Memory maintenance operations
By moving the maintenance operation of the memory system from the memory device to the controller, the problems of insufficient efficiency, reliability and performance of the memory system in the prior art are solved, and more efficient and reliable memory system maintenance is achieved, and the life of the memory device is extended.
Patent Information
- Application Number
- CN202411865773.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Priority Date
- 2023-12-22
- Filing Date
- 2024-12-18
- Publication Date
- 2025-06-24
AI Technical Summary
Existing memory systems cannot effectively improve the efficiency, reliability and performance of the memory system during maintenance operations, resulting in a shortened life of the memory device and an increase in data errors.
Move maintenance operations from the memory device to the controller, determine the bad memory cell group through the controller's circuit system configuration, restore its data, reprogram it to the backup memory cell group, and withdraw the bad memory cell group.
By moving maintenance operations to the controller, the address remapping time of the memory device is reduced, error exposure to the host is reduced, the life of the memory system is extended and its efficiency and reliability is improved.
Smart Images

Figure CN120199311A_ABST
Abstract
Description
Technical Field
[0001] The present disclosure generally relates to semiconductor memories and methods, and more particularly, to memory maintenance operations. Background Art
[0002] Memory devices are typically provided as internal semiconductor integrated circuits and / or external removable devices in computers or other electronic devices. There are many different types of memories, including volatile and non-volatile memories. Volatile memories require power to maintain their data and may include random access memories (RAMs), dynamic random access memories (DRAMs), and synchronous dynamic random access memories (SDRAMs), etc. Non-volatile memories can provide persistent data by storing the stored data when not powered and may include NAND flash memories, NOR flash memories, read-only memories (ROMs), ferroelectric random access memories (FeRAMs), and resistive variable memories such as phase change random access memories (PCRAMs), resistive random access memories (RRAMs), magnetic random access memories (MRAMs), and programmable conductive memories, etc.
[0003] Memory devices can be used as volatile and non-volatile memories for various electronic applications that require high memory density, high reliability, and low power consumption. Non-volatile memories can be used in, for example, personal computers, portable memory sticks, solid state drives (SSDs), digital cameras, cellular phones, portable music players such as MP3 players, and movie players, and other electronic devices.
[0004] Memory devices can be coupled to a host (such as a host computing device) to store data, commands, and / or instructions for use by the host during operation of a computer or electronic system. For example, data, commands, and / or instructions can be transferred between the host and the memory device during operation of a computing or other electronic system. A controller can be used to manage the transfer of data, commands, and / or instructions between the host and the memory device. Summary of the Invention
[0005] In one aspect, the present disclosure relates to an apparatus including: a memory having a plurality of memory cell groups, wherein: no data is stored in one of the plurality of memory cell groups; and data is stored in all of the other ones of the plurality of memory cell groups; and a controller coupled to the memory and having circuitry configured to: determine that one of the plurality of memory cell groups in which data is stored is a defective memory cell group; recover the data stored in the defective memory cell group; program the recovered data into the one of the plurality of memory cell groups in which no data is stored; and retire the defective memory cell group.
[0006] On the other hand, the present disclosure relates to a method of operating a memory, which includes: determining, by a controller coupled to the memory, that one of a plurality of groups of memory cells of the memory is a defective memory cell group; recovering, by the controller, data stored in the defective memory cell group; programming, by the controller, the recovered data into another one of the plurality of groups of memory cells in which no data is stored; retiring, by the controller, the defective memory cell group; and mapping, by the controller, a logical address associated with the defective memory cell group to a physical address associated with the one of the plurality of groups of memory cells into which the recovered data is programmed.
[0007] On the other hand, the present disclosure relates to a system, which includes: a host; and a memory device, wherein the memory device includes: a memory having a plurality of groups of memory cells, wherein: no data is stored in one of the plurality of groups of memory cells; data is stored in all other ones of the plurality of groups of memory cells; and the memory device has different corresponding physical addresses associated with each corresponding one of the plurality of groups of memory cells; and a controller coupled to the memory and having circuitry configured to: determine that one of the plurality of groups of memory cells in which data is stored is a defective memory cell group; recover the data stored in the defective memory cell group; program the recovered data into the one of the plurality of groups of memory cells in which no data is stored; and map a logical address of the host associated with the defective memory cell group to a physical address associated with the one of the plurality of groups of memory cells into which the recovered data is programmed. BRIEF DESCRIPTION OF THE DRAWINGS
[0008] Figure 1 A diagram illustrating a portion of a memory array having several physical blocks according to an embodiment of the present disclosure.
