum active update pool according to the previous scheme, in order to move the space to the new update. The circle 23C schematically illustrates the introduction of the newly configured update block into the pool after the closed update block has been removed to move the space. Figure 24 is a schematic illustration of the use of "5 as illustrated in Figure 22]. -2,, the update zone of the group bear.囷 24B schematically illustrates the closing of the update set in the lowest action according to the previous scheme to move the space to the new update block. Figure 24C schematically illustrates the introduction of a newly configured update block into a pool after the closed update block has been removed to move the space. Figure 25A illustrates the previously described schemes of steps 410 and 23B and Figure 24B of Figure 10, wherein the most recent update block of at least 124919.doc -84 - 200837562 is closed whenever the new configuration exceeds the predetermined limit. Figure 25B illustrates the scenario previously described in step 370 of Figure 10, in which the least recently accessed messy (non-sequential) update block is closed whenever the number of cluttered update blocks exceeds a predetermined limit. Figure 26A schematically illustrates an update pool configured using 5_2, f as illustrated in 囷22. Fig. 26B schematically illustrates the closing of an update block in the update block pool in accordance with the present modification to move the space to the new update block.

Figure 26C schematically illustrates the introduction of a newly configured update block into a pool after the closed update block has been removed to move the space.囷 27A schematically illustrates an update pool using the "5_2" configuration as illustrated in Figure 22. Figure 27B schematically illustrates the closure of an update block in the update block pool in accordance with the present modification, so that The space is moved to the new update block. 囷27C schematically illustrates the introduction of the new messy update block into the pool after the other __(4) update block has been closed and removed. 囷28 Series Flowchart, Description The present improvement is directed to managing a set of limited update blocks during a sequential update in accordance with the first embodiment. Figure 29 is a flow diagram illustrating the management of a limited set of update blocks having two predetermined limits in accordance with a second embodiment. The improvement scheme. Fig. 30 is a flow chart illustrating the present modification of a limited set of update blocks having a hierarchy-based alternative rule. [Main component symbol description] 10 Host ^4919^00 -85- 200837562 / 20 Memory system 100 Controller 110 Interface 120 Programmer 121 Common programmer 122 ROM 124 Programmable non-volatile memory 130 RAM 134 Configuration block Listing 136 Clear Block List 140 Logical to Physical Address Translation Module 150 Update Block Manager Module 160 Erase Block Manager Module 170 Metablock Link Manager 180 Metablock Link Manager 200 Memory Body 210 Group Location Table 220 Scramble Block Index 230 Erase Block List 240 MAP 255 GAT Section 610 Configuration Block List 612 ABL List Erased 614 Open Update Block List 124919.doc -86- 200837562 / l 615 Original Block List 616 Close Update Block List 617 Erased Original Block List 620 CBI Block 700 Logical Group 702 Meta Block 704 Clutter Update Block 720 GAT Block 740 Cleared Block List 750 MAP Block 760 Erase block management section 770 erased block list 772 free block buffer 774 erased block buffer 776 cleared block buffer 780 MAP Segment 782 Source MAP Section 784 Destination MAP Section 1200 Sequential Set Area 1201 Sequential Update Block 1202 Sequential Block 1212 Update Block 1220 Sequential Update Block 1300 Cluttered Set Area 124919.doc -87 - 200837562 1301 Miscellaneous Update Block 1302 messy block 1310 closed messy update block 1320 messy update block C4 messy update block C5 messy update block SI sequential update block S2 sequential update block S3 sequential update block 124919.doc -88-

Claims (1)