[0009] Figure 2 A block diagram of a computing system according to an embodiment of the present disclosure.
[0010] Figures 3A to 3D A table illustrating a conceptual example of a memory maintenance operation performed according to an embodiment of the present disclosure. DETAILED DESCRIPTION
[0011] The present disclosure includes devices, methods, and systems for memory maintenance operations. Embodiments include: a memory having a plurality of groups of memory cells, where one of the plurality of groups of memory cells stores no data and all others of the plurality of groups of memory cells store data; and a controller coupled to the memory and having circuitry configured to: determine that one of the plurality of groups of memory cells in which data is stored is a bad memory cell group; recover the data stored in the bad memory cell group; program the recovered data into the one of the plurality of groups of memory cells in which no data is stored; and retire the bad memory cell group.
[0012] Over the lifetime of a memory device, memory cells can experience wear when being sensed and / or programmed. This wear can eventually cause the memory cells to become unreliable in storing and / or retrieving data, which in turn can reduce the reliability and / or performance of the memory device.
[0013] In previous methods, a host of a memory system utilizing the Compute Express Link (CXL) protocol can retire a group of memory cells (e.g., a page) that has become unreliable (e.g., due to wear) during the lifetime of the memory device of the system to attempt to prevent the entire system from becoming unreliable and failing. Additionally, in such previous methods, the memory device of the system can implement a wear leveling operation (e.g., a scheme), such as on-chip, to relocate (e.g., remap) data currently stored in one physical location (e.g., one physical page) of the memory device to another physical location (e.g., another physical page) of the memory device to attempt to distribute this wear more evenly across the memory device.
[0014] However, such methods may not effectively improve the efficiency and reliability of the memory system. For example, when being read and mapped to different locations (e.g., different logical addresses), a single unreliable (e.g., bad) page can generate several data errors, which can cause applications running on the memory system to be killed. Additionally, the mapping of the location (e.g., address) of the bad page may change during the lifetime of the memory system.
[0015] However, embodiments of the present disclosure can improve the efficiency, reliability, and / or performance of a memory system utilizing the CXL protocol and thus extend the lifetime of the memory system by moving maintenance (e.g., wear leveling) operations from the memory device of the system to the controller of the system. For example, moving the maintenance operation from the memory device to the controller can result in a reduction in the latency of the system because the memory device does not have to perform address remapping. Additionally, the controller can retire bad pages without exposing them to the host of the system, which can reduce the amount of errors returned to the host and increase the lifetime of the memory device.
[0016] As used herein, "a" or "an" or "several" can refer to one or more things, and "a plurality of" can refer to two or more such things. For example, a memory device can refer to one or more memory devices, and a plurality of memory devices can refer to two or more memory devices. Additionally, specifically, the identifiers "N", "M", "R", "S", and "B" used herein with respect to the element symbols in the figures indicate that several specific features so labeled can be included in several embodiments of the present disclosure.
[0017] The figures herein follow a numbering convention where the first or first few digits correspond to the figure number and the remaining digits identify an element or component in the figure. Similar elements or components between different figures can be identified by using similar numbers.
[0018] Figure 1 A diagram showing a portion of a memory array 101 having several physical blocks according to an embodiment of the present disclosure. The memory array 101 can be, for example, a ferroelectric RAM (FeRAM) memory array. However, the embodiments of the present disclosure are not limited thereto. For example, the memory array 101 can be a DRAM array, a NAND flash memory array, or a resistive random access memory array such as a PCRAM, RRAM, or spin transfer torque (STT) array, etc. Additionally, although Figure 1 not shown in the figure, the memory array 101 can be located on a specific semiconductor die together with various peripheral circuitry associated with its operation.
[0019] As Figure 1 shown in the figure, the memory array 101 has several physical blocks 107-0 (block 0), 107-1 (block 1),..., 107-B (block B) of memory cells (such as ferroelectric memory cells). The memory cells can be single-level cells and / or multi-level cells, for example, two-level cells, three-level cells (TLC), or four-level cells (QLC) by way of example. As an instance, the number of physical blocks (such as banks) in the memory array 101 can be 32 blocks, 34 blocks, 128 blocks, 512 blocks, or 1,024 blocks, but the embodiments are not limited to a specific power of 2 or any specific number of physical blocks in the memory array 101.
[0020] Several physical blocks of memory cells (such as blocks 107-0, 107-1,..., 107-B) can be included in a plane of memory cells, and several planes of memory cells can be included on a die. For example, in the Figure 1 example shown in the figure, each physical block 107-0, 107-1,..., 107-B can be part of a single die. That is, Figure 1 the portion of the memory array 101 illustrated in the figure can be a die of memory cells.
[0021] AsFigure 1 As shown, each physical block 107-0, 107-1, …, 107-B includes memory cells coupled to access lines (e.g., word lines) of several physical rows (e.g., 103-0, 103-1, …, 103-R). As an example, the number of rows (e.g., word lines) in each physical block may be 2,097,152, but embodiments are not limited to a specific number of rows 103-0, 103-1, …, 103-R per physical block. Further, although Figure 1 not shown in, the memory cells may be coupled to sense lines (e.g., data lines and / or digit lines).
[0022] Those of ordinary skill in the art will appreciate that each row 103-0, 103-1, …, 103-R may include several memory cell pages (e.g., physical pages). A physical page refers to a programming and / or sensing unit (e.g., several memory cells programmed and / or sensed together as a functional group). In Figure 1 the embodiment shown, each row 103-0, 103-1, …, 103-R includes memory cells of one physical page. However, embodiments of the present disclosure are not limited thereto. For example, in an embodiment, each row may include memory cells of multiple physical pages (e.g., memory cells of one or more even pages coupled to even bit lines and memory cells of one or more odd pages coupled to odd bit lines). Additionally, for embodiments including multi-level cells, the memory cells of a physical page may store data of multiple pages (e.g., logical pages) (e.g., upper page data and lower page data, where each cell in the physical page stores one or more bits towards the upper page data and one or more bits towards the lower page data).
[0023] As Figure 1 shown, a page of memory cells may include several physical sectors 105-0, 105-1, …, 105-S (e.g., a subset of memory cells). The cells of each physical sector 105-01, 105-1, …, 105-S may store data of several logical sectors. Additionally, the data of each logical sector may correspond to a part of the data of a specific page. As an example, the data of the first logical sector stored in a specific physical sector may correspond to the logical sector corresponding to the first page data, and the data of the second logical sector stored in a specific physical sector may correspond to the second page data. Each physical sector 105-0, 105-1, …, 105-S may store system and / or user data, and / or may include overhead data, such as error correction code (ECC) data, logical block address (LBA) data, and metadata.
[0024] Logical block addressing is a scheme that can be used by a host to identify the logical sectors of data. For example, each logical sector can correspond to a unique logical block address (LBA). Additionally, the LBA can also correspond to (e.g., be dynamically mapped to) a physical address, such as a physical block address (PBA), which can indicate the physical location in the memory of the data for that logical sector. The logical sectors of data can be several bytes of data (e.g., 256 bytes, 512 bytes, 1,024 bytes, or 4,096 bytes). However, the embodiments are not limited to these examples.
[0025] It should be noted that other configurations for physical blocks 107-0, 107-1, …, 107-B, rows 103-0, 103-1, …, 103-R, sectors 105-0, 105-1, …, 105-S, and pages are possible. For example, rows 103-0, 103-1, …, 103-R of physical blocks 107-0, 107-1, …, 107-B can each store data corresponding to a single logical sector that can contain, for example, more or less than 512 bytes of data.
[0026] Figure 2 is a block diagram of a computing system 206 according to an embodiment of the present disclosure. The computing system 206 can include a controller 200 having a front-end portion 204, a central controller portion 210, a back-end portion 219, and a management unit 209. The computing system 206 can include a host 208 and memory devices 226-1, …, 226-N coupled to the controller 200.
[0027] The front-end portion 204 includes an interface and interface management circuitry for coupling the controller 200 to the host 208 via input / output (I / O) paths 202-1, 202-2, …, 202-M and circuitry for managing the I / O paths 202. Any number of I / O paths 202 can exist, such as eight, sixteen, or another number of I / O paths 202. In some embodiments, the I / O paths 202 can be configured as a single port.