200837562 X. Patent Application Scope κ A method of storing data in the memory in a non-volatile memory organized into a plurality of blocks, each block is used to store logical data units that can be erased together. The method includes: configuring a maximum number of blocks as (5) to open an update block for storing updates of logical data units; r 2. responding to a write command for writing data in logical order , the data is written to the update block in a logical sequential order; and the time interval should be adjusted for the <closed update block to satisfy the predetermined condition of the sequential logic data, and a new update is configured. The block is = ', 〇海写人' and if the new configuration exceeds the first-pre-number, excellent = any least-accessed update block of non-sequentially stored data - one of the sequentially stored data - the least recently stored Take the update block. Material meter: woven into a plurality of blocks - methods in the non-volatile memory that will be used in the memory - each block is used to store logical lean units that can be erased together, the method includes ·· The first predetermined number of ..v blocks is used as the simultaneous update area ▲, 1 is used to store the update of the logical data unit, and the new = non-sequential storage logical data unit is configured to open up the second predetermined number in the block. The update block, whenever the logic block is introduced, the update block may be exceeded by the latest predetermined number of update blocks - the latest least-access sequence sequence closes the update block introduced by the package; and the update block is ' To give the 124919.doc 200837562 whenever an update block is used to store data in a logically non-sequential order, the introduction of the update block may exceed the second predetermined number of 'closest accesses to the logically non-sequentially closed one of the contained data. The block is updated to move the space to the introduced update block. 3. The method of claim 2, wherein the update block for storing data in a logical non-sequential manner is converted from a data store in a sequential order. f 4. A method of storing data in the memory in a non-volatile memory organized into a plurality of blocks, each block for storing logical data units that can be erased together, the method comprising · Configuring up to a first predetermined number of blocks to simultaneously open update blocks for storing updates of logical data units; - providing one set of up to - maximum number of updated blocks simultaneously for use Storing an update of the logical data unit; 曰i, , and pre-depreciation levels for classifying the update block according to a set of attributes 'each level supports a subset of the updated block of the maximum associated number of updated blocks; i, a corresponding substitution rule set for the replacement S Hai individual sub-chapter F to 6A - » * ^ Dingguo & the update block; by the level of the update area within the 隹P 4 ^ Blocks are grouped into corresponding subsets; and whenever a new block containing the same level of the same level in the lowest role of the associated ones is introduced, the maximum number of closed and moved orders is closed A subset of the update block updates the block, and the removed update block is selected according to the "Hai corresponding replacement rule. 124919.doc 200837562 5 The method of claim 4, wherein the set of attributes comprises storing a block of data in a logical sequential order. 6. The method of claim 4, wherein the set of attributes comprises a block of logically non-sequentially stored data. The method of claim 4, wherein the set of attributes comprises a block storing system data associated with operating the memory. 8. The method of claim 4, wherein the memory system flashes EEPR 〇m. 9. The method of claim 4, wherein the memory has a NAND structure. The method of claim 4, wherein the memory system is on a removable memory card. 11. The method of claim 4, wherein the non-volatile memory has a memory cell using a floating gate structure. 12. The method of claim 4, wherein the non-volatile memory has a memory cell using a dielectric layer structure. The method of any one of the preceding claims, wherein the memory has a memory unit that stores data of one bit. The method of any one of the preceding claims, wherein the memory has a memory unit that can store more than one bit of data. 15. A non-volatile memory comprising: a memory, the system is organized into blocks, and each block is divided into two groups: a memory unit, and each memory unit is used to store a logical data unit. And 'including: when the update area is turned on - the controller is used to control the operation of the blocks to configure up to the first predetermined number of blocks as the same 124919.doc 200837562 block for updating the logical data unit; Responding to a write command for writing data in logical order, writing data to an update block in a logical sequential order; and responding to the update block that is to be closed to satisfy the sequential logic data advance write a predetermined condition, configuring a new update block to continue the write, and if the new configuration exceeds the first predetermined number, prioritize any recent least-accessed update blocks that are stored in a non-sequential manner to be closed A recently accessed least updated update block of data. 16 · A non-volatile memory, comprising: a c memory, which is organized into blocks, each block is divided into memory units that can be erased together, and each memory unit is used to store a logical data And a controller for controlling operations of the blocks, comprising: configuring a maximum of a first predetermined number of blocks as simultaneously opening an update block for updating an update of the logical data unit; Up to a second predetermined number of update blocks in the open update blocks of the logically non-sequential storage logical data unit; the introduction of the update block may be exceeded by the first predetermined number whenever used to store the data in a logical sequential order , in the logical sequential order, close the update block containing the least recently accessed data to move the space to the introduced update block; and whenever the block is used to store the data in a logically non-sequential order The person may exceed the second predetermined number to logically and non-sequentially close the updated block containing the data - the least recently accessed to move the space to the 124919.doc 200837562 Introduced update block. In the case of Shiming, the memory block is used to store data in a logically non-sequential manner. The update block is converted from data stored in a logical sequential order. 1 8 · A non-volatile memory comprising: a carcass, the system is organized into blocks, each block is divided into memory units that can be erased together, and each memory unit is used to store a logical data unit; 'a pool containing up to a first predetermined maximum number of update blocks simultaneously opened for storing updates of logical data units; a set of predefined levels for updating blocks 'levels' according to a set of attribute classifications Supporting a subset of the most associated predetermined maximum number of update blocks; a corresponding replacement rule set for the set of predefined levels to specify the updated block within the individual subsets to be replaced; a set of subsets, which include update blocks by level; and a controller for controlling the operations of the blocks, including: whenever and when another update block of the same level is introduced, closing and removing the inclusion The associated update blocks of a predetermined subset of the maximum number of update blocks are selected, the selected update blocks being selected according to the corresponding replacement rule for the same level. 19. The memory of claim 18, wherein the set of attributes comprises a block of data stored in a logical sequential order. 20. The memory of claim 18, wherein the set of attributes comprises a block of data stored in a logically non-sequential manner. 124. The method of claim 18, wherein the set of attributes comprises a block storing system data associated with operating the memory. 22. The memory of claim 18, wherein the memory system flashes EEPR 〇m. 23. The memory of claim 18, wherein the memory has a structure. ° 24. The memory of claim 8, wherein the memory system is on a removable memory card. 25. The memory of claim 18, wherein the non-volatile memory has a memory cell employing a floating gate structure. 26. The memory of claim 8, wherein the non-volatile memory has a memory cell using a dielectric layer structure. 27· A non-volatile memory, comprising: σ hexamor, which is organized into blocks, each block is divided into memory units that can be erased together, and each memory unit is used to store a logical data unit. a pool containing up to a first predetermined maximum number of update blocks simultaneously opened for storing updates of logical data units; a set of predefined levels for classifying update blocks according to a set of attributes, each level Supporting a subset of up to a predetermined maximum number of updated blocks; and a rule set for the new predefined level to specify the updated area within the individual subsets to be replaced Block; - group subset area 'which contains update blocks by level; and 124919.doc 200837562 Close and remove components, i . The piece is used for when the door is broken into the same level and the block is closed and the one of the subsets of the updated block containing the maximum number of updated blocks is removed. The update block is selected according to the corresponding replacement rule for the same level. The memory of any one of claims 15 to 27, wherein the memory has a memory unit each storing data of one bit. The memory of any one of claims 1 to 5, wherein the memory has a memory unit each storing more than one bit of data. 124919.doc