[0028] In some embodiments, the controller 200 can be a memory controller compatible with Compute Express Link (CXL). The host interface (e.g., the front-end portion 204) is coupled to the host 208 and can be managed using the CXL protocol. CXL is a high-speed central processing unit (CPU) to device and CPU to memory interconnect designed to accelerate next-generation data center performance. CXL technology maintains memory coherence between the CPU memory space on the attached device and the memory, which allows resource sharing for higher performance, reduced software stack complexity, and lower overall system cost. CXL is designed as an industry open standard interface for high-speed communication because accelerators are increasingly being used to supplement the CPU to support emerging applications such as artificial intelligence and machine learning. CXL technology builds on the PCIe infrastructure to provide advanced protocols in several areas using the PCIe physical and electrical interfaces, such as input / output (I / O) protocols, memory protocols (e.g., initially allowing the host to share memory with the accelerator), and coherence interfaces. As an example, the interface of the front-end 204 can be a PCIe 5.0 or 6.0 interface coupled to the I / O path 202. In some embodiments, the controller 200 can receive access requests related to the memory device 226 via the PCIe 5.0 or 6.0 interface according to the CXL protocol.
[0029] The central controller portion 210 can include and / or be referred to as data management circuitry. The central controller portion 210 can control the execution of memory operations in response to receiving requests from the host 208. Examples of memory operations include read operations to read data from the memory device 226 or write operations to write data to the memory device 226.
[0030] The central controller portion 210 can generate error detection information and / or error recovery information based on the data received from the host 208. The central controller portion 210 can perform error detection operations and / or error recovery operations on the data received from the host 208 or from the memory device 226. An example of an error detection operation is a cyclic redundancy check (CRC) operation. CRC can be referred to as algebraic error detection. CRC can include using a check value derived from an algebraic calculation using the data to be protected. CRC can detect unexpected changes in the data by comparing the check value stored in association with the data with the check value calculated based on the data. An example of an error correction operation is an error correction code (ECC) operation. ECC encoding refers to encoding data by adding redundant bits to the data. ECC decoding refers to checking the ECC-encoded data to check for any errors in the data. Generally, ECC can not only detect errors but also correct a subset of the errors it can detect. The central controller portion 210 can also perform maintenance (e.g., wear leveling) operations on the data stored in the memory device 226 and retire groups of bad memory cells in the memory device 226, as will be further described herein.
[0031] The backend portion 219 may include a media controller and a physical (PHY) layer that couples the controller 200 to the memory device 226. As used herein, the term "PHY layer" generally refers to the physical layer in the Open Systems Interconnection (OSI) model of a computing system. The PHY layer may be the first (e.g., lowest) layer of the OSI model and may be used to transfer data via a physical data transfer medium. In some embodiments, the physical data transfer medium may include channels 225-1, …, 225-N. The channels 225 may include various types of data buses, such as a four-pin data bus (e.g., a data input / output (DQ) bus) and a single-pin data mask inversion (DMI) bus, as well as other possible buses.
[0032] The memory devices 226 may be various / different types of memory devices that may each include several memory arrays 101 previously described in connection with Figure 1 For example, the memory devices 226 may each include multiple groups of memory cells, such as multiple memory cell pages (e.g., physical pages), as previously described in connection with Figure 1 No data may be stored in one (e.g., only a single one) of the groups of memory cells of a memory device. For example, one of the groups of memory cells may be reserved as a spare page. Data (e.g., user data) may be stored in all of the other ones of the multiple groups of memory cells of the memory device.
[0033] In some embodiments, different respective physical addresses may be associated with each respective one of the multiple groups of memory cells of the memory device 226. For example, the physical addresses associated with each respective one of the multiple groups of memory cells may correspond to different pins of the memory device 226 (e.g., a first address may correspond to a first pin, a second address may correspond to a second pin, etc.). Additionally, the group of memory cells of the memory device in which no data is stored (e.g., the spare page) may be addressable. For example, a logical address may be mapped to the physical address associated with that group, as will be further described herein.
[0034] In some embodiments, the host 208 may have different respective logical addresses associated with all but one of the owners of the multiple groups of memory cells of the memory device 226 in which data is stored (e.g., one of the groups of memory cells in which data is stored has no logical address associated with it). Additionally, the host 208 may not have a logical address associated with the group of memory cells of the memory device in which no data is stored.
[0035] As an example, the memory device 226 can be a ferroelectric RAM (FeRAM) device including a ferroelectric capacitor that can exhibit hysteresis characteristics. As an example, the FeRAM device can include a ferroelectric capacitor and can perform bit storage based on the voltage or charge amount applied thereto. In such examples, relatively small and relatively large voltages allow the ferroelectric RAM device to exhibit characteristics similar to those of a normal dielectric material (e.g., a dielectric material having a relatively high dielectric constant), but at different voltages between this relatively small voltage and the large voltage, the ferroelectric RAM device can exhibit polarization reversal that results in non-linear dielectric behavior.
[0036] However, embodiments of the present disclosure are not limited thereto, and the memory device 226 can be other types of non-volatile memory devices (e.g., NAND or NOR flash memory devices) or resistive random access memory devices (e.g., PCRAM, RRAM, or spin torque transfer (STT) memory devices), etc. In another example, the memory device 226 can be a dynamic random access memory (DRAM) operating according to a protocol such as low power double data rate (LPDDRx), which can be referred to herein as an LPDDRx DRAM device, an LPDDRx memory, etc. The "x" in LPDDRx refers to any of several generations of protocols (e.g., LPDDR5).
[0037] In some embodiments, the controller 200 can include a management unit 209 for initializing, configuring, and / or monitoring the characteristics of the controller 200. The management unit 209 can include: an I / O bus for managing out-of-band data and / or commands; a management unit controller for executing instructions associated with initializing, configuring, and / or monitoring the characteristics of the memory controller; and a management unit memory for storing data associated with initializing, configuring, and / or monitoring the characteristics of the controller 200. As used herein, the term "out-of-band" generally refers to a transmission medium different from the main transmission medium of the network. For example, out-of-band data and / or commands can be data and / or commands transmitted to the network using a transmission medium different from the transmission medium used for transmitting data within the network.
[0038] The controller 200 (e.g., the central controller section 210) may perform maintenance (e.g., wear leveling) operations on data (e.g., user data) stored in the memory device 226, and retire a group of defective memory cells in the memory device 226. For example, the controller 200 may determine that one of a plurality of groups of memory cells in the memory device 226 in which data is stored is a group of defective memory cells (e.g., a defective page). As used herein, a "group of defective memory cells" may refer to a group of memory cells (e.g., a page) that is unreliable in storing and / or retrieving data (e.g., user data) due to physical damage or corruption. The controller 200 may determine the group of defective memory cells, for example, during a scrub operation performed on the memory device 226 (e.g., on a group of memory cells of the memory device 226). As used herein, the term "scrub operation" may refer to an operation for correcting errors (if any) in data values stored in a group of memory cells and rewriting the error-corrected data pattern to the group of memory cells.
[0039] The controller 200 can recover the data stored in the defective memory cell group by performing an error correction operation on the data stored in the defective memory cell group. For example, the controller 200 can recover the data by performing a redundant array of independent disks (RAID) operation on the data stored in the defective memory cell group. The RAID operation can be performed as part of a RAID data protection and recovery scheme that can partition and / or replicate the data stored in the memory cell group among multiple memory devices (e.g., across memory device stripes) and then use the data in the stripes to recover lost data. As used herein, RAID is a general term for a computer information (e.g., data) storage scheme that partitions and / or replicates (e.g., mirrors) information among multiple memory devices and / or multiple pages of components to help protect the data stored therein, for example. The multiple memory devices and / or components in a RAID array can appear to the user and the computer's operating system as a single memory device (e.g., a disk). RAID can include striping (e.g., splitting) the information such that different portions of the information are stored on different pages of different memory devices and / or components. The portions of more than one device or component that store the split data are collectively referred to as a stripe. In contrast, RAID can also include mirroring, which can include storing duplicate copies of the data on more than one page of more than one device or component. As an example of the former, the written data can be striped across N - 1 of N memory devices and / or components, where the error information can be stored in the Nth memory device or component. A RAID stripe can include user data and parity data (e.g., can be a combination of user data and parity data). The parity data of a RAID stripe can include error protection data that can be used to protect the user data stored in the memory from defects and / or errors that may occur during the operation of the memory. For example, a RAID stripe can protect the user data stored in the memory from defects and / or errors that may occur during the operation of the memory and can thus provide protection against memory failures.
[0040] The controller 200 can program the data recovered from the defective memory cell group into a memory cell group in the memory device 226 in which no data is stored (e.g., such that the recovered data is stored in a spare page of the memory device), and retire the defective memory cell group. Retiring the defective memory cell group can include and / or refer to no longer using the defective memory cell group to store data (e.g., no longer programming data into the defective group or sensing data from the defective group). The controller 200 can map the logical address of the host 208 associated with the defective memory cell group to the physical address associated with the memory cell group into which the recovered data is programmed (e.g., into a page that was previously a spare page), such that the defective memory cell group no longer has a logical address associated with it.
[0041] The controller 200 can then copy data stored in one of the multiple memory cell groups of the memory device 226 to another of the multiple memory cell groups of the memory device 226 in which data is stored (e.g., a group without a logical address associated therewith), erase the memory cell group from which the data is copied, and map the logical address associated with the memory cell group from which the data is copied to the memory cell group to which the data is copied. This erased memory cell group can become a spare page for subsequent maintenance operations and / or bad group retirement. In some instances (e.g., where the memory cells of the memory device 226 are bit-alterable), the erase operation can be optional.
[0042] If a maintenance (e.g., wear leveling) operation is performed on the data stored in the memory device 226, the controller 200 can disable the on-chip wear leveling operation and / or the random data scrambling operation in the memory device 226. The controller 200 can disable these operations using, for example, the configuration data bits in the memory device 226.
[0043] The following example algorithm can be used by the controller 200 (e.g., the central controller portion 210) to perform the wear leveling operations described herein. The following parameters are used in the algorithm:
[0044] N + 1 physical locations from 0 to N
[0045] P: Physical address. P ∈ {0, …, N}
[0046] L: Logical address. L ∈ {0, …, N - 1}
[0047] S: Start address. S is a physical address, S ∈ {0, …, N}
[0048] G: Gap address. G is a physical address, G ∈ {0, …, N}
[0049] Define the logical-to-physical mapping P = f S,G (L). The mapping f S,G is defined by the following rules:
[0050] P = f S,G (L) = (S + L + U[S + L - G - (N + 1)u(S - G)]) mod N+1
[0051] where U[·] is the step function (U[i] = 1 if i ≥ 0, U[i] = 0 if i < 0). At the start (t = 0), S and G are set to 0 and N respectively:
[0052] S(0) = 0, G(0) = N
[0053] At each wear leveling operation, G decreases (cyclically) its value. Thus:
[0054] G(t + 1) = (G(t) - 1) mod N+1
[0055] And if:
[0056] G(t + 1) = S(t)
[0057] Then:
[0058] S(t + 1) = (S(t) + 1) mod N+1
[0059] Otherwise:
[0060] S(t + 1) = S(t)
[0061] Whenever G decreases, move the data D in the location corresponding to the new G to the old G location:
[0062] D G(t) = D G(t+1)
[0063] Figures 3A to 3D Tables 330-1, 330-2, 330-3, and 330-4 are conceptual examples illustrating memory maintenance operations performed in accordance with embodiments of the present disclosure. The memory maintenance operations may be performed, for example, by a controller 200 on a memory device 226, as previously described herein (e.g., in connection with Figure 2 ).
[0064] For example, a memory device may include multiple groups of memory cells (e.g., pages). For instance, in the example shown in Figure 3A Table 330-1, the memory device includes 9 groups of memory cells. However, embodiments of the present disclosure are not limited to a particular number of groups of memory cells.
[0065] Different corresponding physical addresses may be associated with each of the multiple groups of memory cells, as previously described herein. For example, in the example shown in Figure 3A Table 330-1, the first group of memory cells has a physical address of 0, the second group of memory cells has a physical address of 1, the third group of memory cells has a physical address of 2, the fourth group of memory cells has a physical address of 3, the fifth group of memory cells has a physical address of 4, the sixth group of memory cells has a physical address of 5, the seventh group of memory cells has a physical address of 6, the eighth group of memory cells has a physical address of 7, and the ninth group of memory cells has a physical address of 8.
[0066] None of the memory cell groups in a memory device may store data (e.g., it may be a spare group, and data (e.g., user data) may be stored in all of the other of the plurality of memory cell groups of the memory device, as previously described herein. For example, in Figure 3A In the example shown in Table 330-1 in
[0067] In addition, a host (e.g., the host 208 previously described in connection with Figure 2 may have different corresponding logical addresses associated with all but one of the owners in which data is stored in the plurality of memory cell groups (e.g., one of the memory cell groups in which data is stored has no logical address associated therewith), as previously described herein. For example, in Figure 3A In the example shown in Table 330-1 in
[0068] the first memory cell group (e.g., the group with physical address 0) has a logical address of 0, the second memory cell group (e.g., the group with physical address 1) has a logical address of 1, the third memory cell group (e.g., the group with physical address 2) has a logical address of 2, the fifth memory cell group (e.g., the group with physical address 4) has a logical address of 3, the sixth memory cell group (e.g., the group with physical address 5) has a logical address of 4, the seventh memory cell group (e.g., the group with physical address 6) has a logical address of 5, the eighth memory cell group (e.g., the group with physical address 7) has a logical address of 6, and the fourth memory cell group (e.g., the group with physical address 3) is the group without a logical address.
[0068] In addition, the host may not have a logical address associated with the memory cell group of the memory device in which no data is stored. For example, in Figure 3A In the example shown in Table 330-1 in
[0069] the ninth memory cell group (e.g., the spare group with physical address 8) has no logical address.
[0069] The controller may determine that one of the plurality of memory cell groups of the memory device is a bad group, as previously described herein. For example, in Figure 3A In the example shown in Table 330-1 in
[0070] After determining that the third memory cell group is a bad group, the controller can recover the data stored in the third group and program the recovered data into a spare group, as previously described herein. For example, the controller can program the data recovered from the third group into the ninth group (e.g., the group with physical address 8). Additionally, the controller can map the logical address associated with the third (e.g., bad) memory cell group to the physical address of the group into which the recovered data is programmed, as previously described herein. For example, the controller can map logical address 2 to physical address 8, as Figure 3B shown in Table 330-2 as illustrated therein.
[0071] Furthermore, the controller can retire the third (e.g., bad) memory cell group, as previously described herein. For example, the third memory cell group (e.g., the group with physical address 2) no longer has a logical address associated with it, as Figure 3B shown in Table 330-2 as illustrated therein.
[0072] The controller can then copy the data stored in one of the multiple memory cell groups to another of the multiple memory cell groups in which data is stored (e.g., the group that has no logical address associated with it), erase the memory cell group from which the data is copied, and map the logical address associated with the memory cell group from which the data is copied to the memory cell group to which the data is copied, as previously described herein. For example, in the Figure 3C example shown in Table 330-3 as illustrated therein, the data stored in the second memory cell group (e.g., the group with physical address 1) is copied to the fourth memory cell group (e.g., the group with physical address 3), the second memory cell group is erased, and logical address 1 is mapped to physical address 3, such that the second memory cell group no longer has a logical address associated with it.
[0073] This erased memory cell group can become a spare group for subsequent maintenance operations and / or bad group retirement, as previously described herein. For example, in the Figure 3C example shown in Table 330-3 as illustrated therein, the second memory cell group (e.g., the group with physical address 1) has become a spare group.
[0074] In the Figure 3D example of this subsequent maintenance operation shown in Table 330-4 as illustrated therein, the data stored in the first memory cell group (e.g., the group with physical address 0) is copied to the second memory cell group (e.g., the group with physical address 1), the first memory cell group is erased, and logical address 0 is mapped to physical address 1, such that the first memory cell group no longer has a logical address associated with it and becomes a spare group.
[0075] Although specific embodiments have been illustrated and described herein, those of ordinary skill in the art will appreciate that arrangements calculated to achieve the same results may be substituted for the specific embodiments shown. This disclosure is intended to cover adaptations or variations of several embodiments of this disclosure. It should be understood that the above description has been presented in an illustrative rather than a restrictive manner. Those of ordinary skill in the art will appreciate combinations of the above embodiments and other embodiments not specifically described herein after reviewing the above description. The scope of several embodiments of this disclosure includes other applications in which the above structures and methods are used. Accordingly, the scope of several embodiments of this disclosure should be determined with reference to the appended claims and the full scope of equivalents to which such claims are entitled.
[0076] In the "Detailed Description", for the purposes of simplifying the present disclosure, some features are grouped together in a single embodiment. The methods of this disclosure should not be interpreted as reflecting an intention that the disclosed embodiments of this disclosure must use more features than are expressly recited in each claim. Rather, as the appended claims reflect, the inventive subject matter lies in less than all of the features of a single disclosed embodiment. Accordingly, the appended claims are hereby incorporated into the "Detailed Description", with each claim standing on its own as a separate embodiment.
Claims
1. A device comprising: A memory (101) having a plurality of memory cell groups (103-0, 103-1, ..., 103-R), wherein: No data is stored in one of the plurality of memory cell groups (103-0, 103-1, ..., 103-R); and All other of the plurality of memory cell groups (103-0, 103-1, ..., 103-R) have data stored therein; and A controller (200) coupled to the memory (101) and having circuitry (210) configured to: determining that one of the plurality of memory cell groups (103-0, 103-1, ..., 103-R) in which data is stored is a bad memory cell group; restoring the data stored in the bad memory cell group; programming the recovered data into the one of the plurality of memory cell groups (103-0, 103-1, ..., 103-R) in which no data is stored; and The bad memory cell group is retired.
2. The apparatus of claim 1, wherein the one of the plurality of groups of memory cells in which no data is stored is addressable.
3. The apparatus of claim 1, wherein the circuitry is configured to map a logical address associated with the bad group of memory cells to a physical address associated with the one of the plurality of groups of memory cells to which the recovered data is programmed.
4. The apparatus of claim 1, wherein the circuitry is configured to determine the group of bad memory cells during a scrub operation performed on the memory.
5. The apparatus of claim 1, wherein the memory cells of the plurality of groups are ferroelectric memory cells.
6. The apparatus of claim 1, wherein the group of memory cells is a physical page of memory cells.
7. A method of operating a memory (101), comprising: Determining, by a controller (200) coupled to a memory (101), that one of a plurality of memory cell groups (103-0, 103-1, ..., 103-R) of the memory (101) is a bad memory cell group; The controller (200) recovers the data stored in the bad memory cell group; Programming the recovered data into another one of the plurality of memory cell groups (103-0, 103-1, ..., 103-R) in which no data is stored by the controller (200); The controller (200) retires the bad memory cell group; and A logical address associated with the bad memory cell group is mapped by the controller (200) to a physical address associated with the one of the plurality of memory cell groups (103-0, 103-1, ..., 103-R) to which the recovered data is programmed.
8. The method of claim 7, wherein the method includes recovering the data stored in the bad group of memory cells by performing an error correction operation on the data stored in the bad group of memory cells.
9. The method of claim 7, wherein the method comprises: copying data stored in one of the plurality of memory cell groups to another one of the plurality of memory cell groups in which data is stored; and The one of the plurality of groups of memory cells from which the data was copied is erased.
10. The method of claim 7, wherein the method includes disabling on-die wear leveling operations in the memory.
11. The method of claim 7, wherein the method includes disabling random data scrambling operations in the memory.
12. The method of claim 7, wherein retiring the bad memory cell group comprises not storing data in the bad memory cell group.
13. A system (206), comprising: Host (208); and A memory device (226), wherein the memory device (226) comprises: A memory (101) having a plurality of memory cell groups (103-0, 103-1, ..., 103-R), wherein: No data is stored in one of the plurality of memory cell groups (103-0, 103-1, ..., 103-R); All other of the plurality of memory cell groups (103-0, 103-1, ..., 103-R) have data stored therein; and The memory device (226) has a plurality of memory cell groups (103-0, 103-1, ..., 103-R) each corresponding one of which is associated with a different corresponding physical address; and A controller (200) coupled to the memory (101) and having circuitry (210) configured to: determining that one of the plurality of memory cell groups (103-0, 103-1, ..., 103-R) in which data is stored is a bad memory cell group; restoring the data stored in the bad memory cell group; programming the recovered data into the one of the plurality of memory cell groups (103-0, 103-1, ..., 103-R) in which no data is stored; and A logical address of the host (208) associated with the bad group of memory cells is mapped to the physical address associated with the one of the plurality of groups of memory cells (103-0, 103-1, ..., 103-R) to which the recovered data is programmed.
14. The system of claim 13 wherein the circuitry is configured to retire the group of bad memory cells.
15. The system of claim 13, wherein the physical address associated with each respective one of the plurality of groups of memory cells corresponds to a different pin of the memory.
16. The system of claim 13, wherein the host has different respective logical addresses associated with owners other than one of the plurality of memory cell groups having data stored therein.
17. The system of claim 13, wherein the host has no logical address associated with the one of the plurality of groups of memory cells in which no data is stored.
18. The system of claim 13, wherein the bad group of memory cells is a group of memory cells that are unreliable in storing or retrieving data.
